Mid-arm muscle area and anthropometry predict low birth weight and poor pregnancy outcomes in Tanzanian women with HIV

Human immunodeficiency virus (HIV) infection, especially advanced disease without retroviral treatment, has been associated with greater risk of adverse pregnancy outcomes, including low birth weight (LBW) and prematurity and infant mortality [1, 2]. Poor nutrition is one of the factors that has been associated with disease progression and poor survival in HIV-infected women [3, 4]. Maternal undernutrition may also be an important contributor to poor pregnancy outcomes in HIV-infected women.

In low income settings, anthropometric measures of body mass index (BMI) and weight gain have been preferred for assessing maternal nutrition status in pregnancy, because they are field friendly, feasible in clinical settings and are widely used [5]. In some settings however, these indicators and others, including maternal height, weight and arm circumference have shown poor sensitivity and specificity in primary screening for women at risk of LBW and SGA births [6, 7]. Thus, there is need for further evaluation of the effectiveness, sensitivity and specificity of these anthropometric tools as indicators of risk of adverse pregnancy outcomes [7], particularly in HIV-infected women.

Previous studies have examined associations between maternal anthropometric measures of nutrition status and pregnancy outcomes, mostly in HIV-negative women [8]. In studies, poor maternal weight gain, BMI and mid-upper arm circumference (MUAC) in pregnancy, have been associated with preterm, SGA and LBW [8‐10]. Further, the use of maternal anthropometry indicators of symphysio-fundal height and373 abdominal girth as surrogate measures of birth weight have been piloted in Tanzania [11]. Low BMI, poor weight gain during pregnancy and anemia have also been linked to adverse pregnancy outcomes in HIV-positive women [4, 5, 12‐18].

The development and testing of novel anthropometric tools, such as mid-arm muscle area (MAMA) in resource poor settings is important. MAMA is a reliable measure of muscle atrophy and thickness of subcutaneous fat in lean patients [19]. MAMA may therefore be valuable in evaluating nutrition status for HIV positive women, given that muscle wasting is a marker of HIV disease progression [4]. A recent study in Malawi found that women with higher MAMA had lower odds of having LBW infants [20]. Studies have also shown that MUAC, which is strongly associated with MAMA [21], is associated with body weight, malnutrition and increased risk of death in post-partum women [21‐23].

Studies have not extensively evaluated the performance of MAMA as a predictor of adverse pregnancy outcomes. Further, the performance of MAMA in relation to other anthropometric indicators, in predicting poor pregnancy outcomes for HIV-positive women has not been elucidated. In a prospective study, we assess the performance of MAMA as an alternative to traditional maternal anthropometry indicators in predicting poor pregnancy outcomes in a Tanzanian sub-population.

The sample size of the parent study was 3500 HIV-positive women selected to evaluate whether micronutrient supplementation could reduce the risk of mortality and infectious disease morbidity. The analysis sample was composed of 2369 women with available birth outcomes. The flow diagram for the study is shown in Fig. 1 below.

Table 1 shows the key baseline characteristics of the study population. The mean age of women in the study population was 28.3 years (±5) and on average women were enrolled in the study at the gestational age of 24.6 weeks (±5.5). The mean weight for women at enrollment was 59.1 kg (±10.9) and mean MUAC was 26.1 cm (±3.2). Of the women assessed, 22.9% (N = 522) were primiparous, 86.9% (N = 1977) were married or living with a partner, and 80% (N = 1803) had less than 8 years of education. The majority of the infants in the study were born at full term, with mean gestational age at delivery of 39.4 weeks (±2.5) and mean birth weight for infants was 3.1 kg (±0.5) (not shown in results).

Of the deliveries with a record for birth weight, 7% (N = 161) were LBW babies (less than 2500 g), with 1.1% (N = 25) being very low birth weight babies (VLBW, less than 2000 g). 25.5% (n = 587) of the live births were SGA (below the 10th percentile), and 15.6% (n = 359) were preterm infants.

We evaluated the associations of MAMA, MUAC, BMI, maternal weight and pregnancy weight gain with adverse pregnancy outcomes in HIV-infected women in Tanzania. We found that maternal anthropometric indicators consistently showed a protective effect against adverse pregnancy outcomes. Improved anthropometric measures were associated with lower risk of LBW, SGA and preterm, although MAMA was only marginally associated with preterm birth. Thus, MAMA performed comparably to other maternal anthropometric measures in predicting poor pregnancy outcomes. These results are consistent with findings from a previous study that poor maternal nutrition status may be a risk factor for adverse pregnancy outcomes in HIV-positive populations [20].

Our findings further support the use of maternal anthropometry as a tool for assessing nutrition status and screening for adverse pregnancy outcomes including LBW, SGA, and preterm in the developing world, given low costs, resource requirements and ease of implementation. Simpler tools such as attained weight at 16–20 or 24–28 weeks of pregnancy and change in maternal weight during pregnancy may be more practical screening instruments for LBW and intrauterine growth restriction (IUGR) in primary care settings [34]. In addition, maternal BMI during pregnancy also has predictive properties. For example, given that changes in maternal anthropometric measures over gestation, as well as the timing of measurements may influence pregnancy outcomes, we evaluated associations of changes in MAMA, maternal weight, MUAC and BMI throughout pregnancy (overall slope of the curve) and second trimester slopes (MAMA, maternal weight, MUAC) with pregnancy outcomes. Greater overall changes in maternal weight and BMI and changes in weight in the 2nd trimester were protective against SGA. Similar associations has been found elsewhere [35].

However, these tools have limitations. For women with below average pre-pregnancy weight, the greatest effect size for IUGR prediction is for attained weight at 7 lunar months [34], which may not allow sufficient time for intervention on fetal growth before birth [36]. Further, in low income settings equipment for height measurements may not be available, thus the calculation of BMI may not be convenient. It is thus important to consider tools such as MUAC (in second or third trimesters as it is strongly associated with infant mortality in the first week of life [37]) or the proposed MAMA tool.

MAMA was highly correlated with and performed comparatively to established maternal nutrition indicators of weight gain, BMI and MUAC as a predictor of adverse outcomes. Our findings also provide construct validation of MAMA as a measure maternal nutrition status in our location. Thus, MAMA could also be useful as a measure of nutritional status for HIV positive women during pregnancy in low resource settings. While MAMA involves complex calculations, its key advantage over other indicators is its ability to measure muscle mass in lean patients [19]. This attribute makes MAMA a suitable tool for measurements of wasting in HIV-infected women and for evaluating progression of HIV disease. HIV disease is characterized by increased protein catabolism and decreased muscle mass, and lower muscle mass may be an important risk factor for mortality in people with HIV [38]. Further, MAMA has been associated with mortality in chronic diseases such as chronic obstructive pulmonary disease and in hemodialysis patients [39, 40]. The association of MAMA with mortality could be an even more important consideration given high risk for mortality in HIV-positive women.

Similarly, MUAC could also be a useful tool. MUAC is similarly associated with adverse pregnancy outcomes [41], even in HIV-positive populations [20], yet it is not routinely measured in primary care settings. Further, conducting MUAC measurements may be easier than assessing skinfold thickness and calculating MAMA. Follow-up research is required to further evaluate both the use of MAMA and MUAC indicators, as well as comparing sensitivity and specificity of various cut off points for these indicators in HIV positive women.

Previous studies have found mixed results of associations between maternal weight, arm circumference, abdominal circumference, BMI, and weight gain and LBW and SGA in HIV negative populations [42, 43]. Our findings suggest greater utility for maternal anthropometry in HIV positive populations, where the tracking and measuring of maternal nutrition status is of paramount importance for both maternal health and the health of the offspring.

HIV infection in Tanzania remain high, with prevalence of nearly 6% among women of child bearing age [44]. Although the advent of universal HIV treatment and adoption of option B+ treatment, that is the provision of lifelong antiretroviral (ARV) treatment to all pregnant and breastfeeding HIV-positive women (regardless of CD4 count or clinical stage) [45] may have contributed to improved nutrition status, the coverage of programs including prevention of mother-to-child transmission (PMTCT) of HIV has not reached many women [46]. In our study sample the prevalence of ARV drug use was 20.2%. Coverage of antiretroviral therapy in Tanzanian women has since then increased to 78% [47], and ARV use has been associated with weight gain and improved anthropometric status [48]. However, given that knowledge of HIV status remains at 70% [47], we still concerned about maternal underweight during pregnancy and its effects on pregnancy outcomes among untreated, HIV-infected women. Further, in a study in Tanzania, HIV-positive women, the majority of whom were on ART experienced a greater risk of adverse outcomes (fetal death, preterm delivery and LBW) [49]. Thus, while our study may be dated, our findings may still be relevant in the period of option B+ and other HIV treatment regimes.

This may also be true in other contexts. In a Ugandan study, HIV-positive women on combination antiretroviral therapy with low gestational weight gain had increased occurrence of LBW, preterm births, and other poor birth outcomes [50]. Approximately 15% of the women experienced weight loss during pregnancy and 44.9% were anemic [50]. Finally, protein-calorie malnutrition and food insecurity remain barriers in improving nutrition status and may still hasten HIV disease progression [51]. This suggests that malnutrition is still a problem for women on ART and that it can have consequences for child health, hence the need for addressing maternal undernutrition in pregnancy in this population.

Finally, with the advent of treatment, there are additional considerations that need to be studied to fully understand the impact of treatment on this population. This can be evaluated with anthropometric indicators. The initiation of option B+ programs provides an opportunity to integrate MAMA and other anthropometric tools into routine care and treatment of women with HIV. Given evidence in this study and others that MAMA, MUAC, BMI and weight change in pregnancy are predictive of adverse pregnancy outcomes, these tools can be used to identify women at greater risk of adverse outcomes. Secondly, anthropometric indicators may be useful for identifying women to be targeted by nutrition interventions. For these purposes, anthropometric indicators such as MAMA must be further evaluated.

A limitation of our study was that it did not further evaluate the sensitivity and specificity of MAMA and other anthropometric indicators in the study population. This research question can potentially be addressed in future studies now that we have demonstrated a possible role for MAMA and other indicators in this population. Research is further required to evaluate whether MAMA and other evaluated anthropometric indicators could be feasibly used as screening tools for pregnant women with HIV in different locations and environments. Further, there were fewer repeated anthropometric measures in the third trimester of pregnancy. Thus, we could not evaluate associations between anthropometric measures in the third trimester and adverse outcomes. As a result, the study may have excluded important associations for which the third trimester is most important. Finally, we evaluated gestational age using maternal menstrual history and were not able to verify gestational age with the ultrasound measurements which are more accurate, given their high costs and unavailability in low resource areas [52]. Menstrual history was used as in many limited resource settings, because it is easy and inexpensive to collect, however it is prone to measurement error due to poor maternal recall [53]. Other neonatal data methods e.g. Dubowitz and Ballard scoring systems for estimation of child gestational age may also be accurate, however they may be less reliable in malnourished populations and require technical skills which may not be available [54]. Our study design though observational, is prospective in nature. A major strength of this study was the ability to assess various anthropometric measures simultaneously. To our knowledge this is the first study to compare the performance of MAMA with standard measures of anthropometry in a large sample of HIV-infected women.

MAMA performs comparably to MUAC, maternal weight and BMI as a predictor of LBW and SGA in HIV-infected women. The possible role of MAMA and other indicators in screening HIV-positive women at risk of adverse outcomes should be investigated.