Journal of Medical Virology 86:1145–1152 (2014)

Untypable Human Papillomavirus Infection and Risk of Cervical Intraepithelial Neoplasia Among Women With Abnormal Cervical Cytology Arsenio Spinillo,1* Barbara Gardella,1 Marianna Roccio,1 Paola Alberizzi,2 Enrico Maria Silini,3 and Barbara Dal Bello2 1

Department of Obstetrics and Gynecology, University of Pavia, IRCCS Fondazione Policlinico San Matteo, Pavia, Italy 2 Department of Pathology, University of Pavia, IRCCS Fondazione Policlinico San Matteo, Pavia, Italy 3 Department of Pathology, Azienda Ospedaliero-Universitaria, Universita` di Parma, Parma, Italy

The risk of cervical intraepithelial neoplasia and/or invasive cervical cancer associated with untypable human papillomavirus (HPV) infections has been not investigated fully. HPV infection caused by 18 high-risk and 7 low-risk genotypes as detected by the INNO-LIPA genotyping system, was investigated in 4,258 women with abnormal Pap smear referred to a colposcopic service. The prevalence of HPV infection was 76.1%. Rates of cervical intraepithelial neoplasia grade 3þ were 0.88% (9/ 1,017) in HPV-negative subjects, 1.8% (7/380) in subjects with untypable HPV infection, 3.2% (11/343) in subjects with single/multiple lowrisk types, 28.3% (201/709) in subjects with multiple low and high-risk types, 15.2% (162/ 1,069) in subjects with single high-risk types, and 31.2% (229/733) in those with multiple high-risk types. Compared to women without any HPV infection, the odds ratios of cervical intraepithelial neoplasia grade 2þ or grade 3þ in subjects with untypable or low-risk HPV genotypes were 5.73 (95% CI ¼ 2.79–11.78) and 12.4 (95% CI ¼ 6.31–24.5, P ¼ 0.014 compared to untypable) and 3.1 (95% CI ¼ 1.11–8.16) and 7.1 (95% CI ¼ 2.9–17.2, P ¼ 0.07 compared to untypable), respectively. In the subgroup of subjects with cervical intraepithelial neoplasia grade 1 or negative colposcopy/biopsy, the progression to cervical intraepithelial neoplasia grade 2þ at follow-up (median 25 months, range 6–70) was 2% (14/684), 3.4% (7/205), and 5.6% (11/195, P ¼ 0.04 compared to negative) among negative, untypable, and low-risk HPV infection, respectively. The risk of cervical intraepithelial neoplasia associated with untypable HPV infection was higher than that recorded among uninfected women, but lower than the risk associated with low- or high-risk HPV genoC 2014 WILEY PERIODICALS, INC. 

types. J. Med. Virol. 86:1145–1152, 2014. # 2014 Wiley Periodicals, Inc.

KEY WORDS:

human papillomavirus; cervical intraepithelial neoplasia; cervical cancer; colposcopy

INTRODUCTION On the basis of epidemiological, phylogenetic, and mechanistic evidence linking human papillomavirus (HPV) infection to invasive cervical cancer, HPVs have been classified in several categories of risk (carcinogenic, probable carcinogenic, possibly carcinogenic, and not classifiable) reflecting a different carcinogenic potential [Bouvard et al., 2009]. Most commercially available genotyping systems are based on the detection of the more common high and low risk HPVs that cause more than 90% of cervical lesions including invasive cervical cancer. Studies on the worldwide distribution of HPVs in invasive cervical cancer suggest that HPV types 16, 18, 31, 33, 35, 45, 52, and 58 contribute to 91% of the disease [De Sanjose et al., 2010]. HPV infection caused by unknown or untypable genotypes (i.e., HPV infection detected by PCR using consensus HPV primers but

Conflict of interest: The authors declare that there are no conflicts of interest.  Correspondence to: Arsenio Spinillo, MD, Department of Obstetrics and Gynecology, University of Pavia, IRCCS Policlinico San Matteo, P.le Golgi, 19. 27100 Pavia, Italy. E-mail: [email protected] Accepted 11 March 2014 DOI 10.1002/jmv.23938 Published online 1 April 2014 in Wiley Online Library (wileyonlinelibrary.com).

1146

Spinillo et al.

HPV type not identified by genotyping systems) is detected in 3.7% of histological specimens of invasive cervical cancer [De Sanjose et al., 2010; Geraets et al., 2012]. However, the rates of untypable HPV types are much higher in cervical cytological samples obtained from subjects with abnormal Pap smear and in specimens obtained from male external genitalia, suggesting that carriage and diffusion of these HPVs is common [Varnai et al., 2007; Sichero et al., 2013]. Untypable HPV types could reflect viral types other than those detected by the genotyping system, or a very low number of HPV copies, which thus escape genotype identification [Sichero et al., 2013]. Although the carcinogenic role of this type of infection seems limited, there are very few data on the relationship between untypable HPV infection and the risk of cervical intraepithelial neoplasia among women with cervical cytological abnormalities. The counseling of women with untypable HPVs is problematic as the clinical and pathological correlations of these infections are unknown. The purpose of this study was to evaluate the risk of cervical intraepithelial neoplasia and/or invasive cervical cancer among women referred to a colposcopic service because of abnormal Pap smear and with an untypable HPV cervical infection. MATERIALS AND METHODS The initial study patients included all women aged between 21 and 65 years, attending the colposcopy clinic of the Department of Obstetrics and Gynecology of the IRCCS Fondazione Policlinico San Matteo, University of Pavia (Italy) from 2006 to 2012 because of abnormal Pap smear. Women were referred by the cytological screening service of our Department, from private practice and from screening services of external institutions. Exclusion criteria included pregnancy, HPV test, or treatment for cervical neoplasia in the last year, total hysterectomy, lack of a recent (1 month) Pap smear and use of vaginal medication in the previous 2 days. Pap smears had been performed by referring physicians and had not been reviewed in our institution. The Institutional Review Board of the Hospital approved the study and informed consent was obtained from all subjects. All patients were treated according to an established protocol including HPV DNA detection and typing and colposcopy with targeted biopsies. Cervical samples for HPV typing were obtained immediately before colposcopy. After speculum examination, scrapes were taken with a cervix brush, suspended in 25 ml of ThinPrep– PreservCyt Solution (Cytic Corporation, Marlborough, MA), and stored at 4˚C. DNA extraction was performed within 5 days from sampling by lysis and digestion with proteinase K; since 2010 all extractions were standardized on an automatic instrument (Magtration system 12GC, Precision System Sciences, Woerrstadt, Germany). HPV type-specific sequences were detected using the INNO-LiPA HPV genotyping J. Med. Virol. DOI 10.1002/jmv

assay (Innogenetics N.V., Ghent, Belgium) that exploits the amplification of the L1 region with consensus primers SPF-10 and line probe detection. In the years 2006–2009 the V2 version of the assay was used. Since 2010 the EXTRA version of the assay, which allows the simultaneous and separate detection of 18 high-risk (16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 53, 56, 58, 59, 66, 68, 73, and 82), 7 lowrisk HPV types (6, 11, 40, 43, 44, 54, and 70), and 2 unknown risk HPV (69/71 and 74), was introduced. Concurrent amplification of human HPA-DPB1 gene sequences was used as a control for DNA adequacy. Appropriate negative and positive controls were run for each set of reactions. Hybridization patterns were analyzed automatically by the LiRAS system and checked by two independent readers. Specimens testing positive for L1 HPV sequences by PCR but not recognized by the typing probes (HPV-X) were considered as untypable HPV infections. In previous studies, the INNO-LiPa test demonstrated a clinical sensitivity of 95% in the identification of high-grade squamous intraepithelial lesions and proved to be highly suitable for clinical and epidemiological investigations [van Hamont et al., 2006; Safaeian et al., 2007]. A standardized colposcopic examination was performed immediately after cervical brushing for HPV typing by two gynecology residents certified by the Italian Society of Colposcopy and with at least 3 years of colposcopic experience. After speculum placement, 5% acetic acid was applied to the cervix. The entire transformation zone was then examined and any area of abnormality was noted. Multiple targeted cervical biopsies were obtained in all cases in which cervical intraepithelial neoplasia was suspected on colposcopy and in all cases of high-grade squamous lesions, irrespective of colposcopic impression. Endocervical curettage was performed, according to the clinician’s judgment, when the extent of the lesion or the squamo-columnar junction was not entirely visible or in the case of atypical glandular cells on Pap smear. In the analysis of data, either the histological diagnosis of punch biopsy or, when more severe, the diagnosis after cone biopsy obtained by loop electrosurgical or cold-knife excision was used. After the diagnostic workup and treatment, patients referred from the cytological screening service of our Department were enrolled in a 24-month follow-up program including: (a) observation, colposcopy, and/or cytology every 6–12 months for subjects with negative colposcopic impression and/or negative histological findings after an abnormal Pap smear; (b) Observation, colposcopy, and cytology every 6 months for cervical intraepithelial neoplasia grade 1 lesions; (c) HPV test coupled with colposcopy and cytology every 6 months after treatment for cervical intraepithelial neoplasia grade 2 or 3. In the statistical analysis outcome was defined as negative when colposcopy and/or cervical biopsy were negative, or as cervical intraepithelial neoplasia grade 1 to 3, invasive squamous or adenocarcinoma

Untypable HPV Infection

1147

of the cervix. HPV infection was categorized as none, untypable, single/multiple low-risk, multiple low and high-risk, single high-risk, multiple high-risk, and unclassified risk. Data were analyzed by one-way ANOVA and Bonferroni corrected post-hoc test and chi-square test to compare continuous and categorical variables, respectively. The Bonferroni method for multiple comparisons was used to adjust chi-square tests in multiway contingency tables. Multinomial logistic models were constructed to evaluate the association between untypable HPV infection and abnormal colposcopic/pathologic outcomes. Explanatory variables inserted in the models were age (continuous variable), previous treatment for cervical intraepithelial neoplasia (yes, no), human immunodeficiency virus status (negative, unknown, positive), ancestry (Italian, foreigner), parity (yes, no), and category of HPV infection. The analysis was repeated in women referred to the colposcopic clinic due to atypical squamous cells of undetermined significance or low-grade squamous intraepithelial lesions and for those referred for high grade squamous lesions or worse cytological findings. Data were analyzed using the Stata/MP 10 for Windows (StataCorp LP, College Station, TX). Significant differences between regression coefficients obtained from multinomial logistic equations were compared using the Wald test [Scott and Freese, 2006]. RESULTS During the study 4467 women with an abnormal Pap smear were seen at the colposcopic clinic and 4398 (98.5%) agreed to participate in the study. A further 140 (3.1%) women were excluded because of missing or inadequate samples for HPV DNA testing (108 subjects), or inadequate colposcopic examination or biopsy specimens (32 subjects), leaving 4,258 women for analysis. Overall, the mean age of the participants was 37.1  10.8 years; 3,779 (88.8%) were Italian, and 2,895 (68%) were parous. A previous history of squamous cervical lesions (not in the year preceding enrolment) was reported by 287 (6.7%) subjects, whereas 124 (2.9%) women reported HIVseropositivity. Pap smear results at enrolment were

as follows: 2,380 (55.9%) subjects low-grade squamous lesions, 1,274 (29.9%) atypical squamous cells of undetermined significance, 15 (0.35%) atypical squamous cells cannot exclude high grade lesions, 460 (10.8%) high-grade squamous lesions or worse, and 129 (3%) atypical glandular cells of undetermined significance. The relationship between Pap smear at enrolment and final colposcopic/pathologic classification is reported in Table I. Negative subjects included 963 (22.6%) subjects with negative cervical biopsies and 1,155 (27.1%) subjects with negative colposcopy impression in whom cervical biopsy was considered unnecessary. Of the 1,240 cervical intraepithelial neoplasia grade 1 lesions, 793 (64%) were diagnosed on cervical biopsies, and 447 (36%) on excision specimens. Diagnoses of CIN 2–3 and invasive cervical cancer were obtained from either loop electrosurgical excision or cold-knife conization in 595 (73.3%) out of 812 and in 67 (76.1%) out of 88 subjects, respectively. According to the FIGO classification, of the 88 subjects with a diagnosis of invasive cervical cancer, 10 (11.4%) were at stage IA, 52 (59.1%) were at stage Ib–IIa, and the remaining 28 (31.8%) subjects were in more advanced stages of the disease. The rate of HPV infection was 76.1% (3,241/4,258); the distribution of the more common genotypes is reported in Table II. HPV infection was untypable in 380/3,241 (11.7%) specimens whereas multiple genotypes were identified in 1,493/3,241 (46.1%) subjects. Multiple HPV infection included two genotypes in 821 (25.3%), three in 404 (12.5%), and four or more in the remaining 268 (8.3%) samples, respectively. HPV 16, either as a single or multiple infection, was the most common genotype identified in cervical intraepithelial neoplasia grade 2–3 (373/812, 45.9%) and in invasive cervical squamous (25/63, 39.7%), or adenocarcinoma (13/25, 52%). Other common genotypes identified in cervical intraepithelial neoplasia grade 2–3 were HPV 52 (213/812, 26.2%), HPV 31 (203/812, 25%), HPV 18 (119/812, 14.7%), HPV 51 (116/812, 14.3%), and HPV 6 (96/812, 11.8%). Excluding the 380 untypable samples, the proportion of multiple infection among all HPV positive samples was 40.4% (357/884) in negative colposcopic/pathological subjects, 49.1% (545/1,109) in CIN 1, 60.6% (160/

TABLE I. Association between Cytological Findings at Enrolment and Subsequent Colposcopic/Histological Outcome

Cytological results LSIL ASCUS ASC-H, HSIL/invasive cancer AGC

Negative (N ¼ 2,118a)

CIN 1 (N ¼ 1,240)

CIN 2 (N ¼ 280)

CIN 3 (N ¼ 532)

Invasive squamous (N ¼ 63)

Invasive adeno (N ¼ 25)

n. (%)

n. (%)

n. (%)

n. (%)

n. (%)

n. (%)

1,164 818 49 87

(55) (38.6) (2.3) (4.1)

852 310 50 28

(68.7) (25) (4) (2.2)

170 43 66 1

(60.7) (15.4) (23.6) (0.3)

189 95 240 8

(35.5) (17.9) (45.1) (1.5)

2 (3.2) 7 (11.1) 54 (85.7) 0

3 1 16 5

(12) (4) (64) (20)

CIN, cervical intraepithelial neoplasia; LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesions; ASCUS, atypical squamous cells of undetermined significance; ASC-H, atypical squamous cells-cannot exclude HSIL; AGC, atypical glandular cells. a 963 biopsies and 1,155 negative colposcopies.

J. Med. Virol. DOI 10.1002/jmv

1148

Spinillo et al. TABLE II. Frequency of the More Common HPV Genotypes According to Colposcopic/Histological Outcome

HPV

Type of infection

6

Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple

11 16 18 31 33 35 39 44 45 51 52 53 56 66

Negative (N ¼ 2,118)

CIN 1 (N ¼ 1,240)

CIN 2 (N ¼ 280)

CIN 3 (N ¼ 532)

Invasive squamous (N ¼ 63)

Invasive adeno (N ¼ 25)

n. (%)

n. (%)

n. (%)

n. (%)

n. (%)

n. (%)

3 (1.1) 28 (10) 4(1.4) 10 (3.6) 26 (9.3) 69 (24.6) 3 (1.1) 20 (7.1) 9 (3.2) 48 (17.1) 0 16 (5.7) 1 (0.3) 3 (1.1) 2 (0.7) 13 (4.6) 2 (0.7) 6 (2.1) 3 (1.1) 4 (1.4) 12 (4.3) 29 (10.4) 19 (6.8) 55 (19.6) 5 (1.8) 17 (6.1) 1 (0.3) 13 (4.6) 5 (1.8) 23 (8.2)

8 (1.5) 57 (10.7) — 31 (5.8) 74 (13.9) 204 (38.3) 2 (0.4) 94 (17.7) 7 (1.3) 139 (26.1) 4 (0.8) 41 (7.7) 3 (0.6) 17 (3.2) 2 (0.4) 42 (7.9)

— 8 (12.7) — 5 (7.9) 14 (22.2) 21 (33.3)

— 1 (4) 1(4) — 3 (12) 10 (40)

12 (19.1) 1 (1.6) 7 (11.1)

5 (20)

5 (7.9) 1 (1.6) 4 (6.4)

2 (8)

3 (4.8)

3 (12)

30 (5.6) 4 (0.8) 27 (5.1) 5 (0.9) 70 (13.2) 20 (3.8) 119 (22.4) 7 (1.3) 47 (8.8) 2 (0.4) 27 (5.1) 3 (0.6) 2 (0.4)

4 (6.4)

1 (4)

3 (4.8) 3 (4.8) 6 (9.5) 1 (1.6) 9 (14.3) 5 (7.9)

1 (4) 1 (4) 2 (8) 1 (4) 5 (20) 1 (4) 2 (8)

3 (4.8)

1(4)

5 (7.9)

1 (4)

82 74 25 40 105 119 4 45 12 79 7 20 17 14 18 19 10 23 11 10 65 63 45 82 34 38 7 24 25 30

(3.9) (3.5) (1.2) (1.9) (5) (5.6) (0.2) (2.1) (0.6) (3.7) (0.3) (0.9) (0.8) (0.7) (0.8) (0.9) (0.5) (1.1) (0.5) (0.5) (3.1) (3) (2.1) (3.9) (1.6) (1.8) (0.3) (1.1) (1.2) (1.4)

66 113 11 62 147 170 6 69 30 148 4 33 11 14 20 35 12 23 12 19 58 121 75 137 38 53 7 42 23 47

(5.3) (9.1) (0.9) (5) (11.9) (13.7) (0.5) (5.6) (2.4) (11.9) (0.3) (2.7) (0.9) (1.1) (1.6) (2.8) (1) (1.9) (1) (1.5) (4.7) (9.8) (6.1) (11.1) (3.1) (4.3) (0.6) (3.4) (1.9) (3.79%)

1 (4)

CIN, cervical intraepithelial neoplasia.

264) in CIN 2, 72.2% (376/521) lesions, and 66.3% (55/83) in invasive cervical cancer (chi-square for trend ¼ 143.4, P < 0.001). The associations between categories of risk of HPV infection and colposcopic/pathologic outcome are reported in Table III. Single/multiple low-risk HPV types were present in 10.6% (343/3,241) of HPV infections, single high-risk types in 32.7% (1,060/ 3,241), multiple high-risk types in 22.6% (733/3,241),

and multiple low/high-risk types in 21.9% (709/ 3,241); seven HPV infections were caused by HPVs at undefined-risk. In a pairwise comparison of rows, and after the Bonferroni correction, outcomes associated with untypable infection differed significantly from outcomes associated with negative, low-risk and all categories of high-risk HPV infection (P < 0.01 for all comparisons). The mean age was 38.5  10.9 years for HPV-negative

TABLE III. Association Between Classification of HPV Infection and Colposcopic/Histological Outcome

Classification of HPV infection Negative, n ¼ Untypable Single/multiple low-risk Multiple low-high-risk Single high-risk Multiple high-risk Undefined

Negative (N ¼ 2,118)

CIN 1 (N ¼ 1,240)

CIN 2 (N ¼ 280)

CIN 3 (N ¼ 532)

Invasive squamous (N ¼ 63)

Invasive adeno (N ¼ 25)

n. (%)

n. (%)

n. (%)

n. (%)

n.(%)

n.(%)

982 252 172 176 374 160 2

CIN, cervical intraepithelial neoplasia.

J. Med. Virol. DOI 10.1002/jmv

(46.4) (11.9) (8.1) (8.3) (17.7) (7.6) (0.1)

22 109 143 253 445 264 4

(1.8) (8.8) (11.5) (20.4) (35.9) (21.3) (0.3)

4 12 17 79 88 80

(1.4) (4.3) (6.1) (28.2) (31.4) (28.6)

5 6 9 176 136 199 1

(0.9) (1.1) (1.7) (33.1) (25.6) (37.4) (0.2)

2 1 1 20 20 19

(3.2) (1.6) (1.6) (31.8) (31.8) (30.2)

2 (8) 1 5 6 11

(4) (20) (24) (44)

Untypable HPV Infection

1149

cases, 38.9  11.2 for untypable HPV infection, 37.4  10.8 for single/multiple low-risk types, 36.4  10.9 for multiple low and high-risk types (P < 0.01 compared to either HPV-negative or untypable infection), 36.3  10.3 for single high-risk types (P < 0.01 compared to either HPV-negative or untypable infection), and 35.9  10.6 for multiple high-risk types (P < 0.01 compared to either HPV-negative or untypable infection), respectively. The prevalence of untypable HPV infection was significantly higher among women with low-grade lesions in the Pap smear (355/3,648) than in those with high grade lesions or cervical cancer (25/603, P < 0.001). The rates of cervical intraepithelial neoplasia grade 3þ were 0.88% (9/1,017) in HPV-negative subjects, 1.8% (7/380) in untypable HPV, 3.2% (11/343) in single/multiple low-risk, 28.3% (201/709) in multiple low and high-risk (P < 0.001 compared to negative, untypable and low-risk infection), 15.2% (162/1,069) in single-high risk (P < 0.001 compared to negative, untypable, and low-risk infection), and 31.2% (229/ 733) in multiple high-risk genotypes (P < 0.001 compared to negative, untypable, and low-risk infection).

When the analysis was restricted to subjects with atypical squamous cells of undetermined significance or with low-grade cervical lesions on Pap smear, the rates of cervical intraepithelial neoplasia grade 3þ were 0.4% (4/957) in HPV-negative subjects, 1.7% (4/ 355) in untypable infection, 2.5% (8/316) in single/ multiple low-risk infection, 15.9% (87/546) in multiple low and high-risk infection (P < 0.001 compared to negative, untypable, and low-risk infection), 9% (82/ 916) in single high-risk infection (P < 0.001 compared to negative, untypable, and low-risk infection), and 20.1% (112/558) in multiple high-risk infection (P < 0.001 compared to negative, untypable and lowrisk infection). The predictive role of HPV infection classification in subsequent colposcopic/pathological diagnoses adjusting for the confounder effect of age, Human Immunodeficiency Virus seropositivity, ancestry, parity and previous history of squamous lesions was evaluated by multinomial logistic regression. The results are reported in Table IV. Compared to uninfected subjects, untypable HPV infection was associated with an increased risk of cervical

TABLE IV. Likelihood of Colposcopic/Histological Outcome as Obtained by Multinomial Logistic Regression

ASCUS-LSILa (N ¼ 3,648)

Negative (N ¼ 1,980)

Negative, n ¼ 957 Untypable, n ¼ 355 Single-multiple low-risk, n ¼ 316 Multiple low-high-risk, n ¼ 546 Single high-risk, n ¼ 916 Multiple high-risk, n ¼ 558 ASC-H, ACG, HSIL or invasive cancera (N ¼ 603)

Negative (N ¼ 136)

Negative, n ¼ 60 Untypable, n ¼ 25 Single-multiple low-risk, n ¼ 27 Multiple low-high-risk, n ¼ 163 Single high-risk, n ¼ 153 Multiple high-risk, n ¼ 175

Overall (N ¼ 4,251) a

Negative, n ¼ 1,017 Untypable n ¼ 380 Single/multiple low-risk, n ¼ 343 Multiple low/high-risk, n ¼ 709 Single high-risk, n ¼ 1,069 Multiple high-risk, n ¼ 733

CIN 1 (N ¼ 1,158) OR (95% CI)b

CIN 2 (N ¼ 213) OR (95% CI)

CIN 3 (N ¼ 284) OR (95% CI)b

Invasive cancer (N ¼ 13) OR (95% CI)b

Baseline 19.1 (11.7–31.3) 36.5 (22.3–59.6) 65 (40.3–104.7) 53.5 (34–84.5) 75.9 (46.9–122.8)

Baseline 12 (3.2–44.6) 23.3 (6.5–83.7) 113.1(35–365) 59.3 (18.5–189.8) 125.4 (38.7–406)

Baseline 5.2 (1.2–23.3) 15.5 (4.1–59) 163.1 (50.9–522) 72.6 (22.8–231.5) 244.4 (76.4–781.8)

Baseline undefined undefined 21.4 (2.3–195.9) 3.1 (0.19–49.3) 56.9 (6.7–21.4)

CIN 1 (N ¼ 78) OR (95% CI)b Baseline 35.8 (4–322.1) 66 (7.3–596.7) 49.6 (6–409.5) 37.3 (4.7–293) 44.2 (5.5–357.5)

Negative (N ¼ 2,116)

CIN 1 (N ¼ 1,236) OR (95% CI)b Baseline 19.9 (12.3–32.2) 38 (23.6–61.4) 63.6 (39.9–101.3) 52.7 (33.7–82.2) 72.6 (45.5–115.8)

CIN 2 (N ¼ 67) OR (95% CI) 12.4 27.6 56 28.7 48

(1.1–143) (2.6–294.8) (6.5–484.2) (3.4–238.5) (5.7–406.2)

CIN 2 (N ¼ 280) OR (95% CI) 12 24.7 106.6 56.1 117.3

(3.8–37.6) (8.2–74.3) (38.5–295) (20.4–153.9) (42.3–324.8)

CIN 3 (N ¼ 247) OR (95% CI)b 4.3 2.7 135.8 39.6 108.5

(0.5–34.7) (0.2–34.4) (29.4–628.4) (8.8–177.4) (23.9–492.9)

CIN 3 (N ¼ 531) OR (95% CI)b 4.71 10.3 195.5 71 242.2

(1.4–15.5) (3.4–31.2) (79.2–482.3) (28.8–174.7) (98.2–597.6)

Invasive cancer (N ¼ 75) OR (95% CI)b Baseline 1.6 (0.15–16.5) 4.4 (0.6–30.8) 23.3 (6.1–89.3) 14.7 (4.1–53.3) 20.5 (5.5–76.5) Invasive cancer (N ¼ 88) OR (95% CI)b Baseline 0.91 (0.1–8.2) 2.7 (0.49–15) 36 (12.3–105.3) 18.1 (6.3–52.6) 51 (17.6–147.8)

CIN, cervical intraepithelial neoplasia; LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesions; ASCUS, atypical squamous cells of undetermined significance; ASC-H, atypical squamous cells-cannot exclude HSIL; AGC, atypical glandular cells. a Seven subjects infected by HPV 69/71 or 74 were excluded because of undefined prognostic significance. b Odds Ratios (OR) and 95% confidence intervals (CI) were obtained by multinomial logistic regression equations including colposcopic/ histological results as outcomes (negative, CIN 1, CIN 2–3, invasive cancer). Explanatory variables inserted in the models were age (continuous variable), previous SIL treatment (yes, no), ancestry (italina, others), parity (yes, no), human immunodeficiency virus status (negative or unknown, positive), and category of HPV infection (negative, untypable, single/multiple low-risk, multiple low/high-risk, single high-risk, multiple high-risk).

J. Med. Virol. DOI 10.1002/jmv

1150

Spinillo et al.

intraepithelial neoplasia both in overall analysis and among subjects with low-grade cervical cytological lesions at enrolment. Among women referred because of atypical squamous cells or low-grade squamous lesions the likelihood of cervical intraepithelial neoplasia grade 1 among subjects with untypable HPV was lower than that associated with low risk HPV types (P < 0.001) (Table IV). In this subgroup, compared to uninfected subjects, the odds ratios of cervical intraepithelial neoplasia grade 2þ or grade 3þ were 7.3 (95% CI ¼ 2.9–18.5) and 3.9 (95% CI ¼ 0.96–15.5) among subjects with untypable HPV and 16.6 (95% CI ¼ 6.9–39.8, P ¼ 0.03, compared to untypable), and 11.5 (95% CI ¼ 3.4–38.8) in those with low risk genotypes, respectively. In the overall analysis, the odds ratios of the different categories of cervical intraepithelial neoplasia associated with untypable HPV infection were significantly lower compared to all categories of highrisk HPV infection (P < 0.01 for all comparisons). In addition, the odds ratios of cervical intraepithelial neoplasia grade 1 (P < 0.01) or grade 2 (P ¼ 0.06) associated with untypable HPV infection were lower when compared with those associated with low-risk HPV genotypes. Compared to uninfected subjects, the odds ratios of cervical intraepithelial neoplasia grade 2þ and grade 3þ associated with untypable infection and with low risk HPV genotypes were 5.73 (95% CI ¼ 2.79–11.78) and 12.4 (95% CI ¼ 6.31–24.5, P ¼ 0.014 compared to untypable) and 3.1 (95% CI ¼ 1.11– 8.16), and 7.1 (95% CI ¼ 2.9–17.2, P ¼ 0.07 compared to untypable), respectively. During the period of the study, 684 HPV-negative women (669 negative for cervical intraepithelial neoplasia and 15 with low-grade lesions), 205 subjects with untypable HPV infection (132 negative for cervical intraepithelial neoplasia and 73 with low-grade lesions), and 181 subjects (97 negative for cervical intraepithelial neoplasia and 84 with low-grade lesions) with single or multiple low-risk HPV infection attending the screening program of our Department were followed-up for a median of 25 (range 6–64) months. The rates of occurrence/persistence of cervical intraepithelial neoplasia grade 1 were 3.5% (24/684), 5.9% (12/205, P ¼ 0.45 compared to negative), and 14.4% (26/195, P < 0.001 and P ¼ 0.03 compared to negative and untypable) among negative, untypable, and low-risk HPV infection, respectively. The corresponding rates of progression to cervical intraepithelial neoplasia grade 2 or 3 in the same categories were 2% (8 grade 2 and 6 grade 3 lesions out of 684 subjects), 3.4% (4 grade 2 and 3 grade 3 lesions out of 205 subjects, P ¼ 0.83 compared to negative) and 5.6% (6 grade 2 and 5 grade 3 lesions out of 195 subjects, P ¼ 0.04 compared to negative), respectively. DISCUSSION The overall rates of HPV infection in the present study were similar to those reported in other studies J. Med. Virol. DOI 10.1002/jmv

of women attending colposcopic services because of abnormalities on Pap smear, both in Italy and elsewhere [Tornesello et al., 2006; Dickson et al., 2013; Kirschner et al., 2013]. HPV infections caused by untypable HPV types (HPV-X) according to the INNO-LiPA assay account for 11.7% of cases in this setting, most of the infections (95% of cases) occurring in women with negative or low-grade cervical lesions. During a median follow-up period longer than 24 months, the rate of progression to cervical intraepithelial neoplasia grade 2–3 of low-grade cytological squamous lesions associated with undefined HPV infection was low and comparable to that associated with uninfected subjects. One of the potential limitations of the study is related to the population studied. Pap smears were not institutionally reviewed, thus cytological diagnoses were not checked for consistency. In addition, the proportion of subjects with atypical squamous cells of undetermined significance was lower than in other studies [ALTS Group, 2003]. Among women with no HPV test, standard guidelines suggest that women with atypical squamous cells can be treated with follow-up cytology alone whereas for low-grade squamous lesions cases colposcopy is recommended [Wright et al., 2007]. Accordingly, among women attending colposcopic services for cytological abnormalities, lowgrade lesions cases may be overrepresented compared to those with atypical squamous cells. Using HPV SFP10 primers and subsequent LIPA genotyping probes on histological specimens, the prevalence of untypable HPV infection was 3.7% in a large, global series of 8977 HPV-positive cases of cervical cancer and 0.7% in 276 histological specimens of high-grade cervical intraepithelial neoplasia among Danish women [Geraets et al., 2012; Kirschner et al., 2013]. In the present study, the prevalence of untypable HPV infection among women with highgrade cervical intraepithelial neoplasia or cervical cancer (2.2% and 1.1%, respectively) was comparable to that reported in previous investigations [Geraets et al., 2012; Kirschner et al., 2013]. Data on the overall prevalence of untypable HPV infection among women with Pap smear abnormalities are limited. In a series of 5,964 women with cytological abnormalities, using MY09/MY11 consensus primers, Varnai et al. [2007] found 293 subjects (4.9%) with untypable HPV infection. Using SPF10 consensus primers, the overall rate (11.7%) of unknown HPV types among women with cytological abnormalities was significantly higher than that reported in the above study. Comparative studies have shown that HPV detection methods based on MY09/MY11 have intermediate analytical sensitivity, whereas PCR with SPF10 has the highest sensitivity, particularly at very low concentrations of HPV, which is very common among cytological samples [Snijders et al., 2003; Gravitt et al., 2008]. For this reason it is possible that the rate reported by Varnai et al. [2007] may be an underestimate. The results of the present

Untypable HPV Infection

study confirm that HPV cervical infection caused by unknown HPV types is more common among women with mild cytological abnormalities than in those with severe lesions or invasive cancers. However, after adjustment for potential confounding factors, untypable HPV infection was still associated with increased rates of cervical intraepithelial neoplasia when compared to uninfected samples. This association was independent of the cytological diagnosis causing colposcopic referral since it was confirmed both in low and high grade squamous lesions. The magnitude of the risk of high-grade cervical intraepithelial neoplasia associated with untypable HPV infection was significantly lower than that associated with known low-risk HPV types. The follow-up in the present series was limited to subjects attending the screening program of our Department and thus the prospective data are not necessarily representative of the entire population referred to the colposcopy clinic. However, the 5.6% progression rate to cervical intraepithelial neoplasia grade 2 or 3 in mild cervical cytological abnormalities associated with low-risk HPVs during a 2-year median follow-up is comparable to that reported by several large studies [Castle et al., 2011; Katki et al., 2013]. The low rate of progression to high grade cervical intraepithelial neoplasia of subjects with untypable HPV infection is reassuring and suggests that these women could be followed-up with the same protocol used for uninfected subjects. The biological reasons for unknown genotype assignment are multiple. Untypable HPV infections likely reflect different conditions: non-specific amplification of sequences unrelated to HPV (false positives), unknown HPV types, different types from those detected by INNO-LIPA, or molecular variants of known types. It is also possible that samples with a very low copy number did not permit a distinct genotype assignment, despite using a PCR method with high sensitivity and specificity [Geraets et al., 2012; Sichero et al., 2013]. The causes of failure of HPV typing have been investigated in the HPV infection in men (HIM) cohort [Giuliano et al., 2008; Sichero et al., 2013]. In this series, 66% of men tested positive for HPV in genital lesions, but viral types could not be identified in nearly 22% of subjects. Subsequent sequencing and nested PCR analysis of 931 untypable samples originating from the HIM study showed that the rates of false positive (i.e., PCR positive but genotyping negative) results comprised 11% of unclassified HPV infections [Sichero et al., 2013]. In this study it was possible to identify only 54% of undefined HPV infections and most of the a-HPV types identified by sequencing and nested PCR were low-risk HPVs (mostly HPV 6 and 61), which escaped identification by the usual genotyping test probably due to a very low copy number. Notably, in the study by Varnai et al. [2007], who sequenced 293 HPV-positive cervical cytological samples labeled previously as unknown HPV types, HPV

1151

42, HPV 6, and HPV 61 were the most common HPV types identified. It is possible that a low copy number could result from an HPV infection that is about to be cleared by the immunological system of the host. Studies on high-risk HPV suggest that HPV loads are consistently higher among younger women but heterogeneous across HPV genotypes [Ramanakumar et al., 2010], both in persistent and transient infection. Resolving HPV infections are consistently associated with low or declining copy numbers. The higher age of women with untypable HPV infection in the present series suggests that this group of subjects could comprise mostly women clearing an HPV infection caused by HPVs of low or undetermined risk. In the present study multiple infections were found in 46.1% of samples. This rate is significantly higher than that observed in other studies using histological samples [Kirschner et al., 2013] but similar to the prevalence of HPV infection detected in cytological specimens [Cuschieri et al., 2004; Trottier et al., 2006]. The reasons for the high rates of HPV identification in cytological compared to histological samples are probably due to the different pattern of sampling and PCR processing. Identification of HPV infection in histological samples is limited to the areas of cervical intraepithelial neoplasia whereas cytology allows the sampling of the whole cervix including peripheral areas of metaplastic epithelium infected by HPV [Cuschieri et al., 2004; Trottier et al., 2006; Kirschner et al., 2013]. Notably, the rate of multiple infections was directly associated with the increasing severity of intraepithelial neoplasia and invasive cervical cancer. Multiple HPV infection has previously been associated with increasing severity of intraepithelial neoplasia, younger age, and increased HPV viral load in other studies [Ramanakumar et al., 2010]. As studies of cervical intraepithelial neoplasia using laser capture and microdissection suggest that each lesion examined is caused by a single HPV, the higher risk of cervical intraepithelial neoplasia grade 2 or worse associated with multiple high-risk viruses is intuitive [Quint et al., 2012; van der Marel et al., 2012]. Less obvious are the reasons for the excess risk associated with multiple low and high-risk compared to single high-risk HPV infection. Worldwide genotyping studies confirm that the role of single low-risk HPV infection in invasive cervical cancer is limited (less than 1% of cases), but in almost 7% of specimens, cervical cancer tissue contains multiple HPV types [De Sanjose et al., 2010]. Epidemiological studies suggest both cooperative and competitive relationships between various types of HPV, and microdissection studies suggest a potential role of low-risk HPV in multiple infections associated with invasive cervical cancers [Geraets et al., 2012; Guimera` et al., 2013]. Future studies on large samples of cervical cancers associated with multiple HPV infection would be necessary to evaluate the patterns of association J. Med. Virol. DOI 10.1002/jmv

1152

Spinillo et al.

between various types of HPV and to elucidate the carcinogenic mechanisms associated with each HPV type. In conclusion, this study has shown that untypable HPV infection by a commonly used HPV genotyping system is found in more than 11% of subjects attending a colposcopic service because of Pap smear abnormalities. The weak association between untypable HPV infection and high grade cervical intraepithelial neoplasia or cervical cancer suggests that not including less common HPVs in current commercial genotyping assays should have very limited consequences on the evaluation and clinical follow-up of women with abnormal cytological findings. In fact, both in the triaging procedure of women with atypical squamous cells or low-grade squamous lesions or in the follow-up of subjects with previously treated cervical intraepithelial neoplasia, untypable HPV genotypes carry a risk of high grade cervical intraepithelial neoplasia or recurrence similar or lower to that associated with low-risk HPVs. REFERENCES ALTS Group. 2003. A randomized trial on the management of low grade squamous intraepithelial lesion cytological interpretations. Am J Obstet Gynecol 188:1393–1400. Bouvard V, Baan R, Straif K, Grosse Y, Secretan B, El Ghissassi F, Benbrahim-Tallaa L, Guha N, Freeman C, Galichet L, Cogliano V, WHO International Agency for Research on Cancer Monograph Working Group. 2009. A review of human carcinogens— Part B. Biological agents. Lancet Oncol 10:321–322. Castle PE, Gage JC, Wheeler CM, Schiffman M. 2011. The clinical meaning of cervical intraepithelial neoplasia grade 1 biopsy. Obstet Gynecol 118:1222–1229. Cuschieri KS, Cubie HA, Whitley MW, Seagar AL, Arends MJ, Moore C, Gilkisson G, McGoogan E. 2004. Multiple high risk HPV infections are common in cervical neoplasia and young women in a cervical screening population. J Clin Pathol 57:68–72. De Sanjose S, Quint WGV, Alemany L, Retrospective International Survey and HPV Time Trends Study Group. 2010. Human papillomavirus genotype attribution in invasive cervical cancer: A retrospective cross-sectional worldwide study. Lancet Oncol 11:1048–1056. Dickson EL, Vogel RI, Bliss RL, Downs LS Jr. 2013. Multiple type human papillomavirus (HPV) infections: A cross-sectional analysis of the prevalence of specific types in 309,000 women referred for HPV testing at the time of cervical cytology. Int J Gynecol Cancer 23:1295–1302. Geraets D, Alemany L, Guimera N, de Sanjose S, de Koning M, Molijn A, Jenkins D, Bosch X, Quint W, on behalf of the RIS HPV TT study group. 2012. Detection of rare and possibly carcinogenic human papillomavirus genotypes as single infections in invasive cervical cancer. J Pathol 228:534–543. Giuliano AR, Lazcano-Ponce E, Villa LL, Flores R, Salmeron J, Lee JH, Papenfuss MR, Abrahamsen M, Jolles E, Nielson CM, Baggio ML, Silva R, Quiterio M. 2008. The human papillomavirus infection in men study: Human papillomavirus prevalence and type distribution among men residing in Brazil, Mexico, and the United States. Cancer Epidemiol Biomarkers Prev 17:2036– 2043. Gravitt PE, Coutlee F, Iftner T, Sellors JW, Quint WG, Wheeler CM. 2008. New technologies in cervical cancer screening. Vaccine 26:K42–K52.

J. Med. Virol. DOI 10.1002/jmv

Guimera` N, Lloveras B, Alemany L, Iljazovic E, Shin HR, Jung-Il S, de Sanjose S, Jenkins D, Bosch FX, Quint W. 2013. Laser capture microdissection shows HPV 11 as both a causal and coincidental infection in cervical cancer specimens with multiple HPV types. Histopathology 63:287–292. Katki HA, Schiffman M, Castle PE, Fetterman B, Poitras NE, Lorey T, Cheung LC, Raine-Bennett T, Gage JC, Kinney WK. 2013. Five year risk of CIN 2þ and CIN 3þ among women with HPVpositive and HPV-negative Pap results. J Low Genit Tract Dis 17:S43–S49. Kirschner B, Schledermann D, Holl K, Rosenlund M, Raillard A, Quint W, Molijn A, Jenkins D, Junge J. 2013. HPV-genotypes in high-grade intraepithelial cervical lesions in Danish women. Acta Obstet Gynecol Scand 92:1032–1040. Quint W, Jenkins D, Molijn A, Struijk L, van de Sandt M, Doorbar J, Mols J, Van Hoof C, Hardt K, Struyf F, Colau B. 2012. One virus, one lesion. Individual components of CIN lesions contain a specific HPV type. J Pathol 227:62–71. Ramanakumar AV, Goncalves O, Richardson H, Tellier P, Ferenczy A, Coutle´e F, Franco EL. 2010. Human papillomavirus (HPV) types 16,18,31,45 DNA loads and HPV-16 integration in persistent and transient infections in young women. BMC Infect Dis 10:326–338. Safaeian M, Herrero R, Hildesheim A, Quint W, Freer E, Van Doorn LJ, Porras C, Silva S, Gonza´lez P, Bratti MC, Rodriguez AC, Castle P, Costa Rican Vaccine Trial Group. 2007. Comparison of the SPF10-LiPA system to the hybrid capture 2 assay for detection of carcinogenic human papillomavirus genotypes among 5,683 young women in Guanacaste, Costa Rica. J ClinMicrobiol 45:1447–1454. Scott JL, Freese J. 2006. Regression models for categorical dependent variables using Stata. 2nd edition. College Station. TX: Stata Press. Sichero L, Pierce Campbell CM, Ferreira S, Sobrinho JS, Luiza Baggio M, Galan L, Silva RC, Lazcano-Ponce E, Giuliano AR, Villa LL, HIM Study Group. 2013. Broad HPV distribution in the genital region of men from the HPV infection in men (HIM) study group. Virology 443:214–217. Snijders PJ, van den Brule AJ, Meijer CJ. 2003. The clinical relevance of human papillomavirus testing: Relationship between analytical and clinical sensitivity. J Pathol 201:1–6. Tornesello ML, Duraturo ML, Botti G, Greggi S, Piccoli R, De Palo G, Montella M, Buonaguro L, Buonaguro FM, Italian HPV Working Group. 2006. Prevalence of alpha-papillomavirus genotypes in cervical squamous intraepithelial lesions and invasive cervical carcinoma in the Italian population. J Med Virol 78: 1663–1672. Trottier H, Mahmud S, Costa MC, Sobrinho JP, Duarte-Franco E, Rohan TE, Ferenczy A, Villa LL, Franco EL. 2006. Human papillomavirus infection with multiple types and risk of cervical neoplasia. Cancer Epidemiol Biomarkers Prev 15:1274– 1280. van der Marel J, Quint WG, Schiffman M, van de Sandt MM, Zuna RE, Dunn ST, Smith K, Mathews CA, Gold MA, Walker J, Wentzensen N. 2012. Molecular mapping of high grade cervical intraepithelial neoplasia shows etiological dominance of HPV 16. Int J Cancer 131:E946–E953. van Hamont D, van Ham MA, Bakkers JM, Massuger LF, Melchers WJ. 2006. Evaluation of the SPF10-INNO LiPA human papillomavirus (HPV) genotyping test and the Roche linear array HPV genotyping test. J Clin Microbiol 44:3122–3129. Varnai AD, Bollmann M, Ba´nkfalvi A, Griefingholt H, Pfening N, Schmitt C, Pajor L, Bollmann R. 2007. The spectrum of cervical diseases induced by low risk and undefined-risk HPVs: Implications for patient management. Anticancer Res 27:563–570. Wright TC, Jr. Massad LS, Dunton CJ, Spitzer M, Wilkinson EJ, Solomon D, ASCCP-Sponsored Consensus Conference. 2007. 2006 consensus guidelines for the management of women with abnormal cervical screening tests. J Low Genit Tract Dis 11: 201–222.