Dempsey, Amanda F. MD, PhD, MPH*; Koutsky, Laura A. PhD†; Golden, Matthew MD, MPH‡
The burden of HPV-associated illnesses like genital warts and cervical neoplasia may be greatly reduced in the future by the use of effective HPV vaccines. Two candidate HPV vaccines have been developed, and one of these was recently licensed by the US Food and Drug Administration for use in adolescent and young women aged 9–26.1 Both candidate vaccines target the two HPV types most commonly associated with cervical cancer (HPV 16 and HPV 18), and both have been shown to be highly successful in preventing cervical intraepithelial neoplasia.2–4 One of the candidate vaccines also targets two additional HPV types, HPV 6 and HPV11, which cause >80% of cases of genital warts. Clinical trial data of this quadravalent HPV vaccine have shown high-level efficacy for preventing genital warts in young women,4 and trials of this vaccine in young men are currently underway.
Widespread HPV vaccination is likely to result in substantial health benefits to the individuals who are vaccinated. However, the potential impact of these vaccines on health care system utilization has not been extensively examined. Furthermore, having genital warts may prompt many patients to seek complete STD evaluations, and eliminating this disease through HPV vaccination could result in lost opportunities to test and treat individuals for bacterial STDs, herpes, HIV, or other infections that might have been previously unrecognized. In this study, we assess the proportion of STD clinic visits attributable to HPV-related complaints and describe the prevalence of other STD diagnoses among visits motivated by concerns related to HPV infection.
Materials and Methods
Study Population and Design
Since 1994, Public Health-Seattle & King County has maintained an electronic database of medical records from patients seen at a community hospital-based STD clinic. This database contains information collected by clinicians using a standardized medical record form and includes information about presenting symptoms and complaints; sexual, social, and reproductive history; physical examination findings; laboratory test results; and diagnoses and treatment. We performed a cross-sectional analysis of records from all patients seen in this clinic between January 1, 1994, and December 31, 2004.
The University of Washington institutional review board approved all study procedures.
All visit types were initially considered for evaluation. However, as our primary interest was in new STD diagnoses and few such diagnoses are made during follow-up visits for established problems, only visits categorized as “new problem” were considered in the analysis of STD prevalence (eliminated from this analysis were “test results” visits, “follow-up” visits, and visits prompted by referral from an outside provider). New problem visits were divided into those with and without HPV-related concerns as the reason for visit, according to the inclusion/exclusion criteria outlined below.
In our analysis, HPV-related visits (the “HPV group”) were defined as those visits with either: 1) an indication of “anorectal symptoms” and/or “genital lesions” as a reason for visit and a visit diagnosis of genital warts, or 2) “contact to HPV” as the reason for visit, regardless of whether a genital warts diagnosis was obtained. As a comparison, STD diagnosis prevalence was calculated among visits with presenting complaints unrelated to HPV. This group (the “No HPV group”) was composed of the remaining new-problem visits that were not included in the HPV group (Fig. 1). In all cases, the reason for visit was obtained and recorded by the providing clinician using an open-ended question and prespecified response categories. Clinicians could choose as many response categories as needed to reflect the patient's presenting complaints.
We also calculated STD diagnosis prevalence using two other sets of criteria to define the visit categories. In the first of these other analyses, we eliminated from the HPV group those visits where additional presenting concerns that appeared unrelated to HPV (e.g., vaginal discharge or contact to CT) were present. By using these more restrictive criteria to define the HPV group, we could focus on STD diagnosis prevalence among visits where HPV-related concerns appeared to be the sole motivating factor for seeking care, thus eliminating the potential confounding effect of other presenting concerns on STD diagnosis frequency. In the second of these analyses, we limited our calculation of diagnosis prevalence to only those visits where diagnostic testing for that particular STD was performed (e.g., calculating HIV diagnosis prevalence only among visits where HIV testing had been performed). This was done to assess the effect of potential differences in testing patterns between the two-visits groups (HPV and no HPV) on diagnosis frequencies for other STDs.
Diagnostic criteria for 12 different STDs were derived using a combination of lab and clinical data. Our interest was specifically in new STD diagnoses; thus we classified diseases as those that were chronic and therefore could be newly diagnosed only once in a person's lifetime (HIV, hepatitis B, and genital herpes) and those that could be repeatedly newly diagnosed after interval clearing of clinical infection (Neisseria gonorrhea (GC), Chlamydia trachomatis (CT), trichomonas, genital warts, nonspecific pelvic inflammatory disease (PID), nonspecific mucopurulent cervicitis (MPC), nonspecific urethritis (NGU), syphilis and pubic lice). The chronic STDs were counted only once per person when calculating the prevalence of different STD diagnoses.
Diagnostic criteria for genital warts, pubic lice, and trichomonas were based exclusively on the indication of a “presumptive diagnosis” for these diseases in the medical record since these illnesses are detectable by clinical examination findings and/or microscopic examination by providers of genital specimens obtained in the clinic at the time of the examination. Information on trichomonas infection was available only for females. Diagnostic criteria for MPC, PID, and NGU were specific for those cases unrelated to GC or CT and thus required a combination of a presumptive diagnosis of MPC, PID, or NGU in the medical record plus no laboratory evidence of CT or GC infection (described below). Diagnostic criteria for the remaining STDs were derived from a combination of lab and historical data as follows: GC and CT—one or more of the following: 1) positive urethral/cervical swab Gram stain (GC only) and/or, 2) positive culture (from urethra, cervix, throat, or rectum) and/or, 3) positive urine ligase chain reaction (Abbott Labs) or Aptima Combo 2; syphilis—no history of disease plus at least one of the following: 1) newly positive RPR plus a positive treponemal antibody test, 2) newly positive darkfield examination, 3) newly positive VDRL plus a positive treponemal antibody test; HIV—no history of disease and a positive rapid HIV test and/or positive HIV serology; genital herpes (HSV2)—no history of disease plus a positive HSV2 ELISA assay and/or positive Western blot for HSV2; hepatitis B—positive HBsAg.
Because only new problem visits were being considered, and because some STDs could be newly diagnosed more than once in the same individual, clinic visits (rather than unique individuals) were the unit of analysis. In some instances, diagnostic criteria for one of the chronic STDs (HIV, HSV2, or hepatitis B) were present in multiple visits for the same individual. In this case the diagnosis was counted only once when determining diagnosis prevalence.
Demographic and clinical characteristics were determined for the various groups using descriptive characteristics. Prevalence of the different STDs was calculated by dividing the number of visits fulfilling the diagnostic criteria described above by the total number of visits in that group. Differences in disease prevalence among the groups were evaluated using Pearson's χ2 or Fisher exact test under the assumption that each visit was an independent observation. Because the HPV vaccine is currently licensed only for females, and because asymptomatic STD infections have greater clinical impact on women than men, all analyses were stratified by gender. Data were analyzed using Stata8.2 statistical software (StataCorp, LP, College Station, TX).
Follow-Up Visit Analysis
A separate analysis of visit types designated as follow-up was performed to more fully describe the potential impact of widespread HPV vaccination. The follow-up visit data set did not include information about the “reason for visit” and also could not be linked back to specific preceding new problem visits. Because of this, we were unable to directly determine the proportion of follow-up visits attributable to HPV-related concerns. Instead, we calculated the proportion of follow-up visits that resulted in a diagnosis of genital warts. This was done by dividing the number of follow-up visits fulfilling the diagnostic criteria for this disease (described above) by the total number of follow-up visits. In addition, new STD diagnoses among follow-up visits were evaluated using the methods described above. However, analysis of other STD diagnoses among follow-up visits was limited to diseases that did not require information on history as part of the diagnostic criteria (GC, CT, trichomonas, PID, MPC, NGU, hepatitis B, genital warts, and pubic lice), since history data were not available for these visits.
During the study period, 46,963 (71%) of visits were classified as new problem visits, and 22% were for follow-up (Fig. 1), with male visits predominating for both visit types. Among males, 7.9% of 32,263 new problem visits were for HPV-related presenting concerns. In contrast, the proportion of new problem visits attributable to HPV-related concerns was lower among women, comprising only 3.6% of the 14,700 female new problem visits. To more fully describe the potential impact of widespread HPV vaccination on STD clinic resource utilization, we also determined the percentage of follow-up visits attributable to HPV. During the study period, there were 10,289 follow-up visits among men, of which 2595 (25.2%) were for genital warts. Among women, 744 of 4285 follow-up visits (17.3%) were for genital warts. Overall, 10.3% of clinic visits during the study period were for HPV-related concerns, after combining genders and visit types in the analysis.
Compared with the group of new problem visits unrelated to HPV (the No HPV group), “symptoms” and contact to HPV were more common presenting complaints among the HPV-related new problem visits (Table 1). As expected, the distribution of presenting symptom types varied in accordance with differences in the criteria used to define the groups (Table 1). There were no substantial differences in the average age at visit, insurance status, or median number of visits per person between the No HPV group and the HPV-related group of visits for either gender (Table 2). However, compared with visits in the No HPV group, new problem visits for HPV-related concerns were less often by members of racial minority groups, less often by men who have sex with men, and with the exception of genital warts, less frequently associated with a history of prior STDs.
For both genders, more than 75% of the HPV-related visits resulted in a diagnosis of genital warts (Table 3). A few genital warts diagnoses were also found in the No HPV group (2.9% and 1.1% for male and female visits, respectively; Table 3), likely representing patients with small lesions of which they had been previously unaware, or patients with a known prior history of genital warts who came to the clinic for other concerns. Though the percentage of visits resulting in a genital warts diagnosis within the No HPV group was low, this represented nearly 27% of the HPV diagnoses made in new problem visits overall during the study period.
New, non-HPV, STD diagnoses among the HPV-related group of visits were uncommon, with only 242 additional diagnoses found among 2550 male new problem visits, and only 39 additional new diagnoses found among 535 female new problem visits (Table 3). For males, a large proportion of these new diagnoses (42%) were for nonspecific (i.e., nongonococcal, nonchlamydial) NGU. When nonspecific NGU, MPC, and PID were eliminated from the analysis (because these diagnoses may not always be sexually transmitted), we found only 133 new, non-HPV diagnoses among male visits and only 24 new diagnoses among female visits, with GC, CT, and HSV being the most prevalent additional diagnoses for both genders. The prevalence of all non-HPV STD diagnoses was higher in the No HPV group than in the HPV-related group.
When we reanalyzed the STD diagnosis prevalence using only those visits where diagnostic tests for a given STD had been performed, we saw a similar pattern of statistical significance between the HPV and No HPV groups, implying that the differences in STD diagnosis prevalence between these groups were not the result of differential screening practices. We also analyzed the prevalence of additional STD diagnoses among a more restricted subset of HPV-related visits where HPV-related complaints appeared to be the singular reason for seeking care (i.e., eliminating HPV visits where additional, potentially confounding, presenting complaints seemingly unrelated to HPV were present). In this analysis, we found an even lower prevalence of additional, non-HPV, STD diagnoses in both genders with only 112 additional STD diagnoses among 2013 male new problem visits and only 20 additional STD diagnoses among 377 female visits.
Widespread vaccination against HPV is likely to reduce the incidence of cervical neoplasia and to decrease the burden of other HPV-associated illnesses such as genital warts. Such reduction in HPV-associated morbidity has the potential to decrease the clinical burden on public STD clinics; however, this could result in unanticipated negative effects, including missed opportunities to diagnose other STDs in persons seeking care for genital warts or other HPV-related complaints. We determined the proportion of clinic visits attributable to HPV-related complaints and evaluated the prevalence of unrecognized STDs other than HPV in patients presenting to an STD clinic for HPV-related concerns.
We found few additional STD diagnoses among patients with HPV-related concerns. Though risk factors associated with HPV infection have been identified,5–11 the finding that HPV is nearly ubiquitous among sexually active individuals12 supports the notion that infection often occurs outside the setting of “risky sexual behavior.” Because of this, it is not surprising that we found a paucity of other STD diagnoses among the HPV-related visits since many of the individuals with HPV-related concerns may not engage in behaviors placing them at increased risk for other sexually transmitted infections.
Our study suggests that widespread utilization of HPV vaccines that include protection against HPV 6 and HPV 11 could ultimately decrease the demand for STD clinic services by ∼10% without a substantial impact on the diagnosis and treatment of other, non-HPV STDs. This reduction in demand for services could provide STD clinics with the opportunity to shift personnel, financial assets, and other resources to other areas of need in the future as these vaccines become more widely utilized. However, the true population-level benefits of HPV vaccines remain to be determined and will likely be influenced by many factors. These include the expected future availability of two different HPV vaccine products with different spectrums of clinical protection (one with the ability to prevent genital warts, the other without), anticipated changes in age- and gender-related licensure restrictions for HPV vaccines, and federal, state, and private insurance coverage for HPV vaccination, among others. Furthermore, some have expressed concern that behavioral disinhibition resulting from widespread HPV vaccine availability, and a “false sense of protection” against other STDs, could actually lead to higher rates of non-HPV STD acquisition. Many of these issues could be formally explored in future studies once HPV vaccines become more widely utilized.
As with any retrospective study, our analysis was limited by the type of data available. The accuracy of visit type specification and of the characteristics used to define STD diagnoses and the reason(s) for clinic visit could not be confirmed. To circumvent this, we used both less- and more-restrictive criteria to define HPV-related visits but found similar results in terms of other STD diagnosis frequency. An additional limitation of our analysis is that we studied a single STD clinic population, and our findings may not be generalizable to other, non-STD clinic populations, or to STD clinic populations in other geographic locations. Furthermore, the STD clinic evaluated in this study did not provide ongoing management of Pap smear abnormalities, and we were therefore not able to capture this subgroup of patients with “HPV-related concerns.” It will be important to expand future studies on the prevalence of additional STDs in people with HPV-related concerns to the other populations and to include women seeking care specifically for follow-up of abnormal Pap smears.
In summary, we found that while a substantial number of STD clinic patients seek evaluation or follow-up care for complaints related to genital warts, non-HPV STD diagnoses were uncommon among these patients. This suggests that if genital warts were eliminated by widespread use of HPV vaccines, public STD clinics would likely realize substantial gains in availability of clinic resources while sacrificing relatively little in terms of missed opportunities for diagnosing other, unrecognized STDs in this population.
1. U.S. Food and Drug Administration. Product Approval Information-Licensing Action. Gardasil. 2006.
2. Harper DM, Franco EL, Wheeler C, et al. Efficacy of a bivalent L1 virus-like particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: A randomised controlled trial. Lancet 2004; 364:1757–1765.
3. Koutsky LA, Ault KA, Wheeler CM, et al. A controlled trial of human papillomavirus type 16 vaccine. NEJM 2002; 347:1645–1651.
4. Villa LL, Costa RL, Petta CA, et al. Prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in young women: A randomised double-blind placebo-controlled multicentre phase II efficacy trial. Lancet Oncol 2005; 6:271–278.
5. Bauer HM, Hildesheim A, Schiffman MH, et al. Determinants of genital human papillomavirus infection in low-risk women in Portland, Oregon. Sex Transm Dis 1993; 20:274–278.
6. Wheeler CM, Parmenter CA, Hunt WC, et al. Determinants of genital human papillomavirus infection among cytologically normal women attending the University of New Mexico student health center. Sex Transm Dis 1993; 20:286–289.
7. Hildesheim A, Gravitt P, Schiffman MH, et al. Determinants of genital human papillomavirus infection in low-income women in Washington, D.C. Sex Transm Dis 1993; 20:279–285.
8. Winer RL, Lee SK, Hughes JP, et al. Genital human papillomavirus infection: Incidence and risk factors in a cohort of female university students. Am J Epidemiol 2003; 157:218–226.
9. Burk RD, Ho GY, Beardsley L, et al. Sexual behavior and partner characteristics are the predominant risk factors for genital human papillomavirus infection in young women. J Infect Dis 1996; 174:679–689.
10. Peyton CL, Gravitt PE, Hunt WC, et al. Determinants of genital human papillomavirus detection in a US population. J Infect Dis 2001; 183:1554–1564.
11. Ho GY, Bierman R, Beardsley L, et al. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998; 338:423–428.