Methods: We searched CENTRAL; MEDLINE; EMBASE; NIOSHTIC-2; CISDOC and HSELINE up to February 2011.
Results: We included 21 controlled before and after studies with 1447 participants that reported on 29 comparisons. In 15 studies, removal from exposure was compared with continued exposure. Removal increased the likelihood of reporting absence of symptoms (risk ratio (RR) 21.42, 95% confidence interval (CI) 7.20 to 63.77), improved forced expiratory volume (FEV1 %) (mean difference (MD) 5.52 percentage points, 95% CI 2.99 to 8.06) and decreased non-specific bronchial hyper-reactivity (standardised mean difference (SMD) 0.67, 95% CI 0.13 to 1.21). In six studies, reduction of exposure was compared with continued exposure. Reduction increased the likelihood of reporting absence of symptoms (RR 5.35, 95% CI 1.40 to 20.48) but did not affect FEV1 % (MD 1.18 percentage points, 95% CI -2.96 to 5.32). In eight studies, removal from exposure was compared with reduction of exposure. Removal increased the likelihood of reporting absence of symptoms (RR 39.16, 95% CI 7.21 to 212.83) but did not affect FEV1 % (MD 1.16 percentage points, 95% CI -7.51 to 9.84).
Discussion: There is evidence, although of very low quality, that removal from exposure improves asthma symptoms and lung function compared with continued exposure. Reducing exposure also improves symptoms, but seems not as effective as complete removal. However, removal from exposure is associated with an increased risk of unemployment, whereas reduction of exposure is not. The clinical benefit of removal from exposure or exposure reduction should be balanced against the increased risk of unemployment.