HPV genotypes in CIN 2-3 lesions and cervical cancer: A population-based study

Authors


  • Conflict of interest statement: Dr. F.J. Taddeo is an employee of Merck Research Laboratories, a division of Merck & Co., Inc. and holds stock and stock options in the Company.

Abstract

The distribution of human papillomavirus (HPV) varies between countries and continents leading to different effectiveness of upcoming prophylactic HPV vaccines. This study analyses the HPV distribution in CIN 2-3, recurrent CIN 2-3 and cervical cancer in Iceland. About 80% of incident cases with CIN 2-3 lesions in 1990 and 1999, 99% of cancer cases in 1990–1994 and 1999–2003, and cases with recurrent CIN 2-3 after conization in 1990 were tested with PCR analysis for the presence of 12 oncogenic HPV types. About 95% of the CIN 2-3 and 92% of the cancer cases tested positive for the included HPV types. HPV 16 was the most frequent type followed by HPV 33, 31, 52, 35, 18, 58, 56, 39, 45, 59 in CIN 2-3 and by HPV 18, 33 45, 31, 39, 52, 35, 51, 56 in cancer. HPV 16 and 18 were associated with a significantly increased cancer risk and HPV 52 and 31 with decreased cancer risk compared to the risk of CIN 3. The HPV distribution differed between histological cancer types, stages and age groups. The number of HPV types was not a significant predictor of cancer. Oncogenic HPV types were found in all persistent or recurrent CIN 2-3 disease after conization. Vaccination against HPV 16/18 is estimated to achieve a minimum 40% reduced rate of CIN 2-3 and a minimum 60% reduced cancer rate. This rate could, however, be increased to 95% and 92% respectively by including all the 12 HPV types tested for in this study. © 2007 Wiley-Liss, Inc.

Epidemiologic data have established human papillomavirus (HPV) as the central agent of cervical cancer.1, 2 HPV is a heterogeneous group of double-stranded closed circular deoxyribonucleic acid (DNA) viruses with more than 200 genotypes, subtypes and variants. Of these at least 18 are considered oncogenic or probably oncogenic.3, 4 HPV infections reach their peak prevalence soon after sexual debut5 and decline with advancing age through acquired natural immunity, thus allowing for development of prophylactic type-specific HPV vaccines.3

Persistent infection with oncogenic HPV types is considered to be the most important risk factor for the development of preinvasive and invasive cervical disease.6, 7 Most of the preinvasive lesions, however, regress spontaneously and clinical management is based on the severity of these lesions.8, 9 Thus, low-grade cytological (LSIL: low-grade squamous cell intraepithelial lesions) and histological lesions (CIN 1: low grade cervical intraepithelial lesions) are most commonly observed while moderate to high-grade cytology (HSIL: high-grade squamous cell intraepithelial lesions) cases are referred for colposcopy and the higher grade histological lesions (CIN 2-3: moderate to high-grade cervical intraepithelial lesions) referred for ablative therapy.

Meta-analyses of published data on HPV-type specific prevalence have shown that the HPV-distribution is different in CIN 2-310 from that in cancer.11 Thus, HPV16, 18 and 45 were more prevalent in squamous cell carcinoma than HSIL, whereas the reverse was true for other oncogenic types including HPV 31, 33, 52 and 58. These studies also raised the possibility of localized variation in the distribution and malignant potential of particular HPV types,10 which have implications for the effect of different combinations of type-specific vaccines on both the incidence of cancer and the workload of ongoing Pap smear screening programs. The aim of this study was (i) to analyze the distribution of HPV genotypes in CIN 2-3 and cancer in Iceland, (ii) to evaluate the presence of these types in persistent and recurrent CIN 2-3 lesions after conization and (iii) to estimate the effect of prophylactic type-specific vaccines on cancer incidence and CIN 2-3 risk.

Material and methods

The screening population

Iceland is an island situated in the North Atlantic with a population of around 300,000. Pap smear screening started in 1964 and was national-based in the age group 25–69 after 1969. Because of increased rates of CIN 2-3 and invasive cancer after 1980 in women under age 35 the screening guidelines were reformed and coordinated for spontaneous and centralized screening to guarantee a complete follow-up strategy for all women with abnormal screening results. The lower age limit was decreased to 20 in 1988 with 2–3 year screening intervals. The screening program and trends in CIN 2-3 and cancer have been described in earlier publications.12, 13, 14

Standard screening operating procedures

The cytological examination of Pap smears was carried out at 2 cytological laboratories (∼80% at the Cytological Laboratory of the Cancer Society) and defined according to the Bethesda classification.8 Histological examination of biopsies was carried out at 3 pathology laboratories and classified according to the CIN classification.9 All material from the spontaneous and organized screening is collected at the data bank of the Cancer Detection Clinic, which also supervised follow-up and treatment of these women. According to standard operating procedures, women with HSIL, ASCUS-H (atypical squamous cells of uncertain significance - cannot exclude high-grade lesion) or a repeat LSIL/ASCUS smears were referred for colposcopic biopsy. Women with CIN 2-3 and microinvasive cancer were referred for conization. During the study period women with a nonradical cone biopsy (CIN at or within 2 mm from the surgical margins) were referred for repeat Pap smear and colposcopic biopsy 6–12 months after the operation. Women with radical cone biopsies were controlled by a repeat Pap smear 6 months after the operation and thereafter at 12 months.

The study material

The CIN 2-3 cases in this study were gathered from a nation-based data register and limited to cases with no earlier registered CIN lesions. Because of increasing rates of preinvasive and invasive disease after 1980 and organized screening from the age of 20 in 1988 the cases were selected during 2 time-periods from 1990 to evaluate possible changes in the HPV-distribution by time. About 80% of the lesions diagnosed in 1990 (196/244) and 1999 (245/300) were retrieved together with 99% (142/144) of the cervical cancer cases diagnosed in 1990–1994 (2 cases missing in 1994) and 1999–2003. Women operated on with a cone biopsy in 1990 were followed up to the year 2004 for repeat biopsies. The histology of the colposcopic and cone biopsies was reviewed and the area with the most abnormal morphologic picture in either of these biopsies selected for HPV DNA analysis.

HPV DNA analyses

The HPV biopsies were tested at Merck Research Laboratories (Division of Merck & Company, USA) without knowledge of other data. The biopsy material was tested for the high-risk (HPV 16, 18), intermediate risk (HPV 31, 33, 45, 52, 58) and low-risk oncogenic HPV types (HPV 35, 39, 51, 56, 59) using multiplexed PCR-based assays, as described earlier.15 In principle, the assay uses type-specific primers to test for the presence of 3 HPV genes, L1, E6 and E7 in each sample. Amplification of genomic DNA is used as a positive control for DNA quality. Samples were designated “Nonamplified” if amplification of genomic DNA failed in one or more of the type-specific assays.

Statistical analysis

Significance of differences in proportions was tested by the χ2 test. When numbers were low, the conditional binomial exact test was also used to check the result. Logistic regression was used to test prognostic power of the various HPV types on the odds ratio of the result of histological samples being CIN 3 versus CIN 2 and cancer versus CIN 3 adjusted for age and year of diagnosis. The software package used was SPIDA16 and the significance level chosen was 0.05.

Results

HPV analyses were carried out on 92 women with CIN 2, 349 with CIN 3 and 141 with cancer after excluding one widely disseminated small cell tumor NOS (48-year-old woman diagnosed in 1994 with negative HPV tests) from the cancer material. About 92% of cancer, 97% of CIN 3 and 92% of CIN 2 cases tested positive for 1 or more of the 12 HPV types included in this study. The distributions of the single and multiple HPV types in cervical cancer, CIN 3 and CIN 2 are shown in Table I.

Table I. Distribution of HPV Types as Single Types and Combinations of Types in CIN 2 and CIN 3 During the Years 1990 + 1999 and in Cancer During the Periods 1990–1994 + 1999–2003 (All Age Groups)
HPV test resultsCIN 2CIN 3Cancer
071311
162113563
1821518
311262
3392813
35583
39233
45346
51330
527132
56330
58880
59010
16-18133
16-18-31020
16-18-33-45010
16-18-51010
16-31283
16-31-33-52010
16-31-51010
16-31-51-52010
16-31-52030
16-33173
16-33-45100
16-33-51010
16-33-52200
16-33-52-56010
16-33-56010
16-35100
16-35-39010
16-35-51010
16-35-52-56010
16-39133
16-39-51010
16-39-56010
16-45042
16-51021
16-51-52010
16-51-56010
16-520101
16-52-56010
16-52-59010
16-56011
18-31030
18-33100
18-31-33010
18-39-52100
18-51010
18-52020
31-33010
31-33-35-45010
31-33-52001
31-35130
31-39100
31-51-52-58010
31-51-58010
31-52010
31-56-59010
33-39010
33-39-51010
33-45010
33-51010
33-52001
33-56110
33-58110
35-16100
35-45100
35-52100
35-52-58010
39-51010
45-51010
51-52110
52-56020
52-58010
56-58100
58-59100
Nonamp001
Total92349141

In CIN 2 and CIN 3 the distribution of the different HPV types (both single and in combinations with other types) did not differ significantly between the years 1990 and 1999. For the cancer cases the HPV distribution was not significantly different in women with single HPV types. However, when analyzed for single and multiple types combined, HPV 16 was found in a marginally higher proportion (46/76 vs. 34/65; p = 0.045) and HPV 18 was found in a marginally smaller proportion (7/76 vs. 14/65; p = 0.047) of cancer cases in 1999–2003 compared to 1990–1994. Cervical cancer was more frequently diagnosed at stage IA (p = 0.035) in 1999–2003 compared to 1990–1994. The age distributions for CIN 2, CIN 3, squamous and adeno- and adenosquamous carcinoma were similar between years and time periods. In the following analyses, data for individual years and time periods are combined (1990 and 1999 vs. 1990–1994 and 1999–2003).

The distribution of HPV types for CIN 2 and CIN 3 are shown in Table II. The proportion of cases positive for a single HPV type was the same (CIN 2: 70% vs. CIN 3: 71%). About 5% of the CIN 2 (5 cases) and CIN 3 (18 cases) lesions were positive for only 1 or 2 of the genes tested (L1, E6 or E7) and only 1% (4 cases) of the CIN 3 lesions failed in 1 or more of the type-specific assays and thus produced incomplete data. HPV 16 was most prevalent (34% in CIN 2 vs. 56% in CIN 3). In CIN 2 HPV 16 was followed by 33, 52, 58, 35 (17–11% in decreasing order); 18, 31, 39, 45, 56, 51 (5–4%) and 59 (1%) and in CIN 3 followed by 31, 33, 52, 18 (16–8%); 51, 35, 56, 58, 39, 45 (6–3%) and 59 (1%). The distribution of HPV types for women under age 35 and older women (Table III) was not significantly different in CIN 2 but for CIN 3 only HPV 16 (p = 0.006) and 51 (p = 0.03) were significantly more prevalent under age 35.

Table II. Distribution of HPV Types in CIN 2 and CIN 3 Lesions (% Calculated from the Total Number of Cases in Respective Group; All Age Groups)
HPV typesCIN 2CIN 3
TotalSingle typeTotalSingle type
N%N%N%N%
  1. Years 1990 and 1999 combined.

16313421231955613555
185522298154
3155115516267
3316179104914288
3510115516582
39552212331
45553312341
51443321631
521213784212134
56553314431
58111267134103
59110 3110.3
Negative78  134  
Cases total92100647034910024771
Table III. Number (N) of HPV Types (Single Type or in Combination with Other Types) in CIN 2, CIN 3 and Cervical Cancer of Stage IA and Higher Stages According to Age Groups at Diagnosis
HPV typesAge < 35Age 35+
1990 + 19991990–1994 + 1999–20031990 + 19991990–1994 + 1999–2003
  Cancer  Cancer
CIN 2CIN 3Stage IAStage IB+CIN 2CIN 3Stage IAStage IB+
NNNNNNNN
16241461967491540
183193221088
315382101721
33113121518510
35612004403
39510100214
4528023415
51419000201
529291131303
56411001301
58811003200
5902001100
Negative36004729
No DNA amplified       1
Cases total662402411261093076

The distribution of HPV types in the cancer cases is shown in Table IV. The proportion of women with single positive HPV tests was 81%, 1 case (1%) was positive to only 1 of the HPV genes (E6) and 3 cases (2%) failed in 1 or more of the types-specific assays; of those, 2 cases produced incomplete data and 1 case contained no amplifiable DNA. HPV 16 was most prevalent (57%) followed by 18, 33 (15%–13%), 45, 31, 39, 52 (6–4%), 35, 51, 56 (2–1%). The prevalence of HPV 16 was higher (p = 0.04) in women under age 35 (25/35) compared to the older age group (55/106) (Table III).

Table IV. Distribution of HPV Types in Cervical Cancer (% Calculated from the Total Number of Cases in Respective Group; All Age Groups)
HPV typesCancer
TotalSingle type
N%N%
  1. Time periods 1990–1994 and 1999–2003 combined.

1680576344
1821151813
316421
331813139
353232
396432
458664
51110 
525421
56110 
580 0 
590 0 
Negative118  
No DNA amplified11  
Cases total14110011581

The distribution of HPV types in the cancer cases according to stage and histology is shown in Table V. In stage IA HPV 16 was most frequent (63%), followed by 18 (20%), 33 (13%), 31, 39, 45 and 52 (4–2%). In stage IB+, HPV 16 was also most common (53%), followed by 18, 33 (11% each), 45 (8%), 39, 52, 35 (5–3%), 51 and 56 (1%). For all ages the HPV distribution did not differ between stage IA and higher stages (≥IB), but for stage IA HPV 16 was more prevalent (p = 0.03) in women under the age of 35 compared to older women.

Table V. Distribution of HPV Types in Cervical Cancer According to Stages and Histology (% of Total Number of Cases in Each Category; All Age Groups)
HPV typesCancer
Stage IAStage IB+SquamousAdsqAd
N%N%N%N%N%
  1. Time periods 1990–1994 and 1999–2003 combined.

163463465363625461243
1811201011773271139
31362266  0 
3371310111717  14
350 3333  0 
39244555190 
4512784421827
510 1111  0 
52124544  0 
560 1111  0 
580 0 0   0 
590 0 0   0 
02491044  621
No DNA amplified0 1111    
Cases total54100871001021001110028100

In squamous cell carcinoma HPV 16 was the most frequent type (62%) followed by 33 (17%), 18, 31, 39 (5–7% each), 45, 52, 35 (4–3%), and 51, 56 (1% each). In adeno- and adenosquamous carcinoma HPV 16 was most frequent (43 and 46%), followed by 18 (39 and 27%), 45 (7 and 18%), 33 (4 and 0%) and 39 (0 and 9%). HPV 16 was more prevalent (OR: 2.53; 95% CI: 1.13–5.63; p = 0.03) and HPV 18 less prevalent (OR: 0.16; 95%CI: 0.06–0.44; p < 0.001) in squamous cell carcinoma compared to adeno- and adenosquamous carcinoma. No difference was observed between women under age 35 and those older than 35 for their respective histological group (data not shown).

In the CIN 2-3 cases combined, HPV 16, 18 or both were diagnosed as the only involved HPV types in 40% (177/441) and HPV 16, 18, 31, 33, 45, 52, 58 in 74% (327/441) of the cases (Table I). Under age 35 the proportion of HPV 16/18 as the only involved HPV types was 41% (126/306) compared to 38% (51/135) in the older age group (p = 0.51), and the proportion of HPV 16, 18, 31, 33, 45, 52 and 58 was the same in both age groups or 76% (233/306 vs. 103/135) (data not shown). Combined, the 12 HPV types were found in 95% (421/441) of these cases (Tables II and III).

The distribution of multiple HPV infections was not significantly different in CIN 2 and CIN 3 (Table I). Multiple HPV infections were, however, more frequent (p = 0.001) among women under age 35 with CIN 3 and cancer (data not shown). After adjustment for age and time period the number of HPV types was not a significant predictor of cancer compared to CIN 3 (p = 0.16). No interaction was observed between age and the number of HPV types.

In the cancer cases HPV 16, 18 or both were diagnosed as the only involved HPV types in 60% (84/141) and HPV 16, 18, 31, 33, 45, 52 and 58 in 84% (118/141) of the cases (Table I). Under age 35 the frequency of HPV 16/18 was 77% (27/35) compared to 54% (57/106) in the older age group (p = 0.015), and the proportion of HPV 16, 18, 31, 33, 45, 52, 58 was 97% (34/35) versus 79% (84/106), respectively (p = 0.01) (data not shown). Combined, the 12 HPV types were found in 92% (129/141) of these cases (Table III).

The results of the logistic regression analyzing the prognostic power of individual HPV types are shown in Table VI. HPV 31 (OR: 3.53; 95% CI: 1.36–9.19) and 16 (OR: 2.50; 95% CI: 1.53–4.10) were associated with a significantly increased risk of being diagnosed with CIN 3 compared to CIN 2, whereas HPV 58 (OR: 0.26; 95% CI: 0.11–0.52) was associated with significantly decreased risk of developing the higher grade lesion. HPV 18 (OR: 2.66; 95% CI: 1.34–5.30) and 16 (OR: 1.76; 95% CI: 1.07–2.89) were associated with a significantly increased risk of being diagnosed with cancer compared to CIN 3 whereas HPV 52 (OR: 0.23; 95% CI: 0.08–0.70) and 31 (OR: 0.34; 95% CI: 0.14–0.86) were associated with significantly decreased risk (Table VI).

Table VI. Predictive Power of Different HPV Types on the Odds Ratio (or) of Being Diagnosed with CIN 3 Compared to CIN 2 and Cancer Compared to CIN 3
CIN 3 compared to CIN 2Cancer compared to CIN 3
HPV typesOR95% CIp-valueHPV typesOR95%p-value
  1. Univariant analysis adjusted for age and calendar year (all age groups).

162.501.53–4.10<0.001161.761.07–2.890.02
181.520.56–4.100.41182.661.34–5.300.005
313.531.36–9.190.01310.340.14–0.860.02
330.770.41–1.460.43330.690.33–1.400.30
350.850.49–1.460.54350.220.04–1.350.10
390.840.28–2.510.75391.320.38–4.600.66
450.610.20–1.820.38451.880.67–5.250.23
511.490.49–4.570.48510.100.01–1.170.07
520.960.48–1.930.91520.230.08–0.700.009
560.890.30–2.590.83560.100.01–1.050.55
580.260.11–0.520.00258   
590.850.08–8.900.8959   

About 30% (73/245) of the women who underwent conization in 1990 were, according to the standardized clinical protocol, followed for 14 years and referred for a second colposcopic biopsy (63 women after nonradical and 10 after radical cone biopsies). Seven women (7/73: 10%; 7/245: 3%) were diagnosed with a repeat CIN 2-3 lesion (3 within 12 months after nonradical cone biopsies and 4 more than 12 months after radical cone biopsies). All 7 had at that time an abnormal smear and a positive single type HPV results (3 cases with persistent HPV 16, 3 with HPV 31 [one persistent], 1 case with HPV 35). One of the women with a persistent HPV 16 infection was diagnosed 12 years later with advanced cervical cancer.

Discussion

On the basis of meta-analyses of published PCR-based studies, Clifford et al.10, 11 have estimated the average worldwide frequency of HPV to be 84% (regional variation 76–87%) in women with combined higher grade cytological (HSIL) and histological lesions (CIN 2-3) and 86% (83–89%) in women with invasive disease. The population prevalence of HPV infections varies, however, between individual studies, depending on the material and HPV DNA detection assays used. This population-based study showed 96% positivity in CIN 2-3 and 92% in cancer for the 12 high, intermediate and low cancer risk HPV types tested for.

Worldwide meta-analyses have shown fluctuations in the distribution of the oncogenic HPV types between regions and a notably high prevalence of HPV 31, 52 and 58 in some areas outside Europe.10, 11 An increasing prevalence of HPV 33, 39, 52 and 58, decreasing prevalence of HPV 45, and decreasing tendency of HPV 18 prevalence with time have also been reported.17 This study confirmed only for the cancer cases a marginally higher (p = 0.045) proportion of HPV 16 and a marginally lower proportion (p = 0.047) of HPV 18 in 1999–2003 compared to 1990–1994.

In this study, HPV 16 was the most prevalent type in both CIN 2-3 (51%) and cancer (57%) followed in combined CIN 2-3 by 33, 31, 52 (12–15% each) and in cancer by 18, 33 (13–15% each), with other types below 8%. The frequencies of these HPV types were higher than those reported by Clifford et al.10, 11 for Europe: in CIN 2-3 HPV 33 (15% vs. 8.4%), 52 (12% vs. 2.4%) and 31 (14% vs. 10.5%) and in cancer HPV 33 (13% vs. 4.4%). These results can indicate geographical differences but can also depend on the greater variability in the material included in the studies by Clifford et al.

The oncogenic HPV types 16, 18 and 45 have been reported to be more prevalent in cancer than CIN 2-3 whereas the reverse is true for other oncogenic types including HPV 31, 33, 52 and 58.10 This was confirmed in this study and the results were significant for HPV 16, 18, 31 and 52. This study also confirmed that HPV 16 and 31 were more prevalent and HPV 58 less prevalent in CIN 3 compared to CIN 2. The results thus confirm the high-risk potential of HPV 16 and 18 but could not verify the oncogenic potential for HPV 58 and 59.

HPV 16 and phylogenetically related types (HPV 31, 33, 35, 52 and 58) have been reported to be more prevalent in squamous cell carcinoma, whereas HPV 18 and related types (HPV 39, 45, 59 and 68) are more prevalent in adenocarcinoma.11, 17, 18, 19 In this study HPV 16 was significantly (p = 0.03) more prevalent in squamous cell carcinoma and HPV 18 was significantly (p < 0.001) less prevalent in squamous cell carcinoma compared to the other histologic types.

Prior infection by one HPV type has been reported as not promoting the risk of persistent infection by another type.17, 20 The oncogenic types are thus said to act independently and convey a statistically similar risk of developing cervical cancer.18 HPV 16 and 18 are considered high-risk oncogenic types due to the fact that persistence is more likely for these types.21 Current HPV vaccine trials22, 23 are therefore mainly based on these types. These vaccines will, however, with certainty be followed by vaccines that are also including other oncogenic HPV types.

The overall worldwide prevalence of high-risk (HPV 16/18) and intermediate (HPV 31,33,45,52,58) types has been estimated as 52% (regional variation 38–66%) and 30% (26–35%) in HSIL+CIN 2-3,10 indicating that the CIN 2-3 risk could be decreased by 52% by vaccines including HPV 16/18 and up to 82% by a vaccine including both the high-risk and the intermediate risk types. This study indicates that these reductions would be 40% (177/441) for cases only infected by HPV 16/18 and 77% (339/441) for cases only infected with the intermediate and high-risk types (Table I). Including all the 12 oncogenic HPV types included in this study would, however, reduce the CIN 2-3 rate by 95% (421/441), a finding similar to other reports.10

In cervical cancer cases the overall worldwide prevalence of high-risk and intermediate risk HPV types has been estimated as 67% (regional variation 62–77%) and 17% (11–23%), respectively.12 It has thus been estimated that the reduction in the worldwide cancer incidence could theoretically be increased from 68–71% to 86–87% by adding the intermediate types to the current HPV 16/18 vaccines.11, 17, 18 This study indicates that these reductions would be lower or minimum 60% (84/141) for cases infected only with HPV 16/18 and 84% (118/141) for cases infected only with the intermediate and high-risk types (Table I). Including all the 12 HPV types tested for in this study would, however, reduce the cancer rate by 92% (130/141), a finding similar to other reports.11, 17, 18

For the cancer cases the present study confirmed a markedly higher prevalence of the high-risk HPV (77%) and combined high and intermediate risk HPV types (97%) in women under age 35 compared to women older than 35 years of age. This was mainly due to the markedly higher prevalence of HPV 16 in this age group, indicating a higher prophylactic effect of HPV 16 vaccines among the younger women. This was also observed in CIN 3 but not in CIN 2. Although multiple infections were markedly higher in the younger age group, multiple HPV infections did not increase cancer risk compared to infections with single types.

Finally, all the CIN 2-3 cases who underwent conization in 1990 were followed up for 14 years. Only 3% (7/245) had recurrent CIN 2-3 lesions and all of these had persistent or a new oncogenic HPV infection and all had an abnormal smear. The study thus indicates that oncogenic HPV-testing is effective in selecting out women at risk for persistent or recurrent disease after conization.24

The study results lead to the following main conclusions: (i) HPV 16 and 31 are associated with a significantly increased risk of CIN 3 as opposed to CIN 2; HPV 16 is associated with a significantly increased risk of squamous cell carcinoma, mainly at stage IA and in younger women; HPV 18 is associated with increased risk of adeno- and adenosquamous carcinoma; the number of HPV types in a sample was not a significant predictor of cancer; (ii) the results shows that oncogenic HPV types are present in persistent and recurrent CIN 2-3 lesions after conization and, finally, (iii) vaccines based on HPV 16/18 could theoretically reduce the CIN 2-3 risk by minimum 40% and the cancer risk by minimum 60%; the vaccine effect could, however, be increased to 95% for CIN 2-3 and 92% for cancer by including all the 12 HPV types tested for in this study.

Ancillary