Introduction: Obesity is a global public health concern associated with multiple comorbidities and adverse health outcomes. Hematologic changes have been reported in individuals with obesity, including leukocytosis and thrombocytosis, often attributed to increases in pro-inflammatory cytokines. However, the prevalence, magnitude, and clinical significance of these changes remain unclear, with data limited to retrospective studies. It is unknown whether abnormalities in the CBC and WBC differential can be attributed solely to obesity, or if they should prompt further investigation for underlying hematologic disorders. To address this question, we conducted patient-level meta-analysis of clinical trials in overweight and obese patients to characterize changes in the CBC and WBC differential associated with increased body mass index (BMI).

Methods: Individual patient data were obtained from five randomized, double-blind, placebo-controlled trials of topiramate in overweight and obese adults from the Yale University Open Data Access program (YODA). Baseline demographics (age, sex, race), clinical (BMI, smoking status), and hematologic (CBC and WBC differential) parameters were pooled at the time of enrolment, before drug intervention. Associations between BMI and each hematologic parameter were examined using multiple linear regression, adjusting for age, sex, and smoking status as potential confounders. Multicollinearity among predictors was evaluated using variance inflation factors (VIF), with all values near 1, indicating minimal correlation between variables and allowing independent estimation of BMI effects. Model assumptions were assessed using residual, Q-Q plots, and Cook's Distance. Regression coefficients from each trial, representing the change in the laboratory value per 1 kg/m2 increase in BMI, were pooled in a two-step meta-analysis to generate summary effects for each parameter.

Results: A total of 2,904 overweight and obese individuals (BMI range 26.5-51.4 kg/m²) had available hematologic data at enrolment. The cohort included 650 (22.4%) males and 2,254 (77.6%) females with a mean age of 46 (±11) years. Racial distribution was 2,764 (95.18%) White, 107 (3.68%) Black, 17 (0.59%) Asian, and 16 (0.55%) other. Mean CBC values for both sexes were within standard laboratory reference intervals.

In most trial-specific regression models, BMI was positively associated with WBC count, red blood cell (RBC) count, platelet count, absolute lymphocyte count (ALC), and absolute neutrophil count (ANC), and inversely associated with hemoglobin and mean corpuscular volume (MCV). Pooled meta-analysis regression coefficients (per 1 kg/m² BMI) were: WBC +0.06 x10⁹/L; RBC +0.01 x10¹²/L; platelet count +0.84 x10⁹/L; ALC +0.01 x10⁹/L; ANC +0.04 x10⁹/L; hemoglobin -0.04 g/L; MCV –0.05 fL. These coefficients represent the average change in each parameter for every 1 kg/m² increase in BMI; i.e., a 10 kg/m² higher BMI was associated with an estimated WBC increase of 0.6 x10⁹/L, RBC increase of 0.1 x10¹²/L, platelet increase of 8.4 x10⁹/L, ALC increase of 0.1 x10⁹/L, and ANC increase of 0.4 x10⁹/L. Although statistically significant, these BMI-related changes were small in magnitude and most parameters remained within standard reference intervals.

Discussion: This large patient-level meta-analysis of overweight and obese adults is the first to comprehensively evaluate hematologic parameters across multiple randomized trials, demonstrating that higher BMI is associated with statistically significant but clinically modest increases in RBC, WBC, and platelet count, and small decreases in hemoglobin and MCV. While these changes may, in rare cases, contribute to values falling above or below standard laboratory reference interval, the magnitude of effect is generally small and unlikely to explain more significant abnormalities. These findings suggest that obesity-related hematologic changes are modest and more substantial deviations should still prompt further evaluation. Hematologists should exercise caution in attributing abnormalities in the CBC and WBC differential solely to elevated BMI, as significant changes are more likely to reflect alternative or concurrent hematologic disorders. Further analyses will evaluate whether changes in BMI within individuals during trials were associated with corresponding changes in hematologic parameters, and these findings will be presented at ASH.

This content is only available as a PDF.
2025
Sign in via your Institution