CHAPTER 6
The spectrum of heart disease in urban Africans: The Heart of Soweto Study
Melinda Jane Carrington1, Karen Sliwa2, and Simon Stewart1
1 Australian Catholic University, Melbourne, Victoria, Australia
2 University of Cape Town, South Africa; University of the Witwatersrand, Johannesburg, South Africa
6.0 Introduction
This chapter provides an overview of the Heart of Soweto Study [74], now regarded as a landmark study in our attempts to understand the evolving burden of heart disease in sub-Saharan Africa. However, given its limited scope and lack of prospective follow-up data, this study is only the beginning in this regard. The Heart of Soweto Study was built on the premise that while the burden of CVD has stabilized in most high-income countries (with age-adjusted rates of acute and chronic heart disease events continuing to fall [9–11]), in LMIC there was clear evidence it was beginning to rise [75]. In a setting of scarce health care resources, these countries are ill-equipped to cope with the rise of the noncommunicable diseases [38]; particularly when the traditional killers of malnourishment and infection remain prevalent [76]. As discussed in the original reports, there is little scope to tackle new prototypes of heart disease arising from changing risk behaviours due to epidemiologic transition [3]. As also described in Chapter 4, there is clear evidence that urban communities (including Soweto) contain a rising burden of antecedent risk that will fuel a rising burden of noncommunicable forms of heart disease. This begged an important question: How does this emerging threat compete and interact with the historically dominant, communicable forms of heart disease (see Section 4) in sub-Saharan Africa? As shown in Figure 6.1, the Heart of Soweto Study generated a series of reports describing various aspects of the spectrum of heart disease in the world-renowned community of Soweto. Many of these were “first of a kind” reports within the African context. The two most encompassing reports from this study (in chronological order) are initially summarized before a review of the specific diagnoses (from AF to pulmonary arterial hypertension (PAH)/RHF—see Chapter 15) subject to more intensive analysis and interpretation are presented.
Figure 6.1 A family tree of key research reports from the Heart of Soweto Study.
6.1 Initial findings of the Heart of Soweto Study
Sliwa K, Wilkinson D, Hansen C, Ntyintyane L, Tibazarwa K, Becker A, Stewart S. Spectrum of heart disease and risk factors in a black urban population in South Africa (the Heart of Soweto Study): a cohort study. The Lancet 2008; 371:915–22. [50]
As described in the initial report of the rationale and design of the Heart of Soweto Study, this was a prospective clinical registry of all heart disease presentations to the Cardiology Unit of the Chris Hani Baragwanath Hospital in Soweto, South Africa (see Chapter 4). Specifically, all individuals presenting with suspected or confirmed CVD to the hospital were recorded on a prospectively designed registry. The registry was compiled and maintained via a dedicated team of staff and facilities using standardized case report forms (with each patient provided with a unique identifier). A key component of the study was the presence of expert cardiologic review, echocardiography, and 12-lead ECGs for each and every patient attending the purposely built Heart of Soweto clinic that formed the basis for the registry.
6.1.1 Study methods
Wherever possible, the study adhered to the STROBE guidelines for collecting and interpreting observational data [77]. Critically, at the time of the study the hospital provided almost exclusive cardiac services to the community, with many patients bypassing the overstretched primary care services to go straight to the hospital for treatment. The Chris Hani Baragwanath Hospital, therefore, did and still does represent an important barometer (if not complete given a number of caveats) of the heart health of Soweto and, by implication, the heart health of urban communities in sub-Saharan Africa in epidemiological transition.
6.1.1.1 Patient profiling
The Heart of Soweto registry captured basic sociodemographic (including self-reported years of education, origin, and history of any preexisting CVD) in addition to any advanced clinical profiling (described in detail in study reports [50,78,79]) according to clinical need. As noted above, a critical component of the study was the capture of echocardiography (performed and reviewed according to gold-standard criteria [80]), and 12-lead ECG (subject to detailed and blinded Minnesota coding [81]) in each and every presentation to the clinic to facilitate profiling. This reflected the typical lack of “clinical workup” of those presenting de novo with often complex conditions in advance stages of illness. Those of African ancestry were typically of Zulu or Xhosa origin and migrants to Soweto were defined as those not self-reporting to be born in Soweto.
6.1.1.2 Diagnostic classifications
An enormous amount of clinical data from many patients with complex disease states were generated from the registry. Using predominantly European Society of Cardiology guidelines [79] relevant to the time of analyses, these data were independently reviewed and adjudicated (consensus approach) by the principal investigators. Table 6.1 summarizes the broad classification of clinical cases according to communicable versus noncommunicable heart disease, noting that in the case of valvular disease and dysfunction, a new system of classification was devised [7]. Study data were documented on standardized forms and entered into a preliminary database at the Soweto Cardiovascular Research Unit based at the Chris Hani Baragwanath Hospital before being transferred and entered into the definitive trial registry by an Australian team based at the Baker IDI Heart and Diabetes Institute in Melbourne. These data were then used to derive the final classification of cases according to broad classes.
Table 6.1 General classification of Heart of Soweto case presentations.
| Study Classification | Primary Diagnoses | General Characteristics |
| Hypertension and/or type 2 diabetes | Hypertension and type 2 diabetes | These are increasingly prevalent risk factors in urban communities and are strongly associated with lifestyle risk factors. It is considered a primary diagnosis only when no evidence of cardiac dysfunction is found. |
| Historically prevalent heart disease | RHD, nonischemic CMO, pericardial disease, and pulmonary heart disease (10) | These are the main contributors to heart disease in sub-Saharan Africa arising from infectious diseases such as TB, streptococcal infection, and HIV-infection. |
| Noncommunicable heart disease | Hypertensive HF and CAD | These are strongly associated with lifestyle risk factors. |
| CHD | Congenital heart defects and cardiac trauma | Fairly rare |
| Other heart disease | Miscellaneous diagnoses, including cardiac trauma or cases without a definitive diagnosis | Fairly rare |
| Other CVD | Peripheral arterial and cerebrovascular disease | These are almost exclusively linked to lifestyle risk factors. |
6.1.1.3 Study timeline
The initial phase of the Heart of Soweto Study took place from January 1, 2006 to December 31, 2006. During that period, 129,633 inpatients were managed by the hospital’s internal medicine specialists. The estimated case load for the cardiology unit during the same period was 5,000 patients, generating 21,000 patient contacts overall, all with access to gold-standard cardiologic review and advanced diagnostic investigations.
6.1.2 Study findings
As shown in Figure 6.2, 4,506 patients were assessed and entered into the Heart of Soweto Registry in 2006. Of these, 344 (7.6%), who were on average a decade younger and had a similar sex profile, were found not to have underlying CVD and were excluded from further analyses. Therefore, the study population consisted of 4,162 confirmed cases of heart disease. Of these, 1,593 (38.3%) were newly diagnosed patients and 2,569 (61.7%) patients had been previously diagnosed with one or more forms of heart disease. On average, patients with preexisting heart disease were 1.7 years older than the de novo cases.
Figure 6.2 Spectrum of case presentations in the Heart of Soweto 2006 cohort.
6.1.2.1 Cohort profile
As expected, most (n = 1,359; 85.3%) patients were of African ancestry. The study cohort mainly consisted of women (n = 939; 59.0%), and those of African ancestry reported the lowest education levels compared to other ethnic groups. Consistent with an overall picture of premature heart disease, 862 (54.1%) patients were younger than 55 years, and 399 (25.1%) were younger than 40 years. With more than half the patients residing in Soweto (n = 842; 52.9%), only 42 (5.0%) reported residing there for <5 years.
6.1.2.2 Spectrum of risk and disease
A broad spectrum of CVD and risk factors were identified within the study cohort; the most common diagnoses were HF, hypertension, CAD, and VHD (see the following sections). Patients diagnosed with VHD were on average more than a decade younger than those with hypertension or HF. Renal disease, anemia, and concurrent diabetes was also diagnosed in some of the patients in the study (n = 115/1,182 [9.7%]; n = 156/1,185 [13.2%]; and n = 165 [10.4%] respectively), and 74 (4.7%) tested HIV positive. In 146 (9.2%) patients, the most common diagnosis was pericardial effusion as a complication due to HIV/AIDS, TB, or a combination of both (n = 80; 54.8%). Overall, 639 (40.1%) of the newly diagnosed patients had rheumatic (valvular) heart disease, tuberculous pericardial effusion, or CMOs. Although stroke diagnosis was rare, 145 (9.1%) patients reported a family history of stroke and 64 (4%) were diagnosed with an acute stroke. Those patients of African ancestry were more likely to be diagnosed with HF than the rest of the cohort (739 [54%] versus 105 [45%]; OR 1·46, 95% CI 1·11 – 1·94; p = 0·009). On the other hand, they were far less likely to be diagnosed with CAD than any other ethnic group (77 [6%] versus 88 [38%]; OR 0·10, 0·07 – 0·14; p < 0·0001). Although the total proportions of men and women diagnosed with diabetes and hypertension were very similar, proportionately more women than men were diagnosed with VHD (240 [26%] versus 120 [18%]; OR 1·30, 1·11 – 1·52; p = 0·001). Commonly presenting symptoms across all diagnostic groups were angina pectoris/chest pain (n = 451; 28.3%), dyspnoea (NYHA Class III/IV), and peripheral oedema (n = 494; 31%). Echocardiographic and ECG profiling demonstrated that 487 (34%) patients presented with underlying cardiac dysfunction, structural disease, and/or tachycardia. Of the 844 patients with HF, 332 (39.3%) had concurrent valvular dysfunction, 203 (24.1%) had a mixed underlying etiology, and 67 (7.9%) had a primary diagnosis of valve disease. The dilated CMOs (OR 35%; 95% CI 32–38), which include HF secondary to hypertensive heart disease (OR 33%; 30–36), RHF (OR 27%; 24–30), and PPCMO, were the three most common forms of HF.
6.1.3 Study limitations
A critical but understandable limitation of this and other African studies highlighted in this book is the risk of selection bias—in this case predominantly due to the fact that the study’s reliance on presentations to the cardiology unit at the hospital. This meant that milder forms of heart disease, fatal events in the community, and even stroke cases directed toward neurology services would be underreported and influence findings if known. Fortunately, the hospital services the community of Soweto almost exclusively and thus serves as a broad, if not specific, barometer of health and disease in the community. Although systematic ECG and echocardiographic investigations were applied, beyond basic clinical profiling, patients in the study were profiled according to clinical need.
6.1.4 Study conclusions
The initial publications arising from the Heart of Soweto Study generated more questions than answers, providing tantalizing glimpses (via the broad data generated) of important aspects on the specific changes that were/are driving a transition from predominant communicable heart disease toward a mixture of the old and new, the new being noncommunicable heart disease. These data confirmed at least the hypothesis that epidemiological transition was driving Soweto (and other parts of Africa) toward profound changes, not only from a socioeconomic perspective but in health and health outcomes. As previously shown in Figure 6.1, a number of research reports focused on key contributors to the spectrum of heart disease in Soweto; more detailed descriptions of the spectrum of HF (Chapter 13), HIV infection of the heart (Chapter 9), ACS (Chapter 10), RHD (Chapter 7), and RHF associated with PAH (Chapter 15) are described later in the book. The following sections highlight some of the other key findings from the study, noting the same strengths and weaknesses/limitations that applied throughout.
6.2 Nexus between communicable and noncommunicable forms of heart disease in Soweto
Stewart S, Carrington MJ, Pretorius S, Methusi P, Sliwa K. Standing at the crossroads between new and historically prevalent heart disease: effects of migration and socio-economic factors in the Heart of Soweto cohort study. European Heart Journal 2011; 32(4):492–9.
6.2.1 Study background
The initial phase of data derived from the Heart of Soweto Study as outlined above [74] immediately suggested that epidemiologic transition was broadening the spectrum of advanced forms of heart disease in that community. It was immediately apparent that a high burden of complex cases in young individuals and women (a pattern rarely seen in high-income countries) required further investigation [50]. In this context, the intrinsic balance between historically prevalent and emergent forms of heart disease in sub-Saharan Africa was unknown. As noted in the original report [74], with high levels of rural migration and extreme poverty counterbalanced by sufficient consumer demand for new, state-of-the-art shopping precincts, Soweto represented an ideal community to study epidemiologic transition via the Heart of Soweto Study. Collectively, the investigators postulated that the balance between largely communicable versus noncommunicable forms of heart disease [82] had irrevocably changed, and this would be observed in this study cohort, particularly when examining de novo presentation of heart disease at the Cardiology Unit of the Chris Hani Baragwanath Hospital. Specifically, it was hypothesized that in addition to finding an equitable balance between new and historically prevalent forms of heart disease in the study cohort, sociodemographic gradients in the pattern of case presentations as well as a significant role for lifestyle risk factors in contributing to more advanced presentations in those with historically prevalent heart disease would be evident, with the latter representing the confluence of communicable and noncommunicable heart disease.
6.2.2 Study methods
6.2.2.1 Patient enrollment
During the extended period 2006 to 2008, data were subsequently captured on 6,006 de novo presentations at the Cardiology Unit of the Chris Hani Baragwanath Hospital. Of these, 678 cases (11.2%) were found not to have any form of CVD or major risk factors. Of the 5,328 remaining cases, 401 (7.5%) were derived from the emergency case presentations, 367 (7.1%) from external referrals from local primary care clinics, 1,992 (37.4%) from internal referrals as a hospital inpatient, and 2,568 (48.2%) from referrals from other outpatient departments.
6.2.3 Study findings
Overall, 4,626 (86.8%) patients were of African ancestry, with a predominance of women in this group (see Table 6.2).
Table 6.2 Patients’ demographic and clinical characteristics.
| Women (n = 3,168) | Men (n = 2,160) | African (n = 4,626) | Other (n = 702) | |
| Demographic characteristics | ||||
| Age (years) | 51.5 ± 17.9 | 52.8 ± 16.5 | 51.5 ± 17.7 | 56.7 ± 14.1 |
| African ancestry | 2,863 (90.4%) | 1,763 (81.6%) | 4,626 (100%) | — |
| Women | 3,168 (100%) | — | 2,863 (61.9%) | 305 (43.5%) |
| <6-years education | 1,331 (42%) | 933 (43.2%) | 2,018 (43.6%) | 246 (35%) |
| Soweto originated | 1,756 (55.4%) | 1,079 (50%) | 2,807 (60.7%) | 28 (4%) |
| Residing in Soweto (years) | 39.2 ± 18.1 | 38.2 ± 17.7 | 38.8 ± 18.0 | 38.6 ± 15.8 |
| Risk factors | ||||
| Family history | 1,430 (45.1%) | 727 (33.7%) | 1,800 (39%) | 357 (50.9%) |
| Total cholesterol (mmol/L) | 4.4 ± 1.3 | 4.1 ± 1.3 | 4.2 ± 1.3 | 4.8 ± 1.3 |
| History of cigarette smoking | 971 (30.7%) | 1,454 (67.3%) | 200 (4.3%) | 425 (60.5%) |
| BMI (kg/m2) | 29.8 ± 7.6 | 25.7 ± 6.0 | 28.2 ± 7.2 | 28.0 ± 7.5 |
| Clinical characteristics | ||||
| NYHA II, III, or IV | 2,274 (71.8%) | 1,371 (63.5%) | 3,214 (69.5%) | 431 (61.4%) |
| Systolic BP (mm Hg) | 133 ± 27 | 132 ± 28 | 133 ± 27 | 132 ± 26 |
| Diastolic BP (mm Hg) | 75 ± 16 | 76 ± 16 | 76 ± 15 | 74 ± 14 |
| Chest pain/angina pectoris | 385 (12.2%) | 241 (11.2%) | 505 (10.9%) | 121 (17.2%) |
| Peripheral oedema | 1,054 (33.3%) | 648 (30%) | 1,579 (34.1%) | 123 (17.5%) |
| Mean LVEF (%) | 56.3 ± 15.8 | 51.7 ± 16.7 | 54.2 ± 16.5 | 55.8 ± 14.7 |
| LV systolic dysfunction | 597 (18.9%) | 592 (27.4%) | 1,074 (23.2%) | 115 (16.4%) |
| Diastolic dysfunction | 469 (14.8%) | 290 (13.4%) | 676 (14.6%) | 83 (11.8%) |
| Primary diagnosis | ||||
| Hypertensive HF | 732 (23.1%) | 414 (19.2%) | 1,050 (22.7%) | 96 (13.7%) |
| Hypertension | 645 (20.4%) | 343 (15.9%) | 854 (18.3%) | 134 (19.1%) |
| Valve disease | 491 (15.5%) | 233 (10.8%) | 660 (14.3%) | 64 (9.1%) |
| CAD | 239 (7.5%) | 342 (15.8%) | 271 (5.9%) | 310 (44.2%) |
| Idiopathic dilated CMO | 234 (7.4%) | 268 (12.4%) | 470 (10.2%) | 32 (4.6%) |
| RHF/PH | 185 (5.8%) | 160 (7.4%) | 311 (6.7%) | 34 (4.8%) |
| HIV-related heart disease | 321 (10.1%) | 197 (9.1%) | 500 (10.8%) | 18 (2.6%) |
Figure 6.3 highlights the spectrum of 5,328 de novo cases according to sex and ethnicity; the pattern of presentation was also determined by level of education and migrant status. Historically prevalent heart disease was diagnosed in 2,092 (39.3%) patients (comprising 60% women and 93% Africans). Newer, noncommunicable forms of heart disease were found in 35% of the patients (comprising 56% women and 79% Africans; p < 0.001 for both comparisons). A similar proportion of men and women were diagnosed with any other form of CVD (including 87 (60.4%) patients with stroke/cerebrovascular disease and 57 (39.6%) with peripheral arterial disease. An additional 999 (18.8%) patients (almost all with hypertension) presented with a high risk of developing heart disease but with no evidence of the same at that time.
Figure 6.3 Spectrum of de novo patients according to sex and ethnicity.
6.2.3.1 Soweto at the crossroads
Significantly, the proportion of historically prevalent communicable versus newer noncommunicable forms of heart disease was evenly poised (1.8 ± 0.9 versus 2.4 ± 0.8). Adjusting for sociodemographic profile, communicable cases of heart disease were more likely to be younger, of African ancestry, and men. The influence of sex, age, and urban transition with uptake of lifestyles was more likely to lead to noncommunicable forms of heart disease, but less exposure to some forms of communicable disease was evident in the pattern of heart disease in the cohort. Overall, there were more cases of historically prevalent than new heart disease up to the age of 49 years for women and 59 years for men, before this pattern reversed in older age groups. For Sowetan women, new heart disease case presentations continuously increased across all age groups compared to historically prevalent cases, which peaked in those aged 30 to 39 years before declining slowly thereafter. Alternatively, for Sowetan men, new forms of heart disease cases peaked in the age bracket 60 to 69 years and decreased thereafter, while historically prevalent cases peaked in the 50- to-59-year age group before decreasing thereafter. This pattern was very similar for migrant African men. For female migrants however, new heart disease case presentations peaked in those aged 50 to 59 years before declining compared to historically prevalent cases; which decreased across all age groups. Although there was no interaction between age, sex, and presentation in “migrants,” there was a borderline association in native Sowetans (p = 0.059); the probability of a woman aged 20 to 29 years presenting with historically prevalent heart disease was 1.25-fold greater, but this probability was 0.67-fold less in the >70 years age group relative to age-matched men.
6.3 Hypertension and hypertensive heart disease in Soweto
Stewart S, Libhaber E, Carrington MJ, Damasceno A, Abbasi H, Hansen C, Wilkinson D, Sliwa K. The clinical consequences and challenges of hypertension in urban-dwelling black Africans: insights from the Heart of Soweto Study. International Journal of Cardiology 2011; 146(1):22–27.
As outlined in Chapter 4 and Chapter 5, it is highly probable that, for the foreseeable future, hypertension represents the single greatest threat to the development of noncommunicable forms of heart disease in sub-Saharan Africa. As part of the original cohort studied, 761 (84.8%) cases of hypertension among those of African ancestry (comprising 63.3% women) were subject to further scrutiny. On presentation, chest pain, palpitations, and/or dizziness were common. Hypertension was the primary diagnosis in only 35% (n = 266); with a presentation of concurrent non-IHD (54%), concurrent LVH (39%), renal dysfunction (24%), anaemia (11%), and CAD (6.2%). Overall, therefore, 494 (64.9%) patients presented with an advanced form of CVD. An additional 98 (12.9%) patients had a valvular abnormality detected by echocardiography with more men than women presenting with impaired LV systolic function. The majority of the valvular cases (n = 58; 59.2%) were due to underlying structural valve disease, mainly comprising rheumatic valve disease (n = 47; 81%) and degenerative valve disease (n = 9; 15.5%). Although women presented with more clinical symptoms, proportionately, they were less likely to be diagnosed HF and renal disease. The study concluded that more women than men were affected with advanced forms of hypertensive heart disease in the Heart of Soweto cohort. Overall, these data reinforce the potential for untreated and undetected hypertension to ultimately lead to advanced forms of heart disease in many individuals within sub-Saharan Africa. As such, the development and application of sex-specific community-based screening and prevention programs adapted to the African context are urgently required to truncate an almost inevitable rise in the burden of hypertensive heart disease in vulnerable communities in the region.
6.4 A predominance of HF in Soweto
Stewart S, Wilkinson D, Hansen C, et al. Predominance of heart failure in the Heart of Soweto Study cohort: emerging challenges for urban African communities. Circulation 2008; 118(23):2360–7. [79]
The syndrome HF represents a common pathway for most forms of CVD when associated cardiac damage and ventricular dysfunction is left untreated and/or poorly managed. Even in the setting of optimal prevention and management, progressive cardiac dysfunction is inevitable as the body ages. At the time of the Heart of Soweto Study, the single largest HF study had been conducted in Africa in the early 1960s and too few data from previous studies were derived from echocardiography presentations. As part of a more detailed examination of the original 2006 cohort, data from 1,960 HF patients and related CMO’s were analyzed to bridge this critical gap in the African literature (see Section 6) however, demographic and clinical data was collected from all 844 (43.1%) de novo presentations. Overall, 739 (87.6%) patients were of African ancestry, 479 (56.8%) were women, and the mean age (in stark contrast to that seen in high-income countries) was 55 ± 16 years. A complex picture of HF and comorbidity was noted, with the most common diagnoses being hypertensive (n = 281; 33.3%), idiopathic dilated CMO (n = 237; 28.1%), tricuspid regurgitation (n = 234; 27.7%), RHF (n = 225; 26.7%), isolated diastolic dysfunction (n = 180; 21.3%), and concurrent renal dysfunction (n = 172; 20.4%). Figure 6.4 highlights the ethnic differences in HF presentation. Not unexpectedly, given the paucity of CAD overall (see Chapter 5), only 41 (5.6%) patients of African ancestry presented with ischemic CMO compared to 36 (34.3%) in the rest of the cohort (p < 0.0001). Alternatively, 285 (38.6%) patients of African ancestry presented with an idiopathic dilated CMO compared to 13 (12.4%) for the rest (p < 0.0001).
Figure 6.4 Etiology of heart failure according to ethnicity.
Similarly, Figure 6.5 compares the etiology of HF according to sex in those of African ancestry. While there were almost double the number of ischemic CMO cases in men compared to women (13.4% versus 5.9%), the latter were more likely to present with valvular HF, RHF, and hypertensive HF.
Figure 6.5 Etiology of HF according to sex.
6.5 RHF—A not so rare form of HF in Soweto
Stewart S, Mocumbi AO, Carrington MJ, Pretorius S, Burton R, Sliwa K. A not-so-rare form of heart failure in urban black Africans: pathways to right heart failure in the Heart of Soweto Study cohort. European Journal of Heart Failure. 2011; 13(10):1070–77. [83]
In the first year of conducting the Heart of Soweto Study, it was soon clear that one in every five cases of HF involved a component of RV dysfunction/RHF, an otherwise rare form of HF reported in cohorts derived from high-income countries. Consistent with these figures, of the 5,328 heart disease patients presenting during the period 2006 to 2008, 2,505 (47%) patients had any form of HF and 697 (27.8%) of these were diagnosed with a component of RHF (642 [92.1%] patients were of African ancestry). Figure 6.6 shows the diagnostic profile of these RHF patients according to sex. An important component (given the number of presentations overall) of RHF cases was PAH. As such, Figure 6.7 shows the diagnostic profile of these cases according to sex. The most common forms of PAH found in women were idiopathic (n = 32; 34.4%), HIV-related (n = 31; 33.3%), and connective tissue disease-related (n = 25; 26.9%). Among men, idiopathic PAH (n = 29; 60.4%) and HIV-related PAH (n = 11; 22.9%) were predominant. These findings are more fully explored in Chapter 15.
Figure 6.6 Diagnostic profile of RHF according to sex.
Figure 6.7 PAH diagnostic characteristics according to sex.
6.6 Pathways to AF in Soweto
Sliwa K, Carrington M, Klug E, Opie L, Lee G, Ball J, Stewart S. Predisposing factors and incidence of atrial fibrillation/flutter in an urban African community: insights from the Heart of Soweto Study. Heart 2010; 96:1878–82. [84]
At the time of the Heart of Soweto Study, there was a paucity of data on the epidemiology of AF in sub-Saharan Africa. Given the twin epidemics of HF and AF in the aging populations of high-income countries [85], an emerging threat in LMIC elsewhere in the world [86], and the predominance of hypertension and HF in the cohort, a close analysis of 12-lead ECG data was logical. During the period 2006 to 2008, 246 (4.6%) of the 5,328 de novo heart disease patients presented with AF. In contrast to typical presentations of AF in high-income countries, the mean age was 59 ± 18 years. Both those of African ancestry (n = 211; 85.8%) and women (n = 150; 61%) predominated. Perhaps not surprisingly, the most common concurrent diagnosis was any form of HF (n = 138; 56.1%) followed by primary valve disease and/or valvular dysfunction (n = 106; 43.1%). A primary diagnosis of valve disease was made in 71 (28.9%) patients, comprising 51 (20.7%) patients with RHD and 20 (8.1%) patients with degenerative valve disease. Overall, 106 (43.1%) patients presented with clinically significant valve disease/dysfunction. Other diagnoses comprised CAD (n=16; 6.5%), Type 2 diabetes (n = 9; 3.7%), stroke (n = 6; 2.4%) and PPCMO (n = 3; 1.2%). Table 6.3 compares the demographic and clinical characteristics of those presenting in AF according to sex, noting important differences in this regard. Overall, African women were older (average of 4 years) and far more likely to present as obese (73% versus 40%; OR 1.80, 95% CI 1.28–2.52; p < 0.001) and with hypertensive HF (24.7% versus 10.4%; OR 2.37, 95% CI 1.24–4.54; p = 0.006). Alternatively, men were more likely to have a smoking history (OR 2.88, 95% CI 1.92–4.04) and to drink alcohol (OR 2.61, 95% CI 1.83–3.73). Women were less likely to present with a dilated CMO (OR 0.42, 95% CI 0.23–0.76) or CAD (OR 0.38, 95% CI 0.14–1.02).
Table 6.3 AF patients’ demographic and clinical characteristics according to sex.
| Women (n = 150) | Men (n = 96) | All (n = 246) | |
| Demographic characteristics | |||
| Mean age (years) | 60.5 ± 18.5 | 56.2 ± 17.4 | 58.8 ± 18.2 |
| African ancestry | 135 (90%) | 76 (79.2%) | 211 (85.8%) |
| Median (IQR) years in Soweto | 45.5 (37.3–55.0) | 46.0 (30.0–56.5) | 46.0 (35.0–55.0) |
| Risk factor profile | |||
| History of smoking | 46 (30.7%) | 70 (72.9%) | 116 (47.2%) |
| Hypertension | 96 (64.0%) | 52 (54.2%) | 148 (60.2%) |
| BMI (kg/m2) | 29.4 ± 6.7 | 24.6 ± 5.0 | 27.7 ± 6.5 |
| Serum cholesterol (mmol/L) | 4.0 ± 1.2 | 3.9 ± 1.3 | 4.0 ± 1.2 |
| Multiple cardiovascular risk factors | 69 (46%) | 37 (38.5%) | 106 (43.1%) |
| History of alcohol intake | 48 (32%) | 71 (74%) | 119 (48.8%) |
| Clinical presentation | |||
| NYHA Class II or III | 109 (72.7%) | 62 (64.6%) | 171 (69.5%) |
| Dizziness | 93 (62%) | 43 (44.8%) | 136 (55.3%) |
| Palpitations | 96 (64%) | 49 (51%) | 145 (58.9%) |
| Heart rate (beats/min) | 83 ± 20 | 81 ± 21 | 82 ± 21 |
| Systolic BP (mm Hg) | 130 ± 24 | 125 ± 26 | 127 ± 25 |
| Diastolic BP (mm Hg) | 75 ± 13 | 72 ± 15 | 74 ± 15 |
| eGFR | 76 ± 29 | 85 ± 34 | 80 ± 31 |
| Probable aetiology of AF | |||
| Lone AF | 16 (10.7%) | 6 (6.3%) | 22 (8.9%) |
| Valvular AF | 65 (43%) | 42 (43.8%) | 107 (43.5%) |
| Concurrent disease | |||
| Hypertensive HF | 37 (24.7%) | 10 (10.4%) | 47 (19.1%) |
| Idiopathic dilated CMO | 15 (10%) | 23 (24%) | 38 (15.9%) |
| Any type of HF | 87 (58%) | 51 (53.1%) | 138 (56.1%) |
| RHD | 33 (22%) | 18 (18.8%) | 51 (20.7%) |
| CAD | 6 (4%) | 10 (10.4%) | 16 (6.5%) |
| Echocardiography | |||
| LVEF (%) | 53 ± 16 | 48 ± 16 | 51 ± 16 |
| LV systolic dysfunction (LVEF < 45%) | 35 (23.3%) | 41 (42.7%) | 76 (30.9%) |
| LVEDD (mm) | 45 ± 11 | 51 ± 8 | 48 ± 10 |
| LVESD (mm) | 34 ± 11 | 38 ± 10 | 35 ± 11 |
| RVSP >35 mm Hg | 22 (14.7%) | 20 (20.8%) | 42 (17.1%) |
Figure 6.8 shows the age profile of incident case presentations of AF in the Heart of Soweto Study cohort. As expected, there was a sharp rise in case presentations with age. There was <1 case per 100,000 for those aged 15 to 24 years compared with 49.6 cases per 100,000 for those aged ≥75 years. After the age of 25 years, significantly more women presented with AF compared with men (ranging from 46% to 77% in each age group), with the peak incidence rate for men occurring in the 55-to-64–year age group.
Figure 6.8 Age profile of incident case presentations of AF according to sex.
S3.2 Responding to an evolving spectrum of heart disease in Africa
There are many ways to interpret the studies outlined in this section given their profound implications. In particular, the Heart of Soweto Study and the community data supporting it, given its particular and detailed focus on an African urban community in epidemiological transition, requires careful interpretation. Overall, it joins such notable studies as the SABPA Study [22–24], the THUSA Study [20,25,26], and the CRIBSA Study [27–29] in providing critical data to understand the evolving burden of heart disease in sub-Saharan Africa. Notably, the Heart of Soweto represented (and still does) the most comprehensive cohort study of advanced forms of heart disease from sub-Saharan African [87], representing a key barometer for a region in epidemiological transition [36,87]. Dynamic forces such as economic development, erosion of traditional lifestyles, and rural migration are powerful influences among vulnerable communities, and these were evident in the data presented around a “crossroads” between communicable and noncommunicable heart disease in Soweto. Contrary to contemporary expectations [88], the single most prevalent form of heart disease was hypertensive HF. In combination with CAD, the ratio of these “new” communicable forms of heart disease versus historically prevalent forms of the same was almost equal. As noted in the original report [8], these ominous data suggest that if South Africa and other countries in the region continue to experience positive socioeconomic changes with a residual background of infectious disease (including the HIV epidemic—see Chapter 9), there will be a paradoxical increase in the number of relative older individuals (particularly women, but still of working age) who will develop noncommunicable forms of heart disease that will soon surpass the number of relatively younger individuals affected by communicable forms of heart disease. The former will include hypertensive heart disease, CAD, and AF. The latter may well include a latent population of adults who survived the peak epidemic of RHD (noting its close correlation to social deprivation and vulnerability to streptococcal infection) but with residual valve disease and dysfunction in their later years.
The value of these novel data (along with INTERHEART findings) cannot be overstated when compared to the strengths and weaknesses of previously published, population-based studies across the spectrum of CVD [89]. Within the context of a predominant focus on communicable disease and its impact on the burden of disease in sub-Saharan Africa [90] (most notably the impact of HIV infection and HAART [91]), these data highlight the greatest threat will emanate from a complex set of socioeconomic circumstances and the rise of noncommunicable heart disease. Strikingly, unlike high-income countries, those affected by noncommunicable heart disease were relatively young (i.e., working age) and mostly comprised women. A complex interplay between clinical risk and socioeconomic profile (including migrant status) predicates that the pattern and balance of communicable and noncommunicable forms of heart disease will continue to evolve in Soweto and other urban African communities.
Key surveillance studies from two populous (but geographically and culturally) diverse regions of sub-Saharan Africa, confirmed two key factors that will likely drive the future rise of noncommunicable forms of heart disease on the African continent. First, despite the potential threat to their health, African communities remain poorly informed and unaware of cardiovascular risk factors and the threat they pose to their heart health. Second, the potential threat of epidemiological transition toward less healthy lifestyles has already reached many urban communities and now rural communities in the region.
In summary, complex forces are shaping an important transition from communicable to noncommunicable forms of heart disease in Soweto and other urban African communities. Paradoxically, economic development and the control of infectious diseases will still leave a vulnerable portion of the community (notably obese women) at high risk of developing early and aggressive forms of noncommunicable heart disease. Beyond early target organ damage (consistent with previous studies; in an earlier Nigerian study 64%, 41%, and 37% of de novo hypertensive adults demonstrated retinopathy, nephropathy, and LVH, respectively [63]), findings from the Heart of Soweto Study [8,64] support increasing evidence that hypertensive heart disease will emerge as the predominant form of heart disease on the continent. Overall, these data support calls to invest in cost-effective primary and secondary prevention programs/strategies in sub-Saharan Africa [92], with early disease detection and chronic disease programs being an obvious priority [38]. At the same time, there is a need to better understand the underlying dynamics and drivers of epidemiologic transition in different communities across the entire continent.