Maternal MUAC and fetal outcome in an Indian tertiary care hospital: A prospective observational study.

Abstract Studies to date demonstrated the relatedness of mid‐upper arm circumference (MUAC) measurement of pregnant women to their anthropometry/weight. Hence, the objective was to determine whether maternal MUAC at different gestational age predicted birthweight, and if so, to identify which cut‐offs provided the best prediction of low birthweight (LBW) in pregnant women cohort. A total of 928 pregnant women, free of any obstetrical and medical complications known to affect fetal growth, were followed from 20 to 24 weeks' gestation till delivery. Weight, height, and MUAC were determined for the pregnant women, and gestational age along with newborns anthropometry was collected. The mean birthweight was 2.6 ± 0.460 kg. Maternal age, height, weight, MUAC (three time points), gestational age at delivery, and post‐natal weight showed positive correlation with birthweight, crown heel length, and head circumference of the neonates. The cut‐off limit with the best sensitivity–specificity (54.0 and 59.8, respectively) for MUAC was 23 cm, whereas maternal weight of 55 kg had sensitivity and specificity of 62.5 and 59.9 for predicting LBW. Maternal weight of 55 kg and MUAC value of 23 cm had almost similar sensitivity and specificity for predicting LBW. MUAC (≤23 cm) can be considered as a potential indicator of LBW where weighing of pregnant women is not feasible or when presentation for antenatal care is late, especially where pre‐pregnancy weights are not available.

birthweight (BW) should be improved, which in turn is directly dependent on maternal anthropometry (Sen, Roy, & Mondal, 2009). Considerable attention has been focused on the maternal anthropometric measurements as indicators of LBW for identifying women at risk of LBW. Impaired nutritional status of women before conception, short stature, and poor nutrition during pregnancy are important contributing factors of LBW (Sen et al., 2009;Muthayya, 2009;WHO-Provisional agenda item 6.3, 2011).
Maternal anthropometry such as maternal weight (Wt), height (Ht), mid-upper arm circumference (MUAC), and maternal body mass index (BMI) in the first trimester are suggested as good predictors of LBW. Nevertheless, pre-pregnancy body mass index and gestational weight gain are the preferred anthropometric indicators to identify women at risk of producing LBW babies. However, in India, pregnant women begin to visit antenatal clinics after 10 to 12 weeks of pregnancy, and therefore, pre-pregnancy weight may not be available to calculate BMI. Considering various influencing factors, studies have demonstrated that MUAC is closely related to maternal weight, and therefore, it has been suggested as an effective tool for maternal nutrition status screening (Elshibly & Schmalisch, 2008;Tang et al., 2016;Lechtig, 1988). However, knowledge about its changes during the course of pregnancy and the cut-off that could predict LBW is limited (Frison, Kerac, Checchi, & Prudhon, 2016). There is lacuna of established cut-off points of maternal MUAC by taking demographic differences into account. MUAC of <22 cm has been suggested as an indicator of wasting, but MUAC cut-off points to predict LBW are suggested to be different for different regions (Villamor et al., 2002). Yet most studies have measured MUAC just before or after delivery; data on changes in MUAC values during pregnancy are limited (Tang et al., 2016;Ojha & Malla, 2007;Dhar & Bhadra, 2008;Sebayang et al., 2012;Assefa et al., 2012;Shrivastava, Agrawal, & Giri, 2016;Mohanty et al., 2005;Ricalde, Velásquez-Meléndez, Tanaka, & de Siqueira, 1998;López, Calvo, Poy, del Valle Balmaceda, & Cámera, 2011

| Sample size calculation
Assuming a sensitivity of 60% with a similar specificity of 60%, taking a precision of 10% and 95% confidence interval (CI) with 80% power the sample size required was 556. However, expecting attrition of 35% due to long follow-up, the sample size was calculated to 751.

| Participants
Pregnant women in their first and second trimester, willing to participate in the study, were included for the recruitment. Women with gestational age nearing 30 weeks, gestational diabetes, severe anaemia, pre-eclampsia, chronic hypertension, fetal anomaly, rheumatoid arthritis, thyroid and parathyroid disorders, and hepatic or renal or cardiovascular diseases were excluded from the study. Nine hundred and twenty eight (928) pregnant women who fulfilled inclusion criteria were registered after obtaining written informed consent and were followed through during pregnancy till child birth. Of the 928, 615 and 563 turned up for follow-up during 30-34 weeks' and >36 weeks' gestation, respectively; but delivery data such as BW and gestational age were collected from 804 of 928 recruited women. Only singleton deliveries were included in the final analysis of the data (Figure 1). In India, various programmes like "Janani Suraksha Yojana" under the National Rural Health Mission are functional under which women are paid a substantial fund for each ANC (antenatal check) visit to encourage ANCs and institutional deliveries. This enabled us to get good follow-up of the study cohort. In the state where the study was

Key messages
• Though MUAC is identified as a marker to assess nutritional status of pregnant women by WHO, data on optimal cut-off points are lacking especially in Indian context.
• This is the first study to demonstrate relationship between MUAC and maternal weight gain on three serial measurements.
• We found a similar sensitivity-specificity of MUAC for predicting LBW to that of maternal weight, suggesting the validation of MUAC.
• Contributing to validate cut-off point of MUAC measurements to assess poor pregnancy outcomes, which can help in identifying women at risk. conducted, 74.9% of pregnant women had four antenatal care visits to the health centers (Fact sheets-2015-16, NHFS-4, India). All procedures performed in the study were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments.

| Procedures
Anthropometric indicators include maternal weight in kg, height in cm, and MUAC in cm. BMI was calculated by taking pregnant women's weight in kg divided by her height in metres squared. BMI of less than 18.5 was classified as chronic energy deficiency or undernourished.
Wt, MUAC, and skinfolds at four sites were collected at three time points during pregnancy (20-26 weeks' , 30-34 weeks', and >36 weeks' gestation). MUAC was measured in the right arm at the level, midway between acromion and olecranon processes in centimetre, to the nearest decimal place. Triceps, biceps, and subscapular skinfold thickness were measured by trained nutritionists using a Lange skinfold caliper (nearest to 1 mm) according to standardized methods. The average of three measurements was recorded at each site. Maternal body composition was evaluated for all the three times during gestation. Post-natal weight was collected after 24 hr but within 5 days after delivery. Two dedicated project staff, nutritionists were doing the anthropometry after due training and inter/intra rater reliability was done every 3 months to keep the CV below 10%. All the demographic and pregnancy details along with 24-hr diet recall were collected from the participants.
The babies were examined within 24 hr of delivery, and BWs, crown heel length, and head circumference were recorded using the Seca weighing scales (to the nearest 1 g), infanto-metre, and measuring tape (to the nearest 1 cm), respectively. As per the WHO (1995) definition, newborns weighing less than 2.5 kg were considered as LBW neonates. Gestational age was determined by dating the last menstrual period and at the time of recruitment and was corrected by first trimester ultrasonographic findings if the difference exceeded 5 days. SGA neonates were determined by comparing the BW with fetal growth standard by gestation week. Neonates with BW less than 10th percentile of the standard population for gestation week were considered as small for gestational age (Papageorghiou et al., 2014).
SGA is a surrogate marker for identification of newborns with fetal growth restriction.

| Statistical analysis
Data were analysed using SPSS 17.0 for windows IBM, Chicago. Mean and SD were calculated; mothers' weight, height, MUAC, and BMI were ascertained for a BW of 2.5 kg. Odds ratios were computed to assess the risk of LBW between various cut-off points of MUAC and 95% CIs were calculated. Adjusted odds ratios from multivariable regression models were also performed. In addition, sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating characteristic curve (AUC) were calculated for predicting BW outcome for various MUAC cut-offs. Pearson's correlation coefficients between longitudinal measurements of mothers' FIGURE 1 Recruitment flow diagram anthropometry and newborns' anthropometry were done. Significance considered at P value .05.

| Ethical considerations
Institutional ethical committee approval was taken bearing the number 04/I/2014.  Note. Values are mean ± SD. Values in parentheses indicate number of pregnant women. Wt1 and MUAC1-data collected when mean and SD of gestation was 23.6 ± 6.55 weeks; Wt2 and MUAC2-data collected when mean and SD of gestation was 32.6 ± 3.78 weeks; and Wt3 and MUAC3-data collected when mean and SD of gestation was 38.5 ± 2.09 weeks.
Abbreviations: MUAC, mid-upper arm circumference, SGA, small for gestation age; Wt, weight.  Table 3 shows the serial cut-off values and validity indices of weight, height, and MUAC as an indicator of LBW. The best cut-off limit

| Association of MUAC and skinfolds with maternal weight gain and birth outcome
As expected, Wt and MUAC had robust correlations at all the three time points of gestation (r = .890, P = .001; r = .861, P = .001; r = .844, P = .001). Furthermore, Pearson's correlation showed a significant association (r = .554; P = 000) between MUAC change and gestational weight gain during the course of pregnancy in this cohort.
Similarly, there was strong association between MUAC changes and BW. Linear association between maternal MUAC at all three time points during pregnancy (20-24 weeks' , 30-34 weeks', and >36 weeks' gestation) with mean BW of the neonates is depicted in Figure 2. As for skinfolds, Pearson correlation showed a significant association of maternal total body fat and lean body mass at all three time points with neonatal anthropometry such as BW, crown heel length, and head circumference (Table S1).

| DISCUSSION
MUAC cut-off value of ≤23 cm measured during first, second, and third antenatal visits emerged as a good predictor of LBW and SGA neonates in this cohort. Maternal weight and height and as expected gestational age at delivery were related not only to BW but also to crown heel length and head circumference. This is the first study to demonstrate the relationship between MUAC and maternal weight gain on three serial measurements. Pearson's correlation showed association (r = .554; P = 000) between MUAC changes and gestational weight gain during the 15-week period of pregnancy in this cohort.
Pregnant women with lower height, weight, MUAC, lower fat mass, and lean body mass had smaller babies. The best cut-off limit with the highest sensitivity-specificity product for weight was 55 kg and height was 152 cm for LBW.
The incidence of LBW and anaemia in the present study was well within the prevalence range reported from India ( Though weight was the traditional marker, when presenting late for the antenatal checkups and in resource poor settings, MUAC acts as a simple tool to assess poor pregnancy outcomes (Katz, Khatry, LeClerq, West, & Christian, 2010;Lechtig, 1988;Mohanty et al., 2005;Ricalde et al., 1998;Roy & Sen, 2018 (Elshibly & Schmalisch, 2008;Villamor et al., 2002;Ojha & Malla, 2007;Dhar & Bhadra, 2008;Shrivastava et al., 2016;Ricalde et al., 1998;López et al., 2011). Six longitudinal studies that measured MUAC during antenatal visits also reported MUAC as predictor of LBW in pregnant women, similar to the current study. All the six found significantly increased risk of LBW among mothers with low MUAC during pregnancy (Frison et al., 2016;Sebayang et al., 2012;Assefa, Berhane, & Worku, 2012;Mohanty et al., 2005;Kelly et al., 1996;Karim & Mascie-Taylor, 1997 (Dhar & Bhadra, 2008;Elshibly & Schmalisch, 2008;Sebayang et al., 2012). LBW or SGA babies are not only at greater risk of dying than infants of average weight but also at risk of more frequent infections and impaired cognitive development and are more likely to become undernourished children and adolescents (Saugstad, 1981). Evidence is now pointing that LBW/SGA predisposes children to a high risk of diabetes, heart diseases, and other chronic conditions later in life (Barker, 1990). Hence, it is an urgent need to identify pregnant women at risk, to decrease the burden of LBW or SGA babies. MUAC is a simple, easy to conduct marker of maternal nutrition status and therefore has been suggested to identify women at risk of delivering LBW.
BW less than 2.5 kg, defined as LBW, is a poor outcome as a consequence of being born prematurely, having fetal growth restriction or both. Globally, an estimated 20 million births a year are LBW, which is about 20% of all live births, but most of the LBW babies are born in developing countries, and India contributes to 30% of global LBWs (WHO-GNT 2025. There are ongoing nutrition supplementation programmes for adolescent girls and pregnant and lactating women in India. Identifying additional women as high risk based on MUAC might lead to additional allowance of food supplements especially proteinrich foods such as milk and eggs, apart from care and increased follow-up visits for counselling. Knowledge about MUAC changes during the course of pregnancy and the cut-off that could predict LBW is limited (Frison et al., 2016). MUAC of <22 cm has been suggested as an indicator of wasting, but MUAC cut-off points to predict LBW are suggested to be different for different regions (Villamor et al., 2002). Moreover, MUAC collected during any point of time during pregnancy has been proposed to be able to predict LBW.

| CONCLUSION
The current study demonstrated predictive ability of MUAC ≤23 cm at all three time points during pregnancy (20-24 weeks', 30-34 weeks', and >36 weeks' gestation). Maternal MUAC cut-off of ≤23 cm from 20 weeks' gestation up till delivery can be considered for prediction of LBW. However, further studies need to be taken up with a nationally representative sample to validate this method.

ACKNOWLEDGMENTS
We are grateful to the Indian council of Medical Research (ICMR), GoI, for providing the financial support. The authors thank Government Maternity Hospital, Hyderabad, for providing study participants and Mrs. Nancharamma, Bhavani, and Swaroopa Rani for their assistance in sampling. We are grateful to the Indian council of Medical Research (ICMR), GoI, for providing the financial support.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

CONTRIBUTIONS
RH designed the study concept and substantially contributed in executing and writing the manuscript. VD and RB handled the data collection. BV and RS substantially contributed in data analysis. RK, SS, and JJ involved in interpreting the results. RH and VD wrote the manuscript and all the authors reviewed and approved the final manuscript.