Five cohort studies were identified (total procedures 196,819, range 435 to 88,311, with 75,064 knee procedures and 121,755 hip procedures). The number of hospitals in each cohort study ranged from one to 256. Two studies adjusted for confounders (age, sex, American Society of Anaesthesiologists Score, wound class and other factors).
There was a statistically significantly higher risk of severe surgical site infections following knee prosthesis for laminar air flow versus conventional ventilation (RR 1.36, 99% CI 1.06 to 1.74; Ι²=0%; four studies; fixed-effect model); only one individual study had a significant effect. The funnel plot showed no evidence for publication bias for the knee prosthesis analysis. There was a significantly higher risk of severe surgical site infections following hip prosthesis for laminar air flow versus conventional ventilation (RR 1.71, 99% CI 1.21 to 2.41; Ι²=64%; four studies; random-effects model); three individual studies had a significant effect.
A sensitivity analysis that removed the smallest study from the hip prosthesis analysis also found a significantly higher risk of severe surgical site infections for laminar air flow versus conventional ventilation (RR 1.71, 99% CI 1.45 to 2.01; Ι²=25%; three studies; random-effects model) but with much reduced heterogeneity.