Predictors and reference values of CD4 and CD8 T lymphocyte counts in pregnancy: a cross sectional study among HIV negative women in Zimbabwe.

OBJECTIVE

To identify predictors and define reference values for T lymphocyte subsets in HIV negative pregnant black women.

DESIGN

Cross sectional study.

SETTING

Edith Opperman Martenity Hospital, Harare, Zimbabwe.

STUDY POPULATION

1113 HIV negative women 22 to 35 weeks pregnant registering for routine antenatal care.

METHODS

A questionnaire was used to collect demographic and obstetric data. CD4 and CD8 T lymphocyte counts were determined by manual immunocytochemistry. Concentrations in serum, of retinol, beta-carotene, ferritin, folate and 1-antichymotrypsin were also measured. Multiple linear regression analysis was employed to identify and estimate effects of potential predictors.

MAIN OUTCOME MEASURES

CD4 and CD8 T lymphocyte levels, demographic, obstetric data and micronutrient status.

RESULTS

Predictors of CD4 counts were gestational age, serum retinol and season. CD4 counts declined by 25 (95% confidence interval [CI]; 11 to 40; p = 0.001) cells/L for each week's increase in gestation among women with low serum retinol, while low serum retinol was independently associated with lower CD4 counts (-127; 95% CI, -233 to 20 cells/L; p = 0.02) at 35 weeks gestation. The late rainy season was associated with higher CD4 counts (137; 95% CI, 67 to 207 cells/L; p < 0.001). CD8 counts were higher in women with low serum folate (87; 95% CI, 6 to 166 cells/L; p = 0.036) and were slightly higher in gravida 4+ compared to gravida one to three. Reference values of CD4 but not CD8 count and percentage markedly differed from flow cytometry values of pregnant and non-pregnant women in developed and developing countries reported in the literature, even after controlling for the differences in methods of T lymphocyte subset immunophenotyping.

CONCLUSION

Gestational age, gravidity, micronutrient status and season influence T lymphocyte subset levels and need to be considered when designing clinical management and intervention strategies for pregnant women. The data underscores the need for local reference values.