Population estimates and determinants of severe maternal thinness in India

Abstract Objective To examine prevalence, risk factors, and consequences of maternal severe thinness in India. Methods This mixed methods study analyzed data from the Indian National Family Health Survey (NFHS)‐4 (2015–2016) to estimate the prevalence of and risk factors for severe thinness, followed by a desk review of literature from India. Results Prevalence of severe thinness (defined by World Health Organization as body mass index [BMI] <16 in adult and BMI for age Z score < –2 SD in adolescents) was higher among pregnant adolescents (4.3%) compared with pregnant adult women (1.9%) and among postpartum adolescent women (6.3%) than postpartum adult women (2.4%) 2–6 months after delivery. Identified research studies showed prevalence of 4%–12% in pregnant women. Only 13/640 districts had at least three cases of severely thin pregnant women; others had lower numbers. Three or more postpartum women aged ≥20 years were severely thin in 32 districts. Among pregnant adolescents, earlier parity increased odds (OR 1.96; 95% CI, 1.18–3.27) of severe thinness. Access to household toilet facility reduced odds (OR 0.72; 95% CI, 0.52–0.99]. Among mothers aged ≥20 years, increasing education level was associated with decreasing odds of severe thinness (secondary: OR 0.74; 95% CI, 0.57–0.96 and Higher: OR 0.54; 95% CI, 0.32–0.91, compared with no education); household wealth and caste were also associated with severe thinness. Conclusion This paper reveals the geographic pockets that need priority focus for managing severe thinness among pregnant women and mothers in India to limit the immediate and intergenerational adverse consequences emanating from these deprivations.


| INTRODUC TI ON
In most low-and middle-income countries (LMICs), severe thinnessdefined by the World Health Organization (WHO) as body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) below 16 1 -among women of reproductive age (15-49 years) continues to persist in selected regions, along with an increasing prevalence of overweight and obesity. Prevalence of severe thinness is estimated at 1.8% among women aged 20-49 years across 60 LMICs. 2 This estimate is consistent over a 10-year time frame for most LMICs. In India, the prevalence of thinness or severe thinness ranges between 2% and 41%. [3][4][5][6][7][8][9][10][11][12] In a study on Asian adult men and women, mortality risk was twice as high among those with very low BMI (<15) compared with very high BMI (>35). 13 Thinness, either mild (BMI [16][17][18].49) or severe (BMI <16) in pregnant women increases the risk of preterm birth, small for gestational age (SGA) neonates, low birthweight (LBW), and infant mortality. [14][15][16][17] However, there is limited evidence on the burden of and risk factors for severe thinness in pregnancy in India. Therefore, it is crucial to understand the prevalence of and risk factors for severe thinness in pregnancy to develop context-specific preventative policies.
Using data from the National Family Health Survey (NFHS)-4 18 and a desk review of literature with a focus on India, the aim of the present study was to examine prevalence, risk factors, and consequences of maternal severe thinness.

| MATERIAL S AND ME THODS
The geographic scope of the current study is India. The study used a mix of analytical methods including a review of the literature on the prevalence and consequences of severe thinness, and secondary analysis of NFHS-4 18 to estimate the prevalence of and risk factors for maternal severe thinness.

| Review of literature
We undertook a desk review of papers published in India between January 2010 and December 2019. Papers were searched in the PubMed electronic bibliographic database using the search terms: pregnant*, undernutrition, thin*, severely thin, low BMI, and India, using a time limit of 10 years (2010-2019). A standardized Microsoft Excel version 13 (Microsoft Corp) data extraction template was used to extract data including prevalence, causes, and consequences, with details of date published, authors, type of study, location, duration, and outcomes of interest.

| Secondary analyses of NFHS-4 (2015-2016)
The prevalence and determinants of severe thinness in pregnancy as well as in the postpartum period (2-6 months) were estimated through analysis of NFHS-4. 18 The NFHS survey followed a twostage, stratified cluster sample in which there were 699 686 women of reproductive age (15- women in the analytical sample for the present analysis comprised the following: (1) <20 weeks of gestation (based on reported gestational age) to avoid misclassification based on BMI cut-offs; (2) height and weight measurements available for calculating BMI; and (3) BMI not flagged as invalid in the NFHS, i.e. BMI is neither <12 nor >90. Among postpartum women, those under 2 months after delivery were excluded to avoid misclassification based on BMI cut-offs; a maximum of 6 months after delivery was adopted as the cut-off to avoid recall bias about the most recent pregnancy with past pregnancies ( Figure 1).

| Variables
Adult BMI cut-offs for thinness (<18.5) and severe thinness (<16) as per WHO were used to estimate the prevalence of thinness among pregnant and postpartum women (20-49 years). 19 For adolescent females (15-19 years) in the sample, BMI for age Z score (BAZ) < -1 standard deviation (SD) and <-2 SD were used to estimate the prevalence of thinness and severe thinness, as BAZ < -2 SD corresponded to BMI of 16.5 as per a recently published analysis of married adolescents under NFHS-4. 20 Disaggregation of severe thinness prevalence estimates either by age or state was not possible due to the small sample size. However, we present the distribution of several cases by age and district.
Based on existing literature, independent variables used in the analysis consisted of personal and household characteristics. primary, secondary, and higher. Household characteristic included caste, access to drinking water, access to toilets, and wealth quintiles. 20 Additionally, for pregnant women or adolescents, variables related to diet such as consumption of milk/curd, pulses/beans, eggs/meat/fish, and dark green leafy vegetables daily, and weekly consumption of fried food and aerated drinks were added. For postpartum women, variables related to diet and access to antenatal care services were added.
Severe thinness and severe underweight are used interchangeably and for this article they mean the same.

| Statistics analysis
National level sampling weight was used during the analysis to maximize the representativeness of the study population. Descriptive analyses were conducted to present characteristics of the study sample for women in pregnancy and the postpartum period. We developed maps depicting district-wise cases to study the variability in the prevalence of severe thinness at the state level. Two logistic regression analyses were carried out to examine the associations between severe thinness and its correlates after adjusting for other covariates, including gestational age.
Data were analyzed using Stata version 15.1 (StataCorp LLC, College Station, TX, USA). P < 0.01, P < 0.05, and P < 0.10 were considered statistically significant. Table 1 presents the research studies on the prevalence of thinness and/or severe thinness among pregnant women in community and facility settings in India and its consequences in the last decade prevalence ranged from 4% to 12%. 6 The other studies provided estimates of thinness including severe thinness, but the definitions for measuring thinness were inconsistent. Thinness, including severe thinness, was measured using BMI <18.5 in five studies, 3-5,7,10 BMI <19.9 in three studies, 8,9,11 and BMI 19.8 as thinness and BMI <18.5

| RE SULTS
as severe thinness in one study. 12 Only one study measured severe thinness using MUAC <21 cm and thinness using MUAC <23 cm. 6 Among the consequences for mothers, maternal death, rate and types of obstetric complications, gestational weight gain, and anemia were investigated. 4,5,7,8 For child outcomes, LBW was the most investigated outcome. [3][4][5]11,12 The odds of LBW were 1.7-2 times higher among severely thin or thin mothers compared with mothers of normal BMI. 4,5,11,12 The majority of the women in the studies were older than 20 years.

| Study sample characteristics
The majority of women were rural residents (n = 30 090) and Hindu (n = 28 486). Half belonged to the 20-24 years age group. Around one-third of the population belonged to socially disadvantaged groups ("scheduled caste," "scheduled tribe," and "other backward classes," as defined by the Government of India in NFHS-4 18 ). About 80% had access to improved drinking water which was piped in, but less than half had access to improved toilet facilities. Table 2 summarizes the characteristics of the study sample.

| Prevalence of severe thinness
The national prevalence of thinness/severe thinness was 4.3% (95%  Access to household toilets was also associated with lower odds of severe thinness (OR 0.72; 95% CI, 0.52-0.99, P < 0.05) ( Table 3).  The nearly twice higher odds of severe thinness in adolescent pregnancies with an earlier parity compared with nulliparous women indicates a need to accelerate efforts to delay marriage and pregnancy in teenaged girls. Parity was not influential in increasing the odds of severe thinness in older mothers. Among diet-related variables used in our model, the overall intake of a combination of foods (i.e. milk/milk products, protein-rich foods, and dark green leafy vegetables) daily was low. It can explain the lack of association with severe thinness in pregnant adolescents and women. However, the lower odds of severe thinness among pregnant adolescents consuming fried foods weekly is a concern since it is indicative of possible "unhealthy" weight gain in this age group. In another UNICEF-National Centre of Excellence and Advanced Research on Diets (NCEARD) report based on the Comprehensive National Nutritional Survey, it emerged that daily and weekly consumption of fried foods was high in the 15-19 years age group. 24 Our findings also reveal that 36% of pregnant adolescents consumed fried foods weekly. Among mothers aged ≥20 years, level of education, poverty, and access to toilets influenced severe thinness. However, among severely thin mothers, usual Anganwadi services received during pregnancy revealed a correlation with increased odds of severe thinness among those receiving supplementary food and nutrition education. More needs to be done in terms of implementation research and service delivery of the usual package of services when pregnant women are severely thin.

CO N FLI C T S O F I NTE R E S T
The authors have no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
VS, AW and AB conceptualized the paper; TC drafted it with contributions from VS, AW. KD conducted the statistical analysis. All other authors reviewed the manuscript and contributed to the interpretation of findings. All authors agreed to the final version.