BACKGROUND: In adults, heart rate variability triangular index (HRVi) is a highly reproducible measure of heart rate variability (HRV), which makes it more suitable for use in longitudinal studies. Although normative data have been published for newborns, studies in infants with pathological conditions are lacking. METHODS: From 1997 to 2000, within the first 4 days of life, we prospectively evaluated HRVi in cocaine-exposed asymptomatic newborns (N = 97) by Holter monitoring. Their data were compared with infants from two control groups (one with no in utero drug exposure, N = 102; the other with exposure to alcohol, nicotine, or marijuana but no cocaine, N = 111). RESULTS: In assessing concordance between and within operators for HRVi, the intraclass correlations were 0.983 (95% CI: 0.958, 0.994) and 0.997 (95% CI: 0.984, 0.999), respectively. Infants with in utero cocaine exposure had significantly (P < 0.0001) lower HRVi than those exposed to other drugs and to no drugs in utero. If abnormal HRVi is defined as < fifth percentile for the no drug exposed group (HRVi < 8), 10% of the cocaine-exposed newborns, in contrast to 2% in each of the control groups (P = 0.003) had abnormal values. CONCLUSION: HRVi is a reliable measure to study heart rate variability in newborns. Asymptomatic infants with in utero cocaine exposure have lower HRVi. Our study supports the clinical use of an abnormal HRVi as a value < 8 for newborn infants.