was clinically perceived as
substantial, statistically it was moderate in the overall
meta-analysis (I²=38%), and no evidence of heterogeneity
(I²=0%) was reported in the subanalyses of studies on
cardiosurgical patients (fi gure 4), and those assessing
chlorhexidine at concentrations of 0·12% and 2%
(fi gure 3). Heterogeneity is an inherent problem in
systematic reviews and meta-analyses.3