Thirty-two studies, with 10,598 patients, were included in the review; patients were only included if both biopsy and histology results were available (7,873 women). The results of the quality assessment were not reported, but the authors noted that verification bias was a significant issue, as most studies only conducted excision in women with a positive punch biopsy.
The pooled sensitivity for punch biopsy using a threshold of CIN1+ to diagnose CIN2+ disease was 91.3% (95% CI 85.3 to 94.9) and the pooled specificity was 24.6% (95% CI 16.0 to 35.9), based on data from 25 studies. There was significant between-study heterogeneity for both measures.
The pooled sensitivity for punch biopsy using a threshold of CIN1+ to diagnose CIN3+ disease was 91.1% (95% CI 83.7 to 95.4) and the pooled specificity was 18.2% (95% CI 11.3 to 27.9), based on data from 22 studies. The heterogeneity was not reported.
The pooled sensitivity for punch biopsy using a threshold of CIN2+ to diagnose CIN2+ disease was 80.1% (95% CI 73.2 to 85.6) and the pooled specificity was 63.4% (95% CI 50.9 to 76.7), based on data from 32 studies. There was significant between-study heterogeneity for both measures.
The pooled sensitivity for punch biopsy using a threshold of CIN2+ to diagnose CIN3+ disease was 83.6% (95% CI 74.9 to 89.8) and the pooled specificity was 44.5% (95% CI 34.3 to 55.2), based on data from 27 studies. There was significant between-study heterogeneity for both measures.
For studies where less than 70% of punch biopsies were included in the meta-analyses, and for studies where excision biopsy was performed immediately after punch biopsy (during the same surgical procedure), sensitivity was lower and specificity was higher, than for all studies.