Sibling history of MI and sibling history of stroke in this sample of older adults was associated with a greater prevalence of CVD risk factors, subclinical and clinical disease in the participant. Specifically, the prevalence of CHD was associated with a positive sibling history of MI and marginally associated with a positive sibling history stroke. Prevalent MI was associated with a positive sibling history of MI and was marginally associated with a positive sibling history of stroke. Prevalence of angina was associated with a positive sibling history of MI and a positive sibling history of stroke. No significant associations were observed between sibling history of MI or sibling history of stroke and prevalent stroke. Incident CHD and angina, but not incident MI, were associated with a positive sibling history of MI or a positive sibling history of stroke. A combined positive sibling history (either MI or stroke) was marginally associated with incident MI. Sibling history of MI and sibling history of stroke were not associated with incident stroke.
Other studies have shown a similar association between sibling history and CVD in adults, but these reports have not focused on this older age group (10
). Similar to our findings, a recent report combining the Siblings With Ischemic Stroke Study (Swiss cohort) (28
) and the Umeå cohort (29
) observed a lack of aggregation of ischemic stroke subtypes in affected sibling pairs (30
Given the above association between sibling history of MI and participant prevalent subclinical disease, we were interested in assessing if risk factor differences could be observed between positive and negative sibling history. Our data show associations between CVD risk factor levels in these older adults and sibling history of MI. The associations are weaker for sibling history of stroke. Sibling history of MI and sibling history of stroke were associated with a more abnormal risk profile for major CVD risk factors. Similarly, an association between family history of CVD and more abnormal levels of known risk factors has been observed in children (11
) and also in younger and middle-aged adults (7
Our data suggest that the increased prevalence of CHD in participants with positive sibling histories of MI or stroke persisted after adjustment for known CVD risk factors. Other studies of adult populations have observed that the increased susceptibility of adults with a family history of CHD may be independent of lipids and blood pressure levels (32
We did not observe an association between sibling history of stroke and stroke in this elderly population, despite prior suggestion of a genetic contribution to cerebral susceptibility to ischemia (35
). Likewise, recent data from a population-based cohort of patients with recent transient ischemic attack indicated that family history of stroke does not predict future risk of ischemic stroke (38
). The discrepancy between these findings could possibly be explained by different predisposition in subjects with different ethnicity (37
), stroke subtype (36
), and elderly individuals. Few subjects with prevalent stroke were recruited into this study; these subjects were probably at lower risk of having stroke events. It is also possible that due to a limited number of events, our study did not have sufficient power to detect such associations. However, the size of the effect did not appear to be very strong.
There are a number of limitations of this study. A potential drawback is the lack of medical record validation data on sibling’s disease status. However, other investigators have detected a relatively good concordance (78%) between a reported family history of MI and medical record validation, suggesting that, despite some imprecision, the reported history gives a reasonably good estimate of family history for the diseases we have assessed (39
). It is possible that an ascertainment bias may have occurred with those individuals with a positive sibling history of MI having a greater likelihood of diagnoses (hypertension, diabetes, CHD) than participants with a negative sibling history due to increased awareness of the condition by participants and their medical care providers. It is also possible that recall bias might affect the results to some extent. If differential misclassification occurs, due to more active investigations in participants with positive sibling histories, this could have led to some degree of bias. However, risk factors and events in the CHS cohort were sought prospectively, at regular intervals through the study visits, and it is unlikely that participants had different likelihood of being diagnosed with risk factors for CVD or CVD. Our study emphasizes the clinical importance of sibling history as an easily ascertained risk factor for CVD, and a potential feature able to identify subjects amenable to primary prevention.
Given our definition of sibling history, individuals from large families may have a slightly higher risk of being categorized into the positive history group. We investigated confounding by family size by including a covariate for the number of siblings in all fitted logistic and Cox regression. Family size did not appear to confound the associations between the CVD outcomes and sibling history. The relative change in odds ratios was between 1–2% and the relative change in hazard ratios was between 2–4% after adjustment for the number of siblings. The conclusions drawn from the reported results were unchanged. We were also unable to identify half-sibs or step-sibs in our analyses. Given these relationships are aggregated in our data, it is likely the association between sibling history and outcome would be stronger if the half-sibs and step-sibs were excluded from the analysis if the associations observed were genetic in nature. Including non full siblings in the analyses likely bias our results toward the null.
In addition, we did not attempt to obtain family history data on parents and other CVD events. This was done due to concern about the reliability of self-reported data and the comparability of diagnostic methods for ascertaining disease 40–50 years ago.
We have shown that a positive sibling history of MI and a positive sibling history of stroke in older adults were associated with a significantly worse cardiovascular disease prevalence and incidence, more subclinical measures of CVD, and a more adverse risk factor profile. Given that this older cohort of adults represents the “survivors” from early onset of CVD events, it is interesting to note that sibling history of MI and sibling history of stroke remain associated with a more adverse risk factor profile. We have shown that differences were observed in modifiable factors such as lipoprotein levels in older adults with a positive sibling history of disease. The observed differences in prevalent and subclinical CVD remained after adjusting for the major risk factors, suggesting an independent effect of sibling history of MI and sibling history of stroke beyond the direct effects of the risk factors. These data also suggest that these sibling history factors may be markers for other genetic/familial factors that may be poorly or inadequately measured using traditional CVD risk factors in older adults. These data provide additional guidance in the interpretation of siblings' history and could help reduce barriers to the recognition of positive family history (40
). Accurate history taking and increasing awareness of siblings’ cardiovascular disease, even among older adults, might promote the identification of subjects at risk of cardiovascular disease and potentially improve the access to care and motivate subjects to follow a healthier lifestyle.