Dose-response relationship between maternal caffeine intake and low birth weight/IUGR/SGA (n = 7). Adjusted relative risks (RRs) and 95% CIs (dashed lines) are reported. Caffeine intake was modeled with a linear trend (P-value for non-linearity = 0.89) in a random-effects model. The vertical axis is on a log scale. The open circles represent the effect estimates from each study (the number of circles for a study depends on the number of caffeine intake categories in the study) and the size of the circles is proportional to the precision of the estimates. Mills et al.'s  and Sengpiel et al.'s  studies were not included in this graph as they did not provide sufficient results for categories of caffeine intake. CI, confidence interval; IUGR, intrauterine growth restriction; LBW, low birth weight; SGA, small for gestational age.