Systematic review: Exposure to pets and risk of asthma and asthma-like symptoms.

Studies of exposure to pets and risk of asthma have yielded conflicting results. We performed a systematic review to synthesize the evidence of the effect of exposure to pets in the home on the risk of asthma and asthma-related symptoms. We also assessed differences between the studies as sources of heterogeneity of the results. We conducted a MEDLINE search (until the end of 1999) using the following boolean search command: (asthma[all] OR wheez*[all]) AND (domestic animal*[all] OR pets[all]). The outcome was limited to either diagnosis of asthma or the symptom of wheezing. The exposure of interest was domestic animals in the home. Appropriate temporal relationship was defined as present in studies with either pet keeping within the first 2 years of life, in the past, or exposure to pets preceding the outcome. Thirty-two of the 217 retrieved articles fulfilled the eligibility criteria. Inappropriate time sequence of the exposure and outcome information was an important source of heterogeneity and an indication of potential selection bias. Therefore we analyzed studies focusing on early exposure or ensuring appropriate temporal sequence. The pooled risk estimates for both asthma (fixed-effects odds ratio, 1.11; 95% CI, 0.98-1.25; heterogeneity, P =.04; random-effects odds ratio, 1.09; 95% CI, 0.89-1.34) and wheezing (fixed-effects odds ratio, 1.19; 95% CI, 1.05-1.35; heterogeneity, P =.03; random-effects odds ratio, 1.17; 95% CI, 0.95-1.44) indicated a small effect, which was limited to studies with a median study population age of over 6 years (fixed-effects odds ratio, 1.19; 95% CI, 1.02-1.40; heterogeneity, P =.04; random-effects odds ratio, 1.15; 95% CI, 0.86-1.56; fixed-effects odds ratio, 1.29; 95% CI, 1.12-1.48; heterogeneity, P =.31). In younger children the harmful effect disappeared for wheezing (odds ratio, 0.80; 95% CI, 0.59-1.08; P =.38). Exposure to pets appears to increase the risk of asthma and wheezing in older children. The observed lower risk among exposed than among unexposed young children is consistent with a protective effect in this age group but could also be explained by selection bias.