The portals of entry for organisms responsible for most infections which dominate
medicine in tropical countries (as elsewhere) are the skin, and respiratory and intestinal
tracts. A very high proportion of infections of warm climes originates from ingestion
of contaminated water and foodstuffs; many resultant diseases therefore fall into
the subspecialty tropical gastroenterology.1, 2, 3
Most gastroenterological emergencies which occur in a temperate climate also occur
in tropical and subtropical countries. However, there are notable differences in prevalence.
Some are probably ethnically related (although elimination of environmental factors
is often difficult), but the majority are superimposed upon an underlying communicable
(infective) disease; important examples are ileal perforation or haemorrhage resulting
from typhoid (enteric) fever, colonic perforation – and far less often haemorrhage
– in amoebic colitis and shigellosis, and hepatic ‘abscess’ in invasive amoebiasis.
MOUTH AND PHARYNX
The mouth and rectum are the most accessible parts of the gastrointestinal tract from
a clinical viewpoint;
therefore, where endoscopic procedures are impossible (and that applies to many tropical
and subtropical countries), as much information as possible should be derived from
careful examination of these organs.
Viral, bacterial, mycotic and parasitic infections all give rise to oropharyngeal
pathology, which is frequently most pronounced in the presence of associated malnutrition
(especially in infants and children). Herpes simplex virus, Epstein–Barr virus (EBV)
(see Chapter 43) and many enteroviruses can produce a stomatitis; oral ulceration
is also a frequent manifestation of Behçet's syndrome-common in the Middle East and
Japan. Lassa fever (see Chapter 42) and diphtheria (see Chapter 67) are frequently
characterized by severe pharyngeal involvement, and in rabies (see Chapter 44) dysphagia
caused by spasm of the pharyngeal muscles is an important feature of the disease.
In addition to acute bacterial infections, tuberculosis, leprosy, syphilis and yaws
all exert oral manifestations. Candidiasis (exceedingly common in the acquired immune
deficiency syndrome, AIDS) (Chapter 20), histoplasmosis, South American blastomycosis
and coccidioidomycosis can also produce buccal lesions. Acute pharyngitis caused by
infection with young adult Fasciola hepatica (ingested in raw sheep or goat liver
– reported from the Middle East and India – and known locally as ‘halzoun'; Chapter
83) is caused by pentastomids.
Therapeutic agents, such as sulfonamides (included in some antimalarial prophylactics,
e.g. pyrimethamine + sulfamethoxazole, ‘Fansidar') can give rise to the Stevens–Johnson
syndrome, in which oral ulceration is common. Manifestations of specific malnutrition
states (vitamin B and C deficits, and iron deficiency anaemia) are usually obvious,
whereas in kwashiorkor, these are frequently combined with infective complications.
Cancrum oris is a gangrenous condition involving the gums and cheeks and is associated
with Borrelia vincentii and Fusiformis fusiformis infection; it is especially common
in malnourished children,
especially in West Africa. Descriptions of the mouth, especially the tongue, in post-infective
malabsorption (tropical sprue) (see below) were dominant in clinical accounts of this
disease in the nineteenth century (i.e. before the advent of laboratory investigation).
Periodontal disease and dental caries are also a major problem in tropical countries.
Oral submucous fibrosis – a chronic disease of unknown aetiology – may affect any
part of the oral cavity;
most reports are from the Indian subcontinent and Southeast Asia. Fibroelastosis of
the submucous tissues, accompanied by epithelial atrophy, are important sequelae and
are probably premalignant.
Of malignant disease(s), buccal carcinoma is pre-eminent;
Burkitt's lymphoma, ameloblastoma and nasopharyngeal carcinoma (Chapter 35) are other
malignancies that have important geographical distributions in tropical countries.
Hypertrophy of the salivary glands is common in malnourished children; it can also
be associated with Ascaris lumbricoides infection and chronic calcific pancreatitis
Tumours of the salivary glands are probably no more common than in temperate regions.
The most important disease to involve this organ is oesophageal carcinoma
) (Chapter 35); this malignancy possesses an enigmatic geographical distribution.
It has a high prevalence in certain geographical locations:1, 6 central and east-Africa
(western Kenya, Malawi and eastern Zambia have the highest rates), the southern Caspian
littoral (especially north-eastern Iran) and northern China (in and around the Taihang
mountains). Various hypotheses have been advanced to explain the high incidence of
this tumour in these areas (Chapter 35).1, 6
Barium swallow showing oesophageal carcinoma with gross mediastinal invasion.
Megaoesophagus, a feature of chronic Trypanosoma cruzi infection (Chagas' disease),
is described in Chapter 76. Table 10.1
lists some major causes of dysphagia in a tropical environment.
Some causes of dysphagia in tropical countries
South American trypanosomiasis (Chagas' disease)
Candidiasis (usually associated with AIDS)
Rhizopus, Absidia (mucormycosis)
Macronodular cirrhosis (usually varices, postviral)
Portal vein thrombosis
Hyperreactive malarious splenomegaly
Oesophageal varices (Figure 10.2
) usually result from advanced macronodular cirrhosis (see below); however, hepatic
schistosomiasis (caused by Schistosoma mansoni, Schist. japonicum, Schist. intercalatum,
Schist. matthei and Schist. mekongi) are also important (Chapter 82). Portal vein
obstruction (see below) is also common in some parts of Africa and Asia; this probably
results in most cases (although more research is required) from umbilical sepsis in
the neonatal period;
it is occasionally a sequel to hepatocellular carcinoma. A very high splenic blood
flow associated with hyperreactive malarious splenomegaly (HMS; tropical splenomegaly
syndrome) can also give rise to oesophageal varices (see below).
Where and when available, upper gastrointestinal endoscopic sclerotherapy is of enormous
value in the management of oesophageal varices, but an ideal method of dealing with
bleeding varices has yet to appear; in most tropical countries, older methods (see
below) remain extant.
Advanced oesophageal varices in a Zambian woman with severe macronodular cirrhosis
associated with HBV infection; barium swallow examination.
Oesophageal trauma is a major problem in several African countries; foreign bodies
(e.g. kola nuts and fish bones) and corrosive agents – which give rise to strictures
– are also relatively common.
Achalasia, peptic oesophagitis and hiatus hernia are all encountered, but are not
In HIV/AIDS infection, oesophageal candidiasis is a common manifestation; other systemic
mycoses (Chapter 71) can also produce an oesophagitis.
The most common oesophageal lesions in tropical countries are varices (Table 10.1
summarizes the major causes) and carcinoma (see above);
resultant acute complications are upper gastrointestinal haemorrhage and obstruction,
respectively. Hookworm and Ascaris lumbricoides infections (Chapter 85) should not
be neglected in this context.
Of lesser importance, foreign bodies in the oesophagus (e.g. kola nuts) can cause
dysphagia; corrosive lesions can result in stricture formation.
Reported prevalence of bleeding oesophageal varices in tropical countries is unreliable.
Transport facilities are usually exceedingly unsatisfactory; therefore, the majority
of those afflicted die before reaching medical care. Also, high technology (e.g. endoscopic
sclerotherapy) and blood transfusion are less often available; outcome following medical
intervention is therefore frequently less satisfactory than in a Western country.
The cause of upper gastrointestinal bleeding in 131 successive patients admitted with
haematemesis or melaena to a hospital at Harare, Zimbabwe, has been analysed;
in 36 (27%) admissions (mean age 42 years) oesophageal varices were responsible. In
21, conservative management was followed by cessation of bleeding; however, nine suffered
continuous bleeding, and six re-bleeding; five patients died (four within 24 h of
admission) from haemorrhagic shock. Vasopressin infusions were used in four with the
addition of oesophageal tamponade in two.
The pathophysiological mechanisms underlying oesophageal bleeding have been addressed
on numerous occasions.
Both erosive and eruptive bases seem the most likely explanations; in addition, pressure
and variceal size are probably important. In Egypt, endoscopic biopsies obtained from
intervariceal mucosa (within 5 cm of the cardia) in 20 individuals with, and 30 without,
a history of variceal bleeding (most suffered from schistosomal liver disease) were
they showed dilated intraepithelial blood-filled channels within the squamous epithelium
and lamina propria in all of the ‘bleeders’ and in 15 (50%) of the ‘non-bleeders'.
Furthermore, oesophagitis was more pronounced in the bleeders compared with the non-bleeders:
11 (55%) and 7 (23%), respectively.
The role of upper gastrointestinal endoscopy in a developing country has been studied
345 (4%) of 8680 patients examined successively using this technique had evidence
of oesophageal varices, the usual cause being chronic schistosomal liver disease (usually
in Egyptian labourers). By examining 718 successive patients who presented with upper
gastrointestinal bleeding within 24 h of admission, the exact site of the haemorrhage
was delineated in 651 (91%), and the responsible lesion detected in 685 (97%). At
Ibadan, Nigeria, a recent study has indicated that endoscopy gives a superior result
to radiology in the diagnosis of variceal disease, resulting in upper gastrointestinal
endoscopy was successful in 64 (85%), but a barium meal correctly located the source
of bleeding in only 38 (51%) of 75 patients.
Three reports from New Delhi, India, focused on the role of endoscopic sclerotherapy
in the management of bleeding oesophageal varices.14, 15, 16 A total of 79 patients
underwent treatment (with either absolute or 50% alcohol) every 3 weeks, for oesophageal
varices; active bleeding was controlled in 14 of 15 (93%) and 5 of 13 (54%) using
the two fluids, respectively (p< 0.05); the sole disadvantage of absolute alcohol
was that it produced a higher incidence of retrosternal pain. In another study, using
a similar regimen, 5% ethanolamine oleate was compared with absolute alcohol in 47
randomly allocated patients; the latter solution eradicated oesophageal varices earlier
(12.9 vs 8.2 weeks, respectively) (p< 0.001); the mean number of injection courses
and necessary amount of sclerosant were also lower in the alcohol-treated group (p<
0.001), but the frequency of re-bleeding did not differ significantly (p> 0.05). A
total of 31 children with variceal bleeding caused by extrahepatic portal vein obstruction
(19), non-cirrhotic portal fibrosis (5) or cirrhosis (7) were treated by sclerotherapy
using absolute alcohol; arrest of acute bleeding was achieved in 10 by emergency sclerotherapy,
and a 3-week schedule was able to achieve variceal obliteration in all of them. During
a 23-month follow-up period, recurrent varices occurred in three (two with cirrhosis
and one with non-cirrhotic portal fibrosis) patients; a re-bleed was successfully
controlled with emergency sclerotherapy in five, and an oesophageal stricture in four
of them (which was easily dilated) which were the only significant complications.
Although now rarely used in the Western world, oesophageal compression using a Sengstaken
tube is often the only technique available. Intravenous pitressin is of limited value
in acute bleeding. In long-term management, propranolol undoubtedly has a place in
a developing country scenario.
In an attempt to provide clinical guidelines for the prediction of outcome of upper
gastrointestinal bleeding in a developing country, Clamp et al.
carried out a multicentre study based on two centres, in Sikkim and China; in the
former country, 60 (69%) of the patients put into the ‘high-risk’ group (by applying
Bayes' theorem using a computer system) for re-bleeding experienced this event (27
(54%) died), whereas this complication occurred in only six (2%) in the ‘low-risk’
group; furthermore, a simplified scoring system (little computer technology was available
at Sikkim) gave almost exactly the same predictive accuracy. The authors suggest that,
by using one of these systems, patients in remote areas can be categorized in order
that scarce resources (which are available there) can be put to the best use.
The optimal means of managing haemorrhage resulting from extrahepatic portal venous
obstruction is summarized in the section on liver disease (see below).
Presence of histologically diagnosed chronic oesophagitis (using upper gastrointestinal
endoscopy) has been shown to be common in a high-risk population (15–26 years) in
This lesion was significantly associated with: (1) consumption of ‘burning hot’ beverages;
(2) a family history of oesophageal carcinoma (including second-degree relatives);
(3) infrequent consumption of fresh fruit; and (4) infrequent consumption of dietary
staples, other than maize. Associated factors which have been recorded in that population
include: (1) positive cytological smears (568 individuals >30 years of 42 190 had
a positive result); and (2) a high prevalence of pharyngeal carcinoma in free-range
chickens, which lived off domestic scraps
in the local environment.
This tumour often presents late in its clinical course in the heavily affected areas;
in fact, complete luminal obstruction (accompanied by inability to swallow saliva)
is not uncommon at presentation. Passage of a Celestin latex rubber tube (a palliative
technique) is often the only available procedure;
however, blockage is a frequent problem resulting largely from the bulky African (or
other) diet. Chemotherapy and radiotherapy (when available) are of very limited value.
STOMACH AND DUODENUM
Peptic ulceration was at one time considered an unusual cause of abdominal pain in
tropical countries; it was felt by many physicians to be a rare disease.
It is now clear, however, that this is not the case; many difficulties facing the
clinical epidemiologist in a developing country are highlighted by studies of the
geographical distribution of this disease. Because sophisticated methods of diagnosis,
including barium meal and upper gastrointestinal endoscopy, have not until relatively
recently been widely used in developing countries, diagnosis and attempts at establishing
accurate prevalence rates have depended upon recording incidence rates of complications,
especially pyloric stenosis; upper gastrointestinal haemorrhage seems an unusual presentation
overall, but this probably results from the fact that such patients do not reach hospital
before exsanguinations occur. Therefore, serious deficiencies exist in knowledge regarding
the true prevalence of peptic ulceration, and it is currently impossible to draw accurate
conclusions on regional and rural/urban patterns, and also on variations with time,
i.e. during the course of ‘westernization'.
As recently as the 1950s, duodenal ulcer (DU) was considered a rare disease in Africa;
this is not so, because satisfactory radiological, and more recently endoscopic, investigations
have yielded accurate facts on true prevalence rate(s). Prevalence of DU in Africa
has been reviewed using the available literature;1, 19 high-prevalence areas seem
to exist in parts of West Africa, Rwanda, Burundi, eastern Zaire, western Tanzania,
south-western Uganda and the Ethiopian highlands. In southern India
(and Fijians descended from this population
) and Papua New Guinea, the disease also seems relatively common. It has a marked
male predominance; it is frequently post-bulbar, and presentation with pyloric obstruction
is relatively usual. Genetic factors might be important;
the role of diet remains difficult to assess. Whether low rates of presentation resulting
from haemorrhage and/or perforation accurately reflect incidence, or are biased by
the inability to transport a sick patient to hospital, is also impossible to evaluate.
In Lima, Peru there is evidence that the prevalence of peptic ulcer (and also gastric
adenocarcinoma) has declined; between 1985 and 2002 a reduction from 3.15% to 5.05%
Evidence for a causative role for Helicobacter pylori in chronic active gastritis,
peptic ulceration and possibly gastric malignancy has escalated during the last decade;19,
22 however, Koch's postulates have not all been satisfied, and infection rate with
this organism frequently approaches 100% at an early age in an affected population.
In a study carried out in Belgium, South Africa, China and North America, a significantly
higher rate was found in the presence of gastric carcinoma but not duodenal ulceration.
Overall, gastric ulcer (GU) is uncommon in developing countries.
In a study carried out at Kumasi, Ghana, however, perforated duodenal ulcer was less
common than perforated gastric ulcer; the latter was related to the widespread use
of NSAIDs and herbal medicines.
An overall decline in ulcer mortality might be associated with a worldwide reduction
in the occurrence of H. pylori infection.25, 26 When it occurs, it usually has a male
predominance, is most common in the fifth and sixth decades, and afflicts predominantly
the lower social strata. Pyloric obstruction is a common presentation, due frequently
to late-stage disease at presentation. Management of a bleeding peptic ulcer has been
Gastritis, often resulting from alcohol and spicy foods, is a major cause of abdominal
). Infective causes (including tuberculosis) are overall rare, although occasionally
encountered; infections which involve predominantly lower sections of the gastrointestinal
tract (e.g. Salmonella typhi and Shigella spp.) occasionally produce significant gastric
pathology. A heavy infection with hookworm and/or Ascaris lumbricoides can also account
for epigastric discomfort (see below) and should be differentiated from peptic ulceration.
Some causes of severe abdominal pain (without features of intestinal obstruction)
in relation to tropical exposure
Site of pain
Heavy nematode infection (e.g. Ascaris lumbricoides, hookworm)
Mesenteric adenitis (helminthic eggs or tuberculosis)
Acute pancreatitis (helminth related)
Amoebic colitis with perforation (appendix, perforated peptic ulcer or diverticulitis)
Ruptured hydatid cyst
Sickle cell crisis
Recurrent familial polyserositis (familial Mediterranean fever) (Chapter 34)
Hyperinfective syndrome caused by Strongyloides stercoralis
Right upper quadrant
Helminthic infection involving biliary system
Left upper quadrant
Splenomegaly (e.g. hyperreactive malarious splenomegaly [HMS])
Solitary splenic abscess
Right iliac fossa
Anisakis spp. infection
When H2-receptor antagonists (e.g. cimetidine and ranitidine) are used in developing
countries, a possibility exists that they will encourage proliferation of intestinal
pathogen(s) – bacterial and parasitic – for the gastric acid defence mechanism is
available data are, however, presently inadequate for assessing the practical importance
of this. Several studies of gastric acid production indicate that mean acid production
probably varies little in different ethnic groups. Hypochlorhydria is relatively common
in the tropics;
whether it is the cause or consequence of intestinal infection (of bacterial, including
S. typhi, and/or parasitic origin) remains far from clear.
Gastric carcinoma is overall an uncommon malignancy in tropical countries (Chapter
35). At Sura, Fiji, gastric ulcer and carcinoma have been shown to be more common
in Fijians than Indians.
Many facts remain unclear regarding upper gastrointestinal haemorrhage in tropical
countries. For example, DU is apparently common in descendants of southern Indians
in Fiji (see above); however, haematemesis from a chronic DU is more common in Fijians.
Many data suggest that pyloric obstruction is the most common complication of DU in
developing countries. A report from Zaria, northern Nigeria, indicates that at that
location perforation is by no means uncommon;
between 1971 and 1983, 74 (24%) of 302 patients operated for DU, and 29 (58%) of 50
for GU, presented with perforation; furthermore, there was a progressive increase
in the years 1971–1974 to 1979–1983 of from 16% to 45%, respectively. A rare case
report from India has recorded massive haematemesis and melaena from a cholecystoduodenal
fistula secondary to DU in a 24-year-old man;
he was successfully managed surgically.
Ideally, management of the complications of gastritis and peptic ulceration is exactly
the same as in a Western country. In a study carried out at Ankara, Turkey, age, delayed
surgery, presence of shock, status of the anaesthetist, and ‘definitive surgery’ were
significantly associated with a fatal outcome in patients undergoing emergence surgery
for perforated peptic ulcer.
Although usually associated with oesophageal varices, gastric varices also occur alone.
In New Delhi, India, 48 (16%) out of 309 patients with portal hypertension were shown
to have gastric varices;
in six (12%) there was no evidence of associated oesophageal varices. In 11 (28%)
of 40 patients who completed endoscopic sclerotherapy for oesophageal varices, gastric
varices disappeared concurrently with the former, or during the following 6 months.
In the light of their experience, these authors considered that ‘if they persist for
6 months after eradication of oesophageal varices, a combination of paravariceal and
intravariceal sclerotherapy should be attempted for their obliteration'.
Epigastric pain/discomfort is a common presenting symptom in medical practice in tropical
countries (see above);1, 32 this frequently results from heavy small-intestinal helminthic
infections, especially with A. lumbricoides and hookworm. Mesenteric adenitis as a
sequel to the presence of helminthic ova, and tuberculosis, are further causes. Helminth-related
acute pancreatitis is another possibility.
Table 10.2 summarizes some causes of severe generalized abdominal pain. This most
commonly results from peritonitis, which has numerous aetiologies. Right upper quadrant
pain is less likely to result from biliary tract disease than in a ‘temperate’ area
of the world (see below); nevertheless, helminthic infections of the biliary system
are occasionally encountered. Left upper quadrant pain can result from splenomegaly
(following numerous ‘tropical’ infections; see below); an extreme example (HMS) occurs
in most areas which are endemic for human Plasmodium spp. Ruptured spleen is a further
cause of left hypochondrial pain; this event usually presents acutely. Solitary splenic
abscess is by no means an uncommon event in West and Central Africa; the aetiology
Right iliac fossa pain is less likely to be caused by appendicitis (see below) than
in most Western countries. However, an appendix-like syndrome has been recorded in
Yersinia spp., and Anisakis spp. infections and ileocaecal tuberculosis (see below).
Enterobius vermicularis is not infrequently detected in an appendicectomy specimen;
whether there is a cause–effect relationship to acute appendicitis is frequently unclear.
Less common parasites involving the Appendix include Taenia species, Trichuris trichiura
and Angiostrongylus costaricensis (see below). A peripheral blood eosinophilia is
often (but by no means always) present when a helminthiasis is causatively related
to appendicitis. Ileocaecal tuberculosis can account for chronic right iliac fossa
pain; an ileocaecal mass is often palpable clinically (this can be confirmed by ultrasonography
when this technique is available). A colonic amoeboma represents a possible source
of diagnostic confusion.
Tropical enteropathy, subclinical malabsorption and mechanism of diarrhoea
The small-intestinal mucosa of an individual living in a developing country possesses
minor structural differences compared with that in one always resident in a temperate
zone.1, 33, 34 Changes are not related to the clinical syndrome: post-infective malabsorption
(tropical sprue; see below). Although the cause of these changes is not entirely clear,
they seem to result from repeated low-grade viral and bacterial infection(s). Similarly,
marginal xylose and glucose malabsorption has been demonstrated in large numbers of
people indigenous to tropical countries; these abnormalities are certainly greater
in lower socioeconomic groups. Using a breath-hydrogen test, bacterial overgrowth
in the small-intestine was demonstrated in 37.5% of children living in slum conditions
compared with only 2.1% (p< 0.001) controls in urban Brazil.
Subclinical malabsorption exists in many people in developing countries;
xylose and B12 malabsorption have been demonstrated in 39% and 52%, respectively,
of Peace Corps workers living under rural conditions in Pakistan. Apart from repeated
small-intestinal infections, other factors are probably also important.
Xylose, glucose and folic acid absorption have been shown to be impaired in individuals
with systemic bacterial infections, e.g. pulmonary tuberculosis and pneumococcal pneumonia.
Dietary folate depletion also results in xylose malabsorption. Marginal malnutrition
and pellagra have both been suggested as causing subclinical malabsorption, but evidence
The practical importance of subclinical malabsorption is unclear.1, 33, 34 It seems
likely that it significantly contributes to malnutrition in people in developing countries
who subsist on a marginally adequate dietary intake consisting largely of carbohydrate.
Before any rigid conclusions are drawn, however, it should be appreciated that the
small intestine has a very substantial functional reserve, and that the role of the
colon in absorption of carbohydrate (and other substances) (see above) remains unclear.
Diarrhoea resulting from small-intestinal disease consists of two main types;1, 33
(1) profuse watery (e.g. cholera), and (2) steatorrhoeic (exemplified by post-infective
tropical malabsorption (tropical sprue)). Table 10.3
summarizes the most important causes; several of those responsible for the former
type are infective, and then exert their pathogenic effect via an enterotoxin (either
heat stable or heat labile); invasive disease involving the enterocyte is less important.
The role of intestinal hormones – especially vasoactive intestinal peptide – in the
production of watery diarrhoea has become clearer.
The pathogenesis of diarrhoea in AIDS has a multifactorial basis, and is often by
no means clear;
some but not all cases are associated with an opportunistic infection(s), especially
Cryptosporidium parvum (Chapter 79).
The bacteria Escherichia coli, fungi Candida albicans and Histoplasma capsulatum,
and the astroviruses and caliciviruses are also relevant. Other opportunistic infections
in this syndrome include cytomegalovirus, Mycobacterium avium intracellulare, Salmonella
species, and the protozoa Isospora belli, Cyclospora cayatenensis, Sarcocystis hominis
and Microsporidium species infections; in addition, Kaposi's sarcoma (Chapter 35)
causes severe small-intestinal disease.
Watery diarrhoea (large volume, fluid stool(s)):
Traveller's diarrhoea (TD) (turista)
Vibrio cholerae (and other vibrios)
Escherichia coli (enterotoxigenic)
Rotavirus (and other enteric viruses)
(Food poisoning - Staphylococcus, Clostridium perfringens)
Primary - genetically determined
Secondary - resulting from enterocyte damage
Steatorrhoeic diarrhoea (malabsorption) (characteristically large pale, fatty, offensive
stools; microscopy often shows fat globules in faecal smear):
Post-infective tropical malabsorption (‘tropical sprue’)
Trauma - short bowel syndrome (e.g. recovered pigbel disease)
Lymphoma - Burkitt's, Mediterranean lymphomas
Chronic calcific pancreatitis
Acute and chronic liver disease
(Gluten-induced enteropathy (coeliac disease) seems to be uncommon or even rare in
most tropical populations.
Occasionally it can become clinically obvious in visitors from
Western countries to the tropics)
Many of the problems encountered in management, including chemoprophylaxis and chemotherapy,
are exemplified by traveller's diarrhoea (see below).
The clinical syndrome traveller's diarrhoea (TD)1, 38, 39, 40, 41, 42 is arguably
the world's most common disease entity; only rarely is it associated with mortality
(usually in the presence of debility, or at the extremes of life), but the significant
morbidity with which it is associated not infrequently interferes with a crowded schedule
or a leisure or sporting activity. Numerous titles have been applied, including ‘turista',
‘Montezuma's revenge', ‘Hong Kong dog’ and ‘Delhi belly'. One estimate is that 12
million individuals travel annually from an industrialized (Western) country to one
in the tropics or subtropics;
in this group incidence of TD varies from around 20–50%. There is a highly significant
geographical variation in prevalence; high-risk areas include North Africa, sub-Saharan
Africa, the Indian subcontinent, South-east Asia, South America, Mexico and the Middle
East; intermediate ones include the north Mediterranean, Canary Islands and the Caribbean
islands; low-risk ones include North America, Western Europe and Australia. In a retrospective
study carried out in Switzerland, a large group of travellers were asked to complete
a questionnaire after travelling abroad; incidence of the disease varied greatly,
the highest figure (50%) being associated with travel to Tunisia. (No detailed study
exists of TD acquired in a European country.
The disease tends to become less common with advancing years; it is unclear whether
this is due to the fact that older travellers (≥60 years) have a more discerning lifestyle,
or whether relative immunity increases with advancing age.
Individuals resident for substantial periods in areas where TD is common seem to experience
it less frequently than those not previously exposed.38, 39
TD is contracted by ingestion of contaminated water/food; it is characterized by acute-onset
watery diarrhoea (usually of small-intestinal origin);38, 39, 40, 41, 42 when colorectal
involvement exists, diarrhoea is often bloody (see below). Abdominal colic, nausea
and vomiting may be present; fever is unusual, being recorded in 1–10% of infected
individuals. Prostration and resultant dehydration (with electrolyte imbalance) cause
major problems in a severe case. Rarely, symptoms become chronic, and it seems likely
that a small proportion of cases of TD proceeds to post-infective malabsorption (see
Unfortunately, for the investigator, by the time disease has become clinically overt,
the initiating infection(s) has invariably been cleared. Chronic diarrhoea of lesser
severity is a relatively common problem following recovery from acute disease; this
can usually be attributed to (1) tropical enteropathy (in which there is major derangement
of enterocyte structure and function) (see above) or (2) the irritable bowel syndrome
On clinical grounds, an important differential diagnosis is inflammatory bowel disease
presenting for the first time during, or immediately after, tropical exposure.45,
46 In a retrospective review of UK residents presenting at the Hospital for Tropical
Diseases, London, with acute onset/bloody diarrhoea, the majority had inflammatory
bowel disease (usually ulcerative colitis); it was numerically more important than
shigellosis and amoebic colitis.
Acute disease pursues an especially virulent course in certain high-risk groups,38,
39 e.g. those suffering from achlorhydria (Salmonella species and Vibrio species infections
are known to be significantly more common in this group), known inflammatory bowel
disease (see below), previous gastrointestinal tract surgery, a malignancy involving
the gastrointestinal tract, and acquired or congenital immunodeficiency (including
immunosuppressive therapy and HIV/AIDS). In addition, individuals on diuretic therapy
(in whom maintenance of electrolyte balance is precarious) and others at the extremes
of life also fall within the high-risk group. It is important to recognize these factors
when advising chemoprophylaxis (see below).
In 1970, Rowe et al.
recorded results of a study involving British soldiers newly arrived in Aden; in 19
(54.3%) of 33 cases in which a recognized pathogen was not apparent, a ‘new’ serotype
of Escherichia coli was isolated in the acute phase of TD; in a further 14 (40%),
several different E. coli serotypes were also isolated. (B. H. Keane had suggested
in the 1950s (on circumstantial evidence) that bacterial pathogens were implicated.33,
recorded the identity of E. coli serotypes isolated from US Peace Corps volunteers
serving in various countries: Kenya 06:H16, 06H−, 027:H7, 0159:H4 and 0159:H34; Morocco
06:H16, 0128:H12, 027:H20 and 0169:H−; Honduras 08:H9, 015:H49, 015:H− and 027:H20.
Therefore, many common strains of enterotoxigenic E. coli (ETEC) are relevant. Many
other microorganisms are also involved. Salmonella species, Shigella species, Campylobacter
jejuni, enteroadherent E. coli (EAEC) and Vibrio spp. (see Chapter 51); rotavirus
and norovirus (Chapter 45), and Giardia lamblia, Coccidia species (including Cryptosporidium
species, I. belli and Blastocystis hominis) and Entamoeba histolytica (Chapter 79).
Other bacteria which have been implicated include Aeromonas hydrophila, Plesiomonas
shigelloides and Yersinia enterocolitica. The causative agent(s) vary significantly
in different locations, e.g. in an affected individual in Asia, Central America or
Africa the likely organism is different on statistical grounds, although not relevant
to a specific case. Furthermore, more than one organism is frequently present; in
a study involving US Peace Corps workers in Thailand, 33% were infected by two to
four different pathogens.
Although protozoan parasites are usually incorporated in the list of aetiological
agents, the incubation period is usually somewhat longer than is usual in TD; this
applies especially to G. lamblia. When the colorectum is predominantly involved, Shigella
species, enteroinvasive E. coli (EIEC), enterohaemorrhagic E. coli (EHEC) and Ent.
histolytica may be responsible. Rarely, herpes simplex virus and Chlamydia trachomatis
have been implicated. New pathogens will doubtless emerge in future years.
The pathophysiology varies and depends on the site within the gastrointestinal tract
to be involved.1, 38, 39, 40, 41 Whereas in the small intestine toxigenic diarrhoeas
predominate (see above), in the colorectum (see below) invasive disease is more common.
ETEC are characterized by both toxin production and mucosal adherence (via specific
fimbriae); the latter property is required for disease production, for toxin-producing
non-adherent mutants do not cause disease. Enteropathogenic E. coli (EPEC) (probably
not a major cause of TD) adhere to intestinal mucosal cells and although they do not
invade, destroy microvilli. EAEC (detected in up to 15% of patients suffering from
TD) do not belong to classical serotypes of EPEC, but adhere to Hep-2 cells in culture;
they neither produce a toxin nor invade.
EIEC behave similarly to Shigella species and account for up to 5% of cases; the main
site of action is the colorectum, and the major clinical manifestation is therefore
dysentery resulting from epithelial cell invasion and intracellular multiplication;
there is resultant mucosal inflammation and ulceration.
EHEC (an uncommon cause of TD) produces disease via verotoxin production.
Travellers should take maximal care to avoid water/food likely to be contaminated;
common sense is of paramount importance! Use of prophylactic agents is controversial.
Many chemoprophylactics have been used: doxycycline, co-trimoxazole, trimethoprim,
mecillinam, bicozamycin and the fluroquinolone compounds (norfloxacin and ciprofloxacin).
High protection rates (≥90%) have been claimed for co-trimoxazole and the fluoroquinolones;
for trimethoprim a rate of around 50% has been recorded. Most cases of TD therefore
possess a bacterial aetiology. The major problem with antibiotic chemoprophylaxis,
however, is the risk of significant side-effects, dominated by dermatological reactions
(including Stevens–Johnson syndrome) and pseudomembranous colitis (see below); using
co-trimoxazole, a rate of up to 20% of significant skin reactions, necessitating discontinuation
of prophylaxis, has been recorded. Also, the acquisition of resistant faecal E. coli
during chemoprophylaxis has been recorded in several studies; an increase from 21%
to 100% has been recorded using doxycycline in Kenya, and one of 3% to 100% with co-trimoxazole
in Mexico. When chemoprophylaxis is used, either norfloxacin or ciprofloxacin seems
to be the most appropriate, although strains of Campylobacter jejuni rapidly acquire
In a recent study in Egypt, two of 105 individuals on norfloxacin developed TD, compared
with 30 (26%) of 117 given a placebo.
(Ciprofloxacin should be avoided in children because of experimental evidence indicating
cartilaginous damage in young experimental animals; there is no evidence in Homo sapiens.)
Should chemoprophylaxis be recommended widely in this essentially benign clinical
syndrome? In addition to the objections so far outlined (see above), there is a possibility
of inducing a false sense of security, resulting in increased exposure to other infections,
e.g. viral hepatitis.
The following groups should be seriously considered for chemoprophylaxis (for <3 weeks):
Travellers with a bad ‘history’ of TD38, 39, 40, 41
Those in whom hypochlorhydria is proven (or a possibility)
Individuals suffering from inflammatory bowel disease
Those in whom electrolyte balance is precarious (e.g. those receiving diuretic therapy)
and others with chronic renal failure
The ‘elderly’ (not easily defined)
A nebulous group in whom TD is professionally embarrassing (e.g. members of the armed
services, airline pilots, athletes, politicians, businessmen and other professionals
on tight schedules, etc.).
The role of prophylactic antiperistaltic agents is likewise controversial: action
is unphysiological. It has been suggested that they can mask a more serious infection,
e.g. S. typhi, although in this disease diarrhoea is an unusual presenting symptom
(Chapter 52). By delaying excretion of pathogen(s) it is also possible that clinical
disease is prolonged. In children, paralytic ileus is a major complication and has
occasionally precipitated mortality.
Bismuth subsalicylate has a role in prophylaxis; the bismuth moiety possesses antimicrobial
activity and salicylate antisecretory properties.
Early studies in Mexico by DuPont et al.
showed that, given as a suspension (the sheer bulk required precluded its use by travellers),
this agent significantly reduced TD; the same group, also working in Mexico, has demonstrated
that, when given in tablet form (2 tablets 4 × daily for 3 weeks, i.e. 2.1 g daily),
a 65% protection rate can be achieved;
at half that dose, efficacy was greatly reduced. The number(s) of pathogen-positive
TD cases in a group of treated patients was seven of 29, compared with 35 of 59 in
a placebo group; ETEC was present in 3 and 22, respectively, and Shigella species
in two and eight, respectively.
A B-subunit/whole-cell (BS-WC) cholera vaccine has been shown to produce relative
In a study involving Finnish tourists to Morocco, BS-WC induced 52% protection against
diarrhoea caused by ETEC, 65% with mixed infection, 71% when ETEC was present with
another pathogen, and 82% when ETEC and S. enterica were present concurrently. (Sack
has concluded that ‘any advances in prevention and treatment of diarrhoea in travellers
will be directly applicable to the worldwide problem of diarrhoea in children, which
is far more important on a global scale'. This statement does not apply to this BS-WC
vaccine, because protection only lasts for about 3 months.) A further approach under
consideration consists of oral administration of colostrum-derived antibodies against
A recent experimental investigation indicates that lactobacilli, which have the ability
to adhere to the intestinal mucosa, can prevent E. coli colonization. In a limited
clinical study, Lactobacillus GG reduced prevalence of TD by up to 40%.
Treatment (as in cholera, see below) devolves around oral rehydration (see below);
all travellers should carry suitable preparations.1, 41, 53 When properly constituted,
Dioralyte (Rhône-Poulenc Rorer) solution contains glucose 90, Na+ 60, K+ 25, Cl− 45
and citrate 20 mmol/L. Corresponding concentrations for another proprietary preparation,
Rapolyte (Janssen), are 111, 60, 20, 50 and 10 mmol/L. WHO/UNICEF rehydration fluid
contains glucose 111, Na+ 90, K+ 20, Cl− 80 and citrate 10 mmol/L. In a mild case
adequate rehydration can usually be achieved using ordinary mineral water.
The role of chemotherapy in established TD remains controversial. Early work carried
out by DuPont et al.
in Mexico showed that both co-trimoxazole and trimethoprim reduced the length of symptoms.
Recent trials, using antibiotics which have been given for chemoprophylaxis (see above),
have also indicated that the length of symptoms can be shortened; in Mexico, ofloxacin
(600 mg daily for 3 days) produced cure in 77 (95%) of 81, compared with 56 (71%)
patients who received placebo (p= 0.0001).
In a study carried out in Thailand where the causative organism is usually Campylobacter
jejuni, cure rate at 72 h was highest with single-dose azithromycin (96%), compared
with lower rates with 3-day azithromycin and levofloxacin.
Short-course chemotherapy can only be justified in a severe case; this applies at
the extremes of life and in high-risk groups (see above), especially HIV-infected
Cholera (see also Chapter 51) represents the archetypal disease in the context of
small-intestinal secretory (watery) diarrhoea.34, 36, 57
The causative organism, Vibrio cholerae, is not invasive and exerts its effect by
means of an enterotoxin.
If untreated, the disease has a 20–80% mortality; with modern oral rehydration regimens
that figure should be <1%. Death results from dehydration, vascular collapse and renal
Historically, cholera was not confined to tropical countries and involved many temperate
areas, including much of northern Europe. An epidemic in 1854 in London was traced
to contaminated water supplied from the Broad Street pump in Soho. According to legend,
when the handle of the pump was tied down by Dr John Snow, the London anaesthetist,
a rapid decline in the incidence of new cases was recorded.
Cholera is endemic in India, Pakistan, Bangladesh, Afghanistan and many other countries
of South-east Asia. Nosocomial transmission is reported. In recent years, epidemics
have occurred in the Middle East, South America and Africa;
most have been localized. Cholera is endemic along the Gulf Coast of the USA. The
disease is closely associated with poverty, overcrowding and low socioeconomic status.
In former times cholera was spread by population movements such as the annual ‘hajj’
to Mecca; outbreaks involving air travellers have been recorded. Overall, however,
the disease is rare in British travellers.
It tends to affect young people more often than the elderly.
Aetiology and pathogenesis
There is probably a genetic predisposition: blood group O is associated with a higher
infection rate than group A.
Classical cholera is caused by V. cholerae, and is now localized to the Indian subcontinent,
particularly the deltas of the Ganges and Brahmaputra rivers. Elsewhere, the El Tor
biotype, which originated in Indonesia around 1960, and the 0139 strain have been
responsible for most epidemics. Vibrio species are curved, Gram-negative, flagellated
rods approximately 2 mm in length. Each biotype of cholera contains three serotypes:
Inaba, Ogawa and Hikojima. For details of the organism and its pathophysiological
effects, see Chapter 51.
Histologically, the small-intestinal mucosa is intact. Light and electron microscopical
appearances are normal. Following circulatory collapse following gross dehydration,
renal tubular necrosis can be demonstrated.
There are no prodromal symptoms. The incubation period varies from a few hours to
5 days. The disease is similar whichever biotype is involved, but there is a wide
spectrum of severity. When the El Tor biotype is responsible, a higher proportion
of patients are asymptomatic. Onset is sudden, and mild diarrhoea rapidly gives way
to the passage of a large volume of opalescent fluid – the classic ‘rice-water’ stools.
Up to 30 L of fluid, containing a high concentration of Vibrio spp. organisms, may
be passed in 24 h.
Vomiting of fluid of a similar composition is a later feature. Thirst, muscle cramps,
hoarseness and anuria follow.
Clinical signs of severe dehydration may be present by 24 h after onset in an untreated
case. The body temperature is normal or mildly elevated. Circulatory failure and acute
renal failure follow. Confusion, disorientation and hypoglycaemic convulsions may
occur. Mortality rate is directly related to the degree of dehydration. Relative immunity
is short lived. A carrier state, which lasts a few weeks, may occur, and gallbladder
foci have been identified.
Vibrio spp. organisms are easily identified in a faecal specimen; material should
be transported to the laboratory in alkaline peptone water (pH 9.0). A rapid diagnostic
technique for field use has been described. For accurate serological identification
of V. cholerae, rigid criteria are necessary. With classic cholera, organisms are
present during the incubation period and up to 5 days after an attack; in the El Tor
variety, Vibrio spp. can persist for weeks or months.
Faecal samples are isotonic, with a protein concentration of approximately 10 g/L;
pH is about 7.5; typical electrolyte concentrations are: sodium 139 mmol/L, potassium
23 mmol/L, chloride 106 mmol/L and bicarbonate 48 mmol/L. Specimens contain a high
concentration of IgA. Serum IgA and IgM are elevated, the former most markedly in
patients with an El Tor infection. In vitro animal studies, indicate that cholera
toxin enhances IgA secretion from crypt epithelium to ileal lumen.
Serum electrolyte, urea and creatinine concentrations vary with the stage and severity
of the disease. Excessive potassium loss exacerbates metabolic acidosis. Urine is
concentrated; its composition depends on the severity of the disease.
Diagnosis is usually straightforward; however, all other causes of small-intestinal
diarrhoea (with and without vomiting) of acute onset (see below) should be considered.
These include traveller's diarrhoea, E. coli, Staphylococcus species, Clostridium
perfringens, Cl. botulinum, Campylobacter jejuni and viral causes (e.g. rotavirus,
norovirus). Salmonella and Shigella spp. should also be considered. Vibrio parahaemolyticus
(conveyed by infected raw seafood) and other non-cholera Vibrio species can produce
a similar disease. Very occasionally, Plasmodium falciparum malaria presents with
severe watery diarrhoea, especially in infants and children. Food poisoning, caused
by toxic agents, should be added to the list of differential diagnoses.
Basic sanitation and public health procedures should be improved.
Sterility of water supplies is of paramount importance. Contacts of proven cases should
be vaccinated; all faeces and bed linen should be destroyed. Vaccination with inactivated
(dead) Vibrio species organisms gives only limited protection;
0.5 mL and 1.0 mL vaccine should be given at an interval of 1 week, and a 0.5 mL booster
every 6 months.
The 26th Assembly of the WHO recommended, in 1973, that cholera vaccination should
not be compulsory, due to its limited public health value. Despite this, a few countries
continue to demand vaccination before entry. Important progress is being made towards
an effective oral bivalent cholera–typhoid vaccine.
Treatment was revolutionized by the introduction of oral rehydration regimens.64,
65, 66 The enterocyte sodium-glucose carrier system is not affected by cyclic AMP,
and thus glucose (and glycine)-stimulated membrane transport takes place normally.
It is impossible to overload the circulation by the oral route in a previously fit
person. Quantity of ingested fluid should be regulated by faecal loss, best measured
2-hourly. Rehydration should be accomplished within 48 h. In an unsophisticated situation,
sucrose is often more easily obtainable than glucose; results are usually good, although
if severe mucosal damage pre-exists, sucrase concentration is lowered and satisfactory
rehydration is less readily achieved. Cereal-based electrolyte solutions have also
given satisfactory results.
In a severe case, intravenous fluids may be necessary for initial rehydration.
A widely-used formula consists of; sodium chloride 5.0 g, sodium bicarbonate 4.0 g,
potassium chloride 1.0 g, made up to 1 L. Severity of dehydration should be assessed
on clinical grounds; in a case of average severity, 5 L should be given (the first
litre within 10 min) to a 50 kg subject.
Analgesics may be necessary for severe muscle cramps. Intravenous calcium gluconate
is of value for tetany.
Tetracycline hydrochloride, 1 g/day for 5 days, shortens duration of diarrhoea and
clears the luminal content of Vibrio spp. organisms in the case of the El Tor biotype.
A single dose (1 g or 2 g) has also been shown to be effective in V. cholerae infection,
but is associated with asymptomatic bacteriological relapse.67, 68 Tetracycline should
be started several hours after rehydration therapy has begun. Single-dose doxycycline
(300 mg) is probably as effective as tetracycline.
There is clear evidence that in epidemics the El Tor biotype rapidly develops resistance
not only to tetracycline, but also to several other antibiotics (including trimethoprim
plus sulfamethoxazole), and is therefore of very limited value. Recently, Vibrio cholerae
01 biotype El Tor strains have proved resistant to furazolidone and co-trimoxazole.
If cholera is adequately treated, there should be zero mortality, and complete recovery.
A suggestion has been made that individuals who have suffered from cholera might be
predisposed to α-chain disease (see below).
Malabsorption in the tropics
Apart from infective causes, primary hypolactasia (lactase deficiency)1, 71 accounts
for watery small-intestinal diarrhoea in some people indigenous to tropical countries.
A low concentration of this enzyme in the enterocyte brush border is normal for adult
Homo sapiens (as for other species within the mammalian kingdom); the enzyme is under
genetic control. In a minority of the world's population, i.e. northern Europeans,
Africans with an Hamitic ancestry, certain Middle Eastern populations (e.g. Saudi
Arabians) and others in northern parts of the Indian subcontinent, a high concentration
continues into adult life. Secondary hypolactasia results from brush border damage;1,
71 concentration of all disaccharidases (and other digestive enzymes) is reduced,
and slow recovery occurs after the initiating insult has disappeared. Thus, whenever
there is enterocyte destruction (this includes post-infective malabsorption, see below)
Following ingestion of milk or another milk produce, in which lactose is incompletely
hydrolysed, osmotic diarrhoea results; this is accompanied by abdominal colic, distension
and flatulence (‘lactose intolerance'). In a study carried out at Penang, Malaysia,
hypolactasia was demonstrated in all ethic groups, and although there was no clear
association with gastrointestinal symptoms, the authors recommend a low lactose diet
in all Asian countries.
Yoghurt contains adequate bacterial lactase to hydrolyse the lactose component and
is usually well tolerated. Lactic acid production (derived from hydrolysis of lactose
by colonic bacteria) produces irritative diarrhoea, which contributes to the symptoms.
The precise role of the colon in adaptation remains unclear; carbohydrate, in the
form of free fatty acid(s) (and also nitrogen and electrolytes), can be absorbed from
this organ. Investigation of hypolactasia most often utilizes the breath hydrogen
test; lactose ‘tolerance’ test and lactase assay in a jejunal biopsy specimen are
alternatives. In management, milk and all lactose-containing dairy products should
be eliminated from the diet;1, 71 individuals in countries with a high prevalence
of primary hypolactasia can regulate bowel function by varying lactose ingestion.
Post-infective malabsorption (PIM) (tropical sprue)
Relatively little is known about the prevalence and severity of malabsorption in acute
infective conditions of the small intestine (viral, bacterial and parasitic) and the
duration for which it can continue after the specific organism(s) has been eliminated.
In some cases, malabsorption persists in the presence of mixed luminal flora, and
a single infective agent cannot be detected. In others the recognizable initiating
infective cause (or causes) may continue, culminating in a chronic form; a more precise
term is therefore ‘postacute infective’ malabsorption. As with all infective diseases,
the clinical spectrum of disease varies from subclinical to gross pathology (malabsorption).
PIM is of particular clinical significance in tropical countries, where small (and
large) intestinal infections are exceedingly common.
PIM related to tropical exposure has been reviewed by Cook,1, 33, 74 Tomkins,
History and definition
Confusion has existed between PIM and tropical sprue; however, in tropical and subtropical
countries, these entities are synonymous, and the difficulty is primarily one of semantics.1,
33 Patrick Manson first coined the term tropical sprue (derived from a word used by
Dutch workers in the East Indies) in 1880.
The term was rapidly applied to all cases of malabsorption in tropical countries,
undoubtedly including some resulting from tuberculosis and various parasitoses (both
protozoan and helminthic). Historically, chronic diarrhoea accompanied by wasting
was recognized in India before 600 BC; although the Englishman William Hillary is
often credited with the first precise description of tropical sprue at Barbados,
it now seems likely that he described either epidemic G. lamblia infection, or possibly
strongyloidiasis. The clinical syndrome was well known to British physicians in India
during the eighteenth and nineteenth centuries; most descriptions were made in British
expatriate populations. It was in the early 1960s that reports of a high prevalence
of epidemic PIM in indigenous Indians became available.1, 76, 77 Despite early suggestions
that chronic tropical diarrhoea had an insidious onset, it is clear (after careful
assessment) that the vast majority of cases always presented acutely. Confusion has
been compounded further when acute epidemic cases of small-intestinal infection, associated
with gross dehydration (in addition to xylose and fat malabsorption) and acute mortality,
have been designated tropical sprue, as in numerous reports from southern India.
It is essential to include a time factor in the definition of this clinical syndrome,
e.g. chronic diarrhoea and malabsorption, with weight loss, of at least 3–4 months
duration. The term tropical sprue (if used at all) would be better reserved for a
condition where malabsorption of nutrients is quantitatively more important than that
of water and electrolytes. Although the aetiology of PIM is not yet completely clear
(see below), in most cases it undoubtedly follows an acute small-intestinal insult
by either a bacterial, viral or parasitic (or mixed) infection.
Overall, evidence for PIM following a small-intestinal insult is most complete for
bacterial and parasitic infections; those of viral origin might, however, be more
important numerically. Lack of precise data can be largely attributed to the fact
that virology remains a relatively neglected discipline in most developing countries,
where infections of all types are far more common than in the Western world.
The effect of malabsorption on overall nutritional status is largely unknown (see
above); children are especially at risk. The magnitude of energy loss is unclear;
a deficit of 10% of dietary energy (one estimate) is substantial in tropical populations
subsisting on a ‘marginal’ diet. The importance of anorexia in exacerbating associated
malnutrition is also underexplored.
summarizes the geographical localities where PIM has been reported either commonly
or less frequently;33, 34 the map does not include areas where sporadic cases have
been rarely recorded. Although the disease is common (and endemic) in Asia and the
northern part of South America, it is a very unusual condition in tropical Africa.
It remains a problem in travellers to many tropical locations.80, 81, 82 Until recently,
it was a common entity in overland travellers from the UK to Asia; the fact that it
is now rarely seen is probably associated with early antibiotic administration. In
the Middle East and Mediterranean littoral PIM is unusual, but undoubtedly occurs.
World map showing areas where post-infective tropical malabsorption is a significant
There can now be no reasonable doubt that PIM has an infective basis (see above):
it is (1) more common in geographical areas where enteric infection abounds; (2) epidemic
in certain areas, including southern India; (3) the small-intestinal lumen is colonized
by aerobic enterobacteria; and (4) recovery usually occurs rapidly (and dramatically)
following initiation of broad-spectrum antibiotic treatment. Despite this, however,
is of the opinion that in southern India the primary lesion is enterocyte damage resulting
from a ‘persistent’ lesion of the stem cell compartment on a ‘background of tropical
enteropathy'. He further considers that ‘an immunity-conferring agent may be responsible
for the initiating damage'. The widely used definition for this clinical syndrome
in southern India, ‘intestinal malabsorption of at lease two nutrients and the exclusion
of diseases that give rise to secondary malabsorption in a tropical environment',
is inadequate; it does not exclude tropical enteropathy (see above), nor does it introduce
a time (chronicity) factor.
All infective diseases, without exception, have a genetic background. In a limited
study at Puerto Rico, 25 of 27 patients with PIM (not well defined) had at least one
antigen of the HLA-Aw19 series;
the strongest associated link was with Aw31. In India, a high frequency of HLA-B8
HLA-A1, A28 and Bw35 were significantly decreased in the affected group. More data
are undoubtedly required on genetic markers in PIM.
In severe PIM (in the absence of parasites) bacterial colonization has been demonstrated
both within the jejunal lumen and in biopsy specimens. The importance of adhesive
properties of bacteria in pathogenesis is unclear; many bacteria, including E. coli,
S. typhimurium and V. cholerae, possess such properties, mediated by a transmissible
plasmid. In tropical PIM, several groups have demonstrated a higher concentration
of aerobic enterobacteria in relation to the enterocyte compared with luminal fluid.
(In the normal individual, anaerobes outnumber aerobes by about 1000-fold.) It seems
likely that a variety of toxins released by these enterobacteria induce net water
secretion and malabsorption. In the blind-loop syndrome, enterobacteria (which are
invariably obligate anaerobes) do not produce toxins. Several months after tropical
exposure the upper small-intestinal intraluminal bacterial flora (mucosal biopsy or
luminal fluid) remains abnormal;
seven of 11 patients studied had enterobacteria in numbers ranging from 103 to 108/g
or mL. The most common organisms were Klebsiella pneumoniae, Enterobacter cloacae
and E. coli; Citrobacter feundii, Serratia marcescens and Pseudomonas spp. have also
been detected. It seems highly likely, therefore, that these organisms were present
since the onset of disease.
In southern India, a viral aetiology has been sought, but there is little evidence
for this. The origin of continuing overgrowth has not been adequately studied in tropical
PIM; in patients in England with small-intestinal bacterial overgrowth, faecal flora
account for most of the organisms, but salivary flora are probably important in some
Morphological changes are non-specific and range in severity.
Blunting of villi (‘partial villous atrophy') with increased lymphocyte and plasma
cell infiltration (not a feature of tropical enteropathy) are present to a variable
degree; a ‘flat’ mucosa is exceedingly unusual. Although the number of plasma cells
is increased, distribution of IgA-, IgM- and IgG-containing cells is normal.
In untreated gluten-induced enteropathy, T cells expressing T cell receptor g/d heterodimers
are disproportionately raised; this is not so in PIM.
The significance of elevated jejunal surface pH (demonstrated in southern India) is
unclear, but is probably merely an indicator of enterocyte damage. Crypt hyperplasia
has been demonstrated.
Although a predisposing immunological deficit has been postulated in tropical PIM,
there is no good evidence for this; immunological changes (increased IgG, IgE, C4
and orosomucoid, gastric parietal cell antibodies, and lymphopenia with a low peripheral
blood T cell count) seem to be sequelae of mucosal damage, and are not causally related.
In southern India whole-gut transit time (using a radio-opaque marker technique) has
been shown to be unaltered in tropical PIM, despite a striking increase in faecal
weight. Small-intestinal stasis has, however, been well documented in tropical PIM
and might result from excessive enteroglucagon production in response to ileal (and
colonic) mucosal injury (see below).
However, many patients with PIM have received diphenoxylate or loperamide for acute
diarrhoea; both agents produce relative small-intestinal stasis. Both of these agents
interfere with peristalsis and prevent prostaglandin-induced diarrhoea; inhibition
of small-intestinal secretion also occurs. Such stasis is of particular interest because
peristalsis is usually increased in the presence of intraluminal bacteria.
Gut hormones have been studied in tropical PIM in the fasting state and following
a standard meal.
Fasting and postprandial plasma enteroglucagon concentrations (produced by cells in
the distal ileum and colon) and motilin were markedly elevated; furthermore, the elevated
enteroglucagon concentration is significantly correlated with a reduction in small-intestinal
transit (using the H2 breath test). Both enteroglucagon and motilin concentrations
fall after treatment. Concentration of another gut hormone, plasma peptide YY (also
produced by endocrine cells in the ileum and colon and known to delay gastric emptying
and small-intestinal transit, and to reduce gastric and pancreatic secretion) has
been shown to be grossly elevated in PIM;
it seems possible that this results from a change in peptide YY secretion, resulting
from malabsorption, and is a compensatory mechanism in diarrhoea. Patients with PIM
also have a reduced post-prandial rise in gastric-inhibiting polypeptide; gastrin
and pancreatic polypeptide are normal.
Role of the colon
The colonic mucosa, in addition to that of the small intestine, is abnormal in tropical
PIM (‘tropical colonopathy').
Few causes of diarrhoea are strictly confined to one or other of these organs; for
example, shigellosis frequently involves the small intestine, and salmonellosis and
Campylobacter jejuni infection of the colon.
The normal colon is able to absorb 4–7 L of water/24 h,
together with 100–160 mmol carbohydrate (as volatile fatty acid(s)). Failure of the
diseased colon to ‘salvage’ the increased ileal effluent must increase the intensity
Colonic abnormalities have been reported in tropical sprue; using a colonic perfusion
system, impaired water and sodium absorption was demonstrated.
Colonic function has not been investigated in tropical PIM investigated and treated
A clinical syndrome which exhibits very close similarities to PIM has been described
in the German shepherd dog.
Jejunal biopsy specimens show villous atrophy with a variable infiltration of lymphocytes
and plasma cells in the lamina propria. Aerobic bacteria are involved; both clinical
and laboratory recovery take place after broad-spectrum antibiotic therapy.
This is dominated by chronic diarrhoea with large, pale, fatty stools, and sometime
excessive flatulence, usually following an acute intestinal infection.1, 33, 34, 74
Weight loss is sometimes gross and is probably related to anorexia as much as to intestinal
disease. Figure 10.4
shows an affected patient before and after chemotherapy. A wide range of clinical
presentations exists, however, varying from the acute onset type (not strictly post-infective),
described by Baker
as occurring in epidemics (with vomiting and pyrexia in up to 50%) at Vellore, India,
to a far more chronic entity. Other clinical features, such as glossitis (aphthous
ulceration was common in nineteenth-century reports), megaloblastic anaemia, fluid
retention, depression, apathy, amenorrhoea and infertility, occur only after several
(A and C) A 19-year-old Englishman presented in London with post-infective tropical
malabsorption (tropical sprue). Acute diarrhoea started soon after his arrival in
Nepal and he lost approximately 12 kg in weight during the subsequent 2 months. The
total urinary xylose excretion after a 25 g oral load was 2.5 mmol/5 h (normal range
8.0–16.0 mmol/5 h); the 24-h faecal fat was 83 mmol (normal range 11–18 mmol); the
Schilling test result was 0.16% urinary excretion at 24 h (normal >10%) and the 8-h
serum concentration was 0% (normal >0.6%) of the loading dose. Jejunal biopsy histology
showed marked villous blunting with increased lymphocytes in the lamina propria. Parasites
were not found in several faecal samples. Serum albumin 36 g/L; haemoglobin 13.2 g/dL;
mean corpuscular volume 102.9; red blood cell folate 113 ng/L (normal >150 ng/L);
serum vitamin B12 322 pg/L (normal >150 pg/L). He responded rapidly to treatment with
oral tetracycline and folic acid. (B and D) The same man 4 weeks after initiation
of treatment when all investigations were normal.
Table 10.3 summarizes the more important differential diagnoses of chronic malabsorption
in relation to tropical exposure (see below).
There are also many non-infective causes of malabsorption in the tropics and subtropics;
these should be excluded systematically.
During, and immediately after, an acute small-intestinal infection, xylose, glucose,
fat, B12 and folate malabsorption frequently occur (see above). After 4 months or
so, moderate/severe morphological change occurs in the jejunal mucosa; serum folate
and later B12 concentrations decline – often to very low concentrations. Hypoalbuminaemia
and oedema are late signs.
Gastric acid secretion is often depressed, but whether this precedes, or is a sequel
to, the initiating infection is unknown. The role of hypochlorhydria in the production
of small-intestinal infection remains unclear. In a small proportion of cases in southern
India, B12 absorption either improved or became normal with addition of intrinsic
Secondary hypolactasia may be present (see above).
There is no good evidence that PIM predisposes to any gastrointestinal malignancy.
Investigations should include urinary D-xylose excretion, 72-h faecal fat estimation,
a Schilling test and jejunal biopsy; faecal parasites should be excluded (1-h blood
xylose concentration is in practice probably superior to a 5-h urinary collection
in a tropical environment
): serum B12 and red blood cell folate concentrations should be estimated; after 4
months of illness most patients have a low folate concentration. Serum albumin and
globulin concentrations are often depressed. Monosaccharide absorption is impaired
to a greater extent than that of amino acids.
Barium meal and follow-through examination show dilated loops of jejunum with clumping
of barium, in addition to reduced transit rate.
Jejunal mucosal changes are variable, depending on the duration of the disease. By
3 or 4 months, most biopsies are ridged and/or convoluted; a flat mucosa is extremely
unusual and, if present, gluten-induced enteropathy
should be suspected. Submucosal invasion with lymphocytes (predominantly T cells)
and plasma cells is usual.
Ultrastructural changes in jejunal biopsy specimens have been studied;
although lysosomes, peroxisomes and mitochondrial enzymes are not depressed, the organelles
are more fragile. Endoplasmic reticulum is unchanged. A significant reduction in 5-nucleotidase
in the basolateral (plasma) membrane persists after recovery. The latter finding might
reflect an underlying abnormality in the enterocyte of individuals susceptible to
Intestinal permeability has also been investigated;36, 97 abnormalities in urinary
excretion of lactulose and rhamnose following an oral load are similar to results
obtained in gluten-induced enteropathy.
Aetiology and treatment
A hypothesis to account for the aetiology of tropical PIM is summarized in Figure
The ‘vicious cycle’ can be broken by (1) eliminating bacterial overgrowth, and (2)
aiding mucosal recovery (with folic acid supplements). While this hypothesis has been
a satisfactory alternative has not been produced. An adequate diet should be combined
with tetracycline (250 mg three times a day for at least 2 weeks) and folic acid (5
mg three times a day for 1 month). Evidence of susceptibility of the responsible flora
to antibiotics other than tetracycline is limited. Symptomatic treatment may be necessary
in the acute stage of the disease; codeine phosphate (30 mg three times a day), diphenoxylate
(2.5–5 mg four times daily), or loperamide (5–10 mg four times daily) are of value
if stool frequency is excessive. Mild cases respond without treatment, but this may
take several months. Recovery is usually rapid and straightforward;1, 74, 98 in the
pre-antibiotic era a mortality rate of 10–20% was usual.
Hypothetical scheme to illustrate the pathogenesis of post-infective malabsorption.
The open arrows indicate the vicious cycle which, once set in motion, is only broken
by elimination of the abnormal intraluminal flora (†), and hastening of enterocyte
Evidence from south India suggests that response to antibiotics is less satisfactory;76,
77 this has been used as evidence to support a viral rather than a bacterial aetiology
being causative in that locality.
The aetiology of PIM – especially that presenting in association with tropical exposure
– is becoming clearer.
It is probable that several primary insults to the enterocyte (of an infective nature)
are involved. Whereas PIM resulting from most viral, bacterial and parasitic causes
is usually self-limiting, this does not apply to the ‘tropical sprue’ syndrome, when
well established. The reason why only a minority of affected individuals who suffer
an acute small-intestinal infection are susceptible to PIM is unknown; a genetic (or
ethnic) basis for susceptibility seems likely.
Other causes of malabsorption in the tropics
Table 10.3 summarizes some of these. The role of parasitic infection has been highlighted
by AIDS, in which prolonged diarrhoea accompanied by malabsorption and weight loss
can be very troublesome.
Incontrovertible evidence exists that HIV itself causes chronic enteropathy with villous
blunting; crypt hypoplasia results from a direct effect of the viruses on cell replication,
or by an unknown immunological reaction. This is a very common cause of persisting
malabsorption in Africa. In this context, Cryptosporidium parvum and Isospora belli
have recently come to the fore and it is now also clear that these organisms can produce
a self-limiting illness simulating TD in immunocompetent adults and children (see
below). G. lamblia (see below) is undoubtedly the most common cause of parasitic malabsorption.74,
Strongyloides stercoralis (see below), which is widespread in tropical countries,
was until very recently still present in approximately 15–30% of former prisoners
of war in South-east Asia during World War II; it is an underdiagnosed cause.1, 102
Of all causes of malabsorption related to tropical exposure, intestinal tuberculosis
– usually involving the ileocaecal region – is probably that with the lowest index
of suspicion among medical personnel.74, 103 Abdominal tuberculosis can assume several
clinical forms: apart from the hypertrophic ileocaecal form, glandular (involving
the mesenteric glands), peritoneal (sometimes with ascites) and hepatic involvement
(with granulomatous disease) are relatively common. With the first of these presentations,
weight loss and diarrhoea are often accompanied by a low-grade febrile illness; in
severe cases stools are large, pale and bulky. Examination reveals an ileocaecal mass
in 35–50% of cases,
and occasionally enlargement of one or more lymph glands; however, there is often
no clinical abnormality. Late presentation can be as adult kwashiorkor. Anaemia and
hypoalbuminaemia are common.
Chest radiography is usually normal. Absorption tests are frequently abnormal; fat
and B12 absorption are affected most severely. A protein-losing enteropathy may be
present. Pathologically, the disease results either from miliary dissemination, or
follows ileal ulceration. Malabsorption is caused by chronic bile salt loss; unabsorbed
bile salts (normally re-absorbed in the terminal ileum) in turn interfere with colonic
absorption. Barium meal and follow-through examination show ileal strictures,
frequently multiple, in a high percentage of cases; the ascending colon may also be
shortened. The major differential is Crohn's disease, which is statistically much
less common in people indigenous to the tropics. Yersinia infection should also be
considered. Chest radiography is usually normal. The tuberculin test is positive in
70–90% of cases.
A needle liver biopsy specimen occasionally shows hepatic granulomas with caseation.
Diagnostic laparotomy or peritoneoscopy (and peritoneal biopsy) is sometimes necessary
in order to obtain a tissue diagnosis.
Treatment is with an antituberculosis regimen (Chapter 56). Resection of stricture(s)
and occasionally hemicolectomy are sometimes necessary; chemotherapy should be initiated
before surgical intervention.
A further cause of malabsorption in a tropical environment consists of the Mediterranean
(α-chain) lymphoma,104, 105 which occurs sporadically in many parts of the tropics.
If started early, tetracycline usually produces a good result, but not always so.
Although it seems overall uncommon in most indigenous populations in tropical countries
recent reports of coeliac disease have been made from India
and some evidence exists that it might be increasing in prevalence.
Other small-intestinal infections
Significant intestinal protein loss (mean 1.7 g daily) and xylose malabsorption have
been demonstrated in northern Nigerian children with measles (see Chapter 47); approximately
25% also had lactose malabsorption.
Other infections in children caused by enteroviruses and herpes simplex viruses are
also associated with diarrhoea and weight loss; malnutrition may result; the mechanism(s)
(involving enterocyte damage) is probably similar to that in measles.
Volunteers infected with enteric viruses develop small-intestinal morphological lesions
which are not always associated with symptoms.
Jejunal mucosal changes giving rise to severe malabsorption have been well documented
in viral hepatitis;
these may persist for a considerable time after resolution of the hepatic abnormalities.
The norovirus (a 27 nm piconavirus) can also produce mucosal damage and malabsorption.
Rotavirus infections give rise to morphological abnormalities and (especially in children)
These viral infections are invasive, and the resulting diarrhoea and malabsorption
are caused by enterocyte destruction. Malabsorption usually occurs after the virus
has been shed into the intestinal lumen; the villi contain immature crypt-type enterocytes.
In coronavirus infection(s) in piglets, which resemble human rotavirus infections,
glucose absorption is significantly impaired.
This has practical importance in management because sodium and water secretion cannot
be reversed by glucose; oral rehydration fluids, commonly used in small-intestinal
(including travellers') diarrhoea (see above), contain a high glucose concentration
which overwhelms the limited absorptive capacity.
have suggested that coronavirus infections are responsible for at least some cases
of ‘tropical sprue’ in southern India (see above); this might be the case, but asymptomatic
individuals often excrete these viruses and this does not therefore indicate a cause-effect
relationship. Also at Vellore, a search for evidence of Berne virus infection in ‘epidemic
tropical sprue’ proved negative.
Moderate to severe malabsorption is commonplace during acute intestinal infections
of bacterial origin; subnormal absorptive capacity persists for variable periods after
termination of the diarrhoea and apparent clinical recovery. In a study in Bangladesh,
approximately 70% of patients had evidence of xylose malabsorption 1 week after the
diarrhoea had ceased; this was less common after cholera than Shigella species, Salmonella
species and/or Staphylococcus species infections; xylose and B12 malabsorption persisted
for up to 378 and 196 days, respectively, after the diarrhoea had cleared.
Although many different infective insults to the enterocyte are probably important
in PIM (see above), evidence for bacteria being responsible currently has more solid
support than that involving other agents.
These organisms (with varying modes of pathogenicity) produce a spectrum of disease
from TD to malabsorption by enterotoxin production and mucosal invasion-similar to
that caused by Shigella species (Chapter 51). They are frequently food- or water-borne,
and may cause outbreaks of gastroenteritis. Heat-labile enterotoxins exert an effect
by activating adenylcyclase by a mechanism(s) similar to V. cholerae. Both heat-labile
and heat-stable enterotoxins are probably important in TD (see above). A large pool
of resistant E. coli (often showing resistance to multiple antimicrobials) now exists
in the community. Enterotoxin production by E. coli may be transferred simultaneously
with antibiotic resistance (Chapter 53); in a study, 72% and 44% of ETEC isolated
in South-east Asia were resistant to one or more, and four or more antibiotics, respectively.
Enterocyte adhesiveness of E. coli is also a property of some strains and that might
be important in continuing colonization and subsequent malabsorption. The relationship
between adherence and verotoxin production remains unclear.
Attachment of microorganisms to the enterocyte prevents clearance by peristaltic activity;
such mucosal receptors may be determined genetically.
Ultrastructural studies have shown E. coli adherent to mucosal cells, with flattering
of the microvilli, loss of the cellular terminal web and cupping of the plasma membrane
around individual bacteria; intracellular damage was marked in the most heavily colonized
cells. Histological improvement was demonstrated following clearing of E. coli with
neomycin and nutritional support. This mechanism can lead to protracted diarrhoea
in infants. In most cases, resultant malabsorption is short lived.
Malabsorption occasionally follows infection with Salmonella species (Chapter 52),
but the frequency is unknown.
Although unusual, dysenteric disease (bloody diarrhoea) has for long been known to
predispose to tropical PIM;
in addition to shigellosis it is clear that some cases are caused by E. coli (see
above) and others by Campylobacter jejuni (Chapter 51).
Although most cases of Campylobacter jejuni infection are acute, present with gastroenteritis
and are self-limiting, initial symptoms can be prolonged.
The disease is a zoonosis; poultry are frequently contaminated. Many outbreaks have
been traced to infected cow's milk. Dogs also constitute a reservoir of infection.
Although the infection is self-limiting, erythromycin probably hastens recovery when
given early in a severe case. The carrier state is common.
Enteritis necroticans (pigbel disease)
Although described in Germany at the end of World War II (1939–1945), and named Darmbrand,4,
34 this acute infection (Figure 10.6
), which is more common in children than adults, occurs in several tropical countries,
notably the highlands of Papua New Guinea (where it is endemic),
Thailand and Uganda. Recently, enteritis necroticans has been recorded in Khmer children
at an evacuation site on the Thai-Kampuchean border of Thailand; in the former report
36 (58%) out of 62 affected children (10 months to 10 (mean 4) years) died.
It seems likely that a disease termed ‘necrotizing jejunitis’ in rural areas of Bihar,
India – which also affects children – represents the same entity; this condition (‘segmental
necrotizing enteritis') has also been recorded in Jaipur, India, and in Sri Lanka.
Scanty reports of a similar condition have also been made from northern Europe, which
suggests that the disease exists worldwide, but only reaches epidemic proportions
when suitable conditions exist, most importantly for the β-toxin of Clostridium perfringens
type C (ingested in contaminated foodstuffs) to take its toll. Murrell
has suggested (in the light of historical evidence) that the disease was widespread
in medieval Europe when ‘human habitats, food hygiene, protein deficiency and periodic
meat feasting formed the basics of village life as they do in many Third World cultures
today'. Enteritis necroticans is now known to be caused by the ingestion (often at
pig feasts or ‘mumus') of food contaminated by Cl. perfringens type C.
The pathophysiology of the disease is complex, but the presence of a low concentration
of trypsin (resulting from trypsin inhibitors in foodstuffs and chronic protein-energy
malnutrition) allows the β-toxin of Cl. perfringens to survive and produce mucosal
It is sometimes associated with persisting structural changes in the small intestine;
malabsorption may be a sequel.
Gangrenous small intestine at post-mortem in a Papua New Guinean child who had died
from necrotizing enteritis (pigbel disease).
Fluid and electrolyte replacement are essential (see below). Tetracycline or chloramphenicol,
and type C gas gangrene antisera are of value; laparotomy is often indicated. In Papua
New Guinea, immunization against Cl. perfringens type C has given good results;
in a controlled trial, marked reduction in incidence and mortality was demonstrated
in the treatment group. A management strategy has been outlined.
A study carried out in Sierra Leone has indicated that both protozoan and helminthic
infections are particularly common in displacement camps.
The spectrum of disease caused by this flagellated protozoan is broad.1, 74, 100,
101 Symptoms vary from subclinical cases to those with severe malabsorption and malnutrition.
The reason why some individuals are prone to symptomatic giardiasis is not clear;
size of infecting dose, strain variability, genetic predisposition, acquired immunity
factors, achlorhydria, a local secretory IgA deficiency and the presence of blood
group A phenotype have all been considered. An increase in IgE and IgD cell numbers
has been reported in the jejunal mucosa of 20 affected patients;
the former reversed after treatment, when an increase in IgA cell numbers was also
recorded. Genetic characterization has recently been reported from Ethiopia.
The actual mechanism by which the trophozoites cause an absorptive defect is also
unclear. Mucosal injury, with or without invasion, bacterial overgrowth in association
with parasitization, and bile salt deconjugation by bacteria and/or parasites have
all been considered. The extent of jejunal morphological abnormality varies widely.
Clinical presentation is usually between 1 and 3 weeks after infection; contaminated
water and, less commonly, food are the usual sources of infection. Infection occurs
both endemically and epidemically. The disease can probably be contracted from domestic
It is more common in male homosexuals, but is not an opportunistic infection in AIDS
sufferers. Diarrhoea of acute onset, flatus and weight loss may all be present; the
stools have the characteristics of malabsorption. The disease is clinically indistinguishable
from PIM; investigations also give similar results. A full-blown case has all of the
clinical and laboratory features of the classical (historical) reports of ‘tropical
sprue’ (see above). Cysts may be found in a faecal specimen; trophozoites can be detected
in either a jejunal biopsy or jejunal fluid, or with the string test (‘Enterotest').
If mucosal changes and malabsorption exist, circulating antibodies to G. lamblia cysts
can often be detected.
Treatment is with metronizadole (2 g on three consecutive days); alcohol should be
avoided during the treatment period. A single dose of tinidazole (2 g orally) has
been used with success. Two 5-nitroimidazoles – ornidazole and tinidazole (as a single
1.5 g dose) – have been compared;
recurrence of infection during the subsequent 2 months was similar in each case (about
10%). Nimorazole has also been used. An alternative is mepacrine (100 mg three times
daily for 10 days), which is less often used.
The importance of farms as a source of infection has been emphasized in a study from
Importance of domestic pets as a source of infection has also recently been emphasized
in a study carried out in Peru.
Like G. lamblia, this organism produces a broad spectrum of disease; prolonged infection
usually, but not always, occurs in the immunosuppressed (including AIDS) sufferer
where the organism is opportunistic. Diagnosis is similar to that for G. lamblia infection;
oocysts are usually detectable in a faecal sample. Treatment (rarely indicated in
the immunointact) is with spiramycin, but is usually ineffective in the immunosuppressed;
although at least 70 other compounds have been tested, none, including spiramycin,
has proven efficiency in vitro.
The vast majority of small-intestinal parasitic infections do not result in signs/symptoms
unless present at a high concentration.
In a heavy infection, hookworm is responsible for hypochromic anaemia; A. lumbricoides
rarely accounts for obstruction in the small intestine and biliary and pancreatic
ducts (Chapter 85). The major clinical sequel of tapeworm infection is neurocysticercosis
(Taenia solium) (Chapter 87) – a complication unrelated to the intestinal tract.
Although A. lumbricoides, Ancylostoma duodenale and Necator americanus have at various
times been implicated in malabsorption, there is no clear evidence except in rare
or anecdotal case reports.
Diphyllobothrium latum infections are occasionally associated with a low serum B12
concentration; however, this is caused by B12 uptake within the small-intestinal lumen,
and is not an example of true malabsorption.
Clear evidence exists that Strongyloides stercoralis is causally related to malabsorption.1,
34, 74, 102 This helminth can survive in the human host for several decades; some
10–20% of exprisoners of war in South-east Asia during World War II (1939–1945) remained
infected until recently. Onset of diarrhoea is less acute than with G. lamblia. Larvae
can be demonstrated by the ‘Enterotest', and less often by jejunal biopsy. Ova and
larvae can occasionally be detected in faecal specimens. Eosinophilia may be gross;
however, it is often absent. The immunofluorescent antibody test (IFAT) is positive
in approximately 70% of cases; however, cross-reaction with filaria is common. The
enzyme-linked immunosorbent assay (ELISA) test, when available, is more specific.
A negative serological result is common in the immunosuppressed patient. Treatment
is with thiabendazole (1.5 g twice daily on three successive days); repeated courses
may be required. Albendazole (400 mg daily for 3 days) seems less effective. In animal
experiments, cambendazole has given encouraging results; this has also been the case
in limited clinical studies, but the compound has not been officially released for
human use. Other Strongyloides species are important, especially in children. Stongyloides
fülleborni has been implicated in the pathogenesis of severe PIM (see above) in Zambia
and Papua New Guinea, where a significant mortality rate has been recorded.
In the northern Philippines and Thailand, Capillaria philippinensis has been causally
associated with PIM.
It can occur in epidemics. Diarrhoea of acute onset is followed by malabsorption and,
if untreated, infection carries a substantial mortality rate. Protein-losing enteropathy
may also be present. Treatment with one of the benzimidazole compounds has given good
The protozoa Isospora belli and Sarcocystis hominis (usually conveyed by undercooked
pork and beef)
also cause malabsorption. These organisms replicate within the enterocyte. I. belli,
like Crytosporidium parvum, causes a spectrum of disease, from TD to PIM, and is more
common in the immunosuppressed individual. Pyrimethamine + sulfadiazine, and co-trimoxazole
+ nitrofurantoin, have been used with some success. Other protozoan parasites, such
as P. falciparum (in an acute infection) and visceral leishmaniasis (kala azar), can
also produce significant malabsorption. Other protozoa which have assumed practical
importance in the wake of the HIV/AIDS pandemic are Cyclospora cayetanensis,131, 132,
microsporidiosis,134, 135 and Blastocystis hominis.
All can be implicated in a wide range of small-intestinal problems ranging from traveller's
diarrhoea to malabsorption.
Severe dehydration consequent upon secretory watery diarrhoea accounts for enormous
amounts of acute morbidity throughout the tropics; this applies especially to infants
and children. Intravenous replacement therapy has been in use for more than 150 years;
Dr Robert Lewins MD FRCP, of Leith, recorded that he had witnessed Dr Thomas Latta
inject saline intravenously into a patient suffering from cholera (see above) in 1832,
and George Leith Roupell,
a physician at St Bartholomew's Hospital, London, seems to have been an early user
of this technique. It is unlikely, however, that these were the first attempts at
intravenous rehydration (in fact, Sir Christopher Wren, better known for his architectural
achievements, had used the technique experimentally in 1657). Nearly three-quarters
of a century passed before Sir Leonard Rogers, working at Calcutta, demonstrated a
reduction in the mortality rate in cholera patients from 70% to 20% by use of this
technique. Introduction of oral rehydration regimens had to wait much later, in fact
until the latter half of the twentieth century. Introduction of this form of management,
which followed upon important basic applied physiological observations, was, in a
world context, one of the most important medical advances during the twentieth century.
In many acute medical conditions, gastric emptying is delayed; however, this is not
the case in cholera (and presumably other acute small-intestinal infections) and does
not constitute a barrier to oral rehydration, even when fluid and electrolyte loss
(in the stool) is severe.
Oral rehydration therapy remains grossly underused,
however, and infants and children in developing countries with acute gastroenteritis
continue to die unnecessarily because this simple technique is not readily applied.
The authors of this latter article have concluded: ‘the impediment to its wide acceptance
may be that it is counterintuitive for a simpler and much less expensive treatment
to be an improvement over an effective but more complicated technology'!
Enteritis necroticans (pigbel disease)
This acute small-intestinal emergency (see above), which usually affects infants and
children (see above) is characterized by gangrenous changes in the small-intestinal
wall (in patchy distribution); the jejunum is most markedly affected, but the ileum
is also involved. Presentation is usually as an acute abdominal (surgical) emergency,
with abdominal pain, fever and bloody diarrhoea (see above). A chronic stage of the
disease may ensue in which there is narrowing of the small-intestinal lumen (in one
or more places) by a fibrotic stenosis or adhesion; clinical presentation is with
subacute obstruction, often accompanied by malabsorption and malnutrition. Fluid and
electrolyte replacement are vitally important in management; gastric suction is also
required. Penicillin or another antibiotic should be given (see above). Laparotomy
is frequently indicated to confirm the diagnosis and to resect the necrotized, haemorrhagic,
segment(s) of small intestine. Fortunately, active immunization against the β-toxin
has proved effective prophylaxis in Papua New Guinea; hospital admissions for pigbel
in one area of the country fell to less than one-fifth of the previous figure (p<
0.001) when a vaccination programme was introduced.
Morbidity due to this acute abdominal emergency (with a very high mortality rate)
should eventually fall in the seriously affected countries.
Paralytic ileus and acute obstruction
In Pakistan, paralytic ileus has been recorded as a late complication of acute diarrhoeal
disease in infants;
despite rehydration and total parenteral nutrition, the mortality rate was 25%. When
compared with others who did not develop ileus (following acute diarrhoeal disease),
these infants were shown to have had significantly more antimotility agents preceding
the ileus; furthermore, many had a depressed serum potassium concentration. The potential
dangers associated with antiperi-staltic agents, especially in infancy and childhood,
are thus re-emphasized.
Acute intestinal obstruction constitutes a common surgical emergency in both children
and adults in many parts of the tropics, including Africa. Strangulated hernia (usually
of inguinal origin) is usually the most common cause; volvulus and intussusception
are relatively common in tropical Africa; tuberculosis is a further cause due either
to stenosis or to pressure on the third part of the duodenum or jejunum. A heavy A.
lumbricoides infection (especially in children) can also produce small-intestinal
when diagnosed clinically, laparotomy can usually be avoided. Management consists
of intravenous hydration, nasogastric suction and appropriate anthelmintic chemotherapy.
Strangulated hernia, volvulus and intussusception nearly always require laparotomy.
In a report from southern India, 904 children presented with intestinal obstruction;
the most common causes in order of frequency were necrotizing enteritis (see above),
acute intussusception, band obstruction, subacute obstruction, and remnants of the
vitello-intestinal duct. Rare causes of smallintestinal obstruction include: Burkitt's
lymphoma, Mediterranean lymphoma (α-chain disease) (see above) and intestinal schistosomiasis.
Small-intestinal trauma – caused by a road accident or knife, arrow or gunshot wound
– is also important in a tropical context.
Typhoid (enteric) fever
In most areas within the developing world, typhoid (see also Chapters 52 and 53 (and
to a lesser extent tuberculosis) accounts for much small-intestinal disease encountered
in surgical practice;
perforation, obstruction and less often haemorrhage constitute acute surgical emergencies.
This seems especially important in West Africa. S. typhi infection is also an increasing
problem in travellers from industrialized countries to the tropics;
in the USA, 2666 cases (fatality rate 1–3%) of acute enteric fever were officially
notified between 1975 and 1984; 62% of them were imported, the majority of infections
having originated in either Mexico or India. Statistically, surgical complications
are unusual; thus in a series of 82 culture-positive cases in The Gambia there were
no surgical complications;
this was also the case in a series of 192 cases of enteric fever – most caused by
S. typhi – in Thailand.
Despite its relative rarity, however, (perhaps 2–4% of cases worldwide), typhoid perforation
is an extremely serious event, accounting for 20–60% of deaths in this disease (a
statistic which is increased by late presentation, female sex, age ≥40 years and the
presence of multiple perforations). Late perforation is often indistinguishable from
a perforated appendix, amoebic liver abscess, tuberculous peritonitis, an infected
ruptured ectopic pregnancy or intestinal strangulation. The optimal form of management
seems to be surgical, provided the patient is not too shocked to endure such a procedure
(a prolonged period of preoperative resuscitation is often required). There is as
yet no general agreement, however, regarding the ideal type of operative intervention;
simple closure, ulcer excision and closure, wedge excision and closure, ileal resection
and anastomosis, resection and transverse ileotransverse colostomy, and right hemicolectomy
have all found favour. When the perforation is single, simple closure (with or without
excision) is the procedure of choice; an area(s) of impending perforation should not
be oversewn; closure should always be in two layers: an inner one of chromic catgut
and an outer of silk. When there are three or more perforations, bowel resection is
probably advisable. Peritoneal lavage with a copious amount of washing with normal
saline should be carried out. The incidence of postoperative complications is high,
and includes peripheral vascular failure, respiratory infections, anaemia, sepsis,
abscess formation, burst abdomen and intestinal obstruction.
Re-perforation or a new perforation is possible. In a series of 108 consecutive cases
of perforated typhoid enteritis managed in western Nigeria, 100 (93%) underwent ‘debridement
of the perforation and two-layer bowel closure';
35 patients died, usually from overwhelming sepsis. In addition to specific chemotherapy,
although chloramphenicol (1 g four times daily in an average adult, reduced to 1 g
twice daily when body temperature is normal) remains the agent of choice, increasing
numbers of reports of multiple-antibiotic-resistant strains of S. typhi are being
reported (especially from India), metronidazole, and possibly corticosteroids, seem
to improve the prognosis. Alternative chemotherapeutic agents include amoxicillin,
co-trimoxazole, trimethoprim and ciprofloxacin; the last agent is indicated when there
are serious doubts about sensitivity to the other compounds, as is frequently the
case when infection has resulted in Asia. Despite these advances therefore, ileal
perforation in enteric fever remains a potentially lethal complication, especially
Haemorrhage is rarely life-threatening, although recorded;
whereas the majority of cases can be treated conservatively (blood transfusion when
indicated), when selective angiography, fibreoptic endoscopy and high-resolution radionuclide
imaging are available, localization of the bleeding site can be delineated and appropriate
Emergencies associated with helminthiases
Abdominal discomfort (and pain) are common sequelae to heavy small-intestinal nematode
infections (see above), especially ancylostomiasis and A. lumbricoides (see above),
but serious acute complications (see above) are fortunately rare.
Anisakiasis, for example – usually acquired from ingestion of undercooked or raw infected
fish (sushi and sashimi) – can present with an acute appendicitis-like illness.153,
154, 155 Invasive disease caused by this organism is usually localized to the ileocaecal
region; there is no satisfactory parasitological or serological test, and chemotherapy
is not effective. A diagnostic laparotomy is often necessary.
Eosinophilic enteritis is an entity of multiple aetiology.
A report from Townsville, Australia, suggested that Ancylostoma caninum (the dog hookworm)
was responsible for an epidemic (93 cases) encountered there;
nine were subjected to diagnostic laparotomy: eosinophilic infiltration involving
a segment of ileum with indurated thickening of the distal small intestine and proximal
dilatation was the usual underlying pathology. A rare case of acute mesenteric ischaemia
(accompanied by segmental small-intestinal infarction and gangrene) caused by Schist.
mansoni has been reported from Baghdad, Iraq.
The small intestine can also be involved in Schist. japonicum infection; intestinal
obstruction resulting from mesenteric ischaemia, an intussuscepting polypoid mass
or fibrotic stenosis are possible sequelae. Intestinal perforation resulting from
infection with the acanthocephalan helminth Macracanthorhynchus hirudinaceus, a natural
intestinal parasite of the pig, has been described in Bangkok, Thailand
(eight other cases are on record); this infection has also been reported from several
other parts of the world, including China and southern Europe. Fatal gastrointestinal
haemorrhage (associated with fluctuating jaundice, a tender liver, palpable gallbladder
and an eosinophilia) has been attributed to Fasciola hepatica (liver fluke) infection
in Harare, Zimbabwe;
the site of bleeding was probably the biliary tree.
Most cases of colorectal disease occurring in a tropical environment have an infective
basis (Table 10.4
); they are dominated by bacterial (Shigella species
(Chapter 51) (Figure 10.7
), Campylobacter jejuni and invasive E. coli) and protozoan (Ent. histolytica (Chapter
79) and Balantidium coli) infections. Amoebic colitis
and shigellosis present classically with bloody diarrhoea; this should be differentiated
from carcinoma, necrotizing colitis, antibioticassociated colitis and inflammatory
bowel disease (which is overall not very common in tropical countries). Whether or
not amoebic colitis can proceed to inflammatory bowel disease is debatable; however,
misdiagnosis of amoebic colitis as inflammatory bowel disease (with subsequent corticosteroid
therapy) can result in fatality. In AIDS, cytomegalovirus colitis is common; Cryptosporidium
is usually a small-intestinal parasite, but colonic involvement can also occur. In
addition, megacolon resulting from South American trypanosomiasis (Chagas' disease)
(Chapter 76) is another cause of colonic pathology. Of diseases localized to the anal
region, lymphogranuloma is perhaps the most important although bacterial (including
donovanosis, syphilis and gonorrhoea (Chapter 21)) and parasitic (including Ent. histolytica,
Schistosoma species and Enterobius vermicularis) infections constitute differential
Escherichia coli (enteroinvasive)
Schistosomiasis (usually Schistosoma mansoni and Schist. japonicum)
Non-specific ulcerative colitis – inflammatory bowel diseaseb
Irritable bowel syndrome
Characteristically, numerous small stools containing mucus, pus and blood; microscopy
shows pus cells and/or red blood cells in a faecal smear.
Although these diseases are uncommon, or even rare, in most tropical populations,
they can become clinically overt for the first time in visitors from Western countries
to the tropics.
Severe amoebic colitis: operative specimen obtained from an Australian nurse misdiagnosed
as having non-specific ulcerative colitis (inflammatory bowel disease) while working
in Papua New Guinea.
Overall, diseases of the colorectum are far less common in indigenous people in developing
countries compared with individuals in industrialized ones;1, 163 colonic carcinoma
seems, for example, to be an unusual lesion in rural communities. Good evidence now
exists that frequency of these diseases is increasing as urbanization advances, in
Africa especially. Hypotheses to account for these differences include high dietary
fibre consumption in most tropical countries; however, such associations rarely have
a proven cause–effect relationship.
Many data have been collected on colonic function in indigenous inhabitants of developing
it seems likely that mean 24-h faecal weight and volume is higher in Africa, and constipation
unusual. Overall, intestinal transit rate also seems more rapid. Limited evidence
indicates that colorectal histology is mildly different in indigenous people in developing
countries, and is comparable to tropical enteropathy (see above). In PIM in India
(see above) in vivo colonic functional abnormalities have been demonstrated. Whether
colonic pathology is important in a nutritional context remains difficult to evaluate
(see above): evidence now exists that this organ is important in the absorption of
nitrogen and free (volatile) fatty acids.
Inflammatory bowel disease (non-specific ulcerative colitis and Crohn's disease)
is probably less common overall in indigenous people in developing countries compared
with the UK and other Western countries.165, 166 However, a recent study from Lebanon
demonstrated a high prevalence of the disease there.
The aetiology of this disease is unknown, although an infective basis has frequently
been suggested; satisfactory evidence for a viral or bacterial (possibly mycobacterial)
origin is at present lacking. A handful of reports of ulcerative colitis have been
made from African countries, and a few more from Asia.
In individuals in the UK with an ancestry in the Caribbean or Indian subcontinent
this disease clearly exists but is unusual. Such differences also apply to Crohn's
disease, although this disease also is well recognized in Caribbean people in the
UK. Although Crohn's disease behaves very much like intestinal tuberculosis in clinical
practice, response to antituberculous therapy is disappointing. When inflammatory
bowel disease occurs, it seems to behave similarly to that in the indigenous population
of the UK. It is a common cause of bloody diarrhoea in travellers who have returned
to temperate from tropical countries (Figure 10.8
).39, 40, 41 Similarly, appendicitis, diverticular disease and haemorrhoids are overall
less common in a developing country population, where a high-fibre intake has been
implicated in their prevention; a causative association has not, however, been proved.
Barium enema in a 35-year-old English woman who experienced bloody diarrhoea during
a visit to Africa; she had not previously had significant gastrointestinal problems.
Colonic biopsy specimen obtained at colonoscopy confirmed inflammatory bowel disease.
Although irritable bowel (IBD) syndrome (spastic colon)
is extremely common in UK residents (and others) following an intestinal infection
acquired in a tropical country, it seems to be far less significant in indigenous
peoples in Africa
and Asia. Whether this constitutes a genuine difference is unclear because so many
of the latter have more severe symptoms of different origin(s), which might mask symptoms
resulting from IBD. This syndrome does not constitute a single entity;170, 171 although
some cases respond to mebeverine or peppermint oil, many do not. There is no doubt
that recognition of the syndrome in developing countries leaves much to be desired.
More studies are required.
Enterobius vermicularis infection (Chapter 85) is arguably the most common gastrointestinal
infection in the world;
it exists in both tropical and temperate areas.
Colonoscopy is an endoscopic technique which is now available in some, but by no means
all, developing countries; frequently, it is available only at the teaching hospital
and/or other (tertiary referral) centre(s).
Invasive amoebic colitis
Perforation, although a rare event, can complicate this disease, with the production
of amoebic peritonitis;
there may be diffusion of Ent. histolytica from a ‘blotting-paper'-like colon, and
perforation (especially in the rectosigmoid or caecal regions or to the retroperitoneal
tissues) or leakage into a confined space (resulting in a pericolic abscess or internal
intestinal fistula). Management consists of gastric suction and intravenous fluid
replacement; metronizadole, 500 mg 8-hourly (preferably by the intravenous route),
and a broad-spectrum antibiotic should immediately be given. The colon is extremely
fragile; laparotomy is usually best avoided;
overall, mortality is of the order of 50% and after surgery close on 100%. Two reports
have recorded results of surgical intervention in 15 patients with fulminant amoebic
colitis.175, 176 In the first, five out of six patients (four had a subtotal colectomy
with ileostomy, and two a right hemicolectomy and ileostomy) subsequently died (none
was diagnosed either preoperatively or during surgery); in the second, three out of
nine died, all of whom had exteriorization of the cut ends of the bowel following
resection of the necrotic segment (four of those who died had end-to-end anastomoses,
and two peritoneal drainage).
A recent study in China has drawn attention to a significant climatic factor in prevalence.
Although perforation is less common in shigellosis compared with amoebic colitis,
haemorrhage is well documented. The most recent pandemic of this disease in the Western
Hemisphere began in Guatemala in 1969 and ended in 1973. It spread rapidly to Nicaragua,
Belize, Honduras, Costa Rica, Panama and Mexico; with an estimated 500 000 affected,
of whom 20 000 died.
Overall, this entity is less common in developing compared with ‘westernized’ countries.
Nevertheless it certainly exists, and a predominance of appendicectomies in women
has been recorded.
Confusion with an acute gynaecological condition is a real problem and more widespread
use of ultrasound and laparoscopy might be the solution.
In Calabar, Nigeria, 603 consecutive cases were investigated prospectively during
a 5-year period;
there were no major differences from this disease in industrialized countries, and
it constituted the second most common abdominal emergency during the study period,
being less common than acute intestinal obstruction. Many causative agents have been
implicated; in a retrospective review of 2921 appendicectomies carried out at Allahabad,
India, during a 25-year period, 153 produced histological evidence of a specific infection:
tuberculosis (70), Ent. histolytica (17), A. lumbricoides (13), A. lumbricoides and
Trichuris trichiura (2), Enterobius vermicularis (41), and Taenia species (2). This
acute disease should be differentiated from pelvic inflammatory disease, typhoid enteritis,
ruptured ectopic pregnancy, psoas abscess, acute amoebic colitis, and Schist. mansoni
Although the vast majority of cases of appendicitis in developing countries result
from a bacterial cause, helminths, including Schist. mansoni, Strongyloides stercoralis,
Trichuris trichiura and E. vermicularis have also been implicated.1, 184
Volvulus of the colon
This is a disease with clear geographical differences; it is common in much of Central
and East Africa, India and South America;
numerous reports have been made from Uganda and Zimbabwe. Although genetic factors
have been suggested for these high rates; a high-fibre diet, common in most of Africa,
has also been implicated. The major complication is strangulation, and gangrene of
a colonic segment; this should be differentiated from primary volvulus of the small
intestine, compound volvulus (usually ileosigmoid) and internal herniae. Distension
can be relieved with a flatus tube; at laparotomy the nature of the operation, and
extent of resection, depends on the length of gangrenous colon. With simple volvulus,
mortality rate should be low. Zimmerman et al.
have emphasized the value of emergency colonoscopy in the diagnosis of colonic volvulus;
when the mucosa is ischaemic or necrotic, emergency laparotomy is indicated, but when
appearances are normal, relief of flatus (with a flatus tube passed per rectum) together
with medical management followed by elective surgery (resection and anastomosis) 10
days later is recommended.
The common variety, especially in West Africa, is the caecocolic one; although children
may be afflicted, the vast majority are in adults.
The condition has also been reported to be by no means uncommon in Ethiopian adults.
Aetiology, as with that of volvulus, is conjectural; while an intestinal polyp or
amoeboma accounts for some, there is no obvious clue in most cases. Gangrene is about
three times more common with the ileoileal and ileocaecal varieties compared with
the caecocolic type.
Acute colonic dilatation
Several gastrointestinal infections can cause toxic megacolon. These include: Salmonella
species, Campylobacter species and Y. enterocolitica infection; however, there has
been a growing recognition of Shigella species in this potentially lethal condition.
Correct diagnosis is essential; an unnecessary laparotomy can thus usually be avoided.
If the condition is misdiagnosed as ulcerative colitis, and corticosteroids administered,
potentially fatal consequences can ensue. Diagnostically, the causative organism can
usually be identified in a faecal sample. Choice of an appropriate antibiotic is often
difficult; in Shigella species infection, a fluoroquinolone, e.g. ciprofloxacin (200
mg intravenously 12-hourly for 10 days), seems most appropriate. Toxic dilatation
of the colon has also been reported, albeit rarely, in Ent. histolytica infection;
these authors recorded a single case (in which total colectomy, and administration
of metronidazole and emetine, was followed by recovery); they were able to detect
seven cases in the world literature.
Other colorectal lesions
Anorectal infections in relation to tropical exposure have been reviewed.
Trauma to the colon, often resulting from road accidence, constitutes a medical emergency
in most tropical countries.
Necrotizing colitis (the pathology is similar to that of enteritis necroticans; see
above) is rarely encountered. Colonic obstruction is rarely caused by carcinoma (a
rare tumour in the rural tropics
) but is recorded following introduction of a foreign body per rectum. Colorectal
tuberculosis is an unusual cause of stricture formation, which occasionally requires
While the true prevalence of Clostridium difficile infection in developing countries
is considered to be low, many more studies are required.
LIVER AND BILIARY SYSTEM
Liver histology in an individual indigenous to a tropical country differs from that
in one who has spent his/her life in a temperate region of the world.
This organ is subjected to numerous systemic infections – viral, bacterial and parasitic
– and it lies at the distal end of the portal circulation; it is therefore bathed
with portal blood containing viruses, bacteria, parasites, ova, products of digestion
and other antigens. Thus, Kupffer cell hyperplasia and periportal infiltration (with
lymphocytes, plasma cells and eosinophils) are more common, and stellate fibrosis
occurs more frequently. Also, nuclear pleomorphism in hepatocytes and sinusoidal lymphocytes
are frequently prominent; these appearances are unusual in biopsies obtained in a
temperate country. Malaria and schistosomal pigment are often also present. Granulomas
are common (Figure 10.9
) and a large number of differential diagnoses exist; Table 10.5
lists some of them.
Liver biopsy specimen from a 30-year-old Zambian woman. A degenerating Schistosoma
mansoni egg is surrounded by a well-formed granuloma.
Some causes of hepatic granulomas in tropical countries
Viral cytomegalovirus, Epstein–Barr virus
Tuberculosis and atypical mycobacteria, leprosy, syphilis, Q fever, brucellosis
Schistosomiasis, ascariasis, strongyloidiasis, toxocariasis, filariasis, enterobiasis,
Histoplasmosis, coccidioidomycosis, aspergillosis, actinomycosis, candidiasis
Lymphomas – especially intra-abdominal Hodgkin's disease
(sarcoidosis) therapeutic agents – especially sulfonamides
Acute liver infections
Jaundice in a tropical context (Table 10.6
) is most commonly a result of viral hepatitis (types A,193, 194 B (sometimes a combined
infection with D), C,
E196, 197, 198, 199, 200 and F) (Chapter 39), but other causes should also be considered;
Table 10.6 summarizes some of them. An important cause is the jaundice of acute bacterial
infection – most commonly caused by pneumococcal lobar pneumonia or pyomyositis.
The mechanism of this form of jaundice is complex and consists of hepatocellular,
cholestatic and haemolytic elements; the importance of the latter depends on the underlying
prevalence of glucose-6-phosphate dehydrogenase (G6PD) deficiency in the population
under consideration (Chapter 13). It is important to differentiate this form of jaundice
from viral hepatitis, otherwise the appropriate antibiotic will not be administered
for an underlying bacterial infection. In addition to yellow fever, several other
viruses are implicated;
dengue fever, Kyasanur Forest disease, herpes simplex and Coxsackie virus should also
Some causes of jaundice in the tropics
Jaundice of acute bacterial infection: pneumococcal lobar pneumonia, pyomyositis
Marburg and Ebola diseases
Malaria (acute Plasmodium falciparum and P. vivax)
predominantly large-duct obstructive jaundice
Sickle cell disease
Glucose-6-phosphate dehydrogenase deficiency
In AIDS, the liver is affected by many opportunistic organisms. These include viruses;
hepatitis B (HBV) and C (HCV)
infections can be especially virulent. A liver biopsy specimen may also yield evidence
of cytomegalovirus, Mycobacterium tuberculosis, M. avium intracellulare, atypical
mycobacteria, Cryptosporidium parvum, Pneumocystis carinii, Cryptococcus species and/or
Kaposi's sarcoma. Cholestatic features are common. The co-existence of HIV and HBV
should not be underestimated.
In addition to septicaemia, several other infections can produce jaundice;1, 203,
204 leptospirosis is frequently accompanied by renal involvement, while overt jaundice
in typhoid fever ‘hepatitis’ is unusual.205, 206 Melioidosis, plague, tularaemia and
relapsing fever can also produce hepatitis. Of parasitic causes, acute P. falciparum
infection is probably the most important. In acute (Katayama syndrome) and severe
chronic schistosomiasis jaundice may be present, but is rare in invasive hepatic amoebiasis.
Most parasitic infections, including African trypanosomiasis (Chapter 75) and visceral
leishmaniasis (Chapter 77), can produce significant hepatitis and deranged hepatocellular
function – often in the absence of clinical jaundice.
Several parasites produce large duct biliary obstruction; for practical purposes,
A. lumbricoides is the most important to recognize and treat urgently.
Sickle cell disease and haemoglobinopathies (Chapter 13) are important causes of haemolytic
jaundice; they possess a genetic basis. Jaundice in the presence of G6PD deficiency
is frequently precipitated (or worsened) by therapeutic agents and/or toxins. In some
parts of the tropics, especially Indonesia and Papua New Guinea, the Dubin-Johnson
syndrome seems unusually common.
Chronic liver disease
Most cases of chronic active hepatitis in tropical countries result from HBV and HCV
corticosteroids should not be administered for they exacerbate hepatocyte viral infection;
interferon-γ and adenine arabinoside have given encouraging results, but ethnic factors
are probably important. There is no reliable evidence that either malnutrition (including
kwashiorkor) or Plasmodium species infection are aetiologically important, although
such beliefs linger.
In tropical countries most cases of macronodular cirrhosis result from viral hepatitis,
most commonly HBV, and to a lesser extent HCV hepatitis.
The sequence of events is: acute hepatitis → chronic active hepatitis → macronodular
cirrhosis → and, ultimately, hepatocellular carcinoma209, 210, 211, 212 (hepatoma)
(acute viral hepatitis is covered in Chapter 39 and hepatoma in Chapter 35). HBV and
HCV are undoubtedly the most important aetiological factors in hepatoma, but the role
should not be totally disregarded. The true prevalence of autoimmune hepatitis, which
has been studied in Brazil, is unknown.
An important and probably underrated cause of chronic liver disease in a tropical
context is schistosomiasis (Chapter 82).214, 215 Although hepatocellular function
is preserved until late in the disease, portal hypertension and its various complications
(see below) are as important as in the various forms of cirrhosis.
Clinically, cutaneous stigmata of chronic hepatocellular disease are difficult to
detect in brown or black skins;
similarly, other cutaneous stigmata of chronic liver disease may be absent. Diagnosis
is often first suspected by abnormal liver function tests; a needle liver biopsy specimen
is usually diagnostic. Peritoneoscopy is relatively simple and underused in developing
countries; refined diagnostic techniques are rarely available. No treatment is of
any avail in established cirrhosis, but some of the chromolytics in chronic schistosomal
disease of the liver are reversible after treatment (Chapter 82). Major complications
(see below) resulting from portal hypertension are: (1) haemorrhage, from oesophageal
varices (see below); (2) fluid retention, including ascites; and (3) hepatic encephalopathy.
Fluid retention is a major long-term problem, largely the result of a very low serum
albumin concentration. This complication is often difficult to manage, largely because
salt restriction is virtually impossible to impose in a tropical setting; diuretics,
e.g. furosemide (Lasix) (40–120 mg daily) and spironolactone (Aldactone) (100 mg daily),
usually achieve success. Paracentesis abdominis should rarely be undertaken; this
procedure depletes albumin stores further, and electrolyte balance can be seriously
disturbed; tapping ascitic fluid should be reserved for: (1) diagnostic purposes,
to understand whether a bacterial infection, tuberculous peritonitis or hepatocellular
carcinoma is present concurrently; and (2) management of tense ascites, accompanied
by respiratory embarrassment. Hepatic encephalopathy is managed by accepted methods:
oral neomycin (6 g daily) and/or lactulose (20–35 g three times daily); in the presence
of hypolactasia, lactose can be substituted for lactulose.
Other forms of chronic liver disease (with subsequent decompensation) (see below)
include those resulting from excessive alcohol ingestion, Indian childhood cirrhosis,
haemosiderosis and veno-occlusive disease. Wilson's disease (hepatolenticular degeneration)
and other genetically determined forms of cirrhosis are of limited importance numerically
in the tropics, although they too should enter the list of differential diagnoses.
Alcoholic liver disease
Alcohol-related disease (including cirrhosis) is common in both indigenous and expatriate
populations in tropical countries.1, 216 Genetic factors are undoubtedly involved;
HBsAg carriers are especially vulnerable. The liver in chronic alcoholic disease is
classically micronodular, but not always so; liver biopsy histology sometimes shows
characteristic Mallory's hyaline deposits, and haemosiderin may be present in excess.
There are no major differences from the disease in temperate climates. The quantity
of daily alcohol required to produce this disease is not known with accuracy, and
estimates differ widely; an individual variation exists, and women seem to tolerate
chronic alcohol ingestion less well than men. Acute alcoholic hepatitis is underdiagnosed
and possesses a high mortality rate; the role of corticosteroids continues to be disputed;1,
216 any beneficial effect is at best marginal and administration should probably be
confined to severe and advanced cases.
Indian childhood cirrhosis
Indian childhood cirrhosis
is largely confined to India (especially south India, Calcutta and the Punjab) and
surrounding countries; it is frequently familial. Diagnosis is usually made between
1.5 and 3 years of age; members of the upper strata of Hindu society are often affected.
The disease may pursue fulminant, acute or subacute courses, and carries a high mortality
rate. The clinical course therefore varies widely and is comparable to viral hepatitis
(see above), with acute fulminant hepatitis at one extreme of the spectrum and cirrhosis
(with one or all of its classic complications) (Figure 10.10
) at the other. Histologically, there is usually progressive fibrosis, with absence
of regeneration; macronodular and micronodular cirrhosis result. Hepatocellular carcinoma
is an uncommon complication. The disease is associated with a high copper intake;
epidemiological evidence suggests that early weaning followed by milk-feeding from
copper vessels imparts an excessive copper intake.
However, the possibility of an inherited defect resulting in excess copper absorption
and/or metabolism has not been eliminated. There is no adequate treatment; in prevention,
non-human milk for infant and childhood consumption should not be stored in coppercontaining
Indian child suffering from decompensated chronic liver disease – Indian childhood
Haemosiderosis (African or Bantu siderosis) is a disease of southern, and to a lesser
extent other parts of (tropical) East and West Africa.219, 220 Whether it can proceed
to clear-cut cirrhosis is arguable; heavy alcohol intake is commonplace in many geographical
areas where the disease is common; it is frequently impossible to exclude this as
an aetiological factor (as with haemochromatosis). Iron-containing pots for cooking
are commonly used in most areas, such as Zimbabwe, where haemosiderosis is common,
but other factors also seem relevant. Also, chronic pancreatitis is relatively common
in these areas; evidence exists that an excess of iron (and fat) is common.
Although first described in Jamaica, distribution of veno-occlusive disease is now
known to be much wider.
Bush-teas, which contain pyrrolizidine alkaloids (Heliotropium, Crotalaria and Senecio)
are important aetiologically. Veno-occlusive disease occurs in many localized areas
of the tropics, and is certainly not confined to the Caribbean.
Other chronic liver diseases
The liver is involved in most chronic infective diseases; tuberculosis, leprosy, syphilis,
actinomycosis, visceral leishmaniasis and African histoplasmosis are examples. It
is, however, unusual for decompensation (and liver failure) to result. Major space-occupying
lesions involving the liver are amoebic abscess (see below), pyogenic abscess and
hydatid disease; tuberculomas, cysticercosis and melioidosis are of lesser importance.
Of non-infective diseases, sickle cell disease, β-thalassaemia, haemoglobin-H disease,
porphyria and α1-antitrypsin deficiency produce significant hepatic pathology. A change
in disease profile of the Budd-Chiari Syndrome has been recorded over the past three
decades in India.
Portal hypertension1, 223 is a sequel to any form of chronic liver disease; Table
summarizes some causes in a tropical country. Cirrhosis and schistosomal liver disease
(Chapter 82) are numerically very important; however, in the latter entity hepatocellular
function is preserved to a greater extent, and for longer in the course of disease
than in cirrhosis; therefore, fluid retention and more importantly encephalopathy
are less common. A form of non-cirrhotic chronic liver disease, sometimes associated
with portal hypertension, exists in India; despite various suggestions (including
arsenic poisoning), the aetiology remains unclear. Of pre-hepatic causes, HMS (see
Table 10.7) is the most common; portal hypertension results from an increased splenic
blood flow. Portal/splenic vein obstructions, probably resulting from neonatal umbilical
sepsis, are important causes throughout tropical countries, and are undoubtedly underdiagnosed;
hepatocellular function is usually intact. Posthepatic causes of portal hypertension
include (Table 10.7) cardiac failure (usually resulting from chronic rheumatic cardiac
disease), right-sided endomyocardial fibrosis (Chapter 12) and constrictive pericarditis,
usually but not always resulting from tuberculosis. Other causes of portal hypertension
are hepatocellular carcinoma (see above) and various dehydrating diseases, including
dysentery and cholera. Splenomegaly is present whatever the cause of portal hypertension
(which should be distinguished from other causes of enlargement of this organ in a
tropical country). Barium swallow or upper gastrointestinal endoscopy usually confirms
the presence of oesophageal varices. When available, ultrasonography is valuable in
assessing portal vein patency.
Causes of portal hypertension and oesophageal (and gastric) varices, showing those
which are more common in developing countries
Level of obstruction
Hyper-reactive malarious splenomegaly (HMS) (increased portal blood flow)a
Portal vein occlusiona
Splenic vein occlusion
Hepatic macronodular cirrhosisa
Congenital hepatic fibrosis
Cardiac failure (secondary to chronic rheumatic disease)a
Inferior vena caval obstruction
Hepatic vein thrombosis (Budd-Chiari syndrome)
More common in a developing than a developed country.
Biliary tract disease
In tropical countries biliary pathology is largely attributable to parasites,1, 207,
224 ascariasis (Chapter 85), clonorchiasis and opisthorchiasis (Chapter 83); pigment
stones (often intrahepatic) occasionally complicate sickle cell disease. A. lumbricoides
infection (Chapter 85) is an underdiagnosed cause of large-duct obstruction. It should
always be considered in this clinical situation, for it may be confused with pancreatic
carcinoma. Endoscopy, if available, is of value; medical treatment is usually successful.
Clonorchiasis and opisthorchiasis (Chapter 83), acquired from ingestion of raw fresh-water
fish, may result in cholangiohepatitis and biliary obstruction; cholangiocarcinoma
is a late complication of both infections. F. hepatica infection (Chapter 83) can
give rise to tender hepatomegaly accompanied by jaundice; difficulty in diagnosis
from viral hepatitis may be a problem; an eosinophilia is, however, common with this
and all biliary trematode infections. Praziquantel is of no value in treatment; triclabendazole
has now replaced it.225, 226, 227 Overall, cholesterol stones (and associated secondary
infection) are uncommon in rural populations, especially in Africa. Gallbladder infection
by S. typhi can result in the typhoid carrier state (Chapter 52); the focus of infection
is usually intrahepatic. Gallbladder carcinoma is unusual.
Acute hepatocellular failure
Acute liver failure (acute hepatic necrosis) is a major clinical problem in all developing
countries (see above);4, 228 various hepatitis viruses (most commonly B, C, D and
E, and to lesser extent A) are all involved (see above), but some cases are caused
by other viruses, bacteria or toxins. Although acute hepatocellular failure has been
recorded in severe acute P. falciparum infection, this is of very limited clinical
importance; it occurs as a terminal event but is of far lesser importance than other
major organ failure.
The role of several viruses involved in the production of acute liver injury has been
Several reports highlight the aetiological basis of hepatitis in tropical countries;
in Egypt, HBV and hepatitis A virus (HAV) accounted for 47% and 0.7% of cases of acute
hepatitis (there was serological evidence of both viral infections in a further 1.4%),
whereas 14.2% of cases were HBsAg carriers, 31% ‘non-A, non-B’ hepatitis and 6% were
In other locations, however, hepatitis D virus (HDV) is important, especially in southern
America, South-east Asia (and probably India) and northern Africa. Thus in Thailand,
HDV is frequently present in drug abusers; it is also endemic in Chandigarh, India,
and has been described in an epidemic of acute hepatitis in the Himalayan foothills
in south Kashmir.
In India and South-east Asia, hepatitis E virus (HEV) (see above) is responsible for
most cases of the entity previously termed ‘non-A, non-B’ hepatitis; a similar situation
probably pertains in Africa and South America. This virus is transmitted by the faecal–oral
route and is transmitted in contaminated drinking water; the major importance of this
infection is that it produces a high incidence of hepatocellular failure in pregnant
women. HCV also causes severe disease – including acute hepatic failure – similar
to that produced by HBV (Chapter 39).
Many other viruses present in tropical and subtropical regions may also produce acute
hepatic necrosis; these include herpes simplex type 1, herpes virus 6,
Epstein–Barr virus, cytomegalovirus, yellow fever
and the haemorrhagic fever viruses, which include the Lassa fever virus, the Marburg
virus, Ebola virus and Rift Valley fever virus (see above).235, 236 Of bacterial causes
of hepatitis, enteric fever is common, but rarely (if ever) proceeds to hepatocellular
necrosis (see above). The jaundice of systemic bacterial infection
often follows pyomyositis, especially in Africa. P. falciparum malaria causes deranged
liver function tests resulting from centrilobular necrosis (see above). Hepatotoxicity
resulting from herbal remedies is not confined to tropical countries.
Alcoholic hepatitis is a significant clinical problem in both indigenous and expatriate
Tandon et al.
have outlined their experience of acute hepatic failure (resulting from viral hepatitis)
in 145 (>12 years old) patients managed by them using a ‘simple supportive therapeutic
regimen’ during a 5.5-year period at New Delhi, India. Criteria for inclusion were:
Development of hepatic encephalopathy within 4 weeks of onset of symptoms and signs
of acute hepatitis; and
Absence of evidence of pre-existent liver disease.
There were 65 men and 80 women; 46 of them were pregnant and presumably infected by
They used a simple intensive support mechanism; this consisted of:
Isolation in an intensive care room.
Attention to general hygiene and care of a comatose patient.
Intravenous fluid to provide 1000–1500 calories daily using 10% dextrose, supplemented
if necessary, by 20% dextrose.
Nasogastric tube for aspiration of gastric contents and instillation of drugs.
Gut sterilization by ampicillin (1.5 g 6-hourly via nasogastric tube); colonic washes
Liquid antacids (30 mL 2-hourly).
‘Lactisyn' (1 ampoule = Lactobacillus lactus 490 million, L. acidophilus 490 million,
Streptococcus lactus 10 million) three times daily.
Condom or catheter drainage of the urinary bladder.
Maintenance of electrolyte and fluid balance by intravenous supplementation.
Complications were managed as follows:
Infection (diagnosis was based on clinical findings, leukocyte count >15 × 109/L,
and/or chest radiograph abnormality): gentamicin 3.5 mg/kg body weight (as three divided
doses), and/or cephalexin (2 g daily as four divided doses)
Cerebral oedema (criteria for diagnosis were: focal or generalized seizures, abnormal
reactive or unequal pupils, decerebrate posture of the body after minor stimuli, and/or
sudden deterioration of vital signs): intravenous mannitol (200 mL administered during
30 min and repeated three or four times per 24 h).
Gastrointestinal bleeding (diagnosed by aspiration of fresh or altered blood via nasogastric
tube): liquid antacid (30–45 mL every 2 h), gastric lavage (with 100 mL cold saline
containing 8 mg noradrenaline every 30 min) and occasionally cimetidine. (When the
prothrombin time was >7 s compared with a control, fresh frozen plasma was administered.)
Renal failure (the criterion used was: oliguria (urine output <400 mg/24 h, and rising
blood urea) despite adequate hydration): diuretics (judiciously used).
Overall, 42 (28.9%) survived; of those ≤40 years old, 41 (33%) recovered, compared
with only one (4.8%) of those ≥40 years; survival was not affected by pregnancy. Indicators
of poor prognosis were: grade IV coma, presence of HBsAg, serum bilirubin concentration
>20 mg/100 mL and sodium <119 mmol/L. In fatal cases the immediate complications resulting
in death were cerebral oedema (65), bleeding (31), renal failure (11) and infection
(8). The authors concluded that these results were comparable with results from centres
using a variety of complex therapeutic regimens (e.g. exchange blood transfusion,
charcoal perfusion and haemodialysis).
Chronic hepatocellular failure and hepatoma
Cirrhosis, generally resulting from one of the hepatitis viruses (see above), is a
very common problem throughout tropical and subtropical countries. A study carried
out at New Delhi, India, has addressed the problem of survival in young (<35 years
old) and older patients with cirrhosis;
numbers in the two groups were 63 and 106, respectively. Aetiology of cirrhosis in
the young and adult groups was: HBV-related (32 and 51), alcohol-related (10 and 28),
while 19 and 21, respectively, were labelled ‘cryptogenic'; in the former group, one
had Wilson's disease and another α1-antitrypsin deficiency. During the surveillance
period 27 and 47 deaths occurred: 40% and 64% from hepatic failure, and 52% and 26%
from variceal bleeding. The 5-year survival (62% and 56%) and probability of survival
within a similar grade of liver disease (Child's classification) were comparable.
As anticipated, probability of survival was significantly higher in grade A and lowest
in C. Aetiology of cirrhosis did not significantly influence prognosis in this study.
Hepatocellular carcinoma usually presents as a rapidly progressive malignancy; however,
an acute or chronic presentation can occur due to internal necrosis and haemorrhage.
Such a lesion can in fact rupture into the peritoneal cavity, posing problems in differential
In a patient with actively bleeding oesophageal varices, differentiation of the aetiology
of underlying liver disease (from postviral (or another aetiology) cirrhosis and chronic
schistosomal disease) is usually impossible on clinical grounds alone. In a study
carried out at Cairo, Egypt, liver ultrasonography was undertaken in 50 patients who
were undergoing an operation for bleeding oesophageal varices;
ultrasonographic diagnosis was compared with a surgically obtained wedge biopsy specimen.
The authors concluded that ultrasonography gave the more accurate diagnosis; the findings
in schistosomal periportal (pipe-stem) fibrosis were characteristic and were not mimicked
by other liver diseases (including cirrhosis); ultrasonography agreed with the histological
diagnosis in 44 cases.
Role of ultrasonography in management
The overall value of ultrasonographic scanning and scintigraphy in the diagnosis of
chronic liver disease in developing countries has been addressed.
Needle biopsy is frequently necessary to diagnose diffuse disease, but a high degree
of specificity can be anticipated with a space-occupying lesion.
A further problem surrounding ultrasonography has been highlighted:
in Africa and other developing countries, focal lesions ‘often present so late that
lesions revealed by ultrasound are huge and bizarre', and the inexperienced radiologist
may therefore be baffled.
Portal hypertension and its complications
The major causes of portal hypertension (and oesophageal varices) are summarized in
Table 10.7. Some geographical variations have been reviewed.1, 9 While in many parts
of the world cirrhosis is the most common cause, in India non-cirrhotic portal fibrosis
is relatively common.
Indian childhood cirrhosis (see above) also accounts for cases in the younger age
group(s). Extrahepatic portal vein obstruction is common in some countries (including
India);223, 242 however, in Egypt, Africa, the Middle East, South America and China,
Schist. mansoni and Schist. japonicum, respectively, are frequently responsible. In
Jamaica, South Africa, central Asia and the south-western USA, epidemic veno-occlusive
disease (see above) (caused by Heliotropium, Crotalaria, Senecio and other alkaloids;
see above) is important.
Pitressin (vasopressin) forms the basis of management of variceal haemorrhage; if
and where available, upper gastrointestinal endoscopic sclerotherapy is of value,
but this technique usually has to be repeated at 6-monthly intervals. The Sengstaken
tube (for variceal compression) still has a place in developing countries. Haemorrhage
is not a major presenting feature at most tropical hospitals (see above).
Bleeding varices resulting from extrahepatic portal obstruction
The cause of portal vein thrombosis in developing countries remains unclear; it is,
however, a relatively common condition, and neonatal umbilical sepsis is usually cited
as the likely aetiological factor.
During an 8.5-year period, 136 patients with extrahepatic portal hypertension were
treated surgically at New Delhi, India;
in 22 it was carried out as an emergency (for variceal bleeding), and in 114 as an
elective procedure (in 104 for a past haematemesis and in 10 for massive splenomegaly).
The emergency strategy consisted of: splenectomy and splenorenal shunt (14), transoesophageal
variceal ligation (4), splenectomy and gastro-oesophageal devascularization (3) and
mesocaval shunt (1). Elective procedures were: splenectomy and splenorenal shunt (94),
mesocaval shunt (8) and splenectomy and gastro-oesophageal devascularization (12).
Operative mortality was 2 (9%) and 1 (1%), respectively; none of the survivors developed
encephalopathy or postsplenectomy sepsis. One hundred and seventeen (86%) were followed
up for 2–10 years; 17 had a further haematemesis, but 90% and 75% were alive at 5
and 10 years, respectively. Patients experiencing haematemesis are often far from
medical facilities in a developing country; the authors therefore considered that
in this setting operative intervention was more satisfactory than endoscopic sclerotherapy
or management with propranolol (variceal compression was not considered).
Space-occupying hepatic lesions
Invasive hepatic amoebiasis
Amoebic liver abscess is a cause of right upper quadrant pain (and hepatomegaly);
this is usually accompanied by fever, and not infrequently right shoulder-tip pain.
Travellers to infected areas as well as the indigenous population(s) of the tropics
may be affected.1, 243 Pathogenesis is dependent on an oral infection with a potentially
invasive strain (zymodeme) of Ent. histolytica.
The mode of evolution remains unclear.
Diagnosis is based on an appropriate serological technique (IFAT, cellulose acetate
or countercurrent immunoelectrophoresis) and hepatic ultrasonography or computed tomography.
Clinical characteristics in a group of 52 patients suffering from amoebic liver abscesses
have been recorded at Cairo, Egypt;
while 22 (42%) presented with an acute illness (see above), 30 (58%) had a more chronic
illness with dull aching in the right hypochondria, weight loss, fatigue, moderate
to low-grade pyrexia and anaemia. A right-sided pleural effusion, emphysema, ascites
and jaundice were present in three (6%), four (8%), seven (13%) and seven (13%), respectively.
Forty-two (81%) abscesses were solitary and in the right lobe; 29 (43%) were initially
solid or heterogeneous. Response to metronidazole (750 mg three times daily for 10
days) was described as good in 50; in four aspiration was carried out on account of
the large abscess size.
Whether needle aspiration of an amoebic abscess (in addition to satisfactory chemotherapy)
is indicated remains controversial. A prospective, randomized controlled study carried
out at New Delhi, India, has addressed this issue;
in 17 of 37 patients (all received appropriate chemotherapy, 2–4 g metronizadole for
10 days) who completed the study, aspiration was carried out on the day of hospital
admission; clinical improvement (and cure) was similar to that in 20 controls. ‘Abscess’
diameter was slightly lower in those who underwent aspiration (54 vs 72 mm). However,
at Benin, Nigeria, needle aspiration was considered to ‘enhance clinical recovery';
in a non-randomized trial, 19 patients were managed by needle aspiration in addition
to chemotherapy, and 17 were given chemotherapy (metronidazole, diloxanide and chloroquine)
alone; 18 and 10, respectively, experienced complete resolution (as shown by ultrasonography)
after 21 days (p< 0.021), and clinical response was also considered more rapid (p<
0.01), especially when the abscess was >6 cm in diameter. Delay in ultrasonographic
‘recovery’ is not important, there being good evidence that a residual abnormality
after a year or more is compatible with complete, uncomplicated resolution.
Although no in vitro evidence of Ent. histolytica resistance to the 5-nitroimidazole
compounds exists, reports continue to be made from India of drug-resistant cases.
The main problem with such reports is that, in few (if any) has diloxanide furoate
(500 mg three times daily for 10 days) been administered; this is essential for a
definitive cure because it is a far superior luminal amoebicide compared with the
5-nitroimidazole compounds – and therefore kills the cysts (which could belong to
invasive zymodemes). In a prospective randomized study of 50 such ‘resistant’ cases
at New Delhi, four management regimens were used:
(1) a repeat course of conservative therapy (with 1.25 mg/kg dehydroemetine given
intramuscularly daily for 10 days); (2) needle aspiration (under ultrasonographic
guidance); (3) percutaneous catheter drainage (under ultrasonographic guidance); and
(4) open surgical drainage with catheter insertion. The authors concluded that ‘the
most impressive results’ were obtained with regimen 3.
To summarize, in the uncomplicated case, needle aspiration (under cover of a 5-nitroimidazole
compound) is indicated when: (1) the abscess(es) cavity is large and the patient seriously
ill; and (2) the site of the lesion is such that perforation into a nearby viscus
(most importantly the pericardium) seems probable. All cases of invasive amoebiasis
should receive a course of the luminal amoebicide, diloxanide furoate (500 mg three
times daily for 10 days) after metronidazole (800 mg three times daily for 10 days)
or tinidazole (2 g daily for 3 days). If this regimen is omitted, Ent. histolytica
cysts remain in the colonic lumen and, in the event of their being of a pathogenic
zymodeme, further tissue invasion (including liver abscess) might occur.
Spontaneous perforation of an amoebic liver abscess is a serious complication which
is associated with high morbidity and mortality rates;
this applies especially when perforation takes place into the pericardial cavity.
Successful percutaneous drainage (for 7–34 days) of a perforated abscess in five ‘severely
ill’ patients (with a total of 11 lesions) under metronidazole cover has been recorded;
there were resultant abscesses in the subhepatic space, pelvis, chest, right and left
paracolic gutters, lesser sac, retroperitoneum and flank, and associated fistulas
were demonstrated with the bile duct, duodenum and the colon; all healed completely.
No patient required a laparotomy. These authors recommend wider use of catheter drainage
for this serious complication of hepatic amoebiasis.
Pyogenic liver abscess
Although in a tropical context it is far less common than invasive amoebiasis (see
above), pyogenic abscess is a serious disease with high morbidity and mortality, even
when managed in experienced hands.
In most cases, a primary intra-abdominal focus of infection can be detected. Differentiation
from invasive hepatic amoebiasis is usually straightforward, the patient being more
severely and acutely ill; jaundice, septicaemia and renal impairment are common accompaniments.
Ultrasonography is usually diagnostic. In Kuala Lumpur, 25 pyogenic abscesses were
encountered between 1970 and 1985;
during the same period, there were 90 amoebic and one tuberculous abscesses, while
in 89 others the cause of the abscess was not discovered. At Kingston, Jamaica, fever
and abdominal pain were present in 21 (80%) out of 24 cases of pyogenic abscess encountered
between 1977 and 1986;
the most common signs were right upper quadrant tenderness and hepatomegaly; leukocytosis,
elevated alkaline phosphatase and hypoalbuminaemia were common. Reports from London
have given encouraging reports of management by needle aspiration under antibiotic
(usually gentamicin and metronidazole or clindamycin) cover. Another study has also
recorded satisfactory results in 18 of 21 patients using this form of percutaneous
drainage. Other authors have intimated, however, that this form of management should
be reserved for selected patients.
A report from Riyadh, Saudi Arabia, has provided results which were less encouraging.
In Jamaica surgical drainage using a guided percutaneous technique gave comparable
Taking all reports into account, it seems wise to perform a laparotomy and to institute
surgical drainage as soon as possible after diagnosis. Using ultrasonographic control,
a pyogenic abscess can be seen to ‘resolve’ significantly more rapidly than an amoebic
abscess. It should be appreciated, however, that this disease carries a significant
mortality rate; between 1975 and 1986, these authors treated 109 children with pyogenic
liver abscess; the mortality rate was 15%.
There is limited (suggestive) evidence that the overall prognosis is improving.
Hydatid disease and schistosomiasis involving the liver
Only rarely, usually following trauma, does hydatidosis207, 258, 259 present as an
abdominal emergency. Perforation into the peritoneal cavity may produce an anaphylactoid
reaction with hypotension, and/or seeding of daughter hydatid cysts within the peritoneal
cavity. A relatively high prevalence of alveolar echinococcosis has been recorded
Secondary bacterial infection is an unusual event. Chemotherapy is with albendazole
and/or praziquantel (Chapter 86).
is complicated by portal hypertension and oesophageal varices in an advanced case;
however, hepatocellular function is maintained late into the course of disease and
hepatic encephalopathy and ascites occur as advanced (usually terminal) signs. Praziquantel
is the chemotherapeutic agent of choice; evidence of reversal of fibrotic changes
is now available.
The two major diseases involving this organ encountered in tropical countries, and
which differ from those in temperate ones, are (1) ‘J-type’ diabetes, first reported
in Jamaica (Chapter 36) and (2) chronic calcific pancreatitis.
Diabetes, which is not associated with pancreatic calcification in young people, is
encountered throughout tropical countries; those affected are usually thin, and require
high doses of insulin; however, they do not rapidly develop ketosis when insulin is
discontinued. J-type diabetes might have a viral aetiology, a Coxsackie virus being
involved; a raised incidence of antibody to Coxsackie B4 has been demonstrated in
affected patients in India. A suggestion has been made that these patients, especially
those in Africa, are less susceptible to chronic diabetic complications than Europeans;
this now seems unlikely.
A popular Indian and Chinese vegetable, karela (Momordica charantia) possesses hypoglycaemic
properties; these are enhanced by chlorpropamide, a fact that should be taken into
account in the management of diabetes in a number of Asian countries.
A syndrome consisting of pancreatic calcification associated with both exocrine and
endocrine impairment is common in many tropical countries (Figure 10.11
);1, 262, 263 most observations have been made in Africa (East and West), southern
India and Indonesia. The aetiology of chronic calcific pancreatitis remains unknown.
Pancreatic disruption in childhood kwashiorkor can be severe and might be relevant.
Cassava (Manihot esculenta) has also been implicated. Long-standing pancreatic damage
can also follow viral hepatitis. A further hypothesis is that pancreatic ducts blocked
by secretions and inspissated mucous plugs later calcify; this might be more common
after starvation, gastroenteritis and dehydration. Presentation is with weight loss
and malabsorption (in some parts of Africa, this is the most common cause of overt
malabsorption); diabetes mellitus and pancreatic pain are important features. Management
consists of providing pancreatic supplements (e.g. pancreatin BP, 6 g orally with
meals) together with diabetic control.
Pain is often difficult to manage and may be so severe that suicide is a sequel.
Abdominal radiograph showing calcified pancreas in the chronic calcific pancreatitis
syndrome. There was no history of alcohol excess or infant malnutrition; aetiology
was therefore undetermined.
The pancreas can also be involved in many infections including Schist. mansoni and
Schist. japonicum, trichinellosis, cysticercosis and hydatid disease.
Pancreatic duct obstruction, complicated by acute pancreatitis, is most commonly a
sequel to A. lumbricoides infection (see below); tapeworms are rarely implicated.
Clonorchiasis and opisthorchiasis may involve the pancreatic duct system.
Emergencies: pancreas, and biliary system
One of the most widely distributed nematodes in tropical and subtropical countries
is A. lumbricoides. By entering the biliary system (from the duodenum) this parasite
can cause several acute medical and surgical conditions. Reporting from Kashmir, India,
Khuroo et al.
collected 500 cases in which A. lumbricoides involved the liver, biliary tract and
pancreas; biliary ascariasis was present in 171 cases, and in 140 there was hepatic,
in eight gallbladder and in seven pancreatic involvement. These authors recognized
five clinical presentations: acute cholecystitis (64), acute cholangitis (121), biliary
colic (280), acute pancreatitis (31) and hepatic abscess (4). Twenty-seven had a pyogenic
cholangitis, which was treated by decompression and drainage, surgically in two and
endoscopically in 25; removal of adult worms from the ampullary orifice (with extraction
per os) led to rapid relief of biliary colic in 214, and acute pancreatitis in 16;
four patients died, from acute pancreatitis (2), pyogenic cholangitis (1) and hepatic
abscess (1). Worms persisted at 3 weeks in the biliary tree in 12 patients; dead worms
were removed either by surgery (5) or by using an endoscopic basket (7). A. lumbricoides
moved out of the ductal system in 211 cases. The patients were followed-up for a mean
of 48 months; 76 became re-infected and had re-invasion of the biliary tree; in seven
cases intrahepatic duct and bile duct calculi (superimposed on dead worms) were present.
In South-east Asia, the two most common biliary parasites are Clonorchis sinensis
and Opisthorchis spp. Although these cause chronic problems, notably secondary bacterial
and adenocarcinoma of the biliary system, an acute presentation
In most indigenous people of developing countries, gallstones are unusual; when they
occur they are usually of the pigment variety, and often associated with haemolysis.
A report from Saudi Arabia, where the average lifestyle has rapidly become westernized
(with striking changes in diet) over the last few decades, indicates that cholecystectomy
for cholelithiasis is now one of the most common major abdominal operations to be
between 1977 and 1986, for example, 2854 individuals (most of them young Saudis) underwent
this operation at 14 hospitals in the Eastern Province of the country.
Acute pancreatitis is uncommon overall in developing countries, although severe abdominal
pain caused by chronic calcific pancreatitis
can give rise to problems in differential diagnosis. The pain may be severe. Biliary
involvement by A. lumbricoides can result in acute pancreatitis.1, 142 Other helminths,
including Clonorchis sinensis, Opisthorchis and Anisakis species have also been associated
with this condition.
summarizes some causes of splenomegaly in the tropics.
Most of these receive attention in other chapters. The most extreme form of splenomegaly
(HMS) (Figure 10.12
) is covered in Chapters 13 and 72; those caused by various viral, bacterial and parasitic
infections are dealt with under these respective headings.
Some causes of splenomegaly in the tropics
Epstein-Barr virus, cytomegalovirus, viral hepatitis and other virus diseases
typhoid fever, brucellosis, tuberculosis
malaria (especially hyper-reactive malarious splenomegaly (HMS)), schistosomiasis,
visceral leishmaniasis, African trypanosomiasis
Sickle cell disease, thalassaemia
Burkitt's lymphoma, leukaemia, reticuloses
Amoebic; unknown aetiology
Spontaneous haemorrhage and rupture
Papua New Guinea man suffering from hyperreactive malarious splenomegaly (HMS); all
of the features of this syndrome were present. (B) Liver biopsy specimen showing severe
sinusoidal lymphocytosis, a component of the syndrome.
The spleen is an extremely important line of defence against many infections, especially
pneumococcal and Plasmodium species infections. Splenectomized individuals in tropical
countries should receive pneumococcal vaccine; prudent advice regarding malaria prophylaxis
Splenic abscess is a well-documented tropical disease.
Aetiology is usually unknown; underlying viral and parasitic diseases have been suggested,
but not proved. A connection with carriage of the sickle cell gene has also been suggested,
but this has also not been proved. Most reports have been made in West Africa and
Zimbabwe. In most, the aetiology is unknown, but some undoubtedly result from a S.
typhi infection. The clinical history is usually one of 2–3 weeks duration, and consists
of pain/swelling in the left hypochondrium, associated with pyrexia. The splenic swelling
is tender, often exquisitely so, and fluctuant. A radiograph may show gas within the
abscess. Untreated, the abscess can rupture into the peritoneal cavity; splenectomy
therefore has an important role in management. Should the condition become chronic
– an unusual event – splenectomy is also the correct course of management.