Data about the maternal exposure to different risk factors in mothers of all cases and controls was recorded using a pretested questionnaire. Information included sociodemographic profile of the mother and her family; obstetric history of the mother, especially information about previous births/abortions; antenatal services obtained by the mother, including antenatal clinic (ANC) registration, antenatal visits and checkups, tetanus toxoid injections, consumption of iron and folic acid tablets, information relating to heavy work done, rest timings etc exposure to possible risk factors like hypertension, exposure to any form of tobacco, any other systemic disease; results of blood investigations (especially hemoglobin percentage plus any other result) done within 15 days prior to the delivery; weight of the mother before delivery in kg rounded to two decimals; and height in cm. This information was cross checked with the available records such as ANC cards and case sheets to minimize the recall bias.

Adequate antenatal care was considered when the pregnant woman was registered at any time, had at least three antenatal checkups, was adequately vaccinated against tetanus, had consumed at least 100 tablets of iron and folic acid, was not involved in hard work, and had taken adequate rest during pregnancy (minimum 2 hours sleep during day and 8 hours sleep during night). Low socioeconomic status was considered if the family of mother belonged to a below poverty line family. Weight gain was calculated by subtracting weight of the mother at 12 weeks or before from weight of the mother at term, considering negligible weight gain up to 12 weeks of gestation [5]. Maternal exposure to tobacco meant use of any tobacco product such as tobacco or gutkha chewing, cigarette or bidi smoking, or any other form of smoking.

All the pairs were divided into four subclasses depending on the exposure of case and/or control to a risk factor, as follows: (a) number of pairs in which both, case and control, were exposed to the maternal risk factor; (b) number of pairs in which only case and not control was exposed to the maternal risk factor; (c) number of pairs in which only control and not case was exposed to the maternal risk factor and (d), number of pairs in which neither case nor control was exposed to the risk factor. Matched odds ratio with 95% confidence interval was calculated using the method suggested by Liddell [6]. McNemar chi square test, z test and t test were used to compare the qualitative and quantitative exposures in cases and controls, where applicable. Conditional logistic regression analysis was applied to find out the effect of the potential confounders. StatsDirect software was used for analysis of data.

Total babies screened for birthweight were 2382. Number of LBW babies born during the study period was 638 (26.78 %). Out of them 274 LBW babies were included in the study (Fig.1). Table I shows the distribution of various factors among cases and controls.

Fig. 1 Flow chart showing process of case selection.

TABLE I	Maternal Exposure to Various Risk Factors in Low Birthweight Babies and Controls

This multicenter matched pair case control study was done to identify the maternal risk factors associated with full term low birth weight babies. Proportion of low birth weight babies was 26.8%, which is more than the prevalence of LBW (21.5%) observed in National Family Health Survey (NFHS-3) [7]. This was expected as the study was carried out in tertiary care hospitals where many of the pregnant women are referred from the peripheral hospitals because of high risk pregnancy.

This study has shown that full term low birth weight was significantly associated with inadequate antental care, pre-delivery weight ≤55 kg, height ≤145 cm, weight gain ≤6 kg, spacing <36 months, maternal exposure to any form of tobacco hypertension, low socio-economic status, and anemia.

Spacing and weight gain could not be included in the conditional logistic regression analysis because there were only 53 and 82 pairs in which data were available. Imputation method to deal with these missing values was not used, as spacing is not applicable at all for primigravida. There were 192 (70.07%) pairs in which data related to weight gain in case and/or control was missing. The confounding effect of maternal education was probably due to its association with low socioeconomic status.

Thus findings of this study emphasizes the need for improving the quality and utilization of antenatal care, nutritional education to improve the weight gain during pregnancy, spacing, avoidance of tobacco, and prevention and proper management of risk factors like anemia and hypertension.

Acknowledgment: Dr Mrunal Patil, Dean, NDMVP Medical College, Nashik for allowing us to carry out the study and providing the required manpower. Dr Balaji Almale and Dr Sandip Gundare for helping in data collection, feeding and analysis.

Contributors: SSM conceived and designed the study. He also analyzed and interpreted the data; GM and RD coordinated data collection in Government Medical College, Nashik and were involved in manuscript writing; SY coordinated the data collection in Civil Hospital Nashik and MT and KP in NDMVPS Medical College, Nashik. They were involved in critical evaluation and necessary modifications in the contents of the manuscript. All the authors approved the final manuscript.

Funding: None; Competing interests: None stated.