Author + information
Doherty et al. (1) report on the apparently paradoxical observation that, in a cohort of asymptomatic adults, baseline prevalence of coronary calcium was lower in blacks than in whites, whereas blacks suffered a greater number of events over the follow-up. According to the investigators, these findings indicate that coronary calcium has a “different pathobiologic significance in blacks and whites.”
In fact, rather than surprising, these findings are quite consistent with a very fundamental epidemiologic principle that relates to the approximate relationship between prevalence (P) and incidence (I):
where D is the duration (survival) after disease onset.
Thus, the prevalence ratio obtained in a cross-sectional study (e.g., their baseline examination) will have the following approximate relationship with the incidence ratio (relative risk):
If the duration in blacks and whites is not equal, the prevalence ratio will be a biased estimate of the incidence ratio, the so-called prevalence-incidence bias inherent to many cross-sectional studies (2). The substantially higher co-morbidities, levels of risk factors, and lower access to health care treatment and preventive practices may determine that survival after onset of coronary artery disease is shorter in blacks than in whites, that is what Doherty and co-workers found in their prospective analysis (1). Thus, Dblacks < Dwhites and this may explain why the observed prevalence of coronary calcium is lower in blacks (i.e., Pblacks/Pwhites) even if their risk (incidence) of coronary disease is higher (see also Fig. 1).
The findings by Doherty et al. are analogous to an earlier survey showing that tuberculosis was less prevalent in American blacks than in whites (3). Was this an indication of blacks having lower risk of tuberculosis? As subsequent prospective studies demonstrated (3), tuberculosis incidence was indeed much higher in blacks, while their case-fatality rate was also higher. Thus, the earlier baseline finding was simply a product of the incidence-prevalence bias.
Large prospective studies of the natural history and progression of subclinical atherosclerosis in different ethnic groups, such as the ongoing Multi-Ethnic Study of Atherosclerosis (MESA), will provide answers to some of these questions. In the meantime, however, caution should be taken in interpreting complex racial/ethnic differences as “biological” simply because an observed difference persisted after adjustment for standard risk factors and/or surrogates of socioeconomic status. Residual confounding stemming from imperfect or incomplete adjustment (e.g., imperfect measures of socioeconomic status) is an important limitation. In addition, as discussed in numerous publications (4–6), the use of the biological construct “race” defined solely on the basis of skin color is of questionable validity. The marked genetic heterogeneity within groups such as “blacks,” “whites,” or “Hispanics” explains why this practice has been abandoned by anthropologists, even though biomedical scientists persist in ignoring these calls for caution.
- American College of Cardiology
- Doherty T.M.,
- Tang W.,
- Detrano R.C.
- ↵Szklo M, Nieto FJ, Epidemiology: Beyond the Basics. Gaithersburg (MD): Aspen Publishers, 2000:155.
- Comstock G.,
- Sartwell P.
- Muntaner C.,
- Nieto F.J.,
- O’Campo P.