Red blood cell count in short-term prediction of cardiovascular disease incidence in the Gubbio population study.

OBJECTIVE: The Gubbio Study is an Italian population study measuring risk factors and incidence for major cardiovascular diseases. This analysis investigates the association between red blood cell (RBC) count, after preliminarily taking into account haematocrit, and incidence of coronary and cardiovascular events.
METHODS: A population sample of 2,469 men and women aged 35-74 years, free from major cardiovascular diseases and in whom RBC count and haematocrit were measured in 1983 along with other standard risk factors, were followed up for 6 years and incidence was estimated for both fatal and non-fatal coronary heart disease (CHD) and all cardiovascular atherosclerotic (CVD) events. Proportional hazards models were solved for the prediction of these events.
RESULTS: In six years 61 CHD hard criteria, 109 CHD any criterion and 149 CVD events were recorded. Preliminarily, both haematocrit and RBC count, two highly correlated variables, were studied to predict CVD events; however, haematocrit did not contribute multivariately, in the overall population and separately in men and women. Age-adjusted rates per 1,000 of the 3 event categories were computed in sex-specific RBC count quintiles (Q) and a difference was observed between Q5 and Q1 (with 5.21 +/- 0.31 and 4.18 +/- 0.23 x 10(6) per microl, respectively) for CHD any criterion (p < 0.07) and CVD (p < 0.05). P on trends was < 0.05 for both end-points. In multivariate models, adjusted for 7 other risk factors, RBC count contributed a weak statistical significance to predict CVD incidence [relative risk (RR) for a 0.5 x 10(6) per microl difference 1.23 with 95% confidence intervals (CI) 1.00- 1.51], whereas its contribution to predict CHD any criterion (RR = 1.19 with CI 0.93- 1.51) and CHD hard criteria (RR = 1.15 with Cl 0.83-1.58) was not statistically significant. Inclusion of blood glucose and presence of diuretics (11.33% of the population) as possible confounders had no major effect although the latter were, as expected, a significant risk factor (RR = 1.90 with Cl 1.28-2.82) which further diluted the CVD predictive role of RBC count (RR = 1.22 with CI 0.99- 1.50).
CONCLUSIONS: Increased RBC count is independently (yet weakly) associated with risk of CVD events in the 6-year follow-up of the Gubbio Study. Longer follow-up is needed before the contributory role of RBC count can be properly assessed to explain CHD incidence.