Aims/design. Reduction in mean birth weight and increased incidence of low birth weight are both associated with exposure to illicit heroin in pregnancy. Many studies examining neonatal outcomes in pregnant heroin users treated with methadone report improvements in birth weight. As a consequence, methadone treatment has become the ‘gold standard’ for the management of the pregnant heroin user. However, not all studies report significant birth weight increases associated with methadone. We undertook a number of meta-analyses on reduction in mean birth weight, and incidence of low birth weight to estimate more precisely the effect of illicit heroin and methadone. Findings. Results showed mean reductions in birth weight associated with heroin use: 489 g (95% CI 284-693 g), compared with methadone: 279 g (229-328 g). Similarly, the pooled relative risk estimate for low birth weight for maternal heroin use was 4.61 (95% CI 2.78-7,65), compared with 1.36 (0.83-2.22) for methadone. Analysis of data on combined heroin and methadone use produced a pooled mean reduction in birth weight of 557 g (403-710 g), with a pooled relative risk estimate for low birth weight of 3.28 (2.47-4.39). Pooling ‘any’ methadone data, regardless of heroin use, produced an estimated reduction in birth weight of 395 g (311-478 g) and a relative risk estimate for low birth weight of 1.90 (1.29-2.81). Combining all data in an ‘any’ opiate use analysis also produced a mean reduction in birth weight of 483 g (386-583 g) and a relative risk estimate for low birth weight of 3.81 (2,57-5.65). Conclusions. The current findings suggest that heroin use while receiving methadone may counteract the birth weight advantage gained from methadone alone. Whether this is due to fetal exposure to heroin plus methadone, to reduced antenatal care, other behavioural and environmental factors associated with concurrent use of heroin and methadone or a combination of these is unclear. Nevertheless, these results challenge the current belief that the pregnant user is always better off receiving methadone than not, and suggests that methadone may not be the appropriate treatment for the pregnant women who continue to use illicit heroin.