Introduction
The burden of coronary heart disease is reported to be increasing in many developing countries including those in Africa.(1,2) This rise in prevalence has been ascribed to the changes in lifestyle. Increased industrialization and urbanization in developing nations has seen the emergence and increased prevalence of coronary risk factors such as hypertension, diabetes, cigarette smoking, obesity and dyslipidaemia.(3)
Although clinical variables such as age, Killip class (The Killip classification for heart failure quantifies severity of heart failure in acute coronary syndrome (ACS) and predicts 30-day mortality), heart rate and systolic blood pressure have been previously identified as powerful predictors of risk, it is important to recognise other factors that may affect outcome. Although anaemic patients with coronary artery disease have an increased risk of death in the short term (4), the prognostic impact of anaemia in patients presenting with ACS is poorly defined. Anaemia has been previously observed in up to 15% of the patients presenting with myocardial infarction, with a reported peak of 43% in the elderly patients. (5,6) In ACS patients, only few studies have specifically examined the impact of anaemia on clinical and mortality outcomes.(5,6) Moreover, there are no reports detailing the impact of anaemia in ACS patients from the African continent.
This study was therefore aimed to determine the impact of anaemia in patients with ACS admitted to a large urban public hospital in Johannesburg, South Africa. Firstly, we assessed the prevalence of anaemia and its associated clinical characteristics in patients admitted with ACS. Secondly, we assessed the relationship between the presence of anaemia and its impact on in-hospital mortality in these patients.
Methods
The study was a single-centre retrospective assessment of all patients who were 18 years and above hospitalised to CMJAH with a diagnosis of ACS between the periods of January 2010 and June 2012. Patients with percutaneous coronary intervention related myocardial infarction were excluded from the analysis. Approval for the study was obtained from the University of Witwatersrand Ethics committee (M130336).
Data collection
Patient data for the review period was obtained from the patient electronic records stored in the database of the Cardiology Unit at CMJAH. A standard patient data sheet was used to record patient details with regard to the demographic information, cardiovascular history, risk factors and outcomes. Details of general and cardiovascular examination were also recorded. Records of the blood chemistry, lipid profile, cardiac biomarkers and full blood count were included, where appropriate missing data was accessed from hospital records as well as from the National Health Laboratory Service database.
Diagnosis of ACS required the following: presence of specific diagnostic biomarker elevations (troponin I >0.4 ng/mL and CKMB >8.8 ng/mL) and any one of the following; (a) ischaemic symptoms, (b) development of pathologic Q-wave on ECG and (c) electrocardiographic changes indicative of ischaemia (ST-segment elevation or depression).(7) Unstable angina was defined as the presence of angina at rest and lasting for > 20 min.(7)
Anaemia was defined as per criteria of the World Health Organization (haemoglobin <13 g/dl in men and <12 g/dl in women).(8) Optimal medical therapy was defined as those patients with ACS receiving a combination of at least four or more agents (b-blocker, aspirin, clopidogrel, ACE-I/ARBs and statins) where indicated.(9)
Data analysis
All data generated were analysed using computer-based Statistical Package for the Social Sciences [SPSS] version 16.0. Quantitative variables were described using mean and standard deviation. Qualitative variables were presented as percentages and bar charts. The non-parametric test X 2 (chi-squared) or Fisher's exact test was used to test for significance among categorical variables. Multivariate analysis was used to determine predictors of short-term outcomes. Confidence interval of 95% was used and a p-value of <0.05 was regarded as statistically significant.
Results
Baseline clinical characteristics
A total of 431 patients admitted with a diagnosis of ACS were included in the analysis. The mean age of all patients was 58 ± 12.4 years and the majority of the patients were males (72.2%). Hypertension was the most prevalent risk factor, closely followed by cigarette smoking which was present in 40.6% and 37.1% of patients, respectively (see Table 1).
Baseline clinical characteristics.
| Variable | All patients n ≥ 431 (%) | Anaemia n ≥ 81 (%) | No anaemia n ≥ 350 (%) | p-Value |
|---|---|---|---|---|
| Age (years) | 58±12.4 | 63±13.4 | 57±11.8 | <0.0001 |
| Gender-Male | 311 (72.2) | 34 (41.9) | 277 (79.1) | <0.0001 |
| Female | 120 (27.8) | 47 (58.1) | 73 (20.9) | |
| Hypertension | 175 (40.6) | 42 (51.9) | 133 (38) | 0.022 |
| Diabetes | 108 (25.1) | 29 (35.8) | 79 (22.6) | 0.031 |
| Dyslipidaemia | 93 (21.6) | 10 (12.3) | 83 (23.7) | 0.025 |
| Smoking | 160 (37.1) | 18 (22.2) | 142 (40.6) | 0.002 |
| AF | 16 (3.7) | 3 (3.7) | 13 (3.7) | 0.647 |
| Killip ≥ 3 | 54 (12.5) | 19 (23.5) | 35 (10) | 0.001 |
| UA | 47 (10.9) | 10 (12.3) | 37 (10.6) | 0.644 |
| NSTEMI | 174 (40.4) | 35 (43.2) | 139 (39.7) | 0.563 |
| STEMI | 203 (47.1) | 34 (41.9) | 169 (48.3) | 0.305 |
Key: AF = atrial fibrillation, Killip ≥3 = Killip class classification ≥3, UA = unstable angina, NSTEMI = non ST-elevation myocardial infarction, STEMI = ST segment elevation myocardial infarction
Anaemia was found in 18.8% of the study cohort. Anaemic patients were older as compared to non-anaemic patients (63 ± 13.4 years vs. 58 ± 12.4 years, p < 0.0001). In addition, anaemic patients were more likely to have hypertension (51.9% vs. 38%, p < 0.022) and diabetes (35.8% vs. 22.6%, p < 0.031) as compared to non-anaemic patients, respectively. Anaemic patients were also more likely to be female (58.1% vs. 20.9%, p < 0.0001). With regard to clinical presentation, anaemic patients were found to be in a worse functional class (Killip ≥ 3) on admission as compared to their non-anaemic counterparts (23.5% vs. 10%, p < 0.001). With regard to other demographics, anaemic patients were less likely to be smokers (12.3% vs. 23.7%, p < 0.002).
Medications and therapy received
Table 2 highlights the comparison of anaemic and non-anaemic patients with regard to medications prescribed or procedure performed in this cohort of ACS patients. There was no difference in the number of patients who received β-blockers in the two groups of patients. However, anaemic patients were less likely to receive angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (ACE-I/ARBs) as compared to those patients who were not anaemic (46.9% vs. 66.3%, p < 0.001). In addition, anaemic patients as compared to non-anaemic patients were significantly less likely to receive aspirin (81.5% vs. 90%, p < 0.031) or clopidogrel (65.4% vs. 81.4%, p < 0.001). Overall compared to non-anaemic patients, those patients with any level of anaemia were less likely to receive optimal medical therapy (60.5% vs. 77.7%, p < 0.001). No significant differences were observed between the two groups in terms of reperfusion (thrombolytic/primary PCI) therapies.
Anaemia and ACS medications/therapy.
| Variables | All patients n ≥ 431 (%) | Anaemia n ≥ 81 (%) | No anaemia n ≥ 350 (%) | p-Value |
|---|---|---|---|---|
| β-Blockers | 314 (72.9) | 54 (66.7) | 260 (74.3) | 0.165 |
| ACE-I | 270 (62.6) | 38 (46.9) | 232 (66.3) | 0.001 |
| Statins | 380 (88.2) | 67 (82.7) | 313 (89.4) | 0.097 |
| Plavix | 338 (78.4) | 53 (65.4) | 285 (81.4) | 0.001 |
| Aspirin | 381 (88.4) | 66 (81.5) | 315 (90) | 0.031 |
| Thrombolysis | 60 (13.9) | 12 (14.8) | 48 (13.7) | 0.797 |
| Angioplasty | 117 (27.1) | 18 (22.2) | 99 (28.3) | 0.269 |
| Optimal therapy | 321 (74.5) | 49 (60.5) | 272 (77.7) | 0.001 |
Key: AF = atrial fibrillation, Killip ≥3 = Killip class classification ≥ 3, Hb < 11.4 = haemoglobin < 11.4g/dl, PCI = percutaneous coronary intervention
Mortality predictors
Comparison of clinical variables among the deceased and survivors revealed that the prevalence of atrial fibrillation (13.3% vs. 3.4%, p < 0.045) and Killip class ≥3 (40% vs. 11.5%, p < 0.001) was significantly higher among those who died as compared to those who survived (see Table 3). The prevalence of anaemia was significantly higher in those who died compared to survivors (46.7% vs. 1.9%, p < 0.0001). However, on multivariate analysis, only anaemia (Hb level of <11.4 g/dl) was found to independently predict mortality among ACS patients. Those ACS patients with Hb < 11.4 g/dl had more than fourfold risk of death compared to non-anaemic patients (CI – 1.393–13.041: RR 4.262; p < 0.011) (see Table 4).
Mortality and clinical characteristics.
| Variable | All patients 431 (%) | Deceased 15 (%) | Survivors 416 (%) | p-Value |
|---|---|---|---|---|
| Hypertension | 175 (40.6) | 5 (33.3) | 170 (40.9) | 0.559 |
| Diabetes | 108 (25.1) | 4 (26.6) | 104 (25.0) | 0.429 |
| Geder-Male | 311 (72.2) | 10 (66.7) | 301 (72.4) | 0.629 |
| Female | 120 (27.8) | 5 (33.3) | 115 (27.6) | |
| Smoking | 160 (37.1) | 3 (20) | 157 (37.7) | 0.162 |
| AF | 16 (3.7) | 2 (13.3) | 14 (3.4) | 0.045 |
| Killip ≥ 3 | 54 (12.5) | 6 (40) | 48 (11.5) | 0.001 |
| Age > 70 years | 369 (85.6) | 6 (40) | 56 (13.5) | 0.004 |
| Hb < 11.4 | 15 (3.5) | 7 (46.7) | 8 (1.9) | <0.0001 |
| PCI | 117 (27.1) | 2 (13.3) | 115 (27.6) | 0.221 |
| Thrombolysis | 60 (13.9) | 0 (0.0) | 60 (14.4) | 0.113 |
Key: AF ≥ atrial fibrillation, Killip ≥3≥ Killip class classification ≥ 3, Hb < 11.4 ≥ haemoglobin < 11.4g/dl, PCI ≥ percutaneous coronary intervention
Mortality predictors.
| Variables | Confidence interval (CI) | Relative risk | p-Value |
|---|---|---|---|
| Killip ≥ 3 | 0.971–10.766 | 3.234 | 0.056 |
| AF | 0.298–13.341 | 1.993 | 0.477 |
| Hb < 11.4 (g/dl) | 1.393–13.041 | 4.262 | 0.011 |
| Age > 70 years | 0.117–1.217 | 0.377 | 0.107 |
Key: AF ≥ atrial fibrillation, Hb < 11.4 ≥ haemoglobin < 11.4 g/dl, PCI ≥ percutaneous coronary intervention
Discussion
The current study found a relatively high prevalence of anaemia of 18.8%. Anaemia as defined by a haematocrit of less than 40% was reported to be present in 15% of AMI patients in the Thrombolysis In Myocardial Infarction (TIMI) trial.(5) Interestingly, these patients required more blood transfusions following thrombolytic therapy as compared to those who had haematocrits greater than 40% on entry into the study (5). A French study of 1064 patients who were admitted to two French hospitals with ACS, reported a prevalence of 29% using the same definition of anaemia as the current study.(10) A retrospective review of 78,944 Medicare beneficiaries who were 65 years and above reported an overall prevalence of anaemia of 43%, using a haematocrit less than 39% as the definition of anaemia.(6) In an analysis of 39,922 patients enrolled in the TIMI clinical trials of ACS, a prevalence of 15% of anaemia was reported in AMI patients and 30% in patients with ACS using a haemoglobin cut-off of <13 g/dl as the definition of anaemia.(11)
The cause of anaemia in ACS patients is multifactorial. Firstly, anaemia may be related to both overt and occult blood loss particularly that occurring during clinical procedures. However, some pathophysiological changes in AMI may also contribute to the development of anaemia. Anaemia could be secondary to intravascular haemodilutional changes as a consequence of advanced systolic dysfunction. This mechanism may be relevant in our cohort of patients as patients with anaemia presented significantly more frequently with Killip class 3 or greater suggesting impaired left ventricular function.
In the current study of ACS patients, those who were female or elderly had a significantly higher prevalence of anaemia as compared to men and younger patients. It is well recognised that both of these groups of patients have a higher prevalence of haemopoietic disorders that may contribute to anaemia.(12) Anaemia may also be a consequence of renal failure particularly among those with extensive MI leading to the cardiorenal syndrome. Pro-inflammatory cytokines released during AMI have also been shown to inhibit iron metabolism and have been implicated in the development of anaemia.(11)
It has been found that optimal medical therapy is associated with a remarkable impact on the outcomes among the ACS patients.(9,13) A Canadian study has shown that patients with ACS who received optimal medical therapy, after adjusting for other validated prognosticators, had a significantly lower 1-year mortality compared with those given no evidence-based drug or only one evidence-based drug at discharge.(9)
The present study showed that anaemic patients presenting with ACS were significantly less likely to receive optimal medical therapy. It is possible that anaemic patients had contraindications to the use of some of these optimal agents, such as aspirin and clopidogrel. It is well known that the use of ACEI or β-blockers is contraindicated in patients with hypotension (9,14), and although it was not possible to interrogate how many of the anaemic patients in the current study were hypotensive, it is clinically probable that some patients with worse functional class (Killip ≥3) had hypotension. Furthermore, factors such as renal dysfunction and advanced age also determine the choice of medications for these patients. However, a prospective, multicentre observational study involving 6853 patients admitted for ACS, reported that advanced age, female sex, history of heart failure and renal dysfunction were independently associated with lower use of combination therapy, even in the absence of known contraindications.(9) Although the ‘gold standard’ for assessing treatment efficacy is set by randomised controlled trials, observational studies afford unique and valuable insights into treatment effectiveness and importantly whether these are generalisable in clinical practice.(15) For example, a large US registry (CRUSADE) and the international GRACE studies both demonstrate the underuse of proven medical therapies at discharge among patients with ACS, irrespective of the geographic location.(16,17)
In this study Killip class ≥3, atrial fibrillation and haemoglobin (Hb) <11.4 g/dl were significantly associated with mortality. However, on multivariate analysis, only Hb of <11.4 g/dl was found to be an independent predictor of mortality and had more than fourfold increased risk compared to those with a normal Hb. These findings are in accordance with reports in the international literature. A review of the 16 TIMI clinical trials revealed that the likelihood of death, recurrent ischaemia and non-fatal MI was significantly increased in patients with Hb <11 mg/dl compared to those with normal haemoglobin, even after the adjustment for baseline prognostic factors and in-hospital treatment.(11) The large database of elderly Medicare patients showed a powerful relationship between haematocrit on admission and 30-day mortality.(6) These two studies appeared to suggest a dose-response effect with progressively diminished survival rates with more profound degrees of anaemia.(6,11) In the Medicare study the in-hospital mortality was almost 2.5-fold higher in the lowest quintile of haematocrit compared to those patients with a normal haematocrit.(6) In the French prospective registry, the 6-month primary end-point of death, myocardial infarction or hospitalization was also reported to be threefold higher in those diagnosed with anaemia at admission.(10)
In animal models, a high Hb level has been found to prevent ischaemia even in the presence of significant coronary artery stenosis.(18) Anaemia worsens myocardial ischaemia, firstly by increasing myocardial oxygen demand as it initiates a higher stroke volume and heart rate to maintain adequate oxygen delivery to tissues.(19) Secondly, anaemia significantly decreases oxygen delivery downstream of a coronary stenosis to an already compromised myocardium.(20) It is likely that in patients with ACS, a combination of these mechanisms explains the pathophysiology of worse outcomes with lower baseline Hb concentrations. The current study did not stratify the mortality risk by type of ACS presentation. However, a review of the 16 TIMI clinical trials, did suggest that STEMI patients were more likely to be at increased risk of major adverse cardiovascular events at a higher Hb threshold as compared to NSTEMI patients.(11) This difference is probably a reflection of the differing mechanisms by which anaemia predisposes to major adverse cardiovascular events in STEMI and NSTEMI. In STEMI there is usually sudden occlusion of a coronary artery and even mild levels of anaemia may significantly attenuate the ability of coronary collateral flow to alter peri-infarction ischaemia or limit the extent of the infarction. On the other hand, NSTEMI which is often characterised by an incomplete coronary artery occlusion, major adverse cardiovascular events are dependent on a delicate balance between myocardial oxygen supply and demand. With good anti-ischaemic therapy in these patients, a more severe degree of anaemia may be required to cause recurrent ischaemic events.(11)
Limitations
The study was retrospective in design and inherent limitations of retrospective reviews apply to this study as well. Furthermore the study included a relatively small number of patients with ACS. However, it is to our knowledge, the largest dataset of anaemia in ACS in Africa. It is possible that unidentified co-morbidities associated with anaemia could potentially confound our analysis. However, we believe this effect to be small as a comprehensive search was made for confounders.
Conclusion
Anaemia is not uncommon in patients presenting with ACS. It is more prevalent in those who are older, female, hypertensive and in those presenting with a much worse functional class or heart failure. Low haemoglobin, especially below 11 g/dl, is associated with a significantly higher mortality. These findings suggest that a simple measurement of haemoglobin level provides additional prognostic information in patients presenting with ACS. Thus, haemoglobin measurement deserves to be routinely incorporated into the risk stratification toolkit of patients with ACS to improve the risk stratification and event prediction.