• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jcmPermissionsJournals.ASM.orgJournalJCM ArticleJournal InfoAuthorsReviewers
J Clin Microbiol. Oct 2008; 46(10): 3437–3445.
Published online Aug 20, 2008. doi:  10.1128/JCM.00620-08
PMCID: PMC2566086

Comparison of Two PCR-Based Human Papillomavirus Genotyping Methods[down-pointing small open triangle]

Abstract

We compared two consensus primer PCR human papillomavirus (HPV) genotyping methods for the detection of individual HPV genotypes and carcinogenic HPV genotypes as a group, using a stratified sample of enrollment cervical specimens from sexually active women participating in the NCI/Costa Rica HPV16/18 Vaccine Efficacy Trial. For the SPF10 method, DNA was extracted from 0.1% of the cervical specimen by using a MagNA Pure LC instrument, a 65-bp region of the HPV L1 gene was targeted for PCR amplification by using SPF10 primers, and 25 genotypes were detected by reverse-line blot hybridization of the amplicons. For the Linear Array (LA) method, DNA was extracted from 0.5% of the cervical specimen by using an MDx robot, a 450-bp region of the HPV L1 gene was targeted for PCR amplification by using PGMY09/11 L1 primers, and 37 genotypes were detected by reverse-line blot hybridization of the amplicons. Specimens (n = 1,427) for testing by the LA method were randomly selected from strata defined on the basis of enrollment test results from the SPF10 method, cytology, and Hybrid Capture 2. LA results were extrapolated to the trial cohort (n = 5,659). The LA and SPF10 methods detected 21 genotypes in common; HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, -68, and -73 were considered the carcinogenic HPV genotypes. There was no difference in the overall results for grouped detection of carcinogenic HPV by the SPF10 and LA methods (35.3% versus 35.9%, respectively; P = 0.5), with a 91.8% overall agreement and a kappa value of 0.82. In comparisons of individual HPV genotypes, the LA method detected significantly more HPV16, HPV18, HPV39, HPV58, HPV59, HPV66, and HPV68/73 and less HPV31 and HPV52 than the SPF10 method; inclusion of genotype-specific testing for HPV16 and HPV18 for those specimens testing positive for HPV by the SPF10 method but for which no individual HPV genotype was detected abrogated any differences between the LA and SPF10 methods. The LA method detected more carcinogenic-HPV-genotype infections per specimen than the SPF10 method (P < 0.001). In conclusion, the LA method and the SPF10 method with HPV16 and HPV18 genotype-specific detection among ungenotyped HPV-positive specimens were comparable for detection of HPV16 and HPV18, the two HPV genotypes targeted by current prophylactic HPV vaccines. Both approaches are suitable for monitoring the impact of HPV16/18 vaccines in clinical trials.

It is now well established and widely accepted that virtually all cervical cancer and its immediate precancerous lesions are caused by persisting cervical infections by approximately 15 cancer-associated (carcinogenic) human papillomavirus (HPV) genotypes (20, 34, 37). The most important of these HPV genotypes are HPV16 and HPV18, which account for approximately 70% of all cancer of the cervix, with minor variations between continents (29). Prophylactic HPV vaccines targeting HPV16 and HPV18 have shown better than 90% efficacy for preventing persistent infections and cervical precancer caused by these genotypes in women not already infected with these genotypes (7, 8, 21).

Critical-to-large-scale evaluation of these and future HPV vaccines is the genotype-specific measurement of HPV infections, particularly HPV16 and HPV18, for defining genotype-specific endpoints of HPV persistence and cervical precancer (cervical intraepithelial neoplasia grade 2 [CIN2] and grade 3 [CIN3]). We have primarily relied on a PCR method based on SPF10-LiPA25 (16, 17) to measure HPV genotypes in a large, community-based phase III clinical trial of approximately 7,500 women aged 18 to 25 and living in Guanacaste, Costa Rica (14, 25). We recently demonstrated that detection of carcinogenic HPV by the SPF10-LiPA25 method was in good concordance with Hybrid Capture 2 (hc2; Qiagen, Gaithersburg, MD) (25), an FDA-approved test for carcinogenic HPV. However, hc2 does not distinguish which specific carcinogenic HPV genotype(s) is present, and therefore, we could not evaluate the performance of the SPF10-LiPA25 system for detection of individual HPV genotypes.

We were therefore interested in further validating the use of SPF10-LiPA25 in our trial by a post hoc comparison to another PCR-based HPV genotyping method, Linear Array (LA; Roche Molecular Systems), which uses PGMY09/11 L1 consensus primers (10). PCR methods using PGMY09/11 primers have been extensively used in epidemiologic studies of HPV (9, 23, 24, 27, 31, 35). This PGMY-based method detects 37 HPV genotypes, including all known carcinogenic HPV genotypes. We have recently completed several validation studies of this PGMY-based method, showing good analytic and clinical performance (3, 4, 11). Our primary objective in this analysis was to compare HPV genotyping results from these two protocols/methods, one based on the PGMY primers and one based on the SPF10 primers, by testing a set of baseline, prerandomization cervical specimens from a stratified random sample of women enrolled in the vaccine trial.

MATERIALS AND METHODS

Study population.

The primary aim of the Costa Rica Vaccine Trial (CVT) is to independently assess the efficacy of an HPV16/18 vaccine (Cervarix; GlaxoSmithKline, Rixensart, Belgium) for prevention of persistent HPV16/18 infections and HPV16/18-related precancerous lesions (CIN2, CIN3, or adenocarcinoma in situ) as surrogates of risk for invasive cervical cancer. Enrollment began in June 2004 and ended in December 2005. The study participants were women between 18 to 25 years old who were in good general health, had no history of chronic conditions that required treatment, were willing to use birth control during the vaccination period, and were living in the study area without plans of imminent departure. Approximately one-third of the females identified in a previous census fulfilled the inclusion criteria and participated in the study (14). Details of the CVT, including enrollment procedures, exams, and specimen collections, are reported elsewhere in detail (13).

This analysis was based on enrollment, prevaccination specimens from women entering the vaccine trial. All study protocols were reviewed and approved by the NCI and a Costa Rican institutional review board. All participants provided written, informed consent.

HPV detection and genotyping. (i) hc2 testing.

hc2 is a clinical test that collectively targets 13 carcinogenic HPV genotypes (HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, and -68) without distinguishing the HPV genotype present (26). The hc2 assay was performed according to the manufacturer's instructions with residual PreservCyt (Cytyc, Marlborough, MA) samples after they were used for cytology.

(ii) SPF10-LIPA25 testing.

Total DNA was isolated from 200 μl of one of the 500-μl PreservCyt aliquots by using a MagNA Pure LC instrument (Roche Diagnostics, Almere, The Netherlands) and a Total DNA isolation kit (Roche Diagnostics). DNA was eluted in 100 μl of water. Each DNA extraction run contained positive and negative controls to monitor the DNA isolation procedure.

A 10-μl aliquot of extracted DNA was used for each SPF10 PCR (4% of the 500-μl aliquot or 0.1% of the entire specimen). The SPF10 PCR primer set (DDL Diagnostic Laboratory, Voorburg, The Netherlands) was used to amplify a broad spectrum of HPV genotypes as described earlier (16, 17). Briefly, this primer set amplifies a small fragment of 65 bp from the L1 region of HPV. The reverse primers contain a biotin label at the 5′ end, enabling the capture of the PCR amplicon onto streptavidin-coated microtiter plate wells. Captured amplicons are denatured by alkaline treatment and then detected by a defined cocktail of digoxigenin-labeled probes targeting a broad spectrum of HPV genotypes. Incubation steps with an anti-digoxigenin alkaline phosphatase conjugate and then an alkaline phosphatase substrate were used for detection. This method, designated the HPV DNA enzyme immunoassay (DEIA), provides an optical density value. If the SPF10-DEIA test yielded a borderline value (75 to 100% of the cutoff value), the SFP10 PCR and DEIA tests were repeated. Each DEIA batch contained separate positive, borderline, and negative controls. The broad-spectrum SFP10 primers can recognize at least 54 HPV genotypes.

The resultant SPF10 amplicons (from SPF10-DEIA-positive samples) were used to identify the HPV genotype by reverse hybridization on a line probe assay (LiPA25, version 1; Labo Bio-medical Products, Rijswijk, The Netherlands), containing probes for 25 different HPV genotypes (HPV6, -11, -16, -18, -31, -33, -34, -35, -39, -40, -42, -43, -44, -45, -51, -52, -53, -54, -56, -58, -59, -66, -68/73, -70, and -74). Each LiPA25 run contained negative and positive controls. Since the interprimer regions of HPV68 and -73 are identical, the LiPA system cannot distinguish between HPV68 and -73; hence, they are assigned as HPV68/73. SPF10-LiPA25 results were available for all samples.

(iii) Genotype-specific HPV testing.

Because the vaccine trial uses a bivalent HPV16/18 vaccine, genotype-specific PCR primer sets were also used to optimize the detection of these two HPV genotypes. Specimens that tested positive by SPF10 but did not contain HPV16 or HPV18 according to LiPA25 were selectively amplified for HPV16 (TS16) and HPV18 (TS18) (32). The genotype-specific primers were based on those described by Baay et al. (1); they generate amplicons of 92 and 126 bp for HPV16 and HPV18, respectively. Amplicons from the genotype-specific PCRs were detected by DEIA, similar to the method for SPF10 amplicon detection.

(iv) LA testing.

The second 500-μl aliquot of PreservCyt was used for testing by LA, a commercialized, research-use-only, L1 consensus primer-based PCR method that employs a primer set designated PGMY09/11 (10). LA testing was done masked to the all other data. Amplicons were subjected to reverse-line blot hybridization for detection of 37 individual HPV genotypes (HPV6, -11, -16, -18, -26, -31, -33, -35, -39, -40, -42, -45, -51 to -56, -58, -59, -61, -62, -64, -66 to -73, -81 to -84, -82v, and -89) (10, 22).

Because of intellectual property rights, LA does not directly detect HPV52 but combines a set of probes that detects HPV33, -35, -52, and -58 combined (HPVmix). Specimens that test negative for HPV33, -35, and -58 individually but are positive for the HPVmix are considered to be HPV52 positive. The specimens that test positive for HPV33, -35, and/or -58 and the HPVmix have an uncertain HPV52 status, and for this analysis, these specimens were considered to be HPV52 negative. LA was used according to the manufacturer's instructions in the product insert, which involve DNA extraction using a QIAamp MinElute Media kit (Qiagen, Inc., Valencia, CA). The only deviation from the LA product insert protocol was in implementing an automated sample preparation for extraction of up to 96 specimens at a time on the Qiagen MDx platform (using the MinElute Media MDx kit according to the manufacturer's instructions) rather than processing 24 specimens per batch with the manual vacuum method (3, 11). LA testing used 21% of the aliquot or 0.5% of the entire specimen.

It is noteworthy that the primers used for amplification and amplicon detection did not comprise the only differences between the assays. Each assay used different extraction procedures and different amounts of purified DNA in the PCR as well as different primers and procedures for amplicon detection. To reflect that no single step in either assay was comparable to the corresponding step in the other method, we will henceforth refer to MagNa Pure DNA extraction, SPF10 primer PCR amplification, and LiPA25 HPV genotype detection as the SPF10 method and MDx DNA extraction, PGMY09/11 primer PCR amplification, and LA HPV genotype detection as the LA method.

Statistical analysis.

The SPF10 method results were available for all 5,871 sexually experienced women enrolled in the CVT group. At the time that aliquots were selected for testing by the LA method, 5,659 (96%) of the specimens had been tested by the SPF10 method and thus defined our analytic sample. A stratified, random sample of 1,508 specimens was selected on the basis of SPF10 method, hc2, and cytology results as shown in Table Table11 to oversample for cytological abnormalities and discordant test results (e.g., hc2 positive/SPF10 method negative/indicative of low-grade squamous intraepithelial lesion [LSIL] cytology). We sampled the least from common categories, such as specimens that were hc2 negative/SPF10 method negative/indicative of negative cytology and hc2 positive/SPF10 method positive for carcinogenic HPV/indicative of negative cytology. One batch of 81 aliquots selected for LA method testing was lost due to DNA extraction robot failure, and thus, a total of 1,427 specimens (95% of the selected aliquots and 25% of the analytic sample) had results for both the SPF10 and LA methods for a direct comparison. Unless specified, HPV16 and HPV18 genotype-specific PCR testing results were not used in our comparison of the SPF10 and LA methods.

TABLE 1.
Sampling schema used to select specimens for testing by the LA method for comparison with the SPF10 methoda

To account for the use of a stratified random sample based partially on the SPF10 method results to which we were comparing LA method results, we used the sampling fractions to extrapolate the results for both HPV genotyping methods for the 1,427 specimens to the entire analytic sample of 5,659. We calculated the prevalences for (i) the individual HPV genotypes detected by both HPV genotyping methods, (ii) the group of carcinogenic HPV genotypes (HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, and -68/73), (iii) the group of 21 HPV genotypes detected by both methods, (iv) all detected HPV genotypes, and (v) the overall HPV prevalence (for the LA method, we included all genotypes detected; for the SPF10 method, we included all DEIA-positive results, whether or not a specific HPV genotype was detected). We also included the crude paired results for each targeted HPV genotype detected by both methods.

For detection of any carcinogenic HPV genotype by either method, paired extrapolated results were stratified on hc2 status (positive versus negative) and on the cytological interpretation (negative versus nonnegative) of the ThinPrep slide at the clinical site. Nonnegative cytology was defined as the presence of atypical squamous cells of undetermined significance (ASC-US) or greater severity (≥ASC-US).

We explored the design idiosyncrasies of both methods resulting in equivocal HPV genotyping (e.g., the SPF10-DEIA-positive/LiPA25-negative results, the SPF10 method-positive results for HPV68/73, and the LA method-positive results for HPV52) to better understand the meaning of these results. This was accomplished by comparison of the results from one method to the relevant results of the other.

We categorized our HPV genotyping results in two additional ways. First, the numbers of carcinogenic HPV genotypes detected by each method were categorized (zero, one, two, and three or more) and compared. Second, the results for both methods were compared after the results for each were assigned to an HPV risk group according to the following a priori-established cervical cancer risks: (i) positive result for HPV16; (ii) else positive result for HPV18; (iii) else positive result for any carcinogenic HPV genotype and negative result for HPV16 and HPV18 (carcinogenic HPV, excluding HPV16 and HPV18); (iv) else positive result for any noncarcinogenic HPV genotype and negative result for all carcinogenic genotypes, including any SPF10 method-positive result for which no HPV genotype was detected (SPF10-DEIA positive/LiPA25 negative); and (v) negative PCR. The groups, in descending order of risk, are positive HPV16 result > positive HPV18 result > positive carcinogenic-HPV result (excluding HPV16 and HPV18) > positive noncarcinogenic-HPV result > negative PCR.

The results for both HPV genotyping methods were compared by calculating percent overall agreement, percent agreement on testing positive, and kappa and tested for statistical differences (P < 0.05) using an exact-symmetry or McNemar χ2 test. For each analysis, each of the 1,427 results was weighted by the inverse of the probability of sampling each for testing and stratified on hc2 result, cytology, and SPF10 method result (Table (Table1).1). The sampling weights were accounted for in the testing and variance estimation with a leave-one-out jackknife variance estimator (18).

RESULTS

Detection of carcinogenic HPV.

We first compared the results for detection of any carcinogenic HPV genotype by the two HPV genotyping methods as shown in Table Table2.2. There was no difference in the results for detection of one or more carcinogenic HPV genotypes by the two methods (35.3% by the SPF10 method and 35.9% by the LA method; P = 0.5), with a 91.8% overall group agreement, a 79.3% agreement on testing positive, and a kappa of 0.82. There were no significant differences in the overall results for detection of carcinogenic HPV for the two methods as stratified on hc2 test results (negative versus positive) or cytological interpretation (normal cytology versus ≥ASC-US).

TABLE 2.
Comparison of results for detection of carcinogenic HPV genotypes (HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -66, and -68/73) by the SPF10 and LA methods for all womena

Detection of individual HPV genotypes.

The LA method detected significantly more HPV16, HPV18, HPV39, HPV58, HPV59, HPV66, and HPV68/73 and less HPV31 and HPV52 than the SPF10 method, as shown in the extrapolated results in Table Table3.3. The LA method also detected more combined HPV16 and HPV18 (10.7%) than the SPF10 method (9.6%) (P = 0.01). For other genotypes targeted by both methods, the LA method detected significantly more HPV40, HPV42, HPV54, and HPV70 and less HPV11 than the SPF10 method. While the overall agreement levels for individual HPV genotypes targeted by both methods were 95% or greater, the levels of agreement on testing positive ranged from 82.0% (for HPV16) to 6.1% (for HPV42, which was very uncommonly detected by the SPF10 method).

TABLE 3.
Comparison of results for detection of individual HPV genotypes by the SPF10 and LA methods

The testing algorithm for CVT includes HPV16 and HPV18 genotype-specific PCR testing. Upon combining the results for genotype-specific PCR testing and the SPF10 method for detection of HPV16 and HPV18, we found no differences between the results for detection of HPV16/18 by the combined testing and the LA method (10.8% versus 10.7%, respectively) (P = 0.6). The overall agreement was 98.3%, the agreement on testing positive was 85.6%, and the kappa was 0.91.

Of the additional 27 HPV16-positive specimens detected by the LA method compared to the number detected by the SPF10 method in the crude results for the 1,427 paired tests, 19 (70.3%) were confirmed by HPV16 genotype-specific testing. Of the additional 26 HPV18-positive specimens detected by the LA method compared to the number detected by the SPF10 method, 12 (46.2%) were confirmed by HPV18 genotype-specific testing.

The SPF10 method sometimes gave HPV-positive results by DEIA, but no HPV genotype was identified by LiPA25. As shown in Table Table4,4, among the SPF10-DEIA-positive/LiPA25-negative specimens sampled in this analysis, the LA method primarily tested positive for noncarcinogenic HPV genotypes (61.6%), with another 23.0% testing negative for all genotypes, and 15.4% tested positive for any carcinogenic HPV genotype. SPF10-DEIA-positive/LiPA25-negative specimens were more likely to test negative for carcinogenic HPV by the LA method when hc2 was also negative than when hc2 was positive (8.5% versus 31.3%, respectively; P < 0.001). HPV73 (2.7%), HPV51 (2.4%), HPV52 (1.8%), HPV58 (1.8%), and HPV59 (1.8%) were the most common carcinogenic HPV genotypes, and HPV61 (15.7%), HPV84% (11.8%), and HPV89 (10.6%) were the most common noncarcinogenic HPV genotypes detected by the LA method among SPF10-DEIA-positive/LiPA25-negative specimens (data not shown). We further evaluated a random selection of 81 of the 331 SPF10-DEIA-positive/LiPA25-negative specimens by DNA sequencing the SPF10 amplicons (see the table in the supplemental material). We identified an HPV genotype in 63 of 81 specimens (78%). The most common HPV genotypes detected by sequencing SPF10-positive/LiPA25-negative specimens were HPV30 (n = 15; 19%) (an HPV genotype not detected by the LA method), HPV61 (n = 11; 14%) (9 of 11 were confirmed by the LA method), and HPV67 (n = 8; 10%) (8 of 8 were confirmed by the LA method).

TABLE 4.
Distribution of HPV cancer risk groups as determined by the LA method among those specimens that tested positive by the SPF10 method but for which no HPV genotype was detected by LiPA25a

Comparing the two HPV genotyping methods permitted us to examine the indirect strategy used by the LA method to detect HPV52, i.e., specimens that test positive for a pool of HPV genotype probes specific for HPV33, HPV35, and HPV58 and cross-react with HPV52 but test negative for HPV33, HPV35, and HPV58 individually are assumed to be HPV52 positive. We observed that 70.5% of those inferred as HPV52 positive by the LA method were confirmed by the SPF10 method (Table (Table5,5, HPV52 row). The most common genotypes detected by the SPF10 method among those inferred to be HPV52 positive by the LA method but HPV52 negative by the SPF10 method were HPV16 (22.6%), HPV51 (16.1%), and HPV18 (12.9%). We note that the percentage of HPV52-positive specimens, as detected by the SPF10 method with those specimens that were determined to be HPVmix positive/HPV52 negative by the LA method (7.4%), was nonetheless significantly higher (P < 0.001; χ2 test of expected versus observed results) than that for specimens that tested HPVmix negative (2.4%). As expected, some multiple-genotype infections (~5%) of HPV52 and HPV33, HPV35, and/or HPV58 were misclassified as HPV52 negative by the LA method in our analysis and would have been called indeterminate according to the LA product insert.

TABLE 5.
Examination of the indirect method of detecting HPV52 by the LA methoda

Next, we evaluated the relationship of the SPF10 method results for HPV68/73 with individual detection of HPV68 and HPV73 by the LA method (Table (Table6).6). The SPF10 method was more apt to test positive for HPV68/73 among those specimens that tested LA method positive for HPV68 and negative for HPV73 than among those that tested negative for HPV68 and positive for HPV73 (60% versus 35.6%, respectively; P = 0.04). Similar patterns were observed among those with normal cytology and those with nonnormal cytology (≥ASC-US) (data not shown) (nota bene, because hc2 is designed to detect HPV68 but not HPV73, we did not use it to stratify the paired results). However, among the specimens testing positive for HPV68/73 by the SPF10 method, a greater fraction were positive for HPV73 by the LA method alone than for HPV68 (37.5% versus 32.1%).

TABLE 6.
Comparison between detection of HPV68/73 by the SPF10 method and separate detection of HPV68 and HPV73 by the LA methoda

Number of carcinogenic HPV genotypes.

There was 82.4% overall agreement between the two methods for number of carcinogenic HPV genotypes (Table (Table7),7), with a kappa of 0.67 and a linearly weighted kappa of 0.74. The LA method detected more carcinogenic HPV genotype infections per specimen than the SPF10 method (P < 0.001); the LA method detected more infections with two or more carcinogenic HPV genotypes (15.4% versus 11.8%) and more infections with three or more carcinogenic HPV genotypes (5.6% versus 3.5%) than the SPF10 method. A similar pattern was observed when we considered all 21 HPV genotypes targeted by both methods (data not shown). When the data were stratified by hc2 results, there was no significant difference found in the numbers of carcinogenic HPV genotypes among hc2 negatives, while the difference between the two methods was highly significant among the hc2-positive specimens (P < 0.001) (data not shown).

TABLE 7.
Comparison of numbers of carcinogenic HPV types detected by the SPF10 and LA methodsa

HPV risk groups.

When the results were categorized according to the risk groups as described in Materials and Methods (Table (Table8),8), there was an 83.3% overall agreement and an 86.9% agreement found among specimens testing HPV positive by both methods. The LA method had a tendency to categorize women into higher risk groups than the SPF10 method (P < 0.001), primarily because the LA method classified more women as HPV16 (n = 437) or HPV18 (n = 167) positive than the SPF10 method (n = 404 for HPV16 and n = 141 for HPV18). This was mainly the result of reclassification by the LA method of women testing positive for carcinogenic HPV by the SPF10 method as HPV16 positive (n = 33) or as HPV18 positive (n = 37), whereas reclassification by the SPF10 method of the LA method results was relatively uncommon (n = 5 for HPV16 and n = 2 for HPV18). Though restricted to results in which both tests were positive for carcinogenic HPV, the LA method was still more likely to categorize the individual test results into a riskier category (HPV16 or HPV18) than the SPF10 method (P < 0.001).

TABLE 8.
Comparison of SPF10 and LA test results, categorized hierarchically according to HPV cancer riska

DISCUSSION

We evaluated two PCR methods of HPV genotyping based on two well-established primers, SPF10 and PGMY09/11, with the following goals. First, we were interested in further validating the performance of the SPF10 method for detection of individual HPV genotypes. Although we recently completed a comparison of the SPF10 method to hc2 and found that the two methods agree well on the overall detection of carcinogenic HPV (25), hc2 does not provide HPV genotype-specific data to permit comparisons of individual HPV genotypes. PCR methods using PGMY09/11 have been commonly used for HPV genotyping in HPV natural history studies (3, 11, 22, 27), and therefore, the LA method served as a reasonable point of comparison for this analysis. Second, there are few very large comparisons of any two HPV genotyping methods, and such comparisons inform researchers about the impact of choice of HPV genotyping methods on their results. As discussed previously (12) and confirmed here, each method has its own idiosyncrasies. Theoretically, data such as we present here can be used to adjust for differences between methods and better allow data to be combined in meta-analyses when two methods with different strengths and weaknesses are used.

We again emphasize that we compared two HPV genotyping methods, not just different primer systems, and these methods differed in the following ways: (i) DNA extraction and purification (MDx for the LA method versus MagnaPure LC for the SPF10 method), (ii) amount of purified DNA amplified (0.5% for the LA method versus 0.1% for the SPF10 method), (iii) PCR primers (PGM09/11 for the LA method versus SPF10 for the SPF10 method), and (iv) method of amplicon detection (LA for the LA method versus LiPA25 for the SPF10 method). In general, we found the performances for both methods very similar for the overall detection of carcinogenic HPV, as has previously been reported for a smaller set of specimens (33). However, for individual HPV genotypes, the LA method detected more HPV16, HPV18, HPV39, HPV58, HPV59, HPV66, and HPV68/73, while the SPF10 method detected more HPV11, HPV31, and HPV52.

We also observed that the LA method was able to detect more multigenotype infections and a greater number of HPV genotypes per multigenotype infection than the SPF10 method. The increased detection of multigenotype infections by the LA method compared to that by the SPF10 method could be the result of false positives or increased analytic sensitivity. However, 340 of the 577 (58.9%) additional HPV genotypes detected by the LA method were judged (by P. E. Gravitt) to have HPV genotype band strengths of 3 or 4 (on a 1-to-4 scale, with 1 being the weakest and 4 being the strongest), suggesting that at least a portion of the additional pick-ups were true positives. At this time, we cannot attribute these differences in detection between HPV genotyping methods to any one procedural step.

We note that one previous study compared these PCR primer systems by using the same input amount of DNA (0.1% of a PreservCyt specimen) from the same extraction and found results more similar (33) than those observed here, providing more evidence of the importance of standardization and optimization of the front ends (e.g., DNA purity and quantity) of these methods for their performance (6). Yet, in this previous report (33), differences were still found even when DNA purity and quantity were controlled for, with LA detecting marginally more HPV16 and HPV59 and detecting marginally less HPV51 than SPF10-LiPA25. Thus, each step in a PCR-based HPV genotyping method (DNA extraction, DNA input, use of primers, and amplicon detection) influences the test results and must be taken into consideration in any evaluation of method performance. We plan to explore the impact of these parameters on the test performance of the two HPV genotyping methods in subsequent discrepancy analyses, which should also address any possible false-positive/negative results obtained by either assay.

We found that the SPF10 method used in this study detected slightly less HPV16 and HPV18 than the LA method, but the difference, albeit statistically significant, was small. Combining HPV16 and HPV18 genotype-specific PCR assays with the SPF10 method, per the testing algorithm for the vaccine trial, improved the ascertainment of these genotypes and achieved results similar to those for the LA method. The HPV genotype-specific testing confirmed that at least half of the additional detection by the LA method represented true positives; the reason(s) for the less-than-100% confirmation by HPV genotype-specific testing are uncertain. Because the intermethod differences are small, it may be adequate to use either HPV genotyping method as a stand-alone measurement to monitor the outcomes within HPV16 and HPV18 vaccine phase 3 clinical trials or phase 4 surveillance trials or for the next generation of HPV vaccine trials that will include a broader array of HPV genotypes.

As previously reported in a comparison of LA to a line blot assay (3), the indirect method for HPV52 ascertainment by LA led to an underestimation of the prevalence of that genotype when samples positive for HPV33, -35, or -58 were presumptively considered HPV52 negative. As shown in the previous study, this indirect method of detection was in fair agreement only with direct detection by the other method, in this case an SPF10-based method. However, the HPV52 status cannot be discerned when women test positive for one or more of these other HPV genotypes included in the pooled probes individually as well as in the mixed probes. For assessing HPV persistence in a clinical setting, it will likely be necessary to consider all women positive for the mixed probe to be HPV52 positive, regardless of the other HPV genotypes detected, to maximize the detection of persistent HPV52 at the cost of falsely categorizing some women as having persistent carcinogenic HPV. An alternative might be to use an independent HPV52-specific PCR method (5, 30) to confirm all mixed-probe-positive results if the method is cost-effective and user-friendly.

Because the median and mean age of the women participating in this HPV vaccine trial was only 21 years, there were few cases of confirmed CIN2 or worse (≥CIN2) (CIN2, n = 14; CIN3, n = 26), limiting the analytic power to assess clinical sensitivity. Although precancerous lesions in young women tend to be smaller than those diagnosed in older women (28), 97.5% and 100% of the ≥CIN2 subjects tested positive for carcinogenic HPV by the SPF10 method and the LA method, respectively, and 100% of the CIN3 subjects tested positive for carcinogenic HPV by both methods.

The optimal clinical application of HPV genotyping in cervical cancer screening and clinical management decisions has yet to be determined. There is evidence that individual detection of HPV16 and HPV18 may be useful for deciding who among carcinogenic-HPV-positive, cytologically negative women might benefit from immediate colposcopy (HPV16 and/or HPV18 positive) versus a 1-year follow-up (HPV16 and HPV18 negative) (15, 36). There is increasing evidence that detection of viral persistence over a year or two could be clinically useful (2, 19) for identifying women at risk for cervical precancer and cancer, but the best format for those tests (e.g., partial versus full HPV genotyping) requires additional evaluations (2).

Our analysis was limited by the complex sampling based on SPF10 method, hc2, and cytology results, which were used to define the subset of specimens for LA testing, weighting in favor of the most informative specimens. As a consequence, we could estimate only by extrapolation the analytic performance of the LA method for the full cohort. Estimates of the SPF10 method results extrapolated from the 1,427 tests within the sampling scheme to the full cohort of 5,659 agreed well with the empirical testing results for the full cohort of 5,659, which suggests that our estimates for the LA method are fairly accurate. Nonetheless, we acknowledge that our methods may have led to small errors in our evaluation of the LA method.

We conclude that both HPV genotyping methods showed excellent agreement for common HPV genotypes detected in baseline cervical specimens collected from women participating in the HPV vaccine trial. Based on these data, we suggest that the use of both methods will provide an accurate estimate for the main outcomes in HPV vaccine trials, and both will be useful in studies of the natural history of HPV. Each test has its strengths and weaknesses, including differences in detection of individual HPV genotypes, and these differences should be considered when choosing a method of HPV genotyping for a specific application.

Supplementary Material

[Supplemental material]

Acknowledgments

The CVT group is a longstanding collaboration between investigators in Costa Rica and the NCI. Our trial was conducted in agreement with the Ministry of Health of Costa Rica. The NCI and the Costa Rica investigators make the final editorial decisions on this presentation and subsequent publications; GSK has the right to review/comment.

Vaccines were provided for our trial by GSK Biologicals under a Clinical Trials Agreement with the NCI. GSK also provided support for aspects of the trial associated with the regulatory submission needs of the company. Some of the equipment and supplies used in these studies were provided at a reduced cost by Roche Molecular Systems Inc., Alameda, CA. We thank Meera Sangaramoorthy, Dana Ford, and Kennita Riddick for the Roche LA testing.

Our trial was sponsored by the NCI (N01-CP-11005), with support from the NIH Office for Research on Women's Health. This research was also supported (in part) by the Intramural Research Program of the NIH, National Cancer Institute.

The affiliations of the CVT group are as follows: Proyecto Epidemiológico Guanacaste, Fundación INCIENSA, San Jose, Costa Rica (Mario Alfaro, Manuel Barrantes, M. Concepcion Bratti, Fernando Cárdenas, Bernal Cortés, Albert Espinoza, Yenory Estrada, Paula Gonzalez, Diego Guillén, Rolando Herrero, Silvia E. Jimenez, Jorge Morales, Lidia Ana Morera, Elmer Pérez, Carolina Porras, Ana Cecilia Rodriguez, and Maricela Villegas); University of Costa Rica, San Jose, Costa Rica (Enrique Freer, Jose Bonilla, Sandra Silva, Ivannia Atmella, and Margarita Ramírez); U.S. National Cancer Institute, Bethesda, MD (Nora Macklin, Allan Hildesheim, Douglas R. Lowy, Mark Schiffman, John T. Schiller, Mark Sherman, Diane Solomon, and Sholom Wacholder); SAIC, NCI-Frederick, Frederick, MD (Ligia Pinto and Alfonso Garcia-Pineres), Women & Infants' Hospital, Providence, RI (Claire Eklund and Martha Hutchinson); and DDL Diagnostic Laboratory, Voorburg, The Netherlands (Wim Quint and Leen-Jan van Doorn).

Footnotes

[down-pointing small open triangle]Published ahead of print on 20 August 2008.

Supplemental material for this article may be found at http://jcm.asm.org/.

REFERENCES

1. Baay, M. F., W. G. Quint, J. Koudstaal, H. Hollema, J. M. Duk, M. P. Burger, E. Stolz, and P. Herbrink. 1996. Comprehensive study of several general and type-specific primer pairs for detection of human papillomavirus DNA by PCR in paraffin-embedded cervical carcinomas. J. Clin. Microbiol. 34745-747. [PMC free article] [PubMed]
2. Castle, P. E. 2008. Invited commentary: is monitoring of human papillomavirus infection for viral persistence ready for use in cervical cancer screening? Am. J. Epidemiol. 168138-144. [PMC free article] [PubMed]
3. Castle, P. E., P. E. Gravitt, D. Solomon, C. M. Wheeler, and M. Schiffman. 2008. Comparison of linear array and line blot assay for detection of human papillomavirus and diagnosis of cervical precancer and cancer in the atypical squamous cell of undetermined significance and low-grade squamous intraepithelial lesion triage study. J. Clin. Microbiol. 46109-117. [PMC free article] [PubMed]
4. Castle, P. E., M. Sadorra, F. Garcia, E. B. Holladay, and J. Kornegay. 2006. Pilot study of a commercialized human papillomavirus (HPV) genotyping assay: comparison of HPV risk group to cytology and histology. J. Clin. Microbiol. 443915-3917. [PMC free article] [PubMed]
5. Coutlée, F., D. Rouleau, G. Ghattas, C. Hankins, S. Vézina, P. Cote, J. Macleod, A. de Pokomandy, D. Money, S. Walmsley, H. Voyer, P. Brassard, and E. Franco. 2007. Confirmatory real-time PCR assay for human papillomavirus (HPV) type 52 infection in anogenital specimens screened for HPV infection with the linear array HPV genotyping test. J. Clin. Microbiol. 453821-3823. [PMC free article] [PubMed]
6. Dunn, S. T., R. A. Allen, S. Wang, J. Walker, and M. Schiffman. 2007. DNA extraction: An understudied and important aspect of HPV genotyping using PCR-based methods. J. Virol. Methods 14345-54. [PubMed]
7. FUTURE II Study Group. 2007. Quadrivalent vaccine against human papillomavirus to prevent high-grade cervical lesions. N. Engl. J. Med. 3561915-1927. [PubMed]
8. Garland, S. M., M. Hernandez-Avila, C. M. Wheeler, G. Perez, D. M. Harper, S. Leodolter, G. W. Tang, D. G. Ferris, M. Steben, J. Bryan, F. J. Taddeo, R. Railkar, M. T. Esser, H. L. Sings, M. Nelson, J. Boslego, C. Sattler, E. Barr, and L. A. Koutsky. 2007. Quadrivalent vaccine against human papillomavirus to prevent anogenital diseases. N. Engl. J. Med. 3561928-1943. [PubMed]
9. Giuliano, A. R., M. Papenfuss, M. Abrahamsen, C. Denman, J. G. de Zapien, J. L. Henze, L. Ortega, E. M. Brown de Galaz, J. Stephan, J. Feng, S. Baldwin, F. Garcia, and K. Hatch. 2001. Human papillomavirus infection at the United States-Mexico border: implications for cervical cancer prevention and control. Cancer Epidemiol. Biomarkers Prev. 101129-1136. [PubMed]
10. Gravitt, P. E., C. L. Peyton, T. Q. Alessi, C. M. Wheeler, F. Coutlee, A. Hildesheim, M. H. Schiffman, D. R. Scott, and R. J. Apple. 2000. Improved amplification of genital human papillomaviruses. J. Clin. Microbiol. 38357-361. [PMC free article] [PubMed]
11. Gravitt, P. E., M. Schiffman, D. Solomon, C. M. Wheeler, and P. E. Castle. 2008. A comparison of linear array and hybrid capture 2 for detection of carcinogenic human papillomavirus and cervical precancer in ASCUS-LSIL triage study. Cancer Epidemiol. Biomarkers Prev. 171248-1254. [PMC free article] [PubMed]
12. Gravitt, P. E., and R. P. Viscidi. 2004. Measurement of exposures to human papillomavirus, p. 119-142. In T. E. Rohan and K. H. Shah (ed.), Cervical cancer: from etiology to prevention. Springer, Dordrecht, The Netherlands.
13. Herrero, R., A. Hildesheim, A. C. Rodriguez, S. Wacholder, C. Bratti, D. Solomon, P. Gonzalez, C. Porras, S. Jimenez, D. Guillen, J. Morales, M. Alfaro, J. Cyr, K. Morrisey, Y. Estrada, B. Cortes, L. A. Morera, E. Freer, J. Schussler, J. Schiller, D. Lowy, and M. Schiffman. 18 July 2008, posting date. Rationale and design of a community-based double-blind randomized clinical trial of an HPV 16 and 18 vaccine in Guanacaste, Costa Rica. Vaccine [Epub ahead of print.] doi:.10.1016/j.vaccine.2008.07.002 [PMC free article] [PubMed] [Cross Ref]
14. Hildesheim, A., R. Herrero, S. Wacholder, A. C. Rodriguez, D. Solomon, M. C. Bratti, J. T. Schiller, P. Gonzalez, G. Dubin, C. Porras, S. E. Jimenez, and D. R. Lowy. 2007. Effect of human papillomavirus 16/18 L1 viruslike particle vaccine among young women with preexisting infection: a randomized trial. JAMA 298743-753. [PubMed]
15. Khan, M. J., P. E. Castle, A. T. Lorincz, S. Wacholder, M. Sherman, D. R. Scott, B. B. Rush, A. G. Glass, and M. Schiffman. 2005. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J. Natl. Cancer Inst. 971072-1079. [PubMed]
16. Kleter, B., L. J. Van Doorn, L. Schrauwen, A. Molijn, S. Sastrowijoto, J. Ter Schegget, J. Lindeman, B. ter Harmsel, M. Burger, and W. Quint. 1999. Development and clinical evaluation of a highly sensitive PCR-reverse hybridization line probe assay for detection and identification of anogenital human papillomavirus. J. Clin. Microbiol. 372508-2517. [PMC free article] [PubMed]
17. Kleter, B., L. J. Van Doorn, J. Ter Schegget, L. Schrauwen, K. van Krimpen, M. Burger, B. ter Harmsel, and W. Quint. 1998. Novel short-fragment PCR assay for highly sensitive broad-spectrum detection of anogenital human papillomaviruses. Am. J. Pathol. 1531731-1739. [PMC free article] [PubMed]
18. Korn, J. E., and B. I. Graubard. 1999. Analysis of health surveys. John Wiley & Sons, New York, NY.
19. Koshiol, J., L. Lindsay, J. M. Pimenta, C. Poole, D. Jenkins, and J. S. Smith. 2008. Persistent human papillomavirus infection and cervical neoplasia: a systematic review and meta-analysis. Am. J. Epidemiol. 168123-137. [PMC free article] [PubMed]
20. Muñoz, N., F. X. Bosch, S. de Sanjose, R. Herrero, X. Castellsague, K. V. Shah, P. J. Snijders, and C. J. Meijer. 2003. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N. Engl. J. Med. 348518-527. [PubMed]
21. Paavonen, J., D. Jenkins, F. X. Bosch, P. Naud, J. Salmeron, C. M. Wheeler, S. N. Chow, D. L. Apter, H. C. Kitchener, X. Castellsague, N. S. De Carvalho, S. R. Skinner, D. M. Harper, J. A. Hedrick, U. Jaisamrarn, G. A. Limson, M. Dionne, W. Quint, B. Spiessens, P. Peeters, F. Struyf, S. L. Wieting, M. O. Lehtinen, and G. Dubin. 2007. Efficacy of a prophylactic adjuvanted bivalent L1 virus-like-particle vaccine against infection with human papillomavirus types 16 and 18 in young women: an interim analysis of a phase III double-blind, randomised controlled trial. Lancet 3692161-2170. [PubMed]
22. Peyton, C. L., P. E. Gravitt, W. C. Hunt, R. S. Hundley, M. Zhao, R. J. Apple, and C. M. Wheeler. 2001. Determinants of genital human papillomavirus detection in a US population. J. Infect. Dis. 1831554-1564. [PubMed]
23. Richardson, H., M. Abrahamowicz, P. P. Tellier, G. Kelsall, R. du Berger, A. Ferenczy, F. Coutlee, and E. L. Franco. 2005. Modifiable risk factors associated with clearance of type-specific cervical human papillomavirus infections in a cohort of university students. Cancer Epidemiol. Biomarkers Prev. 141149-1156. [PubMed]
24. Richardson, H., G. Kelsall, P. Tellier, H. Voyer, M. Abrahamowicz, A. Ferenczy, F. Coutlee, and E. L. Franco. 2003. The natural history of type-specific human papillomavirus infections in female university students. Cancer Epidemiol. Biomarkers Prev. 12485-490. [PubMed]
25. Safaeian, M., R. Herrero, A. Hildesheim, W. Quint, E. Freer, L. J. Van Doorn, C. Porras, S. Silva, P. Gonzalez, M. C. Bratti, A. C. Rodriguez, and P. Castle. 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. Clin. Microbiol. 451447-1454. [PMC free article] [PubMed]
26. Schiffman, M., R. Herrero, A. Hildesheim, M. E. Sherman, M. Bratti, S. Wacholder, M. Alfaro, M. Hutchinson, J. Morales, M. D. Greenberg, and A. T. Lorincz. 2000. HPV DNA testing in cervical cancer screening: results from women in a high-risk province of Costa Rica. JAMA 28387-93. [PubMed]
27. Schiffman, M., C. M. Wheeler, A. Dasgupta, D. Solomon, and P. E. Castle. 2005. A comparison of a prototype PCR assay and hybrid capture 2 for detection of carcinogenic human papillomavirus DNA in women with equivocal or mildly abnormal papanicolaou smears. Am. J. Clin. Pathol. 124722-732. [PubMed]
28. Sherman, M. E., S. S. Wang, R. Tarone, L. Rich, and M. Schiffman. 2003. Histopathologic extent of cervical intraepithelial neoplasia 3 lesions in the atypical squamous cells of undetermined significance low-grade squamous intraepithelial lesion triage study: implications for subject safety and lead-time bias. Cancer Epidemiol. Biomarkers Prev. 12372-379. [PubMed]
29. Smith, J. S., L. Lindsay, B. Hoots, J. Keys, S. Franceschi, R. Winer, and G. M. Clifford. 2007. Human papillomavirus type distribution in invasive cervical cancer and high-grade cervical lesions: a meta-analysis update. Int. J. Cancer 121621-632. [PubMed]
30. Stevens, M. P., S. M. Garland, and S. N. Tabrizi. 2008. Development and validation of a real-time PCR assay specifically detecting human papillomavirus 52 using the Roche LightCycler 480 system. J. Virol. Methods 147290-296. [PubMed]
31. Tabrizi, S. N., M. Stevens, S. Chen, E. Rudland, J. R. Kornegay, and S. M. Garland. 2005. Evaluation of a modified reverse line blot assay for detection and typing of human papillomavirus. Am. J. Clin. Pathol. 123896-899. [PubMed]
32. van Doorn, L. J., A. Molijn, B. Kleter, W. Quint, and B. Colau. 2006. Highly effective detection of human papillomavirus 16 and 18 DNA by a testing algorithm combining broad-spectrum and type-specific PCR. J. Clin. Microbiol. 443292-3298. [PMC free article] [PubMed]
33. van Hamont, D., M. A. van Ham, J. M. Bakkers, L. F. Massuger, and W. J. Melchers. 2006. Evaluation of the SPF10-INNO LiPA human papillomavirus (HPV) genotyping test and the Roche Linear Array HPV genotyping test. J. Clin. Microbiol. 443122-3129. [PMC free article] [PubMed]
34. Walboomers, J. M., M. V. Jacobs, M. M. Manos, F. X. Bosch, J. A. Kummer, K. V. Shah, P. J. Snijders, J. Peto, C. J. Meijer, and N. Muñoz. 1999. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J. Pathol. 18912-19. [PubMed]
35. Wheeler, C. M., W. C. Hunt, M. Schiffman, and P. E. Castle. 2006. Human papillomavirus genotypes and the cumulative 2-year risk of cervical precancer. J. Infect. Dis. 1941291-1299. [PubMed]
36. Wright, T. C., Jr., L. S. Massad, C. J. Dunton, M. Spitzer, E. J. Wilkinson, and D. Solomon. 2007. 2006 consensus guidelines for the management of women with abnormal cervical cancer screening tests. Am. J. Obstet. Gynecol. 197346-355. [PubMed]
37. Wright, T. C., Jr., and M. Schiffman. 2003. Adding a test for human papillomavirus DNA to cervical-cancer screening. N. Engl. J. Med. 348489-490. [PubMed]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...