Objective. We aimed to assess the impact of cleaner cooking fuels (liquefied petroleum gas, LPG) on childhood neurodevelopment compared with usual cooking practices in Bangladesh.

 Material and Methods. The Poriborton trial was a two-arm parallel cluster randomised controlled trial in Bangladesh that randomised pregnant women to receive LPG cylinders and stoves (intervention) compared with usual cooking practices. In the extension of the trial, a subset of surviving infants and their mothers received continued LPG fuel until 24 months of age. Neurodevelopmental outcomes (cognitive, language, and motor composite scores) were assessed using the Bayley Scales of Infant and Toddler Development-IV at 12 and 24 months of age. Generalised linear mixed models were used to compare the outcomes between groups on an intention-to-treat basis.

Results. Of 1629 children recruited (intervention 816, comparison 813), data from 1622 (intervention 815, comparison 807) were analysed (51% male, 18% preterm; mean [SD] gestational age at birth 38.3 [2.2] weeks). Baseline characteristics were comparable across groups, except that the proportion of bigger households (more than 5 members) was higher in the intervention group (47% vs 38%; p < 0.001). Cognitive (mean [SD] 100.91 [12.63] vs. 100.15 [12.82], p=0.23; 87.46 [10.43] vs 87.12 [10.05], p=0.51) and motor (100.58 [10.71] vs. 101.47 [10.92], p=0.1; 93.24 [9.62] vs. 93.25 [9.79], p=0.97 composite scores, respectively at 12 and 24 months, were similar in the intervention and comparison groups. However, children in the intervention group had significantly lower language scores at 12 months (89.27 [11.05] vs. 90.50 [11.82], mean difference -1.22, 95% CI -2.33 to -0.10, p=0.03) but not at 24 months (95.56 [14.78] vs. 95.85 [14.12], p=0.68) than the comparison group.

Conclusion. LPG for household cooking from early gestation through to 24 months of age did not impact neurodevelopmental status in children at 12 and 24 months.