Leukocyte counts and lymphocyte subsets in relation to pregnancy and HIV infection in Malawian women

Problem We investigated leukocyte and lymphocyte subsets in HIV‐infected or HIV‐uninfected, pregnant or non‐pregnant Malawian women to explore whether HIV infection and pregnancy may act synergistically to impair cellular immunity. Method of study We recruited 54 pregnant and 48 non‐pregnant HIV‐uninfected women and 24 pregnant and 20 non‐pregnant HIV‐infected Malawian women. We compared peripheral blood leukocyte and lymphocyte subsets between women in the four groups. Results Parturient HIV‐infected and HIV‐uninfected women had more neutrophils (each P<.0001), but fewer lymphocytes (P<.0001; P=.0014) than non‐pregnant women. Both groups had fewer total T cells (P<.0001; P=.002) and CD8+ T cells (P<.0001; P=.014) than non‐pregnant women. HIV‐uninfected parturient women had fewer CD4+ and γδ T cells, B and NK cells (each P<.0001) than non‐pregnant women. Lymphocyte subset percentages were not affected by pregnancy. Conclusion Malawian women at parturition have an increased total white cell count due to neutrophilia and an HIV‐unrelated pan‐lymphopenia.


| INTRODUCTION
In the Global Burden of Disease 2010 study, 254 700 deaths were attributed to maternal conditions, accounting for 7.3% of global deaths in women aged 15 to 49 years. Of these, 21 900 were estimated to be due to maternal sepsis 1 and recent reports indicate that as many as 11% of global maternal mortality is due to sepsis. 2  the HIV/AIDS pandemic since the early 1990s, this decline slowed to 1.4% per year for the period 1990-2008. 2 From 1990 to 2000, the maternal mortality ratio (MMR) increased in many countries in sub-Saharan Africa with large HIV burdens; in Malawi, the reported MMR doubled in that decade from 606 to 1397 per 100 000 live births. 3 With the implementation of antiretroviral programmes, this trend has been reversed, and in 2011, the MMR in Malawi had fallen to 422. 3 Nevertheless, despite availability of ART, 20.5% (56 100) of maternal deaths worldwide in 2010 were HIV-related. 1 In South Africa, nationwide confidential enquiries into facilitybased deaths among pregnant women found that the MMR was almost 10-fold higher among HIV-infected women (328/100 000) than non-infected women (34/100 000). 5 A key question pertains to the underlying cause of these HIV-related maternal deaths.
Sepsis and other infections secondary to compromised immunity could be responsible. 6 In the South African series, 53% (2102/3959) of maternal deaths were due to infection, with 28% of these attributed to pregnancy-related sepsis, often post-partum, and 72% to non-pregnancy-related infections such as pneumonia and tuberculosis. Overall, 90% of the pregnancy-related sepsis and 96% of those non-pregnancy-related infections occurred in HIV-infected individuals. 5 Pregnancy itself is a well-recognized cause of altered immunity, particularly cellular immunity 7 affecting CD4 + T lymphocytes, 8 the same cells that are targeted by HIV. We therefore hypothesized that HIV infection and pregnancy may act synergistically to impair cellular immunity, reducing CD4 counts and potentially increasing susceptibility to infection. To explore this possibility among African women, we performed a prospective cross-sectional study of HIV-infected and HIV-uninfected pregnant and non-pregnant Malawians attending a health centre in Blantyre and compared leukocyte and lymphocyte subsets at parturition with those in non-pregnant women.

| Study area and population
We recruited pregnant and non-pregnant women between 26 infected non-pregnant women were aged between 60 and 68, which is above the child-bearing age and were included in the analysis to increase the sample size for this group. All HIV-infected non-pregnant women were not aware of their HIV status prior to being recruited into the study, and thus, none were receiving HAART. However, all HIV-infected parturient women were taking nevirapine for different durations during their pregnancies.

| Investigations
HIV testing was performed using two rapid test kits; Determine (Abbott Laboratories, Japan) and Unigold (Trinity Brotch, Dublin).
Thick and thin blood smears on slides were prepared by standard methods for Plasmodium falciparum parasite detection. Total white cell count (WCC) and percentages and absolute counts of neutrophils, lymphocytes and monocytes were determined using a HMX Haematological Analyser (Coulter, USA).

| Immunophenotyping
Immunophenotyping of blood samples by flow cytometry and lymphocyte subset identification was performed according to Table   S1. In brief, 25 μL of EDTA blood was mixed with 1 μL of each antibody and incubated in the dark at room temperature for 15 minutes.

| Ethical approval
Ethical approval for this study was obtained from College of Medicine Research and Ethics Committee (COMREC) in Blantyre, Malawi.

| Statistical analysis
Participants were grouped according to their pregnancy and HIV status. The median and range were determined for age and absolute and percentage leukocyte and lymphocyte subset proportion in each group and are reported in the table. 10 th and 90th percentiles weredeterminedforuseintheFigures.TheKruskal-Wallisequalityof-populations rank test was used to identify overall differences between groups for each subset. Pairwise differences between two groups were then determined using the two-sample Wilcoxon ranksum (Mann-Whitney) test. A P value of <.05 was considered statistically significant at 95% level of confidence. All statistical analyses were performed using Stata version 14 (StataCorp 2015).    Table 1) were higher. B, NK and γδ T-cell counts were also lower in theHIV-infectedgroup (Fig.S1D,Table1).

| For HIV-uninfected Malawian women, those at parturition have a higher WCC and neutrophil count, but lower lymphocyte count, than nonpregnant women
We

| Among pregnant women at parturition, HIV-infected individuals have a lower CD4 + T-cell count, and a higher CD8 + T-cell count, than HIVuninfected women
To

| For HIV-infected women, pregnant women have higher WCC and neutrophil counts, but lower lymphocyte counts compared to non-pregnant women
We investigated the impact of pregnancy on Malawian women liv- HIV-infected pregnant women had significantly (P<.001) higher WCC and neutrophil counts, but lower (P=.001) lymphocyte counts than non-pregnant HIV-infected women (Table1 and Fig. S3A).

Absolute counts of T cells and T-cell subsets were reduced in HIV-
infected women, as in HIV-uninfected women, but the reduction was less marked and the reduction in CD4 + T cells was not significant.
Pregnant HIV-infected women continued to have lower CD4 + T cells and higher CD8 + T cells than HIV-uninfected women, both as absolute and percentage counts. There were no significant differences in CD4 + T cells, expressed either as a percentage or as absolute counts, between the two groups of HIV-infected women. Overall, when expressed as a percentage, there were no differences between the pregnant and nonpregnant women for all subsets except for B cells (P=.003) (Fig.S3D).

| HIV-infected and pregnant women have higher WCC and neutrophil counts, but lower mean lymphocyte, T-cell, CD4+ T, NK and B-cell counts compared to HIV-uninfected and nonpregnant women
Finally, to investigate the combined effect of simultaneous pregnancy and HIV infection, we compared the absolute cell counts and percentages in HIV-infected parturient women with those in HIVuninfected non-pregnant women. HIV-infected pregnant women had higher (P<.001) WCC and neutrophil counts, but lower lymphocyte countscomparedtoHIV-uninfectednon-pregnantwomen (Fig.S4A, Table 1). A similar picture was observed when these were presented aspercentages (Fig.S4B).WiththeexceptionofCD8 + T-cell counts, which did not differ between the two groups, HIV-infected pregnant women had lower T cells, CD4 + T, NK and B-cell counts compared toHIV-uninfectednon-pregnantwomen (Fig.S4C).Whenpresented as a percentage, HIV-infected pregnant women had a higher proportion (P=.048) of CD8 + T cells compared to the HIV-uninfected nonpregnantwomen (Fig.S4D,Table1).

| DISCUSSION
Our study firstly investigated only pregnancy-associated changes in include non-pregnant women. 25 The study found that CD4 + and CD8 + absolute counts were higher post-partum than in pregnancy, while CD4 + and CD8 + percentages did not change. 25 Recent studies carried out in SSA [11][12][13] have also reported that HIV infection has the same the other previous studies conducted in SSA. [11][12][13]25 It is worth mentioning that CD4 + and CD8 + T cells have been observed to be stable when presented as a percentage of the total lymphocyte counts than when they are presented as totals counts 19,23 and this is similar to what we and others have reported before for HIV-uninfected children during childhood. 26,27 Therefore, percentage CD4 + and CD8 + T-cell values may have better application than absolute counts for assessing immunodeficiency during pregnancy with an added advantage that they can be more easily and accurately performed on flow cytometers. 28,29 Thereareseverallimitationstothiswork.Firstly,weonlyexamined blood from pregnant women at the time of parturition but a longitudinal study tracking changes in individual women throughout pregnancy and post-partum 9 would provide a more comprehensive understanding of immunological changes in pregnancy. Secondly, although we investigated leukocyte and lymphocyte subset numbers and percentages, we did not examine their function and this could be crucial additional information. Thirdly, ideal comparison between the two HIV-infected groups could have been achieved if they were known to be taking similar HAART 30 and for the same period. Lastly, although the overaged (7/20) HIV-infected women might not be considered representative controls for the HIV-infected pregnant group, we had previously shown that the different leukocyte and lymphocyte subsets do not differ greatly between the ages of 18 years and beyond 60 years. 26

| CONCLUSION
We have examined the separate and combined effects of pregnancy and HIV infection on the cellular compartment of the immune system in Malawian women by measuring leukocyte and lymphocyte subsets.
Although changes are seen that can be attributable to the interaction of both factors, our general finding is that the effect of pregnancy is the dominant factor for each individual immune parameter, apart from CD4 + and CD8 + T-cell counts. Contrary to our initial speculation, pregnancy does not appear to exacerbate the CD4 + T-cell lymphopenia of HIV infection, and therefore, simple immunophenotyping, such as that undertaken in this study, cannot easily explain the high levels of infection and mortalityamongpregnantwomeninSSA.Furtherstudiesarerequired to explore this and will ideally follow longitudinally a cohort of women of childbearing age and study cell percentage, numbers and function.

ACKNOWLEDGMENTS
We thank Grace Mwimaniwa, Meraby Mfunsani and Paul Pensulo for collectingbloodsamplesfromtheparticipants,FelanjiSimukondafor doing the full blood counts and all individuals who participated in this study. We are also grateful to the Gates Malaria Partnership, funded bytheBillandMelindaGatesFoundation,andtheWellcomeTrustfor funding this study.

AUTHORS CONTRIBUTIONS
WLM, JM and CAM conceived the study. WLM, JM and EG performed the investigations. JM and WM analysed the data. WLM, MEM, JM and CAM wrote the report. All authors contributed to the study design and reviewed the report.