101 - Neonatal Skin-to-Skin Care and Autism Screening Risk in Very Preterm Infants

Publication Number: 1093.101

Molly F.. Lazarus, Harvard University, Concord, MA, United States; Virginia A. Marchman, Stanford University School of Medicine, Stanford, CA, United States; Melissa Scala, Stanford University School of Medicine, San Francisco, CA, United States; Fargol Yeganeh Fathi, Weill Cornell Medicine, New York City, NY, United States; Heidi M. Feldman, Stanford University School of Medicine, Palo Alto, CA, United States; Katie Travis, Weill Cornell Medicine, New York City, NY, United States

Background: Preterm infants often experience reduced positive tactile input in the NICU and are at elevated risk for autism. Skin-to-skin care (kangaroo care) supports neurodevelopment and bonding, yet natural variation in its use has not been examined in relation to autism risk.

Objective: To test whether cumulative exposure to skin-to-skin care or swaddled holding during NICU hospitalization is associated with autism risk at 18-30 months in very preterm infants.

Design/Methods: We retrospectively studied 94 infants born < 32 weeks' gestation and admitted to a single level IV NICU. Cumulative minutes of skin-to-skin care and swaddled holding, recorded in the medical chart, were averaged per NICU day. Skin-to-skin eligibility was determined by standardized clinical readiness criteria and systematically charted by trained bedside nurses. Autism risk was defined as a positive Modified Checklist for Autism in Toddlers-Revised with Follow-up (M-CHAT-R/F; score ≥3). Logistic regression tested associations between tactile exposures and autism risk, adjusting for gestational age, medical risk, insurance status, and sex. Sensitivity analyses controlled for COVID birth cohort and concurrent cognitive/language delay.

Results: Twenty-three infants (24%) screened positive for autism risk (Table 1). Elevated-risk infants received significantly less skin-to-skin care than low-risk peers (mean 15.2 vs 26.6 min/day, p < .05). In adjusted models (Table 2), greater skin-to-skin care was associated with lower odds of a positive autism screen (β = -1.74, p = .017, OR = 0.175, 95% CI: 0.04-0.50). Overall, each tenfold increase in daily exposure (e.g., 0 to 10 min/day or 5 to 50 minutes/day) was linked to an 82% reduction in autism risk (OR = 0.175, 95% CI: 0.038-0.50). Figure 1 illustrates the predicted probability of autism risk by skin-to-skin exposure. Results remained significant in sensitivity analyses. Swaddled holding was not associated with autism risk.

Conclusion(s): Very preterm infants who received more skin-to-skin care were less likely to screen positive for autism risk at 18-30 months, independent of clinical and sociodemographic factors. Findings highlight early tactile experience as both a potentially modifiable protective factor and an early behavioral marker of autism risk. This study should be considered preliminary, given the modest sample size, correlational design, and reliance on a screening tool rather than diagnostic evaluation. Further research using larger, prospective cohorts and gold-standard autism assessments is needed to establish causality, clarify mechanisms, and guide intervention strategies.

Table 1. Clinical and Socio-demographic Variables for Very Preterm Infants Divided by M-CHAT-R/F Risk group status at 18-30 months of age (n=94).
(a)Presence of bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), sepsis and/or necrotizing enterocolitis (NEC), all serious complications of preterm birth; b Adjusted for degree of prematurity; (c) Standard score on composite language and cognitive assessments ≤ 85, n=72. Note. Comparisons were conducted using Wilcoxon rank-sum tests for continuous variables due to non-normal distributions (gestational age, birthweight, skin-to-skin care, and swaddled holding)
* p<.05


Table 2. Logistic regression predicting autism risk (n=94)
(a) Presence of bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), sepsis and/or necrotizing enterocolitis (NEC), all serious complications of preterm birth; (b) Skin-to-skin exposure was log-transformed because Box-Tidwell testing suggested some non-linearity in the relationship between raw exposure values and the log-odds of autism risk (p = .07; values < .05 indicate violation).

Figure 1. Predicted Probability of Autism Risk by Skin-to-Skin Exposure.