Reply to Comparison of cervical cancer screening results among 256,648 women in multiple clinical practices.

We appreciate Dr. Castle's interest in our recent article,1 which evaluated 8.6 million women screened for cervical cancer in >10,000 clinical practices. Our intent was to elucidate the real‐world performance of screening options for biopsy‐proven cervical intraepithelial neoplasia of grade 3 and cancer in women aged 30 to 65 years, recognizing that adherence to guidelines in cervical cancer screening is incomplete.2

The advantage of our study compared with the ATHENA (Addressing the Need for Advanced HPV Diagnostics) trial3, 4 and the Kaiser Permanente study5 is that it better represents true clinical practice. Because our study included women in whom a biopsy was preceded by a Papanicolaou test and human papillomavirus (HPV) test, our patient population was composed of women presenting for rescreening as well as screening that might be outside of recommended guidelines. Thus any bias, as suggested by Dr. Castle, is mitigated. Bias would also be mitigated by the fact that nearly equal numbers of women with negative Papanicolaou test results (66,478 women) and HPV‐negative test results (64,870 women) were evaluated.

We agree with Dr. Castle that the risks and benefits of cervical cancer screening should be evaluated over a “screening lifetime.” The small number of cancer cases (8 cases) identified in the ATHENA trial limits its use in assessing lifetime cancer risks. The Kaiser Permanente study,5 in which >400 cervical cancer cases were identified, offers a more robust data set for such analysis. It demonstrated a statistically significant increase of 57% in the 3‐year cumulative risk of cervical cancer associated with a negative HPV‐only test (0.11%) versus a negative cotest (0.07%) (P=.03).6 Understanding how this increased risk observed with a one‐time HPV‐only test will translate into lifetime risk is important before new screening guidelines are adopted.

In our study, HPV‐only testing missed 98 of 526 cervical cancers (18.6%), which is significantly more than were missed by cotesting (29 cancers; 5.5%) (P <.0001).1 In this context, it is important to note that verification bias is a limitation of all routine clinical practice cancer screening data sets that tends to increase the apparent sensitivity of the screening tests, including HPV‐only and cytology.7 Therefore, our results1 could actually be an understatement of the risk inherent in HPV‐only testing.

Finally, we agree with Dr. Castle that a single cotest will be more sensitive and provide a higher level of protection from future disease than either test alone. Our real‐world study supports cotesting as the most effective cervical cancer screening method for women aged 30 to 65 years.8, 9

FUNDING SUPPORT

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

Drs. Blatt, Rabin, Kennedy, and Luff are full‐time employees of Quest Diagnostics.