The authors would like to thank Siegler and Mackuli for their comments on our meta-analysis and for the contribution of their data. The data they present reinforces the finding shown in the prospective studies included in the meta-analysis that the colposcopically directed punch biopsy does have a non-ignorable false-negative rate.[3-6] This false-negative rate was underestimated in the majority of retrospective studies because, typically, only women with a punch-biopsy-proven abnormality would go on to have an excisional biopsy. Moreover, Siegler and Mackuli demonstrate the occurrence of undercalling cervical intraepithelial neoplasia (CIN) lesions in punch biopsies that appear to be more severe in subsequent excisions performed because of suggestive symptoms or discordance between cytology and the histological interpretation of the punch biopsy. This observation is in agreement with findings of the meta-analysis showing a lower pooled sensitivity of CIN2+ punch biopsy for an outcome of CIN2+ (80.1%, 95% CI 73.2–85.6%) compared with the sensitivity of CIN1+ punch biopsy for an outcome of CIN2+ (91.3%, 95% CI 85.3–94.9%).
We echo the appeal for the collection of more data on the performance of colposcopy and the colposcopically directed punch biopsy so as to accurately determine the level of missed disease and to investigate whether this has a clinical implication. Coordination and linking between population-based cytology, colposcopy, pathology and cancer registries would be extremely helpful in avoiding certain biases inherent to clinical patient series.