Effect of HPV Assay Choice on Perceived Prevalence in a Population-based Sample : Diagnostic Molecular Pathology

Journal Logo

00019606-201306000-00004ArticleDiagnostic Molecular PathologyDiagnostic Molecular Pathology© 2013 by Lippincott Williams & Wilkins.22June 2013 p 85–90Effect of HPV Assay Choice on Perceived Prevalence in a Population-based SampleOriginal ArticlesCuschieri, Kate PhD*; Kavanagh, Kim PhD†; Sinka, Katy PhD‡; Robertson, Chris PhD†,§; Cubie, Heather PhD*; Moore, Catherine MSc*; Donaghy, Martin PhD‡*Scottish HPV Reference Laboratory, Royal Infirmary of Edinburgh, UK†Department of Mathematics and Statistics, University of Strathclyde‡Health Protection Scotland, Health Protection, Glasgow, Scotland§International Prevention Research Institute (iPRI), Lyon FranceFunded by Health Protection Scotland, National Services Division.K.C. and H.C. have received project funding from Qiagen. The other authors declare no conflict of interest.Reprints: Kate Cuschieri, PhD, Scottish HPV Reference Laboratory, Royal Infirmary of Edinburgh, Edinburgh, UK (e-mail: [email protected]).AbstractHuman papillomavirus (HPV) immunization programs clearly have considerable potential to reduce HPV-associated disease; they are also resource-intense; so, it is essential that their effectiveness is determined accurately and in a timely way. Measuring circulating HPV types in a population can provide an early measure of vaccine impact. We assessed the impact of HPV assay on the observed population prevalence of HPV in women who provided samples as part of a National HPV Immunisation Surveillance Exercise. A total of 1145 liquid-based cytology samples, 326 self-taken swabs, and 371 urine samples were tested with a line-blot assay (the Digene reverse hybridization HPV genotyping assay) and a luminex-based assay (the Mulitmetrix HPV genotyping assay). Assay agreement was determined for the different sample types. Positivity (according to assay) was compared at different levels ranging from positive for HPV 16 and/or 18 to positive for any one of the 18 HPV types common to both assays. The luminex assay consistently detected a higher prevalence of HPV—up to 10% for HPV types common to both assays. In addition, disagreement for HPV 16 and/or 18 was observed in around 9% of the overall sample, with an associated κ score of 0.74. These data indicate that assay choice has a significant impact on observed prevalence of HPV, including vaccine types. The impact of any change of assay during longitudinal surveillance programs should thus be taken into account to avoid confounding the assessment of any vaccine-induced changes.Human papillomavirus (HPV) immunization programs have commenced across the world and it is clearly essential that we monitor their effectiveness. One measure of success that can provide answers in advance of an observed reduction in HPV-associated cancers is determination of HPV type-specific prevalence in the population over time, after characterizing a pre-vaccine baseline.Obtaining representative samples to give a population-based measure can be challenging, particularly for HPV where suitable biospecimens are not routinely available for the whole population. Cervical cytology specimens from screening programs provide good coverage but miss those who do not attend the program and cannot provide any information about males, necessitating alternative collection strategies and biospecimens for investigation of type-specific prevalence in these groups. Stored samples can be a useful source of information, although suboptimal storage conditions can lead to biospecimen deterioration that could affect observed prevalence.1 Biospecimen type can also affect associated HPV detection, as we observed in a previous epidemiological study of Scottish women (who provided either a urine or self-taken swab) and in a technical study where the influence of urine versus cervical liquid-based cytology (LBC) sample was assessed.2,3Furthermore, over the last decade HPV assay development has become a highly dynamic field, with tests becoming more automated, sensitive, and cheap in the face of increased demand and market competition.4 For national epidemiological studies that necessarily take a view of prevalence over several years—and for intercountry comparison—it is important that the impact of HPV detection assay on observed HPV prevalence is quantified in addition to the impact of sample type, associated processing, and storage time.In Scotland, HPV immunization with the bivalent vaccine was introduced in 2008 for 12- to 13-year-old girls, with an initial 3-year catch up for girls up to age 18. In concert with this, a National HPV Epidemiology and Surveillance strategy was developed that included a plan to measure HPV prevalence in successive birth cohorts of women attending for their first smear (aged around 20) over ∼12 years and initiating in 2009, to measure the baseline prevalence in an unvaccinated cohort of women. The test used initially in 2009 was a reverse hybridization (RH) line-blot assay that detects 18 high-risk or putatively high-risk types. Consequently, a higher throughput assay was developed for surveillance that incorporated more HPV types (including low-risk types). This article presents a comparison of the 2 assays used to detect HPV in samples taken from a general population sample of young women, and discusses consequent implications for longitudinal surveillance.MATERIALS AND METHODSSpecimen Collection and Initial Processing/ExtractionIn 2009, just over 2000 residual cervical LBCs from women attending for their first smear and 725 self-taken samples (either urine or self-taken swabs) from nonattendees of a similar age were collected from across Scotland for National HPV surveillance. No preprocessing of LBC samples before extraction was required, and the protocol for preprocessing of urine and self-taken swabs has been described previously.2 Briefly, 1 mL of sample was subjected to automated Nucleic acid extraction using the Qiagen MDX platform and the QIAamp media kit that incorporates proteinase K digestion and silica-based purification (Qiagen Ltd, Crawley, UK). As this automated platform incorporates batch extraction of samples within a 12×8 grid, 10% of all samples constituted HPV-negative internal quality control material, the grid reference of which varied across batches—as a contamination check. In addition, previously tested positive-control material was also included within each batch. Nucleic acid was eluted in 100 μL of AE buffer (Qiagen Ltd.) and 10 μL of extract was used in subsequent polymerase chain reactions.HPV DetectionNucleic acid extracts derived from the 2009 collection were tested by the research use only (RUO) Digene RH HPV genotyping assay (Qiagen Ltd), according to manufacturers’ instructions. Briefly, this assay incorporates a polymerase chain reaction stage using GP5+6+ primers. Amplicons are then hybridized to strips with 18 immobilized probes corresponding to HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68, 26, 53, 66, 73, and 82. A minimum of 1 positive and 1 negative (from extraction) control was included in each assay run. This version of the assay did not contain an endogenous cellular control.Subset for ComparisonNucleic acid from a subset of 1445 LBC samples, 371 urine samples, and 326 self-taken swabs (which had been stored at −80 for up to 2 y) were retested by the multimetrix HPV genotyping assay (Progen Biotechnik GmbH, Heidelberg, Germany). The multimetrix assay incorporates modified GP5+6+ primers (GPS) and utilizes a luminex system to detect 24 HPV types including 6 low-risk types and incorporates a hybridization control to adjudicate successful conjugation of target to reporter and also a β-globin control to assess sample cellularity/amplification eligibility.5 Only valid samples (as indicated through the presence/absence of the β-globin control) were included in the subset comparison. Beyond validity, the subsets allocated for analysis were chosen to achieve power; no other exclusions were imposed which makes the subset representative of the larger population.Statistical Analysis (Proportion Positive and Agreement)HPV positivity according to both assays was determined at 5 levels:HPV positive or negative according to assayHPV positive or negative (for 1 or more of the 18 types common to both assays)HPV positive or negative for HPV 16 and/or 18HPV positive or negative for HPV 16, 18, 31, and 45HPV positive or negative for 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68 [ie, the 12 HPV types considered group 1 carcinogens and the 1 type considered a group 2A carcinogen (HPV 68)—according to IARC]In addition, raw agreement of the assays (RH vs. multimetrix), split by positive, negative, and overall agreement, was performed for levels 2 to 5. Agreement was assessed separately for each of the 3 samples types (LBC, urine samples, and self-taken swabs).As raw agreement does not take into account results agreeing by chance. Cohen κ coefficient was calculated along with the corresponding 95% confidence interval.6 Positivity and agreement were assessed for each sample type and κ values determined for each of the 18 types common to both assays. Differences in HPV positivity between the 2 assays, stratified by sample type, was examined for levels 2 to 5 using interaction plots that allow for comparison of the slope of the difference between sample types. Statistical significance of these interactions was evaluated using a mixed-effects model, with study ID as a random effect to account for the paired nature of the data, and the interaction term between assay type and specimen type as a fixed effect. All statistical analysis was conducted in R version 2.14.1.In samples where types were detected that were common to both assays, we also compared the number of samples that showed multiple infection (according to each assay and for each sample type).RESULTSOverall Prevalence Measured by Both AssaysA total of 2142 samples, tested with both assays, were included in the overall analysis (1445 LBC samples, 371 urine samples, and 326 self-taken swabs). Table 1 depicts prevalence according to assay and specimen type. As expected, prevalence of HPV in this population of young unvaccinated women was high. The multimetrix assay detected a higher overall prevalence of HPV compared with the RH (51% vs. 43.8% for all 2142 samples; 56% vs. 47.6% for LBC samples; 35.3% vs. 32.3% for the urine samples; and 46.9% vs. 39.9% for the self-taken swabs. As the multimetrix assay detects more types than RH, this was expected. However, higher prevalence associated with the multimetrix test was consistent when analysis was confined to the 18 types in common to both assays: 51.2% vs. 43.7% for LBC samples, 31.0% vs. 20.2% for urine samples, and 41.1% vs. 31.0% for the self-taken swabs.JOURNAL/dimp/04.03/00019606-201306000-00004/table1-4/v/2021-02-17T200059Z/r/image-tiffProportion HPV Positive According to Sample Type and AssayAssay AgreementTables 2–4 present raw agreement and κ scores associated with the assays at the different levels of comparison and according to specimen type. In relation to the LBC samples, overall agreement between the RH and multimetrix assay was around 90% (range, 87.6 to 91.2), with the lowest and highest level of agreement associated with detection of the 18 types common to both assays and detection of HPV 16/18, respectively. The range of agreement was 84.4 to 92.3 for urine samples and 86.8 to 89.7 for self-taken swabs, and as observed with the LBC samples, the highest agreement was associated with the narrower range of types (16/18 or 16/18/31 and 45). However, it should be noted that in all sample types negative agreement was consistently higher than positive agreement, with the lowest positive agreement evident in urine samples, for all levels of analysis (ie, from the 18 types in common to both assays to agreement for HPV 16/18). Positive agreement for the 18 types in common; 13 high-risk types; 16/18; and 16, 18, 31, and 45 was 80.7, 82, 74.5, and 80.9 for LBCs; 71.6, 71.8, 59.0, and 67.4 for the self-taken swabs; and 57.4, 62.9, 57.8, and 61.8 for urine samples, respectively.JOURNAL/dimp/04.03/00019606-201306000-00004/table2-4/v/2021-02-17T200059Z/r/image-tiffProportion of Positive, Negative, and Overall Agreement (RH Assay vs. Multimetrix Assay) According to Sample TypeAssessment of κ (Tables 3, 4), which takes into account that agreement may occur due to chance, showed higher values (indicating greater assay agreement) for the LBC samples specifically: 0.75 (18 types common to both assays) 0.77 (16/18), and a value of 0.78 for the 13 high-risk types and 16, 18, 31, and 45 compared with self-taken swabs: (respective values of) 0.72. 0.74, 0.67, and 0.72 and urine samples: 0.60, 0.66, 0.62, 0.61.JOURNAL/dimp/04.03/00019606-201306000-00004/table3-4/v/2021-02-17T200059Z/r/image-tiffκ Agreement (RH Assay vs. Multimetrix Assay) According to Sample TypeTable 4 depicts κ’s for the 18 HPV types common to both assays in all samples; this analysis was not disaggregated into separate sample types given associated power issues. Types associated with κ values <0.5 were HPV 35, 53, 68, and 82. Although interesting, overextrapolation of these findings should be avoided in view of the low overall positivity of these types. For example, the confidence interval for HPV 68 contains negative κ scores implying more disagreement than expected due to chance. However, the width of this interval is driven by the low positivity observed [0.61% positive in the new test (13 samples) and 0.14% positive in the old test (3 samples)].JOURNAL/dimp/04.03/00019606-201306000-00004/table4-4/v/2021-02-17T200059Z/r/image-tiffκ Values for all 18 HPV Types Common to Both AssaysImpact of Sample Type on Assay PositivityFigure 1 shows the interaction between sample type and assay for 4 different “positivity” categories (any comparable HPV type, HPV 16/18, HPV 16/18/31/45, and any high-risk type). There was some evidence of a greater assay discrepancy (in positivity) for the self-taken swabs compared with urine and LBC samples. This was most evident when analysis was restricted to types 16/18 (P=0.05) and types 16/18/31/45 (P=0.126).JOURNAL/dimp/04.03/00019606-201306000-00004/figure1-4/v/2021-02-17T200059Z/r/image-jpegInteraction plots examining differences in positivity between the 2 assays and LBC versus self-taken samples (self-taken swabs or urine samples). Results are shown for 4 categories of positivity—any comparable HPV type—HPV types 16 or 18, HPV 16 or 18 or 31 or 45, and any high-risk HPV type. If the differences in percentage positive by test are same for different specimen types, then the lines would be expected to be parallel. HPV indicates human papillomavirus; LBC, liquid-based cytology.Detection of Multiple InfectionsThe proportion of samples with multiple infection is presented in Figure 2. Analysis was confined to samples that tested positive for one of more of the 18 types common to both assays and was performed for each of the different specimen types. The multimetrix test detected more multiple infection (ie, 2 or more types) in all sample types compared with the RH: 28% versus 24% in the LBC samples, 25% versus 13% in the self-taken swabs, and 14% versus 9% in the urine samples.JOURNAL/dimp/04.03/00019606-201306000-00004/figure2-4/v/2021-02-17T200059Z/r/image-jpegExtent of multiple infection as measured by the RH and multimetix assays. LBC indicates liquid-based cytology; RH, reverse hybridization.DISCUSSIONThis work emphasizes how the measurable impact of the vaccine on HPV prevalence could be affected by assay choice. There are natural caveats to our study and analysis; type-specific reproducibility of the assays was not assessed (samples were tested only once) and our analysis also did not take into consideration the potential for “false” positives by either assay. This said, it is unlikely that these technical considerations would account for the extent of the differences we observed. One of the outcomes of immunization effectiveness will be the reduction of vaccine specific types and HPV types associated with cross-protection (most demonstrably HPV 31 and 45). Considering the entire sample, that is, LBCs, self-taken swabs, and urine combined, overall disagreement for HPV 16/18 was observed in around 9% of samples, and when confining this analysis to the (Mutlimetrix) positive samples, 30% of samples showed disagreement for HPV 16/18. With respect to samples positive for 16/18/31/45 by the Multimetrix assay, agreement was observed in 77.1% of samples (∼23% disagreement). In addition, the multimetrix assay consistently detected a higher prevalence of HPV (up to 10% for types common to both assays). Our data also indicate that assay-driven differences may be affected/exacerbated by specimen type as the urine and self-taken swabs showed a greater disparity in assay agreement compared with LBC. Furthermore, it should be noted that our comparison was performed using samples that had been stored for up to 2 years and subjected to 2 freeze-thaw cycles. If tests had been performed contemporaneously, it is likely that the difference would have been larger. Scotland has a 12-year program of HPV surveillance that includes yearly sampling of the screened population. Consequently, if the RH assay had been simply replaced by the multimetrix assay (without evaluation/comparisons), we would have observed increases in overall and vaccine-type HPV prevalence that could have confounded assessment of real changes.In a recent US study, Unger et al7 performed a comparison of the prototype and (subsequent) RUO version of the linear array HPV genotyping kit (Roche Molecular Systems, Pleasanton, CA) on 3001 self-collected, population-based vaginal swabs. The authors showed that the newer assay detected HPV in approximately double the number of samples and concluded that assay differences could “complicate the interpretation of HPV prevalence in studies when the HPV testing platform changes.” Although our study did not show as profound a difference as the US study, the general observations and implications are consistent: for any country involved in longitudinal HPV surveillance, it will be essential to quantify assay-driven influences. Another related issue is that as HPV vaccine types (and feasibly related types) reduce in immunized populations, other types that were present below the level of detection of the surveillance assay may become unmasked in the face of less competition for assay resources. It is well established that low-level targets within mixed infections are underrepresented in genotyping assays that use an initial consensus approach (such as those described in this article). This unmasking could lead to a perceived “increase” of certain types, and caution should be exercised to avoid false claims of type replacement.To conclude, it is a unique and exciting time in the development of HPV assays and HPV immunization—the challenge will be how to best apply the former to accurately monitor the latter given the changing pattern of infection and associated disease over time. Furthermore, we agree strongly with Unger and colleagues in that the type of studies required to address this will only be possible through the ongoing collation and appropriate storage of well-annotated samples to form population-based archives. To this end, Scotland is committed to a longitudinal program of HPV surveillance where knowledge of vaccine status is interpreted alongside HPV type-specific prevalence.REFERENCES1. Castle PE, Solomon D, Hildesheim A, et al. Stability of archived liquid-based cervical cytologic specimens. Cancer. 2003;99:89–96[Context Link][CrossRef][Medline Link]2. Sinka K, Lacey M, Robertson C, et al. Acceptability and response to a postal survey using self-taken samples for HPV vaccine impact monitoring. Sex Transm Infect. 2011;87:548–552[Context Link][Full Text][CrossRef][Medline Link]3. Cuschieri K, Nandwani R, McGough P, et al. Urine testing as a surveillance tool to monitor the impact of HPV immunization programs. J Med Virol. 2011;83:1983–1987[Context Link][Full Text][CrossRef][Medline Link]4. Poljak M, Kocjan BJ. Commercially available assays for multiplex detection of alpha human papillomaviruses. Expert Rev Anti Infect Ther. 2010;10:1139–1162[Context Link][Full Text][Medline Link]5. Schmitt M, Bravo IG, Snijders PJ, et al. Bead-based multiplex genotyping of human papillomaviruses. J Clin Microbiol. 2006;44:504–512[Context Link][CrossRef][Medline Link]6. Cohen JA. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37–46[Context Link][CrossRef][Medline Link]7. Unger ER, Steinau M, Lin JM, et al. Impact of HPV assay on observed population prevalence. Diagn Mol Pathol. 2011;20:101–104[Context Link][Full Text][CrossRef][Medline Link]HPV; immunization; surveillance00019606-201306000-0000400002808_2003_99_89_castle_stability_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e468_citationRF_FLOATING))|11065213||ovftdb|SL000028082003998911065213citation_FROM_JRF_ID_d1815e468_citationRF_FLOATING[CrossRef]10.1002%2Fcncr.1105800019606-201306000-0000400002808_2003_99_89_castle_stability_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e468_citationRF_FLOATING))|11065405||ovftdb|SL000028082003998911065405citation_FROM_JRF_ID_d1815e468_citationRF_FLOATING[Medline Link]1270468800019606-201306000-0000400115314_2011_87_548_sinka_acceptability_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING))|11065404||ovftdb|SL0011531420118754811065404citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING[Full Text]00115314-201112000-0000500019606-201306000-0000400115314_2011_87_548_sinka_acceptability_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING))|11065213||ovftdb|SL0011531420118754811065213citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING[CrossRef]10.1136%2Fsextrans-2011-05021100019606-201306000-0000400115314_2011_87_548_sinka_acceptability_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING))|11065405||ovftdb|SL0011531420118754811065405citation_FROM_JRF_ID_d1815e503_citationRF_FLOATING[Medline Link]00019606-201306000-0000400004807_2011_83_1983_cuschieri_surveillance_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING))|11065404||ovftdb|SL00004807201183198311065404citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING[Full Text]01445456-201111000-0001600019606-201306000-0000400004807_2011_83_1983_cuschieri_surveillance_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING))|11065213||ovftdb|SL00004807201183198311065213citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING[CrossRef]10.1002%2Fjmv.2218300019606-201306000-0000400004807_2011_83_1983_cuschieri_surveillance_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING))|11065405||ovftdb|SL00004807201183198311065405citation_FROM_JRF_ID_d1815e538_citationRF_FLOATING[Medline Link]2191587400019606-201306000-0000401190727_2010_8_1139_poljak_papillomaviruses_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e573_citationRF_FLOATING))|11065404||ovftdb|SL0119072720108113911065404citation_FROM_JRF_ID_d1815e573_citationRF_FLOATING[Full Text]01436931-201010000-0001700019606-201306000-0000401190727_2010_8_1139_poljak_papillomaviruses_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e573_citationRF_FLOATING))|11065405||ovftdb|SL0119072720108113911065405citation_FROM_JRF_ID_d1815e573_citationRF_FLOATING[Medline Link]2095488000019606-201306000-0000400004688_2006_44_504_schmitt_papillomaviruses_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e601_citationRF_FLOATING))|11065213||ovftdb|SL0000468820064450411065213citation_FROM_JRF_ID_d1815e601_citationRF_FLOATING[CrossRef]10.1128%2FJCM.44.2.504-512.200600019606-201306000-0000400004688_2006_44_504_schmitt_papillomaviruses_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e601_citationRF_FLOATING))|11065405||ovftdb|SL0000468820064450411065405citation_FROM_JRF_ID_d1815e601_citationRF_FLOATING[Medline Link]1645590500019606-201306000-0000400001111_1960_20_37_cohen_coefficient_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e636_citationRF_FLOATING))|11065213||ovftdb|SL000011111960203711065213citation_FROM_JRF_ID_d1815e636_citationRF_FLOATING[CrossRef]10.1177%2F00131644600200010400019606-201306000-0000400001111_1960_20_37_cohen_coefficient_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e636_citationRF_FLOATING))|11065405||ovftdb|SL000011111960203711065405citation_FROM_JRF_ID_d1815e636_citationRF_FLOATING[Medline Link]00019606-201306000-0000400019606_2011_20_101_unger_population_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING))|11065404||ovftdb|SL0001960620112010111065404citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING[Full Text]00019606-201106000-0000600019606-201306000-0000400019606_2011_20_101_unger_population_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING))|11065213||ovftdb|SL0001960620112010111065213citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING[CrossRef]10.1097%2FPDM.0b013e3181f56fa500019606-201306000-0000400019606_2011_20_101_unger_population_|00019606-201306000-00004#xpointer(id(citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING))|11065405||ovftdb|SL0001960620112010111065405citation_FROM_JRF_ID_d1815e658_citationRF_FLOATING[Medline Link]21532491Proportion HPV Positive According to Sample Type and AssayProportion of Positive, Negative, and Overall Agreement (RH Assay vs. Multimetrix Assay) According to Sample Typeκ Agreement (RH Assay vs. Multimetrix Assay) According to Sample Typeκ Values for all 18 HPV Types Common to Both AssaysInteraction plots examining differences in positivity between the 2 assays and LBC versus self-taken samples (self-taken swabs or urine samples). Results are shown for 4 categories of positivity—any comparable HPV type—HPV types 16 or 18, HPV 16 or 18 or 31 or 45, and any high-risk HPV type. If the differences in percentage positive by test are same for different specimen types, then the lines would be expected to be parallel. HPV indicates human papillomavirus; LBC, liquid-based cytology.Extent of multiple infection as measured by the RH and multimetix assays. LBC indicates liquid-based cytology; RH, reverse hybridization.Effect of HPV Assay Choice on Perceived Prevalence in a Population-based SampleCuschieri Kate PhD; Kavanagh, Kim PhD; Sinka, Katy PhD; Robertson, Chris PhD; Cubie, Heather PhD; Moore, Catherine MSc; Donaghy, Martin PhDOriginal ArticlesOriginal Articles222p 85-90

You can read the full text of this article if you:

Access through Ovid