Introduction
Low birth weight (LBW), defined as birth weight <2500 g, is a leading driver of neonatal mortality globally, accounting for an estimated 60%–70% of all neonatal deaths [1]. Beyond mortality, LBW is strongly associated with impaired neurodevelopmental outcomes, childhood morbidity, stunting, and adult‑onset noncommunicable diseases [2, 3]. An estimated 15.5% of all births worldwide are LBW, representing >20 million births annually, with >95% occurring in low‑ and middle‑income countries (LMICs) [1, 4]. Preterm birth and small‑for‑gestational‑age (SGA) are the two principal etiological pathways to LBW [4, 5].
SGA is defined as birth weight below the 10th percentile or >2 standard deviations below the population mean, for a given gestational age [6]. In LMICs, the burden of SGA disproportionately exceeds that of preterm birth, yet context‑specific evidence to guide prevention remains limited [5, 7].
Established maternal risk factors for SGA include undernutrition and low prepregnancy weight [8], iron‑deficiency anemia [9], chronic hypertension, and hypertensive disorders of pregnancy [10, 11]. Short maternal stature is recognized as a biological marker of intergenerational nutritional deprivation [12]. Emerging evidence implicates maternal overweight in SGA risk, potentially mediated through vascular and metabolic complications [13].
In Sri Lanka, substantial gains in maternal mortality have been achieved over recent decades; however, maternal and child malnutrition persist, particularly in the North Central, Northern, and Eastern provinces [12, 14, 15]. The Anuradhapura district recorded an LBW prevalence of ~17% in 2011 (1966 of 11,560 live births) [16] and consistently reports the highest LBW rates nationally [14]. Despite this burden, no matched case‑control study has previously characterized independent maternal risk factors for SGA in this population.
This study aimed to identify maternal anthropometric, obstetric, and antenatal risk factors independently associated with SGA births in Anuradhapura district, Sri Lanka, to inform targeted, evidence‑based antenatal care strategies.
Methods
Study design and setting
A retrospective case‑control study was conducted using pregnancy records maintained by Public Health Midwives (PHMs) in 13 areas across two Medical Officer of Health (MOH) areas in Anuradhapura district, Sri Lanka. PHMs are community‑based maternal and child health workers each responsible for populations of 3000–5000 persons. Data were extracted from PHM records spanning January 2014 to December 2017. This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [17].
Case and control definitions
Cases were neonates with birth weight below the 10th percentile for gestational age (SGA). Controls were neonates with birth weight between the 10th and 90th percentile (AGA). All SGA infants registered with participating PHMs during the study period were eligible as cases. Neonates from multiple pregnancies, those born after 43 completed weeks, large‑for‑gestational‑age infants, and stillbirths were excluded. Two AGA controls were individually matched to each case (1:2 ratio), yielding 136 cases and 272 controls (N = 408). Matching was performed by gestational age category: extremely preterm (<30 + 0 weeks), preterm (30 + 0–36 + 6 weeks), term (37 + 0–41 + 6 weeks), and postterm (>= 42 + 0 weeks). Within each category, controls born in the same gestational week as the case were preferentially selected, and all controls were restricted to those born within one year of the corresponding case.
SGA ascertainment
Gestational age was calculated from the first day of the last menstrual period, with corroborating ultrasound measurements used when available (<20 weeks). Birth weight percentiles were determined using a computer‑based reference chart based on the Hadlock fetal weight equation [18, 19] calibrated to a Sri Lankan population mean birth weight of 3140 g at 40 weeks of gestation [20].
Data extraction
A structured data extraction form was developed based on a systematic review of global literature on SGA determinants. Variables captured included maternal sociodemographic characteristics (age, education, occupation, parity, and marital status), anthropometric measurements (prepregnancy weight, height, body mass index (BMI), and gestational weight gain), obstetric history (previous LBW delivery, miscarriage, subfertility, cesarean section, and consanguinity), and antenatal factors (folic acid supplementation, anemia, gestational hypertension, preeclampsia, antepartum hemorrhage, and mode of delivery). All records were assigned anonymized identification numbers.
Statistical analysis
Data were entered and analyzed using IBM SPSS Statistics version 24. Categorical variables were examined using Pearson chi‑square test or Fisher exact test when expected cell counts were <5. Variables with P < 0.10 in unadjusted analyses were advanced to multivariable logistic regression. Pairwise Spearman rank correlations among candidate variables all yielded coefficients <0.20, confirming mutual independence.
Given collinearity between prepregnancy weight, height, and BMI, two separate multivariable models were constructed: Model 1 included prepregnancy weight and height; Model 2 included prepregnancy BMI. Both models adjusted a priori for infant sex, maternal age, parity, and educational level. Adjusted ORs with 95% CIs are reported. Model fit was evaluated using the Hosmer–Lemeshow goodness‑of‑fit test. Statistical significance was set at P < 0.05 (two‑tailed).
Per the AMA Manual of Style, all P values are reported without a leading zero, with results reported as statistically significant when P < 0.05 and as statistically nonsignificant when P >= 0.05.
Ethical considerations
Ethical approval was granted by the Ethics Review Committee of the Faculty of Applied Sciences, Rajarata University of Sri Lanka (approval date: January 30, 2017; permit number: ERC/FAS/2017/01). Individual informed consent was waived as the study involved secondary analysis of de‑identified, routinely collected PHM records, in accordance with applicable national ethical guidelines. All procedures complied with the Declaration of Helsinki (2013 revision).
Results
Study population
A total of 408 PHM records were analyzed (136 SGA cases and 272 AGA controls). Of all infants, 51.7% (n = 215) were male and 46.4% (n = 193) were female (1.9% gender data missing). Participants were drawn from MOH Area A (53.7%; n = 219) and MOH Area B (46.3%; n = 189). All mothers were married. The SGA prevalence in the study population was 5.4%. Of SGA infants, 57.4% (n = 78) were below the 10th percentile, 14.7% (n = 20) below the 5th percentile, and 27.9% (n = 38) below the 3rd percentile. Approximately 11% of SGA infants were preterm; no extremely preterm or postterm infants were identified. Maternal and neonatal characteristics are presented in Table 1.
Table 1
Maternal and neonatal clinical characteristics of the study sample (N = 408).
| CHARACTERISTIC | CONTROLS (AGA), N = 272 | CASES (SGA), N = 136 | TOTAL, N = 408 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| MEAN (SD) | MIN | MAX | MEAN (SD) | MIN | MAX | MEAN (SD) | MIN | MAX | |
| Birth weight (g) | 2806 (319) | 1446 | 3600 | 2257 (329) | 700 | 2760 | 2623 (413) | 700 | 3600 |
| Gestational age (weeks) | 38 (2) | 31 | 41 | 39 (2) | 30 | 41 | 38 (2) | 30 | 41 |
| Maternal age (years) | 28 (6) | 16 | 41 | 27 (5) | 17 | 44 | 28 (6) | 16 | 44 |
| Prepregnancy weight (kg)* | 52.7 (10.5) | 34.0 | 85.0 | 49.1 (11.5) | 30.5 | 83.6 | 51.6 (10.9) | 30.5 | 85.0 |
| Maternal height (cm)** | 155 (6) | 142 | 172 | 152 (6) | 139 | 180 | 154 (6) | 139 | 180 |
| Prepregnancy BMI (kg/m2) | 21.9 (4.2) | 13.6 | 37.3 | 21.1 (4.8) | 12.7 | 37.2 | 21.7 (4.4) | 12.7 | 37.3 |
| Gestational weight gain (kg) | 9.7 (4.1) | 1.0 | 23.6 | 9.5 (4.5) | 1.7 | 22.0 | 9.7 (4.2) | 1.0 | 23.6 |
[i] Abbreviations: AGA, average‑for‑gestational‑age; BMI, body mass index; SD, standard deviation; SGA, small‑for‑gestational‑age. *48 records (11.8%) missing prepregnancy weight. **12 records (2.9%) missing height data.
Unadjusted associations
Chi‑square analysis identified significant unadjusted associations between SGA and pre‑pregnancy weight (P < 0.001), maternal height (P = 0.007), and prepregnancy BMI (P = 0.003). A history of a previous LBW infant was also significantly associated with SGA (P < 0.001). Mode of delivery differed significantly between groups (P = 0.012), with cesarean section more frequent among SGA cases (41.9%) than AGA controls (29.4%). Anemia (P = 0.067) and absence of folic acid supplementation (P = 0.062) showed borderline associations. No significant associations were found for parity, previous miscarriage, gestational weight gain, gestational hypertension, or antepartum hemorrhage (all P > 0.10). Complete results are presented in Table 2.
Table 2
Unadjusted univariate analysis of maternal and antenatal factors associated with SGA (N = 408).
| VARIABLE | TOTAL (N = 408) N (%) | SGA CASES (N = 136) N (%) | AGA CONTROLS (N = 272) N (%) | P VALUE |
|---|---|---|---|---|
| Maternal anthropometric factors | ||||
| Prepregnancy weight (kg) | <0.001a | |||
| <50 kg | 165 (45.8) | 68 (59.1) | 97 (39.6) | |
| >= 50 kg (reference) | 195 (54.2) | 47 (40.9) | 148 (60.4) | |
| Maternal height (cm) | 0.007a | |||
| <= 150 cm | 118 (29.2) | 52 (38.8) | 66 (24.4) | |
| 151–160 cm | 229 (56.7) | 69 (51.5) | 160 (59.3) | |
| >160 cm | 57 (14.1) | 13 (9.7) | 44 (16.3) | |
| Prepregnancy BMI (kg/m2) | 0.003a | |||
| <18.5 (underweight) | 93 (26.0) | 41 (36.0) | 52 (21.3) | |
| 18.5–24.9 (normal; reference) | 185 (51.7) | 45 (39.5) | 140 (57.4) | |
| >= 25 (overweight/obese) | 80 (22.3) | 28 (24.6) | 52 (21.3) | |
| Obstetric history | ||||
| Parity (primiparous) | 170 (40.9) | 63 (46.3) | 107 (39.3) | 0.177 |
| History of subfertility | 13 (3.1) | 7 (5.1) | 6 (2.2) | 0.136 |
| Previous LBW infant | 74 (17.8) | 40 (29.4) | 34 (12.5) | <0.001a |
| Previous miscarriage | 64 (15.4) | 21 (15.4) | 43 (15.8) | 0.923 |
| Previous cesarean section | 52 (12.5) | 14 (10.3) | 38 (14.0) | 0.294 |
| Consanguinity | 13 (3.1) | 7 (5.1) | 6 (2.2) | 0.136 |
| Antenatal and delivery factors | ||||
| Anemia in pregnancy | 111 (26.7) | 45 (33.8) | 66 (25.1) | 0.067 |
| Folic acid supplementation (yes) | 274 (65.9) | 83 (61.0) | 191 (70.2) | 0.062 |
| Gestational weight gain | 0.471 | |||
| Below recommended | 170 (47.8) | 58 (51.3) | 112 (46.1) | |
| Within recommended | 133 (37.4) | 37 (32.7) | 96 (39.5) | |
| Above recommended | 53 (14.9) | 18 (15.9) | 35 (14.4) | |
| Antepartum hemorrhage | 9 (2.2) | 4 (2.9) | 5 (1.8) | 0.489b |
| Gestational hypertension | 11 (2.6) | 4 (3.0) | 7 (2.6) | 1.000b |
| Mode of delivery (cesarean section) | 137 (33.6) | 57 (41.9) | 80 (29.4) | 0.012a |
[i] Abbreviations: AGA, average‑for‑gestational‑age; LBW, low birth weight; LSCS, lower segment cesarean section; SGA, small‑for‑gestational‑age. aP < 0.05 (statistically significant). bFisher exact test applied.
Multivariable logistic regression
Both multivariable models confirmed independent associations between maternal anthropometric factors and SGA (Table 3). In Model 1, prepregnancy weight <50 kg (adjusted OR 2.18; 95% CI, 1.28–3.69; P = 0.004) and maternal height <= 150 cm (adjusted OR 1.98; 95% CI, 1.14–3.45; P = 0.015) were significant independent predictors. In Model 2, both underweight BMI <18.5 kg/m2 (adjusted OR 2.24; 95% CI, 1.27–3.94; P = 0.005) and overweight BMI >= 25 kg/m2 (adjusted OR 1.95; 95% CI, 1.04–3.64; P = 0.036) were significantly associated with SGA. Across both models, a prior LBW delivery was the strongest predictor (adjusted OR 3.87; 95% CI, 1.98–7.57 in Model 1 and 2.01–7.47 in Model 2; both P < 0.001). Consanguinity, folic acid supplementation, and anemia did not reach significance in either adjusted model.
Table 3
Multivariable binary logistic regression: adjusted odds ratios for maternal risk factors of SGA (N = 408).
| MODEL | VARIABLE | ADJUSTED OR | 95% CI LOWER | 95% CI UPPER | P VALUE |
|---|---|---|---|---|---|
| Model 1 | Prepregnancy weight <50 kg (ref: >= 50 kg) | 2.18 | 1.28 | 3.69 | 0.004a |
| Height < = 150 cm (ref: 151‑160 cm) | 1.98 | 1.14 | 3.45 | 0.015a | |
| Height > 160 cm (ref: 151–160 cm) | 0.82 | 0.37 | 1.82 | 0.615 | |
| Previous LBW infant | 3.87 | 1.98 | 7.57 | <0.001a | |
| Consanguinity | 1.95 | 0.69 | 5.58 | 0.210 | |
| Folic acid supplementation | 1.44 | 0.84 | 2.49 | 0.186 | |
| Anemia in pregnancy | 1.37 | 0.77 | 2.41 | 0.285 | |
| Model 2 | BMI < 18.5 kg/m2 (ref: 18.5–24.9 kg/m2) | 2.24 | 1.27 | 3.94 | 0.005a |
| BMI >= 25 kg/m2 (ref: 18.5–24.9 kg/m2) | 1.95 | 1.04 | 3.64 | 0.036a | |
| Previous LBW infant | 3.87 | 2.01 | 7.47 | <0.001a | |
| Consanguinity | 1.61 | 0.59 | 4.44 | 0.354 | |
| Folic acid supplementation | 1.33 | 0.78 | 2.28 | 0.295 | |
| Anemia in pregnancy | 1.32 | 0.76 | 2.32 | 0.325 |
[i] Abbreviations: BMI, body mass index; CI, confidence interval; LBW, low birth weight; OR, odds ratio; SGA, small‑for‑gestational‑age. Both models adjusted a priori for infant sex, maternal age, parity, and educational level. Reference categories: weight >= 50 kg; height 151–160 cm; BMI 18.5–24.9 kg/m2. aStatistically significant (P < 0.05).
Discussion
This retrospective case‑control study identified five independent maternal risk factors for SGA births in Anuradhapura district, Sri Lanka: low prepregnancy weight, short maternal stature, underweight BMI, overweight BMI, and a prior LBW delivery. These findings are broadly consistent with global evidence linking maternal body composition to fetal growth restriction [21–24] and align with earlier Sri Lankan research documenting maternal undernutrition and pregnancy‑induced hypertension as determinants of adverse birth outcomes [8].
Associations between maternal undernutrition (low prepregnancy weight, short stature, and underweight BMI) and SGA are mechanistically well‑established: inadequate prepregnancy energy and nutrient reserves restrict placental development and fetal substrate delivery [8, 21, 22]. Short maternal stature is a recognized biological marker of cumulative intergenerational nutritional deprivation and childhood poverty, both prevalent in the Anuradhapura district [12, 25].
The finding that overweight BMI (>= 25 kg/m2) also elevated SGA risk warrants careful interpretation. While some studies report null or protective effects of maternal overweight on SGA [23, 24], others document increased risk mediated through hypertensive disorders, gestational diabetes, or placental vascular insufficiency [13, 25]. Our study lacked statistical power to examine these mediators directly, but the observed association is consistent with the double burden of malnutrition increasingly documented in South Asian LMICs, where overweight coexists with micronutrient deficiency and compromised metabolic health [26].
The strongest predictor in both models was a prior LBW delivery (adjusted OR approximately 3.87). This reflects both genetic determinism and recurrent socioeconomic and nutritional disadvantage driving intergenerational SGA risk [24, 27]. It reinforces a biological–socioeconomic feedback loop in which maternal growth restriction perpetuates adverse birth outcomes across generations [12, 28].
Consanguinity, anemia, and absence of folic acid supplementation showed borderline univariate trends that did not persist in adjusted models [29–32], likely reflecting insufficient power to detect modest effects. The higher cesarean section rate among SGA cases (41.9% vs 29.4%; P = 0.012) reflects obstetric decision‑making in response to identified fetal growth restriction rather than a causal antecedent of SGA.
Strengths and limitations
Strengths include the matched case‑control design with gestational age category matching, use of a Sri Lankan population‑specific birth weight reference, inclusion of two demographically diverse MOH areas, and the relatively large community‑based sample. Retrospective use of PHM records provided real‑world programmatic data.
Limitations include missing data (11.8% missing prepregnancy weight; 2.9% missing height), potential gestational age misclassification from last menstrual period‑based dating, and inability to capture dietary assessment, socioeconomic status, maternal workload, or micronutrient status. Residual confounding cannot be excluded. The retrospective design precludes causal inference, and restriction to one district limits generalizability.
Policy implications
These findings support an integrated, life‑course approach to maternal health. Preconception and early antenatal screening should routinely assess weight, height, and BMI, with standardized referral for both underweight and overweight women to nutritional counseling and supplementation. Women with a prior LBW delivery are a high‑priority group for enhanced antenatal surveillance. Structural drivers of chronic maternal undernutrition in this predominantly rural district require intersectoral policy action addressing food security, female education, and agricultural workloads.
Conclusions
Maternal undernutrition, short stature, overweight BMI, and a prior LBW delivery are significant independent predictors of SGA in Anuradhapura district, Sri Lanka. The co‑occurrence of undernutrition and overweight as independent risk factors reflects the double burden of malnutrition in this LMIC setting. Intergenerational risk pathways, particularly a prior LBW delivery, represent the most actionable target within existing antenatal care infrastructure. Prospective studies with larger samples and comprehensive nutritional and socioeconomic data are needed to build the causal evidence base for scalable interventions.
Acknowledgments
The authors sincerely thank all Public Health Midwives and Medical Officers of Health of the Anuradhapura district for their generous support and cooperation. We are grateful to all research assistants who contributed to data collection and management, and to the Department of Health, Sri Lanka, for facilitating access to community health records.
Funding
This research received no specific funding from any public, commercial, or not-for-profit funding agency. The study was self-funded by the authors. No funder had any role in study design, data collection, analysis, interpretation, or the decision to publish.
Competing Interests
The authors have no competing interests to declare.
Data Accessibility Statement
The datasets underlying the results of this study are derived from de‑identified, routinely collected Public Health Midwife pregnancy records in Anuradhapura district, Sri Lanka, and are not publicly available owing to institutional data governance requirements. Anonymized data may be made available from the corresponding author upon reasonable request, subject to the approval of the relevant institutional authority. Requests should be directed to: dumindaguruge@gmail.com.
Author Contributions
GGND: Conceptualization, methodology, project administration, resources, supervision, validation, writing, review, and editing.
JE: Conceptualization, data curation, formal analysis, investigation, methodology, software, writing original draft, writing, review, and editing.
OSJ: Data curation, investigation, visualization, writing original draft.
KA: Writing, review, and editing (critical revision for important intellectual content).
HL: Conceptualization, methodology, resources, supervision, validation, writing, review, and editing.
All authors read and approved the final manuscript.