Abnormal platelet count is an independent predictor of mortality in the elderly and is influenced by ethnicity

Abstract

Even though alterations in platelet counts are presumed to be detrimental, their impact on the survival of patients has not been studied in large cohorts. The prevalence of thrombocytopenia and thrombocytosis was examined in a large inner city outpatient population of 36,262 individuals aged ≥65 years old. A significant association with shorter overall survival was found for both thrombocytopenia (HR=1.45; 95% CI: 1.36-1.56) and thrombocytosis (HR=1.75; 95% CI: 1.56-1.97) when compared to the survival of patients with normal platelet counts. This effect persisted across all ethnic groups. However, African-Americans (non-Hispanic Blacks) with either thrombocytopenia or thrombocytosis were at significantly lower risk compared to non-Hispanic Caucasians (HR=0.82; 95% CI: 0.69-0.96 and HR=0.70; 95% CI: 0.53-0.94, respectively). Furthermore, Hispanics with thrombocytosis were found to have a lower mortality risk compared to non-Hispanic Caucasians with thrombocytosis (HR=0.60; 95% CI: 0.44-0.81). A value of <125,000 platelets per microliter was a better prognostic marker for non-Hispanic Blacks and these subjects with this platelet count had similar overall survival to that of Caucasians with a value of <150,000 per microliter. In conclusion, thrombocytosis and thrombocytopenia are independently associated with shorter overall survival in elderly subjects and this effect is modified by ethnicity. Using different thresholds to define the association of thrombocytopenia and thrombocytosis with overall mortality risk among non-Hispanic Blacks may, therefore, be warranted.

Figure 1.

Overall survival curves stratified by platelet count after adjustment for all covariates including gender, ethnicity, anemia, neutropenia and age-adjusted Charlson comorbidity index. Both thrombocytopenia and thrombocytosis independently and significantly decreased overall survival compared to that of patients with normal platelet counts (P<0.01).

Figure 2.

LOESS regression model showing the association between platelet count (horizontal axis) and the log relative hazard ratio for overall mortality (vertical axis) for each ethnicity after adjusting for gender, anemia, neutropenia and age-adjusted Charlson comorbidity index. The shaded areas around each curve delimit the 95% confidence intervals.