Establishing pregnancy-specific haematological reference intervals in Ghana; a three-center cross-sectional study

Abstract

Background: Abnormal intra-pregnancy haematological variables are associated with adverse feto-maternal outcomes. However, the reference intervals (RIs) employed in sub-Saharan Africa to inform clinical decisions are generally imported. Since RIs are influenced by age, geographical location, and race, we hypothesized that context specific RIs should be established in Ghana to contextualize intra-pregnancy decision making.

Methods: This cross-sectional study retrospectively retrieved data of 333 pregnant women with no known clinically determined intra-pregnancy complications; 22 participants in their first trimester (T1; 1-13 weeks), 177 in their T2 (14-27 weeks), and 132 in T3 (28-41 weeks). RIs for haematological parameters were non-parametrically determined at 2.5th and 97.5th percentiles in accordance with CLSI guidance document EP28-A3c. Two-sample comparisons were undertaken using Wilcoxon rank-sum tests whereas more than two-sample comparisons were undertaken using Kruskal-Wallis test. Statistical significance was set at p <0.05 under the two-tailed assumptions.

Results: In accordance with WHO trimester-specific haemoglobin cutoffs, anaemia prevalence was a moderate (T1: 36.4%; 8/22 & T2: 31.6%; 56/177) to severe (T3:68.0%; 90/132) public health problem. Additionally, 9.3% (31/333) individuals had high gestational haemoglobin levels (Hb >13.0 g/dL). Moreover, haemoglobin (T2: 8.6-14.3 vs T3: 7.5-13.6 g/dL), MCH (T2: 22.5-69.8 vs T3: 21.6-31.9 pg), MCHC (T2: 30.2-51.8 g/L vs T3: 30.5-37.9 g/L), TWBC (T2: 4.0-13.4 vs T3: 4.1-13.0 x 109/L) required trimester specific RIs, compared to RBC (2.8-5.1 x 1012/L), MCV (66.2-100.2 fL), and platelet counts (106.3-388.3 x 109/L) that each required combined reference intervals.

Conclusions: The intra-pregnancy haematological RIs determined have appreciable lower limits; there is the need to determine context-specific thresholds for haematological variables predictive of positive and/or adverse maternal and infant health outcomes.

Fig 1. Haemoglobin levels of participants stratified per socio-demographic variables.

Fig A compares the median haemoglobin levels of participants across the trimesters. Fig B compares the haemoglobin levels of all participants per age categories. Fig C compares the haemoglobin levels of participants in trimester 2 per age categories. Fig D compares the haemoglobin levels of participants in trimester 3 per age categories. Fig E compares the haemoglobin levels of all per the region. Fig F compares the haemoglobin of participants in trimester 2 per study site. Fig G compares the haemoglobin levels of participants in trimester 3 per study site. [All statistical comparisons were undertaken using the Kruskal-Wallis test with Dunn’s multiple comparison correction test; All boxplots were constructed with whiskers indicating minimum and maximum data points].