45
views
0
recommends
+1 Recommend
0 collections
    0
    shares
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      Raw Pig Blood Consumption and Potential Risk for Streptococcus suis Infection, Vietnam

      brief-report

      Read this article at

      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Abstract

          We assessed consumption of raw pig blood, which is a risk factor for Streptococcus suis infection in Vietnam, by using a mix-method design. Factors associated with consumption included rural residency, age, sex, occupation, income, and marital status. We identified risk groups and practices and perceptions that should be targeted by communication programs.

          Related collections

          Most cited references14

          • Record: found
          • Abstract: found
          • Article: not found

          Streptococcus suis: an emerging human pathogen.

          Streptococcus suis infection is acquired through exposure to contaminated pigs or pig meat. Over the past few years, the number of reported S. suis infections in humans has increased significantly, with most cases originating in Southeast Asia, where there is a high density of pigs. Increased awareness, improved diagnostics, and the occurrence of outbreaks have contributed to this increase. Meningitis and sepsis are the most common clinical manifestations of S. suis infection; hearing loss is a frequent complication. In this article, we provide an overview of the emergence and clinical manifestations of S. suis infection.
            Bookmark
            • Record: found
            • Abstract: found
            • Article: not found

            Human Streptococcus suis Outbreak, Sichuan, China

            Streptococcus suis is a zoonotic microbe that can exist in pigs without causing illness but can occasionally cause disease. Serotype 2 is a dominant pathogenic serotype ( 1 ). Types 2 and 5 have been isolated from purulent lesions in the lungs and other extramammary sites in cattle, sheep, and goats ( 2 ). Infection may cause death in weaning piglets as well as growing pigs ( 3 ). The bacterium is isolated from an increasingly wide range of mammalian species, including horses, dogs, cats, and birds ( 4 ). Sporadic cases of S. suis infection may occur in humans, the most common clinical manifestations included purulent meningitis, septicemia, arthritis, and endocarditis; some infections lead to sequelae such as deafness and ataxia ( 5 – 7 ). To date, ≈200 human cases have been reported in areas of intensive pig rearing (the Netherlands, Denmark) or areas where large quantities of pork are eaten (Hong Kong, Thailand, Vietnam) ( 8 ). Most reported cases of human S. suis infections were associated with contact with pigs or pork products ( 5 ). Human S. suis infections do not normally cause major outbreaks. On July 11, 2005, a local hospital in Ziyang Prefecture of Sichuan Province reported a suspected case of hemorrhagic fever with renal syndrome. The patient was a 46-year-old male farmer with acute onset of high fever, lethargy, vomiting, and generalized purpura. The day before illness onset, the farmer slaughtered a pig that had died of an unknown cause. The farmer rapidly lapsed into coma. On further investigation, we identified 4 other patients with similar circumstances in the same hospital and more patients from other hospitals in the area. S. suis was isolated from blood cultures in some of these cases. We began an investigation of this outbreak to describe its epidemiologic, clinical, and microbiologic characteristics ( 9 ). Methods Epidemiologic Investigation We reviewed medical records in all health care facilities in Ziyang Prefecture for patients admitted since June 10, 2005 (2 weeks before the onset of the first known case), with a diagnosis of septic shock or meningitis. On July 19, enhanced surveillance was introduced to include healthcare facilities in Ziyang and the 5 surrounding prefectures. We ordered all healthcare facilities to immediately report all new patients with clinical sepsis, meningitis, arthritis, or endocarditis and fever >37.3°C, who had epidemiologic risk factors (contact with sick pigs or any part, such as meat, skin, organs, or tissue, of pigs that had died of undetermined causes). Beginning on July 21, we made public announcements to encourage reporting. Public health workers interviewed patients, or surrogates for deceased patients, by using a questionnaire to collect demographic, clinical, and exposure information. Specimens (blood, cerebrospinal fluid [CSF], or postmortem tissue) were collected for laboratory investigation. We reviewed medical records to obtain supplementary clinical information. We placed all close contacts of case-patients, including family members and attending healthcare workers, under medical surveillance. We extended these surveillance procedures to all of Sichuan Province on July 25. A probable case of S. suis infection was a compatible clinical illness (sepsis, meningitis, arthritis, or endocarditis), without laboratory evidence of infection by another organism, with history of contact with sick or dead domestic livestock (pigs, goats, or sheep) or another case-patient within 7 days before onset of symptoms. A confirmed case was defined as a compatible clinical illness regardless of exposure and the verification of S. suis isolated from a normally sterile site. We stopped the enhanced surveillance on August 18, two weeks after onset of the last case. To assess the extent of underreporting, notification statistics on meningococcal meningitis were reviewed. Meningococcal meningitis is a statutorily notifiable disease in China. Because laboratory diagnosis was not routinely performed for all meningitis patients, the figures on suspected meningococcal cases could reflect nonmeningococcal meningitis caused by other bacteria, including S. suis. In this investigation, statistics in the 12 affected prefectures of Sichuan Province from January 2003 to August 2005 were reviewed. Laboratory Investigation Specimens of blood, CSF, or postmortem tissue from human patients and blood or postmortem tissue from affected pigs were injected onto sheep blood agar and infusion broth (REF 237500, Oxoid, Basingstoke, UK) in 0.5% CO2 at 37°C. S. suis strains were identified by examining growth colony shape, followed by biochemical reactions with Vitek2 compact and API 20 strep (bioMérieux, Inc., Beijing, China) according to the manufacturer's instructions. Polymerase chain reaction (PCR) was used to characterize selected genes of S. suis serotype 2. PCR products were purified by using the QIAquick PCR purification kit (Qiagen, Valencia, CA, USA), according to the manufacturer's protocol, and sequenced with an ABI Prism 3700 DNA instrument (Applied BioSystems, Foster City, CA, USA). The following genes were sequenced, in accordance with methods published elsewhere: 1) genus-specific gene segments, tuf sequence, and 2) the species-specific gene coding for 16S rRNA of S. suis, the gene coding for the capsule of S. suis serotypes 2 (cps2J) and 1/2, the muramidase-released protein gene (mrp), suilysin (sly), and the extracellular factor gene (ef) ( 10 – 12 ). Sequence data were analyzed by using a basic local alignment search tool (BLAST) search performed against sequences published by the National Center for Biotechnology Information (Bethesda, MD, USA). Automated PvuII and PstI ribotyping was performed by using the RiboPrinter microbial characterization system (DuPont China, Shenzhen, China), with bacterial isolates grown overnight on brain-heart infusion blood agar. Template preparation, restriction enzyme digestion, gel electrophoresis, and Southern hybridization with an Escherichia coli rrnB rRNA operon probe were carried out with the RiboPrinter system. Images were developed with a charge-coupled-device camera and analyzed by using the RiboPrinter's customized software. Statistical Analysis Percentage, proportion, and case-fatality ratios were calculated. The χ2 and rank sum tests were used to compare the case-fatality ratios and the median duration from exposure to onset between 2 groups, respectively, by using Stata version 8 (StataCorp LP, College Station, TX, USA). All probabilities were 2-tailed, and p 2 cases of S. suis infection occurred in patients who did not know each other and had no common exposure to a pig. Figure 3 Geographic distribution of Streptococcus suis cases in Sichuan Province, China, relating to (A) all reports, and (B) 66 laboratory-confirmed cases alone (as of August 18, 2005). Clinical Patterns All 215 affected persons were previously healthy adults. We observed 3 distinct clinical symptom settings (Table 2). First, 28% of patients had sepsis characterized by acute onset of fever, chills, headaches, dizziness, malaise, abdominal pain, and diarrhea. In some severe cases, patients became comatose. Second, 48% of patients had meningitis characterized by headache, stiff neck, and other signs of acute meningitis. Some meningitis patients also had disseminated intravascular coagulation and coma. Third, 28% of patients had STSS that has been described in other forms of streptococcal infections ( 13 ) and met the criteria established by Centers for Disease Control and Prevention ( 14 ). Patients with STSS had a 62% case-fatality ratio compared to 0.6% for other clinical forms of streptococcal infections. Fatal STSS cases progressed from onset to death in a median of 25 hours (range 8 hours to 10.5 days). Discharge records available on 85 survivors in all 3 clinical symptom settings showed a median interval from onset to recovery of 15 days (range 5–36 days). Table 2 Clinical symptoms and case-fatality ratio of 215 Streptococcus suis patients in Sichuan Province, China, 2005* Clinical symptom Probable Confirmed Total No. (%) No. deaths (case-fatality %) No. (%) No. deaths (case-fatality %) No. (%) No. deaths (case-fatality %) Sepsis 45 (30) 0 7 (11) 0 52 (24) 0 Meningitis 69 (46) 1 (1) 33 (50) 0 102 (48) 1 (1) STSS 35 (24) 23 (66) 26 (39) 15 (58) 61 (28) 38 (62) Total 149 (100) 24 (16) 66 (100) 15 (23) 215 (100) 39 (18) *Streptococcal toxic shock syndrome (STSS) is defined according to the 1996 criteria established by the Centers for Disease Control and Prevention, Atlanta, GA, USA, which include hypotension (systolic blood pressure 2 of the following: renal impairment, coagulopathy, liver involvement, acute respiratory distress syndrome, generalized erythematous macular rash that may desquamate, soft-tissue necrosis, including necrotizing fasciitis or myositis, or gangrene. Difference in case-fatality ratio between STSS and other clinical symptoms p 37.3°C) 61 (100) 154 (100) 215 (100) Chills 48 (79) 128 (83) 176 (82) Headache 38 (62) 110 (71) 148 (69) Myalgia 30 (49) 73 (47) 103 (48) Vomiting 41 (67) 80 (52) 121 (56) Abdominal pain 24 (39) 33 (21) 57 (27) Diarrhea 28 (46) 22 (14) 50 (23) Coma 16 (26) 26 (17) 42 (20) Petechiae, ecchymosis† 37 (61) 12 (8) 49 (23) Neck rigidity 4 (7) 50 (32) 54 (25) Kernig positive 1 (2) 27 (18) 28 (13) Brudzinski positive 2 (3) 17 (11) 19 (9) Hypotension (blood pressure 10 × 109/L, no. (%)† 26 (52) 97 (71) 123 (66) Mean leukocyte count (range) 12.3 (1.0–47.8) 14.5 (2.0–64.0) 13.9 (1.0–64.0) Thrombocytopenia, 2× upper limit of normal.
§Creatinine >177 μmol/L for adults or >2× upper limit of normal for age.
¶CSF, cerebrospinal fluid, protein >0.45 g/L, glucose 8 leukocytes × 106/L. Postmortem examination of 4 STSS patients (2 confirmed and 2 probable infections) showed features of disseminated intravascular coagulation. Evidence of multiple organ damage was observed, primarily involving kidneys, adrenal glands, lungs, liver, pancreas, and heart. Histologic findings included microthrombosis (hyaline thrombus) in organ capillaries; necrosis of parenchymal cells; and congestion, exudate, and hemorrhage of interstitial vessels of kidneys, lungs, and other organs. Microbiologic Investigations We collected 348 specimens of blood (271), CSF (53), and tissue (24) from postmortem examination of 172 case-patients. We isolated S. suis from blood (36), CSF (27), and postmortem liver, spleen, and heart tissues (3). Fifty-five (83%) of the confirmed patients were diagnosed after July 23, 2005, representing 42% (55/130) (data not shown) of all tested samples, compared to that of 26% (11/42) (data not shown) before enhancement of surveillance. Isolates from 66 patients and 3 diseased pigs featured pure growth of tiny α-hemolytic colonies on sheep blood agar. The colony shape and biochemical reactions by Vitek2 compact and API-Strep were all compatible with that of S. suis. In the investigations after July 23, 2005, the results matched the suggested key indicators for the pathogen, including Voges-Proskauer negativity, hydrolysis of esculin, trehalose positivity, negativity for growth in 6.5% NaCl, and absence of β-hemolysis on sheep blood agar. PCR on all isolates showed gene coding for tuf, species-specific 16S rRNA of S. suis, genes coding for the capsule of S. suis serotypes 2 (cps2J) and 1/2, and mrp, sly, and ef. The PCR products of virulence genes of all 69 strains were sequenced. These were identical to the sequences published by the National Center for Biotechnology Information. On further investigation, only a single ribotype for either PvuII or PstI restriction was identified among the 25 S. suis type 2 isolates, including 22 from patients and 3 from diseased pigs (Figure 5), indicating a single clonal strain as the source of the infection ( 1 ). Figure 5 Ribotyping of Streptococcus suis serotype 2 isolates by PvuII restriction. Lane 1, deceased pig isolate SC5; lane 2, deceased pig isolate SC16; lane 3, patient isolate SC154; lane 4, patient isolate SC160; lane 5, patient isolate SC175; lane 6, patient isolate SC179; lane 7, patient isolate SC204; lane 8, patient isolate SC206; M, molecular size standard. Discussion We report an unprecedented outbreak of human S. suis serotype 2 infection involving 215 patients in Sichuan Province during the summer of 2005. This outbreak was characterized by the large number of patients involved, distinctive clinical manifestations, and the major challenges facing public health authorities in both surveillance and control. This is the largest recorded outbreak of S. suis infection in humans. From late July to early August 1998, an outbreak of 25 cases with 14 deaths occurred in Jiangsu Province, China, from serotype 2 infection ( 15 ). The size of the outbreak in our study was probably related to a local farming practice. In Sichuan, a sizable swine population was found in small backyard farms; each family kept only a few animals. Farmers habitually slaughtered sick pigs for human consumption. Beginning mid-June 2005, a major outbreak of S. suis killed 647 pigs in almost the same areas as the outbreak in humans ( 16 ). The outbreak in swine peaked around July 20 with ≈4 dead pigs in each affected village. S. suis caused 98% of the deaths of these pigs in Sichuan during this period ( 17 ). All infected pigs came from backyard farms, usually with 1 sick pig in the herd. A pathogenic strain could have spread by distributing infected piglets to the backyard pig farms and then propagated among healthy pigs. The single ribotype identified in the investigation lends support to this theory. The main risk factor for S. suis infection in the outbreak was direct involvement in slaughtering sick pigs and preparing carcasses of pigs that died of unknown causes. Unlike professionals in modern abattoirs, the local farmers did not wear protective gear or gloves. Normally 1–2 persons carried out the procedure, which involved bloodletting though a neck artery, manually inflating the carcasses, scalding the pigskin with ≈80°C water, and splitting and shaving the skin with large knives. Scalding and shaving were often performed together. The farmers then sliced the meat into smaller pieces before cooking for food. The complete process of slaughtering could take >1 hour. Our study demonstrated that all patients had been infected during direct contact with blood or tissues of sick or dead pigs. Often this may have occurred through direct exposure of skin wounds. Droplet exposure may also have occurred during slaughter or processing of carcasses, but we could not document this occurrence. The observed risk factors were consistent with those reported in other studies ( 5 , 6 ). No evidence of infection from eating cooked pork from these pigs was observed. The uncooked meat was shared with neighboring families, but these villagers normally do not eat raw meat or raw animal viscera. Person-to-person transmission was highly unlikely since we found no disease in family members, neighbors, or healthcare workers who had not been exposed to sick or dead pigs. One missing link in the outbreak is the exact relationship between the dead goat in the early cluster of patients and the subsequent propagation of the S. suis infections. We could not confirm these 2 cases microbiologically. We speculate that S. suis caused these 2 early human infections because they occurred shortly after exposure to the dead goat, and because clinical manifestations in the 2 patients were similar to those of others in the outbreak. Human S. suis infection after exposure to sick goats has not been reported, despite isolation of the organism from these animals ( 2 ). In backyard farms where different animals are kept together, S. suis infection could have been transmitted between pigs and goats. Animal surveillance would help establish the role of animals other than pigs in carriage of the bacteria and the potential for causing human infections. Clinically, 3 distinctive forms of human S. suis infection occurred, namely, STSS, sepsis, and meningitis. STSS has not been reported from S. suis infection, although it has been previously described in other streptococcal infections ( 13 ) and Staphylococcus aureus infections ( 18 ). Unusual STSS-like illnesses brought the outbreak to the attention of local health authorities. A dose effect may explain the relatively high proportion of STSS in this outbreak. While other explanations like comorbid conditions, e.g., asplenia, diabetes mellitus, alcoholism, and malignancy ( 19 ), have been reported, this was not the case for the outbreak in our study, which involved previously healthy adults. Laboratory examination confirmed virulence factors in the S. suis isolates from this outbreak. They include mrp, sly, and ef, although their precise clinical role has not yet been shown. These isolates (mrp+, ef+, sly+) are related to European strains that are considered to be more virulent than North American strains ( 1 ). Genome analysis could determine if novel virulence genes were involved. Finally, the high number of deaths due to STSS is a cause for concern. Prompt institution of effective responses to human S. suis outbreaks is a public health challenge, especially in rural China. Timely diagnosis is difficult. In our study, only 31% of the cases were laboratory confirmed. Suboptimal access to health services, personal delay in seeking treatment, underutilization of blood cultures in local hospitals, and self-administration of antimicrobial drugs may explain the relatively low proportion of culture-positive human cases, especially during the early phase of the outbreak. Many patients died without having sought treatment from any health facility. As for the investigation process, the existing surveillance system that covers meningitis is not robust enough to alert public health personnel of the impending threat of S. suis. Control measures included prohibiting by law of slaughtering, eating, selling, and transporting deceased or sick pigs. Subsidies were offered to families to support hygienic handling of deceased or sick pigs and to patients for medical care. Village heads were held accountable for illegal slaughtering in their village. These measures were supplemented with disinfection of affected backyard farms. Public education campaigns were staged to increase awareness of how to prevent and control human S. suis infections. In the long run, the prevention and control of swine infection should form the more strategic component of the public health program. Surveillance systems should be established to alert farmers and the general public if an infection outbreak in pigs is recognized ( 19 ).
              Bookmark
              • Record: found
              • Abstract: found
              • Article: found
              Is Open Access

              Epidemiology, Clinical Manifestations, and Outcomes of Streptococcus suis Infection in Humans

              Streptococcus suis is a neglected zoonotic pathogen that has caused large outbreaks of sepsis in China ( 1 , 2 ) and has been identified as the most common and the third leading cause of bacterial meningitis in adults in Vietnam and Hong Kong, respectively ( 3 – 5 ). S. suis infection is acquired from pigs, either during slaughtering or by handling and eating undercooked pork products. It is potentially preventable ( 3 , 6 ). Epidemiology of the infection differs between Western and Asian regions ( 7 ), and the role of high-risk eating habits (i.e., ingesting raw or undercooked pig parts, including pig blood, organs, and meat) in some Asian communities recently has been recognized ( 6 , 8 , 9 ). Rates of S. suis infection are low in the general populations of Europe and North America, and cases are concentrated among occupationally exposed groups, including abattoir workers, butchers, and pig breeders ( 10 , 11 ). In a 2009 review, ≈700 S. suis infections were reported worldwide by 2009, mostly from China and Vietnam ( 12 ). Clinical characteristics of this infection have been reviewed ( 12 , 13 ) and include meningitis, sepsis, endocarditis, arthritis, hearing loss, and skin lesions. Treatment of S. suis infection requires ≈2 weeks of intravenous antimicrobial drugs ( 12 ). The death rate varies, and deafness is a common sequela in survivors. Although substantial new data on the incidence, clinical and microbiological characteristics, and risk factors for S. suis infection have accumulated during recent years, the prevalence of this infection has not measurably decreased. We conducted a systematic review and meta-analysis to update the evidence and summarize the estimates of epidemiologic and clinical parameters to support practitioners’ and policy makers’ efforts to prevent and control this infection. Methods We conducted the review in accordance with recommendations of the PRISMA statement ( 14 ). The protocol for this review has been registered at PROSPERO International prospective register of systematic reviews (no. CRD42011001742). Search Strategy and Selection Criteria We systematically and inclusively searched 5 main electronic databases (PubMed, Scopus, ISI Web of Science, Science Direct, and Google Scholar) for all studies published until the end of December 2012 (Figure 1). The following search term was used as a text word in each database, as follows: PubMed—“streptococcus suis” in all fields, limited to humans; Scopus—“streptococcus suis” in all fields, excluding veterinary medicine articles; ISI Web of Science—“streptococcus suis” in topic with exclusion of veterinary science areas; Science Direct—“streptococcus suis” in all fields, with articles in veterinary medicine journals excluded; and Google Scholar—“allintitle: ‘streptococcus suis’” Figure 1 Flow diagram of the search and review process for this review of Streptococcus suis infection. We also searched using the same search term “streptococcus suis” in other databases, including Virtual Health Library, SIGLE, WHOLIS, LILACS, IMSEAR-HELLIS, and China Academic Journals Full-text Database and checked the reference lists of retrieved articles. We did not restrict the types of studies and publication languages, and non-English papers were translated for review. Publications were excluded if they did not report any human cases of S. suis infection, reported data that overlapped with already included articles and provided no additional information, reported cases without information indicating the location of the patients, or reported data that could not be reliably extracted. Data Extraction Two reviewers (N.H. and V.T.L.H.) independently screened the titles and abstracts, and examined the full-text publications by using identical selection criteria and data abstraction forms. The results of data extraction showed a high degree of agreement between the reviewers (κ>0.90 for the main variables). Any disagreements were resolved by discussion and consensus between the reviewers and other authors (N.T. Huy, H.W., P.H., K.H.). We emailed the original authors of the articles that contained ambiguous data (1 email attempt per author) for clarification, and the ambiguous data were excluded from analyses if we did not receive a response. Data extracted included year of publication, year of data collection, study design, data collection approach, country of origin, hospital where the patients were recruited, patient characteristics, clinical manifestations, methods of diagnosis, clinical and laboratory parameters, outcomes, and histories. Analyses We described the relevant epidemiologic and clinical factors using count for number of cases, proportions with 95% CIs for categorical factors (sex, occupation, exposure, history), and mean with SD for continuous factors (age, duration, and laboratory parameters). Event rates are presented as proportions with 95% CIs for signs, symptoms, and outcomes. We defined an event rate as the ratio of number of events to the number of all patients assessed in each study. We pooled all single cases from the publications that reported 5 cases). We report the values of reviewed factors in 3 sets: summary values from the single-case dataset, median values (range) of the large studies, and pooled values from the meta-analysis as appropriate. Meta-analysis was conducted by using Comprehensive Meta-analysis software version 2 (Biostat, Englewood, NJ, USA; http://www.Meta-Analysis.com) when >2 studies reported the reviewed factor. We tested heterogeneity using the Q statistic and I2 test ( 15 ). Pooled values and 95% CIs were generated from a fixed-effects model or from a random-effects model, and each was study weighted by the inverse of that study’s variance. We used the fix-effects model when heterogeneity was not significant and a random-effects model when heterogeneity was evident ( 16 ). Median (range) was converted to mean (SD) by using proposed formulas ( 17 ), and interquartile ranges to SDs and subgroup values to total values by Cochrane suggested methods ( 18 ). We assessed publication bias using funnel plots and the Egger’s regression test (if >10 studies were included in the meta-analysis). If publication bias was found, the Duvall and Tweedie trim and fill method was performed to add possible missing studies to improve the symmetry and calculate the adjusted pooled values ( 19 ). We used subgroup analyses (when >10 studies were included) and bivariate meta-regression (when >20 studies were included) to examine the effect of study-level variables, including year of publication (2005 and earlier vs. after 2005 [because the Sichuan outbreak occurred in 2005]), study design (case series, outbreak investigation, cross-sectional), location (China mainland, Hong Kong, Thailand, Vietnam, and others), data collection (retrospective vs. prospective) and meningitis rate ( 90%) on the main outcomes. The general linear model was used for meta-regression, with adjustment for multiple comparisons by using the Bonferroni method and weighting by study sample size. Results Systematic Review We used 177 publications that met inclusion and exclusion criteria for data extraction and final analyses (Figure 1; Technical Appendix Table 1). The studies were diverse in terms of design, data collection, and reporting approaches. We identified 20 case series (8 from South-East Asia region, 8 from the Western Pacific region, and 4 from Europe) and 21 cross-sectional studies (9 SouthEast Asia, 8 Western Pacific, and 4 Europe). Five articles about 3 outbreaks (in Sichuan and Jiangsu, China; and Phayao, Thailand) were classified as outbreak investigation reports. The only prospective case–control study was conducted in Vietnam (Table 1). Table 1 Characteristics of 177 articles in a systematic review of Streptococcus suis infection Characteristic Articles, no. (%) Cases reported, no. (%)* Geographic region† Europe 98 (55) 168 (11) Western Pacific 47 (27) 836 (53) SouthEast Asia 24 (14) 572 (36) Americas 8 (5) 8 (0.5) Type of study design Case report 130 (73) 151 (7) Case series 20 (11) 511 (25) Cross-sectional 21 (12) 761 (37) Outbreak investigation 5 (3)‡ 532 (26) Case–control 1 (1) 101 (5) Data collection approach Retrospective 159 (90) 1299 (63) Prospective 15 (9) 697 (34) Both§ 3 (1) 60 (3) Language of publication¶ English 130 (74) 1947 (95) Spanish 13 (7) 15 (1) French 12 (7) 13 (1) Other# 22 (12) 81 (4) Year of publication 1968–1980 13 (7.5) 18 (1) 1981–1990 27 (15) 95 (5) 1991–2000 32 (18) 119 (6) 2001–2005 28 (16) 115 (6) 2006–2010 55 (31) 1052 (51) 2011–2012 22 (12.5) 659 (32) *Case duplicates were removed in the counts for the geographic region subheading (totaling 1,584 cases, no duplicates). Duplicates were not removed in the counts for other subheadings (totaling 2,056 cases, with duplicates).
†Geographic regions as defined by the World Health Organization.
‡Includes 3 articles reporting about the patients in the Sichuan outbreak in China; each was included for analysis of different factors.
§Included in the prospective groups in subsequent analyses.
¶Almost all large studies were published in English. Most reports in languages other than English were case reports.
#German (7 articles); Dutch (4); Czech, Italian, and Japanese (2 each); Chinese, Polish, Serbian, Swedish, and Thai (1 each). Epidemiology By the end of 2012, a total of 1,584 cases had been reported in the literature (including 189 probable cases identified in 3 outbreaks), mainly from Thailand (36%), Vietnam (30%), and China (22%). More than half (53%) were in the Western Pacific region; 36% were in the South East Asia region, 10.5% in the European region, and 0.5% in the Americas. The highest cumulative prevalence rate was in Thailand (8.21 cases/million population), followed by Vietnam (5.40) and the Netherlands (2.52) (country population data for 2008–2012 by World Bank [ 20 ]) (Figure 2). Figure 2 Global cumulative prevalence of Streptococcus suis infection through 2012. The pooled mean age of the patients was 51.4 years, and 76.6% were men (Table 2). All case-patients were adults, except 1 female infant reported in Thailand ( 21 ). The pooled proportion of case-patients with occupational exposure was 38.1% (95% CI 24.4%–53.9%); this proportion was higher for industrialized countries than for other countries (83.8% [95% CI 73.4%–90.7%] for the United Kingdom, Netherlands, and Japan together). Recent contact with pigs or pork was reported for 15.5% of single cases but for 33.9% (95% CI 21.1%–49.5%) in the meta-analysis. History of eating meals containing pork was reported mainly in Asia (Thailand and Vietnam); the pooled estimate was 37.3% (95% CI 20.2%–58.3%). For Thailand only, the proportion was 55.8% (95% CI 33.7%–75.9%). In other countries, only 1 patient in France was reported eating artisanal dry sausage ( 22 ), and 1 patient in the United States ate raw pork while traveling in the Philippines ( 23 ) before the infection. Table 2 Epidemiologic factors of patients with Streptococcus suis infection included in a systematic review Variable Single-case dataset, %* Large studies, median
(range), % Meta-analysis, pooled mean (95% CI), %† No. studies meta-analyzed, %‡ Mean age, y, n = 156 49.4 50.5 (37.0–61.2) 51.4 (49.5–53.2) 25 Male sex, n = 155 83.2 77.5 (37.5–100) 76.6 (72.2–80.6) 26 Pig-related occupation 58.6 25.0 (0–100) 38.1 (24.4–53.9) 21 Contact with pig/pork 15.5 33.3 (2.4–100) 33.9 (21.1–49.5) 14 Eating of high-risk food 4.0 53.3 (5.9–88.7) 37.3 (20.2–58.3) 9 Skin injury 19.5 16.0 (9.5–71.4) 25.1 (15.1–38.7) 8 Drinking of alcohol 8.6 23.0 (4.8–83.9) 29.7 (17.2–46.3) 13 Concurrent diabetes§ 2.9 7.2 (3.2–25.0) 8.0 (4.6–13.7) 9 *N = 174 unless otherwise indicated.
†Random-effects model unless otherwise specified.
‡Include the single-case dataset and the large studies (Technical Appendix Table 2).
§Other less common underlying conditions are listed in Technical Appendix Table 3. Skin injury was shown for one fourth of case-patients, and alcohol use was evident in approximately one third of case-patients. However, a case–control study in Vietnam did not identify alcohol use as an independent risk factor after adjustment for other risk factors and confounders ( 6 ). The most commonly reported preexisting condition was diabetes. Other conditions included underlying heart disease, hypertension, cirrhosis, and cancer (Technical Appendix Table 3). Smoking was mentioned in 5.2% of patients in the single-case dataset. Microbiological Diagnosis Blood and/or cerebrospinal fluid culture were the most common reported diagnostic methods among the case reports (Technical Appendix Table 4). Molecular diagnosis was more common in the large studies (11 studies) than in case reports. The most prevalent strain was serotype 2 (86.5%), followed by serotype 14 (2.3%), and serotype 1 (0.6%) of all 1,156 patients with serotype information mentioned in the articles. Serotypes 4, 5, 16, and 24 also were reported (1 patient per serotype). Misdiagnosis of S. suis infection was not uncommon, either by conventional biochemical tests or commercial identification systems. The bacteria were often reported as viridans streptococci in initial cultures. In fact, up to 70% of all viridans streptococci cases in Thailand were confirmed as S. suis infections in the follow-up investigations ( 24 ). A total of 62.5% of S. suis–infected patients in another study in Thailand ( 25 ) and 20% in a study inthe Netherlands ( 10 ) were initially reported to be infected with viridans streptococci. Misidentification of the infectious agent as S. bovis (2 patients), S. pneumoniae (1 patient), and S. faecalis (1 patient) also was reported in the Netherlands series. Tsai et al. ( 26 ) showed large variations between the 2 commercial systems (Phoenix Identification System, Beckon Dickinson, Sparks, MD, USA; and Vitek II GPI Card, bioMérieux Vitek, Hazelwood, MO, USA), and misidentification of S. suis as S. acidominimus was common when the Phoenix system was used. S. suis is mostly sensitive to penicillin; resistance was reported in only 2 patients ( 21 , 27 ). After cessation of antimicrobial drug treatment, infection relapsed in a small proportion of patients, including those for whom the organism tested highly sensitive to penicillin ( 28 , 29 ). The pooled relapse rate was 4.4% (Table 3). Table 3 Main clinical and laboratory parameters at admission of the patients with Streptococcussuis infection in a systematic review Variable Single-case dataset Large studies, median value (range) Meta-analysis, pooled mean (95% CI)† No. studies meta-analyzed‡ Clinical syndrome, %§ Meningitis 69.5 64.5 (30.2–100) 68.0 (58.9–75.8) 26 Sepsis¶ 19.5 23.8 (11.8–39.4) 25.0 (20.5–30.2) 12 Arthritis 2.87 16.7 (1.5–50.0) 12.9 (6.0–25.6) 12 Endocarditis 8.6 14.3 (1.9–39.0) 12.4 (6.7–21.9) 10 Endophthamiltis 2.9 4.5(1.5–28.6) 4.6 (2.8–7.4)# 9 Spondylodiscitis 4.6 1.9 (1.5-2.4) 3.7 (2.1-6.6) 4 Toxic shock syndrome 2.9** 37.7 (28.9–64.0) 25.7 (9.8-52.6) 4†† Mean duration, d Onset to admission, n = 90 7.3 3.5 (2.0–11.4) 4.1 (2.7–5.4) 7 Hospitalization, n = 68 20.5 17.4 (13.0–19.2) 17.2 (15.6–18.9)# 5 Symptoms, % Meningeal sign‡‡ 49.4 66.7 (12.5–95.1) 67.1 (54.9–77.4) 18 Skin rash 10.9 12.5 (0–52.0) 15.4 (8.6–25.9) 10 Shock 8.6 11.8 (1.3–64.0) 11.9 (6.3–21.5) 12 Respiratory failure 5.2 20.0 (8.3–35.8) 16.7 (8.6–29.9) 6 Acute renal failure 5.2 8.3 (1.3–28.0) 7.1 (2.2–20.5) 5 Disseminated intravascular coagulation 10.3 6.0 (2.4–57.1) 10.3 (5.4–18.8) 9 Relapse 2.9 7.3 (2.9–8.3) 4.4 (2.4–7.8)# 5 Laboratory values (mean)§§ Leukocytes, 109 cells/L, n = 98 17.4 15.1 (13.9–18.2) 15.8 (14.6–16.9) 9 Hemoglobin, g/L, n = 22 106.7 – – – Platelets, 109/L, n = 41 121.0 182.4 (115–241.5) 164.9 (132.9–197) 7 Blood glucose, mg/dL, n = 32 147.8 – – – C-reactive protein, mg/L, n = 36 349.7 – – – Cerebrospinal fluid Leukocytes, cells/mm3, n = 88 3,166 2029 (450–3253) 2330 (1721–2939)# 7 Protein, g/L, n = 74 3.20 2.35 (1.7–4.18) 2.45 (1.91–2.99) 7 Glucose, mg/dL, n = 70 20.9 8.60 (1.7–25.6) 12.6 (3.5–21.7) 6 *N = 174 unless otherwise indicated. –, not applicable because no large study reported these data.
†Random-effects model unless otherwise indicated.
‡Includes the single-case dataset and the large studies (Technical Appendix Table 2.
§Other less common syndromes included peritonitis, myositis, pneumonia, sacroiliitis, abdominal aortic aneurysm, hemorrhagic labyrinthitis, gastroenteritis, vertebral osteomyelitis, lymphadenopathy, cellulitis, and vertigo.
¶Case-patients with toxic shock syndrome in China and in Thailand not included in this sepsis category.
#Mixed-effects model.
**Counted if the author described the case as toxic shock syndrome.
††Include 3 large studies reporting toxic shock syndrome, including 2 outbreaks in China (2,30) and 1 prospective study in Thailand (24).
‡‡Mainly reported with neck stiffness.
§§Reference values may differ among laboratories. Commonly used reference values for presented laboratory blood tests are as follows: leukocytes 4.0–10 × 109 cells/L; hemoglobin 140–170 g/L (for male patients) and 120–160 g/L (for female patients); platelets 150–350 × 109/L; blood glucose (fasting) 70–100 mg/dL; C-reactive protein 0–8.0 mg/L. Reference ranges for cerebrospinal fluid are as follows: leucocytes 0–5 cells/mm3; protein 0.15–0.60 g/L; glucose 40–80 mg/dL. (Source: http://im2014.acponline.org/for-meeting-attendees/normal-lab-values-reference-table/) Clinical Syndromes Meningitis was the most common clinical syndrome (pooled rate 68.0% [95% CI 58.9%–75.8%]), followed by sepsis (25.0% [95% CI 20.5%–30.2%]), arthritis (12.9% [95% CI 6.0%–25.6%]), endocarditis (12.4% [95% CI 6.7%–21.9%]), and endophthalmitis (4.6% [95% CI 2.8%–7.4%]) (Table 3). Toxic shock syndrome also was reported as a distinct severe clinical feature at high rates in 2 outbreaks in China (64.0% and 28.9% of patients) ( 2 , 30 ) and in Thailand (37.7%) ( 24 ) but at a rate of only 2.9% among the case reports. We found evidence of publication bias in the meta-analysis of meningitis rates (Figure 3) (significant Egger’s test result). The adjusted rate, based on the Duvall and Tweedie trim and fill method, was 56.0% (95% CI 45.2%–66.2%). Our meta-regression analysis showed that meningitis rate was significantly associated with country of publication, study design, and data collection approach (Technical Appendix Table 5), although only country of publication remained significant in a multivariable model. All patients in the 4 studies conducted in Vietnam had meningitis. When we excluded these studies, the pooled rate of meningitis was reduced to 59.4% (95% CI 51.1%–67.1%), and the adjusted value after we used the trim and fill method was 54.8% (95% CI 46.0%–63.4%). In contrast, if we excluded the 2 outbreak investigations in China, because sepsis dominated these outbreaks, the pooled meningitis rate increased slightly to 72.2% (95% CI 62.4%–80.2%). Figure 3 Funnel plot showing evidence of publication bias among 26 studies in a meta-analysis of meningitis rates in Streptococcus suis infection. Each blue circle represents each study in the meta-analysis, forming an asymmetric funnel plot with a pooled log event rate (blue rhombus). Eight missing studies (red circles) added in the left side through the trim and fill method to make the plot more symmetric and gave an adjusted log event rate (red rhombus), which was lower than the original one. Case-Fatality Rates The pooled case-fatality rate (CFR) for S. suis–infected patients was 12.8% (95% CI 9.0%–18.0%) (Table 4). This rate varied by country; reported rates were lowest in Vietnam (Figure 4). However, country of publication was not significant in the bivariate meta-regression after adjustment for multiple comparisons (Technical Appendix Table 5). Instead, only meningitis rates remained significant in explaining between-study variations in CFR. Meningitis rates correlated negatively with CFRs among the included studies (Figure 5). Studies with meningitis rates 90% (mean CFR difference 20.3%, p = 0.001). The pooled CFR was 4.0% (95% CI 2.2%–7.0%), estimated for the studies in which all patients had meningitis ( 3 , 4 , 9 , 10 , 31 – 33 ), whereas the pooled rate for the other studies was 17.1% (95% CI 12.3%–23.4%). CFRs were higher for outbreaks than for nonoutbreaks (21.6% [95% CI 6.4%–52.5%] vs. 11.5% [95% CI 7.9%–16.7%]). Table 4 Summary rates of the main clinical outcomes among patients with Streptococcus suis infection included in a systematic review Variable Single-case dataset, n = 174 Large studies, median (range) Meta-analysis, pooled mean (95% CI) No. studies meta-analyzed* Death 10.3 8.9 (0.0–56.0) 12.8 (9.0–18.0) 25 Hearing loss† 44.8 38.7 (6.0–100) 39.1 (31.0–47.8) 26 Recovery from hearing loss ‡ 5.0 (0.0–52.3) 15.4 (5.3–37.3) 8 Vestibular dysfunction§ 16.7 25.0 (3.3–60.0) 22.7 (15.6–32.0) 13 Visual impairment 4.0 – –¶ – *Includes the single-case dataset and the large studies (Technical Appendix Table 2).
†Studies included if case-patients were reported to have any degree of hearing impairment (unilateral or bilateral, temporary or permanent).
‡Reliable data could not be extracted for the majority of the case reports.
§Studies included if case-patients were reported to have ataxia, vertigo, loss of balance, or vestibular dysfunction.
¶Dashes indicate not applicable because no large study reported these data. Figure 4 Forest plot of subgroup meta-analysis (random effects) for the case-fatality rates by country reported in the 25 studies included in a review of Streptococcus suis infection. For each study, the event rate of the death outcome and 95% CI are presented, with size proportional to study weight. The red rhombus indicates the pooled event rate for each country group. Figure 5 Meta-regression scatter plot showing the correlation between case-fatality rate and meningitis rate in a review of Streptococcus suis infection. The logit event rate was calculated for case-fatality rate as follows: logit event rate = ln[event rate/(1 − event rate)]. Each circle represents a study in the meta-analysis, and the size of the circle is proportional to study weighting. Studies with higher meningitis rates tended to report lower death rates. Clinical Outcomes Among the survivors, hearing loss (pooled rate 39.1% [95% CI 31.0%–47.8%]) and vestibular dysfunction (22.7% [95% CI 15.6%–32.0%]) were the most common sequelae (Table 4). Reported rates for both sequelae varied widely, even within a country such as Thailand, (Technical Appendix Figures 1–4). Similar to CFRs, there was a marginally positive correlation between hearing loss and meningitis rates (p = 0.05) (Technical Appendix Table 5). The pooled hearing loss rate for studies in which all patients had meningitis was 55.3% (95% CI 36.2%–72.9%), compared with 34.0% (95% CI 26.0%–43.1%) for the remaining studies. For the vestibular dysfunction, none of the investigated study-level factors were significant. Among the Asian countries, the reported rate of vestibular sequelae was lowest in Vietnam (4.0%). Limited information described how hearing loss and vestibular dysfunction were evaluated during and after infection. Eight of 25 large studies reporting hearing loss indicated whether hearing loss was permanent after hospital discharge. Only 4 described their follow-up processes; follow-up time ranged from 2 months to 30 months ( 4 , 8 , 28 , 31 ). On the basis of these limited data, we estimated a comparatively low median rate of recovery from hearing loss of 5.0% (range 0%–52.3%) and the pooled rate of 15.4% (95% CI 5.3%–37.3% (Table 4). Hearing loss might be mediated by adjunctive corticosteroid treatments, as was shown in a trial in Vietnam ( 34 ). Of the S. suis patients, 12.3% had deafness in at least 1 ear in the dexamethasone treatment group (n = 57), compared with 37.7% in the placebo group (n = 53). Discussion We have updated estimates of the global prevalence, epidemiology, and clinical characteristics of S. suis infections in humans. After possible duplicates were removed, the total number of S. suis infections by 2012 was close to 1,600 cases, doubling the figure published in 2009 ( 12 ). Most of the increase comprised cases from Thailand and Vietnam, placing both countries in the highest disease prevalence stratum in the world. In contrast, only a few cases have been reported from the Americas, particularly the United States, the second largest producer of pigs worldwide ( 35 ). This low number might be attributable to the high industrialization of pig farming systems in the region. Nevertheless, we saw far more cases in Europe, a region where modern farming operations are presumably similarto those in the Americas. Other plausible explanations include the lower virulence of North American bacterial strains ( 36 ) or different slaughtering practices. We counted only published cases; the actual number of cases would be considerably higher, particularly in areas to which S. suis is endemic, such as Asian countries with extensive pig rearing. The problem of underestimation is further exacerbated by the fact that S. suis infection is not a notifiable disease in many countries. In addition, lack of diagnostic capacities or limited disease awareness in local laboratories can result in undiagnosed or misdiagnosed cases. Meningitis is the main syndrome in approximately two thirds of S. suis–infected patients, although this finding varied by country. The syndromic distribution of the reported cases may depend on study design and case ascertainment methods. All major studies in Vietnam ascertained S. suis cases from the population of patients with central nervous system diseases, which could lead to underrepresentation of S. suis patients with clinical syndromes other than meningitis. Only 1 patient without meningitis (diagnosed as spontaneous bacterial peritonitis with serotype 16 infection) has been reported in this country ( 37 ). Nevertheless, whether the existing strains infecting humans in Vietnam mainly cause meningitis remains unclear. In fact, lumbar puncture is performed regularly for all S. suis–infected patients, including those with severe sepsis, at a hospital for tropical diseases in Vietnam, and almost all had exhibited typical characteristics of bacterial meningitis in cerebrospinal fluid. On the other hand, meningitis might not be diagnosed or reported from other countries, therefore reducing the global S. suis meningitis estimate. The difference in CFR between case-patients with meningitis and case-patients with severe sepsis has been documented in both outbreak and nonoutbreak situations in China and Thailand ( 1 , 2 , 24 ). Significantly more deaths were reported among S. suis patients with systemic infection, including hypotension, septic shock, multiorgan failure, and disseminated intravascular coagulation in these studies. In the Sichuan outbreak in 2005, the CFR reached 62% for patients classified as having streptococcal toxic shock syndrome ( 1 ). Several hypotheses have been suggested; however, the pathologic mechanisms underlying this high CFR remain to be elucidated ( 7 , 12 ). Regarding meningitis cases, the pooled CFR is lower than that for other common causes of adult bacterial meningitis, such as S. pneumoniae (19%–37%) ( 38 ) and Neisseria meningitidis (10%) ( 39 ). However, the rates of sequelae caused by S. suis tend to be higher than those caused by other agents reported in a recent meta-analysis ( 40 ). We were unable to establish pooled risk estimates for different risk factors because of a lack of studies with appropriate designs. In the Netherlands, the annual risk for S. suis meningitis among abattoir workers and pig breeders was 1,500 times higher than that in the general population ( 10 ). In Vietnam, S. suis–infected patients were more likely to have eaten high-risk foods (odds ratio [OR] 4.38), to have pig-related occupations (OR 5.52), and to have pig exposure while having skin injuries (OR 15.96) than community controls ( 6 ). The lower proportions of patients with occupational exposure in Thailand and Vietnam than in Europe shown in our meta-analysis supports the hypothesis that other risk factors, including food consumption practices, may play a major role in the epidemiology of S. suis infection in Asia. This review is not without limitations. The included studies were highly heterogeneous in quality and in the factors reported, which reduced the number of studies included in each meta-analysis. The summary values of the single-case dataset should be interpreted with caution because the patients in this merged “sample” were heterogeneously “recruited” from different populations, with different assessment protocols. In addition, the studies were mainly retrospective; data could have been easily missed on recall or by re-collecting from the existing data records. We were unable to assess the extent to which this misinformation could affect the overall estimates. However, data collection approach was not significantly associated with the main outcomes examined under this review in our meta-regression analyses. This review helps to highlight areas in which additional research is needed. Geographic gaps obviously exist in the data on S. suis cases, especially in the pig rearing countries in the Americas, Eastern Europe, and Asia, such as Mexico and Brazil, Russia, and the Philippines, respectively. Second, much uncertainty remains in understanding sequelae of S. suis infection and recovery from these conditions over time. Careful prospective assessments of these debilitating outcomes and associated social and economic impacts are essential for understanding and reducing the effects of S. suis infection. More studies also are needed to assess the treatment effects of adjunctive corticosteroid on hearing loss or other neurologic sequelae. The effects of S. suis infection are mainly in Asia; occupational exposure and eating possibly contaminated foods containing undercooked pig tissues are prime risk factors. Further research in Asia should focus on the factors pertinent to local risk for infection, including the practices of unsafe handling and consumption of pork. Prevention of human infections needs to be tailored for different risk groups, and studies are needed to assess the feasibility and effectiveness of those tailored programs. Additional work is needed to assess the fraction of S. suis cases that can be attributed to different risk factors (the population-attributable fraction) and the proportion of S. suis cases that might be preventable if specific risk factors were reduced. Technical Appendix Details of publications in a review of the epidemiology, clinical manifestations, and outcome of Streptococcus suis infection in humans.
                Bookmark

                Author and article information

                Journal
                Emerg Infect Dis
                Emerging Infect. Dis
                EID
                Emerging Infectious Diseases
                Centers for Disease Control and Prevention
                1080-6040
                1080-6059
                November 2014
                : 20
                : 11
                : 1895-1898
                Affiliations
                [1]Oxford University Clinical Research Unit, Hanoi, Vietnam (V.T.L. Huong, P. Horby, J.E. Bryant, H.F.L. Wertheim); Oxford University Clinical Research Unit, Ho Chi Minh City, Vietnam (N.T. Hoa);
                [2]University of Oxford, Oxford, UK (V.T.L. Huong, N.T. Hoa, P. Horby, J.E. Bryant, H.F.L. Wertheim);
                [3]National Hospital for Tropical Diseases, Hanoi (N.V. Kinh);
                [4]Hanoi Medical University, Hanoi (T.K. Toan)
                Author notes
                Address for correspondence: Vu Thi Lan Huong, Wellcome Trust Major Overseas Programme, Oxford University Clinical Research Unit, National Hospital for Tropical Diseases, 78 Giai Phong, Dong Da, Hanoi, Vietnam; email: lanhuongcgfed@ 123456gmail.com
                Article
                14-0915
                10.3201/eid2011.140915
                4214319
                25340391
                6e31d2cd-9ca0-427e-97e0-1a3379f8aef7
                History
                Categories
                Dispatch
                Dispatch
                Raw Pig Blood Consumption and Potential Risk for Streptococcus suis Infection, Vietnam

                Infectious disease & Microbiology
                raw pig blood,consumption,practices,perceptions,zoonoses,risk factor,streptococcus suis,bacteria,vietnam

                Comments

                Comment on this article