Maternal malnutrition, including both undernutrition and overnutrition, adversely affects maternal and fetal health and requires reliable assessment during antenatal care. Conventional use of body mass index (BMI) for nutritional assessment in pregnancy has limitations and can be difficult to implement in low-resource, community settings. The mid-upper arm circumference (MUAC) is a simple, inexpensive and easily interpretable anthropometric measure that may be better suited for community-based screening. This study aimed to assess the diagnostic accuracy and potential utility of MUAC for screening maternal malnutrition in a rural community setting in eastern India.
This investigation is a secondary analysis of data from a community-based cross-sectional study. The specific objective reported was to evaluate the diagnostic performance of MUAC compared with BMI for detecting maternal undernutrition and overnutrition among pregnant women.
Data were collected at selected health sub-centres (Ayushman Arogya Mandir) within the rural field practice area of a medical college in Ranchi district, Jharkhand. A total of 337 pregnant women were randomly selected from antenatal care registers and enrolled after providing informed consent. Pregnant women with chronic medical conditions such as hypertension, diabetes mellitus and hypothyroidism were excluded from participation.
Anthropometric measurements, including MUAC and parameters required to calculate BMI, were obtained using standardised procedures. Nutritional classification based on MUAC was compared to classification by BMI. The analysis evaluated MUAC as a screening test against BMI as the reference measure, with diagnostic performance quantified by standard metrics and statistical measures described below.
The diagnostic evaluation reported in the source included sensitivity, specificity, overall accuracy, positive predictive value (PPV), negative predictive value (NPV), Cohen Kappa for agreement, simple linear regression and receiver operating characteristic (ROC) curve analysis to identify optimal MUAC cut-offs for undernutrition and overnutrition.
According to BMI-based classification, 21.4% of the enrolled pregnant women were underweight. MUAC-based classification identified 38.3% of participants as having moderate malnutrition and 0.6% as having severe acute malnutrition. These descriptive results demonstrate a higher proportion classified as malnourished by MUAC categories compared with BMI-defined underweight prevalence.
For the task of detecting undernutrition (as defined by BMI), MUAC demonstrated the following diagnostic performance:
These values indicate that MUAC was highly sensitive in this sample for identifying women who were underweight by BMI, with a high NPV and moderate specificity and PPV.
Receiver operating characteristic curve analysis identified an optimal MUAC cut-off of 22.8 cm for detecting undernutrition in this cohort. For identifying overnutrition, the ROC analysis identified a MUAC cut-off of 24.8 cm. These cut-offs were derived from the study data and are presented as the thresholds that optimised discrimination in this sample.
Agreement between MUAC and BMI classifications was assessed using Cohen Kappa. The observed agreement was described as moderate, with a Cohen Kappa value of 0.509 and a p-value < 0.001, indicating statistically significant concordance between the two methods but not perfect agreement.
The authors conclude that MUAC may serve as a useful and effective screening tool for maternal malnutrition in community and low-resource settings where conventional anthropometric assessment is challenging to perform. The high sensitivity and NPV suggest MUAC could be particularly valuable for ruling out undernutrition when used in antenatal community screening programs. The simplicity and low resource requirement of MUAC make it a practical option for field use.
The authors note the need for further studies to establish robust MUAC cut-offs across different populations and gestational ages. The present report is based on a secondary analysis of cross-sectional data from a single rural region in eastern India; generalisability to other populations and settings was not addressed in detail in the source. The source states that all data produced in the study are available upon reasonable request to the authors.
Ethical approval for the original study was granted by the Institutional Ethics Committee, Rajendra Institute of Medical Sciences, Ranchi, Jharkhand, India. The authors declared no competing interests in the source document.