High risk human papillomavirus in women with normal cervical cytology prior to the development of abnormal cytology and colposcopy

Authors


Correspondence: Professor Th. J. M. Helmerhorst, Academic Hospital Rotterdam, PO Box 2040, 3000 GA Rotterdam, The Netherlands.

Abstract

Objective To study the significance of the presence of high risk human papillomavirus (HPV) in women with initially normal cervical cytology for the development of abnormal cytology and an abnormal colposcopic impression.

Design Prospective, observational study

Participants and methods Sixty-eight women with cytomorphologically normal smears and at least one positive HPV test result were evaluated every six months by cytology, colposcopy and HPV testing. The endpoint of the study was abnormal cervical cytology.

Results The median time of follow up from the first positive HPV test was 34 months. A total of 17 women developed abnormal cytology, of whom 16 (94%) had persistence of a high risk HPV infection. Women with persistent high risk HPV were more likely to develop abnormal cervical cytology than women without high risk HPV (hazard ratio 28.2, 95% CI 3.72–215.2); they also had an increased risk of developing an abnormal colposcopic impression (hazard ratio 4.4, 95% CI 1.69–11.7). Among the 17 women with abnormal cytology, high grade dysplasia was histopathologically demonstrated in eight women.

Conclusion Persistent presence of high risk HPV in normal cervical smears is associated with a significantly increased risk of developing abnormal cytology and to a lesser degree with developing an abnormal colposcopic impression.

INTRODUCTION

Many studies have already shown the association of a human papillomavirus (HPV) infection with the development of pre-invasive and invasive cervical disease1–4. Extensive prospective studies have been advised to substantiate this observation. Since the period between the acquisition of an HPV infection until the development of severe cervical intraepithelial neoplasia (CIN 3) was expected to be many years, two studies that could be performed simultaneously were initiated. In one study women with abnormal cytology who were referred for colposcopy were regularly evaluated without intervention until CIN 3 developed. The cytological, colposcopic and virological results of this study are reported elsewhere5. The results of the other study, investigating the presence of a HPV infection before the development of abnormal cervical cytology, are described in this report. Women with normal cervical smears and at least one positive HPV test for either a low risk or a high risk HPV genotype were included and regularly evaluated by cytology, colposcopy and HPV testing until abnormal cytology was found.

METHODS

Women with a cytomorphologically normal cervical smear and a positive HPV test for either low risk or high risk HPV genotypes were enrolled. The database of the pathology department of the Onze Lieve Vrouwe Gasthuis in Amsterdam from 1990 to 1992 were searched for these keywords after approval of the study protocol by the hospital's research and ethic board. For research purposes HPV testing was routinely performed on every Pap smear. Among the 6965 women identified with normal cervical cytology, 647 (9.3%) had a positive HPV test. The clinical records of these 647 women, if aged between 18 and 55 years, were individually checked for the exclusion criteria: a history of cervical pathology; prenatal diethylstilbestrol exposure; and concomitant immunological or malignant disease. A written invitation to participate in the study was sent to the remaining 354 women (54.7%), followed by a reminder to nonresponders two months later. Of the 84 women who consented to be enrolled, four were excluded as a result of additional information from the intake questionnaire that could not be derived from the patient's file: three women had abnormal cervical smears that were evaluated in another hospital and one woman was suffering from an auto-immune disease. Ultimately, 80 women volunteered and were enrolled in the study, after giving informed consent. After the first evaluation, 12 women did not return despite two written reminders. Consequently, prospective data of 68 women were available and formed the study population.

Participating women were evaluated every six months by cytology, colposcopy and HPV testing. Colposcopist, laboratory personnel and participating women were not aware of the HPV test results during follow up. The endpoint of the study was abnormal cervical cytology (Pap 3 A twice or ≥ Pap 3 B once). Follow up ended after completion of a three year's colposcopic follow up or when there was colposcopic suspicion of invasive disease, repeated invisibility of the squamo-columnar junction or when the woman decided to withdraw. In the case of abnormal cytology and an abnormal colposcopic impression, a biopsy was taken of the most serious lesion, analogue to clinical practice. Histopathology was considered as the final diagnosis.

The women included in analysis were seen at least three times: 1. at the time of selection (t−1: cytological and virological data were available); 2. at the start of the study (t0: cytological, colposcopic and virological data were available); and 3. at the end of follow up (t1: cytological, colposcopic and virological data were available). The cytological and colposcopic findings during follow up in women with persistent high risk HPV between t−1 and t0 were contrasted with the equivalent findings in women without persistent high risk HPV at t0. The group of women without persistent high risk HPV at t0 consisted of HPV negative and low risk HPV positive women. To study the development of abnormal cytology and development of an abnormal colposcopic impression among HPV negative women both at t−1 and t0 would require a very large cohort of volunteers and a follow up time of many years, as can be concluded from the study of Rozendaal et al.6. Therefore, for practical and economical reasons, the development of abnormal cytology or an abnormal colposcopic impression among women with persistent high risk HPV was contrasted with a negative high risk HPV test at t0 but with a known history of either low risk or high risk HPV at t−1.

At each visit a cervical cytological sample was taken with a cervix brush (International Medical Products, Zutphen, The Netherlands). Cervical smears were classified by a modified Papanicolaou method as used in The Netherlands: Pap 1 (normal smear); Pap 2 (very mild dyskaryosis, including atypical squamous cells of unknown significance); Pap 3 A (suggesting mild or moderate dyskaryosis); Pap 3 B (suggesting severe dyskaryosis); Pap 4 (suggesting carcinoma in situ); and Pap 5 (suggesting at least micro-invasive carcinoma). Review of the smears of the first evaluation by an expert pathologist who was not aware of the first assessment, did not change the diagnosis.

HPV testing was performed by enzyme immunoassay polymerase chain reaction using HPV-general-primer-mediated polymerase chain reaction with the general primers GP5+/6+. This test has been described earlier7,8. All 14 high risk HPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66 and 68) were tested for in one assay. In addition, polymerase chain reaction amplimers were analysed for the individual high risk HPV types.

All colposcopic examinations were performed by the same colposcopist. The colposcopic impression was recorded as the prediction of the histopathological diagnosis of the most serious lesion: no CIN; CIN 1; CIN 2; or CIN 3. An abnormal colposcopic impression was defined as the impression suggesting at least CIN 1. No biopsies were taken for verification of the colposcopic impression to avoid any interference with the natural course of the disease until the protocol criteria were met.

At the end of follow up a colposcopically directed biopsy was taken in women with an abnormal cervical smear. Treatment was performed according to the hospital's standard protocol. The histopathological diagnoses of the colposcopically directed biopsies in women with abnormal cytology are shown in Table 1. Review of all biopsy specimens by an experienced pathologist according to clinically used criteria did not change the diagnoses.

Table 1.  Characteristics of women who developed abnormal cytology during follow up. CIN = cervical intraepithelial neoplasia; HPV = human papillomavirus; t1= at time of selection; t0= at the start of the study; t1= at the end of follow up.
  Cytological classification*      
Patient numberAge (years)t1t1Persistent high risk HPV-typeFollow up HPV (months)Colposcopic impressions at t0Follow up colposcopy (months)Colposcopic impressions at t1Histopathology
  1. *3A1 = mild dyskaryosis, 3 A2 = moderate dyskaryosis, 3 B = severe dyskaryosis, 4 = carcinoma in situ.

  2. No random biopsy was taken; in analysis this woman was not considered to have dysplasia.

  3. The reported metaplasia was confirmed histopathologically.

12213A1166CIN 03CIN 2CIN 2
23523A11612CIN 04CIN 1CIN 1
33213A11814CIN 19CIN 1CIN 2
430241815CIN 18CIN 2CIN 3
52813A11619CIN 113CIN 2CIN 1
62723B1621CIN 09CIN 2CIN 3
73523A123CIN 012CIN 1CIN 1
84623A15623CIN 113CIN 1CIN 1
93813A23123CIN 117CIN 1CIN 1
103323A25624CIN 012CIN 0
113513A21626CIN 120CIN 3CIN 3
123513A23327CIN 021CIN 1CIN 1
134013A21635CIN 227CIN 2CIN 3
144723A11638CIN 030CIN 1CIN 1
152923A21839CIN 028CIN 3CIN 3
163313A21641CIN 121CIN 2CIN 2
174623A21848CIN 124CIN 0CIN 0‡

Statistical methods

Follow up started at t−1. The cumulative proportion of abnormal cytology and of an abnormal colposcopic impression in high risk HPV positive and negative women were estimated using the Kaplan-Meier method and the relative hazard ratio with 95% confidence intervals was estimated using Cox's regression analysis. Analyses were performed using SPSS, version 9.0.

RESULTS

The median follow up time from t−1 for all 68 women was 34 months (range 5–77). Three years of colposcopic follow up were completed by 20 women, while 31 women decided to withdraw from the study after a median follow up time of 17 months (range 6–27). The median time between two consecutive visits was six months (range 4–8). For 17 women an abnormal Pap smear defined the end of follow up after a median follow up time from t−1 until abnormal cytology of 23 months (range 6–48) (Table 1). No women were withdrawn for colposcopic reasons.

At the time of the first clinical visit (t−1), 59 women had a positive high risk HPV test and nine women had a negative high risk HPV test. During follow up between t−1 and t0 the high risk HPV infection persisted in 32 women, while 27 women no longer tested positive for high risk HPV. No new high risk HPV infections were acquired. Consequently, for analysis, the persistent high risk HPV positive group consisted of 32 women (mean age 33.5 years) and the group without persistent high risk HPV of 36 women (mean age 39.3 years) (Table 2).

Table 2.  Number of women with and without presence of high risk human papillomavirus (HPV) at the time of selection (t1) and at the start of the study (t0). Values are given as n.
 High risk HPV
 PositiveNegative
t1599
t03236

The cumulative proportion of abnormal cytology after 48 months was 68.6% in women with persistent high risk HPV and 2.9% in women without persistent high risk HPV. The Kaplan-Meier curves are represented in Fig. 1. Women with persistent high risk HPV were more likely to develop abnormal cervical cytology than women without high risk HPV (hazard ratio 28.2, 95% CI 3.72–215.2).

Figure 1.

Cumulative proportion (vertical axis, 100%) of abnormal cytology during follow up (horizontal axis, months) in women with persistent presence of high risk HPV (HR-HPV positive) and women without persistent presence of high risk HPV (HR-HPV negative).

In 16 of the 17 women who developed abnormal cervical cytology, the high risk HPV test was persistently positive during follow up. In one woman the test for high risk HPV was negative in the two smears assessed as Pap 3 A. The last positive test result for HPV-33 was found five months before the first abnormal cervical smear. Thus it is concluded that she no longer had a HPV-33 infection. The histopathological diagnosis was CIN 1 (Table 1, patient no. 7).

For the detection of the cumulative incidence of an abnormal colposcopic impression among the groups studied, 24 women with an abnormal colposcopic impression at the first colposcopic evaluation (t0) were excluded from the total of 68 because the time between the first detection of the abnormal colposcopic impression and t0 was not available. Among the women excluded were 21 women with a colposcopic impression suggesting CIN 1 and three women with a colposcopic impression suggesting CIN 2. For the remaining 44 women who had a normal colposcopic impression at t0, it was assumed that no colposcopic lesion had been present from t−1. Among these 44 women, 19 had a persistent high risk HPV infection during follow up and 25 women did not.

The cumulative proportion of women showing an abnormal colposcopic impression in the 48 months from t−1 onwards was 82.9% in women with persistent high risk HPV in contrast to 48.7% in women without persistent high risk HPV. The Kaplan-Meier curves are represented in Fig. 2. The hazard for developing an abnormal colposcopic impression in women with persistent presence of high risk HPV compared with absence of high risk HPV was 4.4 (95% CI 1.69–11.7).

Figure 2.

Cumulative proportion (vertical axis, 100%) of an abnormal colposcopic impression during follow up (horizontal axis, months) in women with persistent presence of high risk HPV (HR-HPV positive) and women without persistent presence of high risk HPV (HR-HPV negative).

DISCUSSION

This study was conducted to gain insight into the natural history of CIN in relation to HPV. Women with normal cervical smears were regularly evaluated by HPV detection and colposcopy until abnormal cytology was assessed. The cumulative incidence of abnormal cervical cytology among women with persistent presence of high risk HPV was substantially higher than among women without high risk HPV. The Cox's proportional hazard ratio associated with persistence of high risk HPV in cervical smears for developing abnormal cytology, compared with the absence of high risk HPV since the start of the study was 28.2. Developing an abnormal colposcopic impression was not that strongly correlated with persistent presence of high risk HPV (hazard ratio 4.4).

We studied a gynaecological outpatient population with a history of at least one positive HPV test for either a high risk or a low risk genotype. Our rationale for selecting women with a history of HPV infection, instead of selecting a HPV negative control group, was the assumption of an increased risk of development of abnormal cytology in the presence of HPV3,8–10. To study the development of abnormal cytology among HPV negative women would require long term follow up of a large cohort of healthy women. Moreover, recruiting volunteers for follow up was found to be very difficult. Despite many efforts on our part to motivate women to participate in the study only 23% agreed to enrol. As a consequence, this study was performed among 68 women. Follow up was completed by only 39% of the women enrolled with permanent normal cervical smears. The small number of women studied should be taken into consideration when interpreting the results and making comparisons with results from other studies. Furthermore, it should be borne in mind that the high risk HPV negative group in this study consisted of women with a low risk HPV genotype and women who became free of low risk or high risk HPV DNA during follow up. Therefore, when contrasted to a real HPV negative control group, the estimated risks for developing abnormal cytology and colposcopy will be even larger.

The results of our study agree with the findings of others. Recently, Ho et al.9 reported the results of a natural history study of HPV infection of almost similar design among students. The main risk factor for the development of an abnormal cervical smear was the persistent presence of high risk HPV. The authors found a relative risk of 37.2 (95% CI 14.6–94.8) for the development of abnormal cervical cytology when a high risk HPV infection persisted for at least one year. In our study a similar increased risk was found after persistence of the high risk HPV infection for a median time of two years. This difference could be attributed to differences in the study populations. In our study women in the control group had a history of at least one positive HPV test, while the women in the study by Ho et al. were HPV negative. This might reduce any contrasts between the groups studied in our study. Furthermore, the mean age of the population studied was higher in our study (36 vs 20 years, respectively). Others found that the prevalence of HPV was higher in younger women11 and HPV infections were more transient in younger women9. Koutsky et al.10 revealed a relative risk for high grade cervical dysplasia of 11.0 (95% CI 4.6–26) for women with an infection with HPV 16 or HPV 18 compared with women without HPV. They studied a cohort of women from a clinic whose main focus was in sexually transmitted diseases. All cases of cervical dysplasia were detected within two years after the first detection of HPV. In our study high grade dysplasia was diagnosed in the biopsies of 43% (8/17) of women with abnormal cytology.

During follow up women were evaluated colposcopically every six months. The development of an abnormal colposcopic impression in relation to the persistent presence of a high risk HPV infection was studied. The cumulative proportion of an abnormal colposcopic impression was not substantially higher among women with persistent high risk HPV than in women without high risk HPV. The number of women studied was very small. However, it was found that an abnormal colposcopic impression in the absence of an abnormal Pap smear is associated with only minor, if any, dysplastic cervical changes, even in the presence of high risk HPV. In a screening population in Finland, abnormal colposcopic impressions were seen in 38% of women with normal cytology and a positive test for HPV-1612. Schneider et al.13 reported the colposcopic findings in 150 women with normal cervical cytology, a positive HPV test and no history of abnormal smears. An abnormal transformation zone, mainly characterised just by aceto-whiteness, was registered in 69%. In colposcopically directed biopsies ‘only minor changes’ were demonstrated. The authors suggested that those changes might represent a precursor lesion of a flat condyloma because all women were found to have an HPV infection. They agreed with others that a HPV infection cannot be distinguished colposcopically from mild dysplasia14.

CONCLUSION

We conclude that the persistent presence of high risk HPV in cervical smears results in an significantly increased risk of the development of abnormal cytology compared with the absence of high risk HPV. An abnormal colposcopic impression correlates less with the persistence of high risk HPV, and lacks specificity for clinical implication in screening.

Acknowledgements

This study was conducted in the Free University Hospital and the Onze Lieve Vrouwe Gasthuis, Amsterdam. We would like to thank Dr Sj. de Blok and Dr J. L. G. Blauwgeers for the opportunity to conduct the described study in the Onze Lieve Vrouwe Gasthuis in Amsterdam, and for reading the manuscript.

Ancillary