OUP user menu

High-Risk HPV Testing in Women 30 Years or Older With Negative Papanicolaou Tests
Initial Clinical Experience With 18-Month Follow-up

Michael J. Thrall MD, Donna K. Russell CT(ASCP), Michael S. Facik CT(ASCP), Jorge L. Yao MD, JulieAnn N. Warner MD, Thomas A. Bonfiglio MD, Ellen J. Giampoli MD
DOI: http://dx.doi.org/10.1309/AJCPAZV88VIFZSFD 894-898 First published online: 1 June 2010

Abstract

Cervical screening with combined cytology and high-risk human papillomavirus (HR-HPV) detection has been approved for women 30 years or older. We investigated the clinical use of cotesting for women with negative Papanicolaou tests. Follow-up cytology, HR-HPV test, and biopsy findings were identified during an 18-month period. In 1 year, 2,719 cotests from 2,686 women were identified; 146 were positive for HR-HPV. Among women with positive HR-HPV testing, 120 had follow-up, including 70 with repeated cotesting, and 3 had high-grade dysplasia identified (2.5% of women with follow-up). In 1,334 women with initial double-negative cotest results who had repeated cytologic testing within 18 months, 2 high-grade dysplasias were found (0.1%).

The vast majority of cotest results are double-negative. Among tests that show HR-HPV positivity, the prevalence of underlying high-grade dysplasia is low. About half of all women who undergo cotesting receive follow-up that is not in accord with published guidelines.

Key Words:
  • Human papillomavirus
  • HPV
  • Papanicolaou test
  • Negative for intraepithelial lesion or malignancy
  • NILM

The emergence of robust, affordable, and widely available human papillomavirus (HPV) assays for high-risk types, most notably the Hybrid Capture II (HCII) test (Digene, Gaithersburg, MD), has opened up new possibilities for the detection of women at risk for cervical cancer despite negative Papanicolaou (Pap) tests. Since high-risk (HR)-HPV testing is considerably more sensitive than Pap testing for the presence of cervical dysplasia, although also much less specific, a combination of the 2 tests can be used to identify women at low risk for cervical cancer who do not need annual screening. In 2004, the US Food and Drug Administration approved the use of HPV cotesting along with cervical cytology for women 30 years or older. Age 30 years was used as a cutoff because infections are less frequent and less transient in older women as compared with women in their teens and twenties,1 increasing the positive predictive value of a positive HR-HPV test result for high-grade dysplasia in older women. Because of the high sensitivity of HPV testing,2 women with negative results for both tests could move to a 3-year screening interval with minimal risk owing to the 99% negative predictive value of combined negative cytologic and HR-HPV results for high-grade dysplasia.3 Owing to the increased costs associated with cotesting, additional screening of women with double-negative results before 3 years had passed was discouraged.3 Follow-up guidelines from the American Society for Colposcopy and Cervical Pathology recommend repeated cotesting after 1 year for women with positive HR-HPV testing and negative cytology,4 although the initial interim guidelines recommended an interval of 6 to 12 months.3 Women with positive results for either repeated test should undergo colposcopy to rule out dysplasia.4

Although published guidelines have been made widely available regarding the use of HR-HPV testing, in actual practice, its application has been uneven. Many practitioners never use HR-HPV cotesting, whereas others use it consistently and still others use it on a case-by-case basis. Furthermore, practitioners who request cotesting often ask for it for women younger than 30 years or for women who already have had double-negative results within the past 3 years.

At the University of Rochester Medical Center (URMC; Rochester, NY), we decided to systematically monitor the follow-up cervical testing of women who had cytology interpreted as negative for intraepithelial lesion or malignancy (NILM) and HR-HPV cotesting to quantify these observations in our laboratory. We began offering cotesting in January 2006, and herein we report our findings, including 18-month follow-up, for the first year of NILM Pap tests with HR-HPV testing in women 30 years or older.

A question that frequently arises in the minds of cytologists and practitioners alike is the following: Do Pap tests interpreted as NILM with positive HR-HPV results really lack evidence of dysplasia, or was the dysplasia simply missed during screening? This worry may prompt cytologists to consider reviewing the slide and/or changing the interpretation to atypical squamous cells of undetermined significance (ASC-US). We systematically rescreened a subset of NILM Pap tests with positive HR-HPV results to determine the usefulness of looking back at those cases. Concern about this scenario on the part of practitioners may also prompt them to be more aggressive about follow-up for an isolated finding of HR-HPV positivity than is advised by guidelines. We therefore also sought to find out whether HR-HPV+ results have a noticeable impact on clinical follow-up patterns in our patient population.

Materials and Methods

The URMC laboratory processes approximately 70,000 Pap tests per year, of which more than 95% are ThinPrep liquid-based preparations (Hologic, Marlborough, MA), with the remainder consisting of conventional smears and SurePath liquid-based preparations (BD, Franklin Lakes, NJ). We receive specimens from a mix of clinical settings that include low-risk populations (privately insured women) and high-risk populations (prisoners and recipients of subsidized women’s health services). Specimens were examined by a team of 12 cytotechnologists and 4 cytopathologists during the study period. The atypical squamous cells/squamous intraepithelial lesion (ASC/SIL) ratio for the laboratory was 1.3 in 2006. Cervical biopsy specimens were reviewed by the surgical pathology department that included 19 pathologists.

HR-HPV testing was performed with the HCII method using a mixed probe targeting 13 cancer-associated HR-HPV types. The test is performed by the microbiology section of our department based on the manufacturer’s protocol. Specimens without sufficient material remaining in the vial, approximately 4 mL, were excluded from this study. In the cotesting scenario, we request HR-HPV testing only after screening and only for Pap tests interpreted as NILM or ASC-US.

The computerized laboratory records for URMC were reviewed for the 1-year period from January 1, 2006, to December 31, 2006, to find all Pap tests interpreted as NILM with HR-HPV cotesting. Follow-up Pap test, HR-HPV test, and biopsy results were also retrieved for the period extending from January 1, 2006, to July 31, 2008. The original biopsy slides were not reviewed.

Our laboratory decided to rereview all HR-HPV+ NILM Pap tests as a quality assurance measure. The first 45 such cases were reviewed by a senior cytotechnologist (D.K.R.) who frequently performs rescreening in the laboratory. The senior cytotechnologist then showed these 45 cases to a panel of pathologists, including 3 practicing cytopathologists (T.A.B., J.L.Y., and J.N.W.) and a cytopathology fellow (M.J.T.), who arrived at a consensus revised interpretation for each case. This review found several cases that by consensus were thought to merit upgrading of the cytologic interpretations, but the validity of this finding was in question because there was no way of knowing how often such upgrades might occur in HR-HPV– cases. To test the validity of upgrading on rereview, a set of 180 slides, with all marks removed, was independently given to a senior cytotechnologist (M.S.F.) and a cytopathologist (E.J.G.), neither of whom was involved in the original consensus group, for rescreening. All 180 cases were originally interpreted as NILM; of these, 90 tested HR-HPV+ (including the 45 cases seen by the consensus group) and 90 tested HR-HPV–. Rescreening interpretations were compiled for this set of 180 slides for both reviewers. These 180 cases make up a subset of all cases reported for this study and were included in the study regardless of the findings during rereview.

The institutional review board of URMC authorized this study. Statistical calculations used the Yates correction of the χ2 test or the 2-tailed Fisher exact test, depending on sample sizes, using an online resource.5 We considered findings to be statistically significant at P values less than .05.

Results

In 2006, there were 2,856 requests for HR-HPV cotesting. Some of these were eliminated from the study because they were requested for women younger than 30 years (108 [3.8%]) or because the HR-HPV test was not sent owing to insufficient volume of residual fluid (29 [1.0%]). Of the 2,719 NILM Pap tests with HR-HPV testing in women 30 years or older, 146 (5.4%) were HR-HPV+. The Pap tests were from 2,686 different women, although deidentification of the final data precluded elimination of the small number of cotests in women with a previous cotest in the same year. Among women who tested HR-HPV+, the age range was 30 to 78 years, the mean age was 45 years, and the median age was 42 years. For the HR-HPV– women, the age range was 30 to 92 years, with a mean of 47 years and a median of 46 years. For the cohort of women aged 30 to 39 years, 57 (8.2%) of 691 were HR-HPV+; among women 40 to 49 years, 43 (4.6%) of 926 were HR-HPV+; among women 50 years or older, 46 (4.8%) of 956 were HR-HPV+.

The follow-up rates for women with HR-HPV+ NILM Pap tests were quite different from the rates for women who were HR-HPV– Table 1 and Table 2. In the HR-HPV+ group compared with the double-negative group, follow-up with cytology or biopsy was significantly more likely (P < .0001). Furthermore, follow-up tended to be more extensive, with 36 (33.3%) of 108 women with a follow-up Pap test having multiple follow-up Pap tests within 18 months following an NILM HR-HPV+ result, as opposed to 48 (3.6%) of 1,334 women with an initial double-negative result (P < .0001). Proportionately more cytologic and/or histologic abnormalities were found in the women with follow-up Pap testing and/or biopsy who were originally HR-HPV+ (P < .0001). However, among women who underwent biopsy, the frequency of finding cervical intraepithelial neoplasia (CIN) overall (P = .7845) or CIN 2 or 3 (P = .5832) did not statistically significantly differ between HR-HPV+ and HR-HPV– women. All 5 of the high-grade dysplasias (defined as grade 2 or 3) found during follow-up in this study occurred in women 39 years or younger. Of the 5 high-grade dysplasias, 1 was vaginal intraepithelial neoplasia grade 3 (in the NILM HR-HPV– group), and the other 4 were CIN grade 2.

On initial consensus review of the subset of 45 NILM HR-HPV+ slides that were used to investigate the usefulness of rereview of such cases, 5 (11%) were found to have cytologic abnormalities consistent with an interpretation of ASC-US or atypical squamous cells, cannot exclude high-grade SIL (ASC-H). This proportion is considerably higher than the percentage of cases upgraded during routine 10% random review of NILM mandated by the Clinical Laboratory Improvement Amendments of 1988. Based on this preliminary finding, we decided to investigate further with a more rigorous method. We gave 90 cases (NILM HR-HPV+) and 90 controls (NILM HR-HPV–) without marks to be rescreened to a senior cytotechnologist (M.S.F.) and a cytopathologist (E.J.G.) not involved in the original consensus group. One reviewer upgraded 7 HR-HPV+ and 5 HR-HPV– cases (P = .7665). The other reviewer upgraded 8 HR-HPV+ and 3 HR-HPV– cases (P = .2119). Although both reviewers upgraded more HR-HPV+ cases, the upgrade frequency was not statistically significantly higher.

View this table:
Table 1
View this table:
Table 2

Discussion

We have found, in concordance with others, that the great majority of women 30 years or older with NILM Pap tests have negative test results for HR-HPV by the HCII method Table 3. The HR-HPV+ rate found in our laboratory is similar to that in previous reports. Recent studies in the United States have found rates of 4.0%6 and 8.2%7 in this age group. Studies from Europe found a range from 3.2% to 9.6%.812 Not all of these studies are entirely comparable with ours; however, it seems that our result of 5.4% is reasonably in accord with expected values.

Our study expands on previous work in that it documents the follow-up patterns and pathologic findings in a cohort of women undergoing testing as a part of routine clinical care. We found that many of the physicians requesting HR-HPV testing did not follow the protocols established in published national guidelines.4 Many HR-HPV tests were ordered for women younger than 30 years, amounting to 3.8% of the total requests. More significant, a very large proportion of the women for whom HR-HPV testing was initially ordered appropriately did not have the expected follow-up.

If guidelines were strictly followed, none of the women with double-negative results should have had follow-up Pap testing, HR-HPV testing, or biopsy in the ensuing 18 months. We found the following actual rates of follow-up in women with NILM HR-HPV– Pap tests: 51.8% had another Pap test performed within 18 months, of which a substantial subset (13.4% of the total) had another Pap test and another HR-HPV test, with the findings leading to a small number of biopsies (0.007% of the total) that revealed 2 high-grade lesions (0.0008% of the total).

Guidelines in effect during this study (2006) recommended that women with an initial cotest result of NILM with a positive HR-HPV result should be followed up with repeated cotesting in 6 to 12 months.3 We found 2 common patterns of deviation from this norm: 18.4% of these women did not have any follow-up within 18 months, and another 20.4% went directly to biopsy without first undergoing repeated Pap and HR-HPV testing. Only 47.6% of the women in our study had a Pap and HR-HPV cotest as the initial follow-up. Among these cotested women, only low-grade dysplasia was found in the biopsy follow-up for these cases. All 3 high-grade dysplasias were found in women who underwent biopsy immediately following the initial NILM HR-HPV+ result.

For both groups of women, HR-HPV+ and HR-HPV–, only about half had follow-up consistent with published recommendations. Although we do not know for certain why actual practice has diverged so frequently from expected practice, there are 2 likely scenarios worth considering. The first is that the practitioners ordering HR-HPV testing are not yet accustomed to using the data they receive and integrating it into their practice patterns. This leads to confusion and redundant testing. If this hypothesis is correct, actual practice should increasingly converge with guidelines over time. The other possibility is that practitioners and patients are seeking to maximize the sensitivity of cervical cancer screening by ordering both tests frequently and going directly to colposcopy and biopsy after a single positive HR-HPV result, in essence using HR-HPV testing for primary screening. This tendency is common throughout medical practice, particularly in the United States, because physicians have a strong desire to do everything possible for their patients and have little incentive to consider the economics of more testing and more follow-up in pursuit of rare lesions. Many patients may also be unwilling to wait for rescreening once they learn that they are potentially at risk of having cancer, regardless of how low that risk may be. If this hypothesis is correct, we would expect to see ongoing continuation of the current pattern. Only additional studies to look at future rates of guideline adherence will tell us which of these 2 scenarios is most prevalent.

View this table:
Table 3

Cytotechnologists and cytopathologists are often reluctant to sign out HR-HPV+ cases as NILM. For laboratories that do HR-HPV cotesting before screening, where the result is known before the slide is examined, the temptation to upgrade to ASC-US is especially intense. Because the number of such cases is relatively small, noticeable changes in the ASC/SIL ratio or HR-HPV positivity rate in ASC-US may not occur. Nevertheless, upgrading many or all such cases to ASC-US will prompt considerable additional workup with a relatively low likelihood of finding clinically significant disease. Indeed, this practice would be tantamount to implementing primary HR-HPV screening. In our study, a few high-grade cervical lesions were found by colposcopy immediately following an NILM HR-HPV+ result (2.0% of the total and 5.0% of women who underwent biopsy). This result should be kept in perspective by comparison with the rate of finding high-grade lesions in HR-HPV+ ASC-US cases. In a previously published study from the same patient population, although with women younger than 30 years included and covering a different period, we showed a rate of high-grade lesions on follow-up for HR-HPV+ ASC-US that was more than twice as high (5.1% of the total and 11.3% of women with biopsies).13 Furthermore, presumably, these high-grade dysplasias would have been detected by repeated cotesting after 1 year if guidelines had been followed. For these reasons, as well as the integrity of the cytologic screening process, Pap test results should not be upgraded on the basis of a positive HR-HPV test alone.

For laboratories, like ours, that send for HR-HPV testing after the completion of screening, the arrival of a positive result may prompt a desire to rereview the slide to ensure nothing was missed. This is not unreasonable, because women positive for HR-HPV undoubtedly have a higher likelihood of cytologic abnormalities than does the general population. Our initial review based on consensus analysis of a small number of HR-HPV+ NILM slides, with no controls, found a number of cases that we thought could have been upgraded. However, a larger and more careful study using matched cases and controls found that HR-HPV– Pap tests were not much less likely to be upgraded on rereview than HR-HPV+ ones, with no statistically significant difference between the 2 groups. Although a larger study might find a significant difference, our data indicate it would probably not be large. The key point to take away from this aspect of the study is that the great majority of HR-HPV+ cases called NILM on initial review will rightly remain NILM after rescreening, and of the few that are upgraded, hindsight bias will account for a substantial proportion.

The implementation of cotesting for HR-HPV has added a new dimension to cervical cancer screening. Guidelines have been put in place regarding the proper follow-up for women with cytologically negative but HR-HPV+ results. This study supports the guidelines in the sense that we have shown a low rate of discovery of high-grade lesions in these women, but it also shows the degree to which the guidelines are being disregarded in practice.

Notes

Upon completion of this activity you will be able to:

  • state the American Society for Colposcopy and Cervical Pathology guidelines regarding the use of high-risk HPV cotesting.

  • determine whether a laboratory has high-risk human papillomavirus (HR-HPV) positive rates in cotests that are in accord with published findings.

  • discuss the utility of rereviewing slides that were initially interpreted as negative for intraepithelial lesion or malignancy by cytology but tested positive for HR-HPV.

The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.

The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Questions appear on p 979. Exam is located at www.ascp.org/ajcpcme.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
View Abstract