- Email: [email protected]
18–Associated Cervical Lesions: Differences by Area-Based Measures of Race and Poverty
ARTICLE IN PRESS
RESEARCH ARTICLE
Human Papillomavirus 16/18−Associated Cervical Lesions: Differences by Area-Based Measures of Race and Poverty Monica M. Brackney, MS,1 Julia W. Gargano, PhD,2 Susan E. Hannagan, MPH,1 James Meek, MPH,1 Troy D. Querec, PhD,3 Linda M. Niccolai, PhD1
Introduction: This analysis evaluates trends in cervical lesions with human papillomavirus 16/18 detected by area-based measures of race, ethnicity, and poverty during 2008−2015.
Methods: Trends in the proportion of lesions with human papillomavirus 16/18 detected among residents of New Haven County, Connecticut were examined by area-based measures of race, ethnicity, and poverty. Area-based measures are aggregate descriptors of census tract characteristics useful for measuring differences in health outcomes in the context of where people live. Multivariable logistic regression modeling was conducted, adjusted for individual-level race, ethnicity, and insurance status to assess the independent effects of area-based measures. Data were analyzed in 2018−2019. Results: Among women aged 21−24 years and 25−29 years, significant declines in the proportion of lesions with human papillomavirus 16/18 were observed. Among women aged 21−24 years, declines began earlier and were greater in magnitude in areas of lower poverty (OR=0.55, 95% CI=0.36, 0.85 for 2010−2012 vs 2008−2009 and OR=0.30, 95% CI=0.18, 0.51 for 2013−2015 vs 2008−2009) compared with higher poverty (OR=1.66, 95% CI=0.86, 3.21 and OR=0.48, 95% CI=0.19, 1.20). Similar patterns were observed for women aged 25−29 years, and for area-based measures of race and ethnicity. Conclusions: Differences were observed in declines in the proportion of human papillomavirus 16/18 lesions by area-based measures since the introduction of human papillomavirus vaccines, with greater and earlier declines in areas with fewer residents living in poverty and racial minorities. Ongoing human papillomavirus vaccine impact monitoring is necessary to track differences by sociodemographic characteristics. Am J Prev Med 2020;000(000):1−9. © 2020 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
H
uman papillomavirus (HPV) is the primary cause of cervical and other anogenital cancers, oropharyngeal cancer, and genital warts.1 In 2006, a safe and efficacious quadrivalent HPV vaccine was licensed for use in the U.S. by the Food and Drug Administration.2 In 2007, the Advisory Committee on Immunization Practices recommended routine vaccination for female youth aged 11−12 years with catch-up vaccination through age 26 years.3 In 2011, the Advisory Committee on Immunization Practices extended the routine recommendation to male youth aged 11−12 years with catch-up vaccination up to age
21 years.3 All 3 HPV vaccines that are approved in the U.S. target HPV 16 and 18, which are responsible for 70% of cervical cancer cases.4 From the 1Connecticut Emerging Infections Program, Yale School of Public Health, New Haven, Connecticut; 2Division of Viral Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia; and 3Division of High Consequence Pathogens and Pathology, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia Address correspondence to: Monica M. Brackney, MS, Connecticut Emerging Infections Program, Yale School of Public Health, 1 Church Street, 7th Floor, New Haven CT 06520. E-mail: [email protected]. 0749-3797/$36.00 https://doi.org/10.1016/j.amepre.2019.12.003
© 2020 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.
Am J Prev Med 2020;000(000):1−9
1
ARTICLE IN PRESS 2
Brackney et al / Am J Prev Med 2020;000(000):1−9
It is important to monitor HPV vaccine impact at the population level, especially when vaccination coverage is suboptimal, as it is in the U.S. Although HPV vaccination coverage has increased over time, the results from the 2018 National Immunization Survey−Teen showed that coverage for at least 1 dose was 69.9% in female youth and 66.3% in male youth.5 This coverage is substantially lower than coverage for other adolescent vaccines, including tetanus, diphtheria, and pertussis vaccine at 88.9% and meningococcal vaccine at 86.6%.5 In Connecticut, coverage increased from 45.0% for at least 1 dose for adolescent girls aged 13−17 years in 2008 to 70.5% for at least 1 dose for adolescent girls and boys aged 13−17 years in 2018.5,6 Despite this moderate coverage, declines in precancerous cervical lesions caused by HPV have been observed in the U.S.7−9 Monitoring trends in precancerous cervical lesions by race, ethnicity, and poverty is important to assess whether vaccine impact is occurring similarly by these sociodemographic and economic factors. Before vaccine impact, differences in HPV type distribution by race and ethnicity had been observed, with black and Hispanic women having a lower prevalence of HPV 16/18 in cervical lesions.10,11 Since the HPV vaccination has been introduced in the U. S., declines in HPV 16/18-associated cervical lesions have also differed by race and ethnicity with declines being observed in non-Hispanic white and non-Hispanic black women, but not in Asian or Hispanic women.12 Analysis of census tract−level area-based measures of sociodemographic characteristics is another way to monitor health outcomes by characteristics of communities.13 Area-based measures can reflect the immediate surroundings in which people live, such as access to health services, resources, socioeconomic advantage/disadvantage, and other living conditions. These measures have been demonstrated to be strong predictors of health outcomes.14,15 For example, neighborhood poverty has been demonstrated to be associated with poorer health outcomes, including cervical cancer.16,17 Early data from the HPV-IMPACT surveillance program has shown differences in HPV type distribution by area-based measures of poverty with women living in areas of high poverty being less likely to have HPV 16/18−associated lesions after controlling for individual-level race and ethnicity.18 Furthermore, since the introduction of HPV vaccines, greater declines in the incidence of cervical lesions have been observed among women living in areas with lower proportions of black, Hispanic, and people living in poverty.19 To the best of our knowledge, no previous studies have examined the extent to which HPV 16/18−associated cervical lesions have declined by area-based measures of race, ethnicity, and poverty since the introduction of HPV vaccines. Therefore, the objective of this study is to
determine if the proportion of HPV 16/18−associated lesions declined similarly across census tract area-based measures of race, ethnicity, and poverty while adjusting for individual measures of race, ethnicity, and insurance (a proxy for poverty).
METHODS The HPV-IMPACT program, established in 2008, is a collaboration of the Centers for Disease Control and Prevention (CDC) and 5 Emerging Infections Program sites to monitor the population-level impact of the HPV vaccine on precancerous cervical lesions. Methods of HPV-IMPACT surveillance have been described previously.20 In Connecticut, cervical intraepithelial neoplasia (CIN) with grades 2 or higher and adenocarcinoma in situ (collectively known as CIN2 +) has been a reportable condition since January 1, 2008. This program has since collected data from >30 laboratories statewide. Enhanced surveillance, conducted for women aged 18−39 years residing in New Haven County, includes the collection of data from medical chart reviews, patient interviews, and residual specimen tissue that is sent to CDC for HPV DNA typing. This work has been deemed public health surveillance by the university, state, and federal IRBs and thus exempt from the requirement of human subjects approval. For this study, HPV typing was performed on 1 archived block from each case using previously described methods established for this surveillance system.10 Briefly, after histologic verification of the lesion in the formalin-fixed, paraffin-embedded tissue sections, DNA was extracted and tested using the Linear Array HPV Genotyping Assay (Roche Diagnostics). Samples with inadequate or negative results were retested with the INNO-LiPA HPV Genotyping Extra Assay (Innogenetics), with a smaller amplicon.
Study Population This analysis included CIN2+ women residing in New Haven County aged 21−39 years (n=2,779). According to the 2018 U.S. Census Bureau, New Haven County has approximately 857,620 people: 77.6% white, non-Hispanic; 15.0% black; 18.6% Hispanic; and <5% Asian or American Indian.21 Women aged 18−20 years were excluded because updated guidelines beginning in 2009 no longer recommended cervical cancer screening in this age group. Women who did not have a lesion typed (n=1,420) or could not be geocoded (n=23) were excluded. Reasons lesions were not typed included: specimen tissue could not be retrieved from the pathology laboratory, no residual tissue was available, or pathology review at CDC did not confirm the presence of a lesion.
Measures Each woman diagnosed with CIN2+ and reported to the surveillance system was geocoded to the census tract level using the residential address and the Federal Financial Institutions Examination Council Geocoding/Mapping System database. Census tract−level area-based measures of race, ethnicity, and poverty were obtained from the 2010−2014 American Community Survey provided by the U.S. Census Bureau.22,23 Geocoded cases were matched to the corresponding area-based measures from the American Community Survey. Areabased measures were examined by dividing them into 2 groups: high, defined as ≥20% of the census tract, and low, defined as <20% of the www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9 census tract. These categories were adapted from the cut points used in the Public Health Disparities Geocoding Project with the 3 lower levels combined into 1 category to increase the statistical power for analyses.24 The surveillance period was divided into 3 periods: 2008−2009, 2010−2012, and 2013−2015. Multiple years were combined for statistical power to permit a more robust analysis of trends. The first period, 2008−2009, was selected because declines in precancerous cervical lesions had not yet occurred in Connecticut.19 The second and third periods were divided evenly with 3 years in each group. Individual race and ethnicity variables were combined into the following categories: white, non-Hispanic; Hispanic; black, non-Hispanic; and other race, non-Hispanic. Categories that were too small to evaluate independently (Asian, American Indian, Pacific Islander, other race, multiple races) were combined into a single category, “other.” Information on individuallevel race, ethnicity, and insurance status were collected from laboratory reports, medical chart reviews, and case interviews. Insurance status was categorized into 2 groups: private insurance and no private insurance.
Statistical Analysis The primary outcome of interest was the proportion of CIN2+ lesions with HPV 16/18 detected, the oncogenic types included in all vaccines used in the U.S. since 2006. Analyses were stratified by 4 age groups: 21−24, 25−29, 30−34, and 35−39 years. Unadjusted logistic regression analysis was conducted to examine changes in the proportion of lesions with HPV 16/18 detected with the study period as the primary exposure variable of interest. Only age groups that demonstrated significant changes over time were analyzed further by area-based measures. Unadjusted and adjusted logistic regression models were created to examine the proportion of CIN2+ with HPV 16/18 detected over time. To determine whether declines were evident in all areas, separate models were developed for each level (≥20%, high and <20%, low) of the 3 area-based measures (race, ethnicity, and poverty), in the 2 age groups (21−24 years and 25−29 years) that showed overall declines in the proportion of cases with HPV 16/18. A previous Connecticut HPV-IMPACT analysis demonstrated that both area-based measures and individual demographic characteristics were associated with the rate of CIN2+.25 Therefore, all models also included individual-level race/ethnicity and insurance status. Data analysis was completed in 2019 using SAS, version 9.4. The declines of each level, high (≥20%) or low (<20%), for each of the area-based measures were compared by examining when/if significant declines began and the magnitude of these declines. Earlier declines were defined as declines that began in the second period (2010−2012); later declines were defined as declines that began in the third period (2013−2015). Larger declines were defined as a difference in AORs over time that were greater in magnitude.
RESULTS A total of 4,222 women aged 21−39 years were reported to have CIN2+ in New Haven County from 2008 to 2015. Included in the analysis were 2,779 women (65.8%) who
& 2020
3
were geocoded and had cervical biopsy tissue that was typed successfully. Women who were included in the analysis were significantly more likely to be white, have private health insurance, and live in areas with low proportions of black, Hispanic, and poor residents compared with those who were excluded (p<0.05 for each). During the study period, HPV 16/18 was detected in 1,075 (38.7%) lesions. The unadjusted logistic regression analysis demonstrated a significant overall decline in the proportion of CIN2+ lesions with HPV 16/18 detected when comparing 2013−2015 with 2008−2009 (OR=0.73, 95% CI=0.60, 0.89) (Table 1). The stratified analysis showed that changes in the HPV 16/18 proportion over time differed by age groups, with women in the younger age groups, 21−24 and 25−29 years, demonstrating significant decreases in the later periods but women aged 30−34 and 35−39 years showing no significant decreases (Table 2). When examining the effect of individual unadjusted characteristics of race/ethnicity, declines in white, non-Hispanic women aged 21−24 years began in the second period, 2010−2012 (OR=0.62, 95% CI=0.39, 0.98), and continued through the third period, 2013−2015 (OR=0.24, 95% CI=0.13, 0.43). Black women aged 21−24 years demonstrated significant declines beginning in the third period (OR=0.26, 95% CI=0.08, 0.92). Declines in white, non-Hispanic women aged 25−29 years began in the second period (OR=0.62, 95% CI=0.41, 0.94) but were not significant in the third period. No declines were seen in black or Hispanic women aged 25−29 years in either time period (data not shown). The subsequent analysis focused on area-based measures to address the primary research question. Within the 2 younger age groups with significant declines in the proportion of HPV 16/18 overall, declines in the proportion with HPV 16/18 began at different times and to different degrees of magnitude in areas with low versus high proportions of each area-based measure (Figure 1A and B). To quantify these changes in the proportion of HPV 16/18 over time in the context of area-based measures, ORs were estimated for the latter 2 periods (2010−2012, 2013−2015) compared with the first period (2008−2009), stratified by area-based measures and adjusted for individual-level characteristics (Table 3). Among women living in areas with lower poverty, significant declines in the proportion of HPV 16/18−associated lesions were observed in women aged 21−24 years in 2010−2012 (AOR=0.55, 95% CI=0.36, 0.85) and 2013 −2015 (AOR=0.30, 95% CI=0.18, 0.52) when compared with 2008−2009 (Table 3). A similar pattern was seen in women aged 25−29 years. In high-poverty areas, no significant declines were observed in any of the age groups.
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
4
Table 1. Proportion of HPV 16/18 Detected in CIN2+ Lesions Among Women Aged 21‒39 Years, 2008‒2015 Variable Period of diagnosis 2008‒2009 2010‒2012 2013‒2015 Age groups, years 21‒24 25‒29 30‒34 35‒39 Diagnosis CIN2/CIN2/3 CIN3/AIS Race/ethnicitya White Black Hispanic Other race Insurancea Private Not private Area-based measures Percent black <20% ≥20% Percent Hispanic <20% ≥20% Percent poverty <20% ≥20%
Total N=2,779
Number 16/18-positive (%) n=1,075
Number 16/18-negative (%) n=1,704 Unadjusted OR (95% CI)
783 1,138 858
333 (42.5) 442 (38.8) 300 (35.0)
450 (57.5) 696 (61.2) 558 (65.0)
1.00 0.86 (0.71, 1.03) 0.73 (0.60, 0.89)
757 983 678 361
273 (36.1) 413 (42.0) 265 (39.1) 124 (34.4)
484 (63.9) 570 (58.0) 413 (60.9) 237 (65.7)
1.08 (0.83, 1.40) 1.39 (1.08, 1.78) 1.23 (0.94, 1.60) 1.00
1,968 811
615 (31.3) 460 (56.7)
1,353 (68.8) 351 (43.3)
1.00 2.88 (2.44, 3.41)
1,575 384 523 126
635 (40.3) 120 (31.3) 202 (38.6) 47 (37.3)
940 (59.7) 264 (68.8) 321 (61.4) 79 (62.7)
1.00 0.67 (0.53, 0.85) 0.93 (0.76, 1.14) 0.88 (0.61, 1.28)
1,694 1,031
655 (38.7) 396 (38.4)
1,039 (61.3) 635 (61.6)
1.00 0.99 (0.84, 1.16)
1,946 833
787 (40.4) 288 (34.6)
1,159 (59.6) 545 (65.4)
1.00 0.78 (0.66, 0.92)
1,759 1,020
679 (38.6) 396 (38.8)
1,080 (61.4) 624 (61.2)
1.00 1.01 (0.86, 1.18)
2,017 762
782 (38.8) 293 (38.5)
1,235 (61.2) 469 (61.6)
1.00 0.99 (0.83, 1.17)
Note: Boldface indicates statistical significance at p<0.05. There were 1,420 women excluded because no lesion was typed, and 23 women excluded because they could not be geocoded. a Total numbers may not sum because of missing data; missing race/ethnicity n=171 (6.3%); missing insurance n=54 (2.0%). AIS, adenocarcinoma in situ; CIN, cervical intraepithelial neoplasia; CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
Among women living in areas with low proportions of black residents, significant declines in the proportion of HPV16/18−associated lesions were observed in women aged 21−24 years in 2010−2012 (AOR=0.61, 95% CI=0.40, 0.95) and 2013−2015 (AOR=0.37, 95% CI=0.22, 0.63) when compared with 2008−2009 (Table 3). A similar pattern was seen in women aged 25−29 years. However, in areas with high proportions of black residents, significant declines in the HPV 16/ 18−positive proportion did not begin until 2013−2015 for those aged 21−24 years, and not observed at all for those aged 25−29 years. Finally, in analyses stratified by proportion Hispanic, significant declines began in 2013−2015 for women
aged 21−24 years in both low- and high-Hispanic areas; declines were similar but only significant in the low-Hispanic areas among those aged 25−29 years.
DISCUSSION The results from this study contribute to the growing body of literature demonstrating HPV vaccine impact. These results demonstrated significant declines in the proportion of HPV 16/18−associated lesions in young women, aged 21−24 and 25−29 years, whereas no declines were seen in women aged 30−34 and 35−39 years. Women who were in the younger age groups were more likely to be vaccinated, which strengthens evidence of vaccine impact. www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
5
Table 2. Proportion of HPV 16/18 Detected in CIN2+ Lesions and Unadjusted ORs, 2008‒2015 Variable Age 21‒24 years (n=757) 2008‒2009 2010‒2012 2013‒2015 Age 25‒29 years (n=983) 2008‒2009 2010‒2012 2013‒2015 Age 30‒34 years (n=678) 2008‒2009 2010‒2012 2013‒2015 Age 35‒39 years (n=361) 2008‒2009 2010‒2012 2013‒2015
Totals N=2,779
Number 16/18 positive (%)
Number 16/18 negative (%)
Unadjusted OR (95% CI)
252 318 187
111 (44.1) 124 (39.0) 38 (20.3)
141 (56.0) 194 (61.0) 149 (79.7)
1.00 0.81 (0.58, 1.14) 0.32 (0.21, 0.50)
259 395 329
128 (49.4) 160 (40.5) 125 (38.0)
131 (50.6) 235 (59.5) 204 (62.0)
1.00 0.70 (0.51, 0.96) 0.63 (0.45, 0.87)
167 284 227
57 (34.1) 114 (40.1) 94 (41.4)
110 (65.9) 170 (59.9) 133 (58.6)
1.00 1.29 (0.87, 1.93) 1.36 (0.90, 2.07)
105 141 115
37 (35.2) 44 (31.2) 43 (37.4)
68 (64.8) 97 (68.8) 72 (62.6)
1.00 0.83 (0.49, 1.43) 1.10 (0.63, 1.90)
Note: Boldface indicates statistical significance (p<0.05). There were 1,420 women excluded because no lesion was typed, and 23 women excluded because they could not be geocoded. CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
Vaccines can reduce disease morbidity, mortality, and disparities. However, unequally distributed vaccine uptake can increase disparities, especially when vaccination coverage has not yet reached levels high enough to produce herd immunity. Building on previous findings of area-based differences in the prevalence of HPV 16/ 18 in cervical lesions and area-based differences in the declines in rates of CIN2+ overall in young women in Connecticut,18,19 this analysis further demonstrates differences in declines in the proportion of cervical lesions with HPV 16/18 detected. After adjusting for individuallevel race, ethnicity, and SES, earlier and greater declines in the proportion of lesions with HPV 16/18 were detected in areas of lower poverty and, to a lesser extent, in areas of lower proportions black and Hispanic residents. A recent analysis of data from the 5 HPV-IMPACT sites through 2014 demonstrated significant declines in the proportion of CIN2+ lesions with HPV 16/18 among white, non-Hispanic and black, non-Hispanic women aged 18−39 years, but no significant declines in Hispanic or Asian women.12 The present analysis of data from the Connecticut site of HPV-IMPACT demonstrated that similar results by individual-level race/ethnicity, but in addition to this, declines differed across area-based measures of poverty and to a lesser extent, race, even when controlling for individual-level race and insurance status, a proxy for income. Although vaccination coverage by area-based measures was not available, differences in patterns of vaccine & 2020
uptake such as overall coverage and age at vaccination initiation may be responsible for these differences. Furthermore, high cost may have been a contributing factor for uninsured young adult women. The HPV vaccine has been covered by the Vaccines for Children Program, which has covered the cost of vaccines for children aged ≤18 years who are uninsured or underinsured since 2006. Clinicians in low-income communities may be less likely to offer vaccines to women aged >18 years if they perceive them to be unaffordable.26 Additional possible reasons include lower access to preventive healthcare services and differences in sexual networks.
Limitations This analysis has some limitations worth noting. First, individual-level vaccine status was not assessed. Limited statistical power may have affected significance testing for some comparisons owing to stratification by both area-based measures and age groups, resulting in small numbers in some groups. In particular, the strata with higher proportions black, Hispanic, and living in poverty are smaller than the areas with lower proportions, and differences in observed timing of declines may be attributable to differences in precision. Second, this analysis does not account for changes in cervical cancer screening recommendations over the study period, which included increasing the age of screening initiation, longer screening intervals, and adding HPV co-testing.27 However, even if screening changes were adopted
ARTICLE IN PRESS 6
Brackney et al / Am J Prev Med 2020;000(000):1−9
Figure 1. Proportion of HPV 16/18 detected in CIN2+ lesions, among women aged 21‒24 years (A) and 25‒29 years (B), stratified census track area-based measures and year of diagnosis, 2008‒2015. HPV, human papillomavirus; CT, census tract; CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ.
www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
7
Table 3. Adjusted Associations Between Years of Diagnosis and Proportion HPV 16/18 Detected in CIN2+ Lesions, 2008‒ 2015 Year of diagnosis 2008−2009 Variable Age group 21‒24 years Area-based poverty <20% (n=498) ≥20% (n=198) Area-based black <20% (n=473) ≥20% (n=223) Area-based Hispanic <20% (n=433) ≥20% (n=263) Age group 25‒29 years Area-based poverty <20% (n=665) ≥20% (n=254) Area-based black <20% (n=636) ≥20% (n=283) Area-based Hispanic <20% (n=595) ≥20% (n=324)
2010−2012
2013−2015
AOR (95% CI)
1.00 1.00
0.55 (0.36, 0.85) 1.66 (0.86, 3.21)
0.30 (0.18, 0.52) 0.48 (0.19, 1.20)
1.00 1.00
0.61 (0.40, 0.95) 1.22 (0.64, 2.31)
0.37 (0.22, 0.63) 0.24 (0.09, 0.65)
1.00 1.00
0.65 (0.41, 1.03) 0.98 (0.55, 1.75)
0.31 (0.17, 0.56) 0.42 (0.19, 0.87)
1.00 1.00
0.64 (0.43, 0.95) 0.84 (0.46, 1.53)
0.62 (0.41, 0.93) 0.74 (0.38, 1.41)
1.00 1.00
0.61 (0.41, 0.91) 0.99 (0.54, 1.83)
0.61 (0.40, 0.91) 0.78 (0.40, 1.50)
1.00 1.00
0.73 (0.48, 1.10) 0.62 (0.36, 1.09)
0.64 (0.41, 0.98) 0.64 (0.35, 1.17)
Note: Boldface indicates statistical significance (p<0.05). All analyses adjusted for individual-level race/ethnicity and insurance. There were 849 women excluded because no lesion was typed and 18 women excluded because they could not be geocoded. CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
differently for residents of relatively advantaged and disadvantaged neighborhoods, these would have affected the number of detected lesions but it seems unlikely that they would have impacted the proportion testing positive for HPV 16/18. Finally, the results may not be generalizable beyond New Haven County or Connecticut. However, the demographics of sex, race/ethnicity, and age are similar to the U.S. as a whole.
CONCLUSIONS Results from this analysis extend previous research showing declines in CIN2+ lesions and the proportion of lesions with HPV 16/18 in the era of HPV vaccination.7,12 Overall declines are encouraging; however, the delayed impact seen in neighborhoods with large proportions of black and poor residents indicates a need for continued monitoring of these trends and continuing efforts for optimal vaccine uptake. Residing in areas of concentrated poverty is associated with several poor health outcomes, including cervical cancer.14−16,28 Previous studies demonstrating higher late-stage cervical cancer diagnoses and higher rates of death from cervical & 2020
cancer in areas of high poverty indicate that cervical cancer prevention efforts, including vaccination and screening, need to be stronger in these communities to eliminate disparities.16,17 In addition, efforts to vaccinate all eligible adolescents and young adults would also contribute to driving down these disparities by increasing herd immunity. Efforts to increase vaccine uptake should include both a focus on the administration of vaccine at the recommended ages of 11−12 years for optimal impact, as well as catch-up vaccine for those aged 13−26 years who were not previously vaccinated. Furthermore, this analysis demonstrates the utility of census tract area-based measures that allows public health officials, advocates, and community partners to guide efforts that effectively target interventions at the community level. Future research should consist of continued population-based monitoring of HPV-related outcomes, vaccination uptake, and screening. In addition, in-depth studies at the individual level of missed opportunities would provide valuable insight that may help explain the underlying cause of socioeconomic disparities identified herein.
ARTICLE IN PRESS 8
Brackney et al / Am J Prev Med 2020;000(000):1−9
ACKNOWLEDGMENTS The Connecticut Emerging Infections Program would like to thank James Hadler, Elizabeth Unger, and Lauri Markowitz for their consultation on this manuscript. The Connecticut Emerging Infections Program would also like to thank the data collection efforts of Kyle Higgins and the many students in the Yale School of Public Health over the years that have contributed to the surveillance system that provided data for this analysis. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. This work was supported by a Centers for Disease Control and Prevention cooperative agreement (5 NU50CK000488-0200) with the Connecticut Emerging Infections Program. The study sponsor had a role in the study design, collection, analysis, and interpretation of data, or writing the report. MMB contributed to the conception and design, acquisition of data, analysis, and interpretation of data, and writing and revision of the manuscript. LMN, JWG, SEH, and JM contributed to the conception and design, analysis and interpretation of data, and review and revision of the manuscript. TDQ contributed to the acquisition of data, analysis and interpretation of data, and the review and revision of the manuscript. Linda M. Niccolai has served as a scientific advisor for Merck. No other financial disclosures were reported by the authors of this paper.
9.
10.
11.
12.
13.
14.
15.
REFERENCES 1. Van Dyne EA, Henley SJ, Saraiya M, Thomas CC, Markowitz LE, Benard VB. Trends in human papillomavirus-associated cancers United States, 1999‒2015. MMWR Morb Mortal Wkly Rep. 2018;67 (33):918–924. https://doi.org/10.15585/mmwr.mm6733a2. 2. Centers for Disease Control and Prevention (CDC). FDA licensure of quadrivalent human papillomavirus vaccine (HPV4, Gardasil) for use in males and guidance from the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep. 2010;59(20):630–632. https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5920a5.htm. Accessed October 9, 2018. 3. Markowitz LE, Dunne EF, Saraiya M, et al. Human papillomavirus vaccination: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep. 2014;63(RR-05):1–30. https:// www.cdc.gov/mmwr/preview/mmwrhtml/rr6305a1.htm. Accessed February 7, 2019. 4. de Martel C, Plummer M, Vignat J, Franceschi S. Worldwide burden of cancer attributable to HPV by site, country and HPV type. Int J Cancer. 2017;141(4):664–670. https://doi.org/10.1002/ijc.30716. 5. Walker TY, Elam-Evans LD, Yankey D, et al. National, regional, state, and selected local area vaccination coverage among adolescents aged 13−17 years - United States, 2018. MMWR Morb Mortal Wkly Rep. 2019;68(33):718–723. https://doi.org/10.15585/mmwr.mm6833a2. 6. Centers for Disease Control and Prevention (CDC). National, state, and local area vaccination coverage among adolescents aged 13−17 years—United States, 2008. MMWR Morb Mortal Wkly Rep. 2009;58 (36):997–1001. https://www.ncbi.nlm.nih.gov/pubmed/19763075. Accessed October 9, 2018. 7. Gargano JW, Park IU, Griffin MR, et al. Trends in high-grade cervical lesions and cervical cancer screening in 5 states, 2008−2015. Clin Infect Dis. 2019;68(8):1282–1291. https://doi.org/10.1093/cid/ciy707. 8. Flagg EW, Torrone EA, Weinstock H. Ecological association of human papillomavirus vaccination with cervical dysplasia prevalence
16.
17.
18.
19.
20.
21.
22.
23.
in the United States, 2007-2014. Am J Public Health. 2016;106 (12):2211–2218. https://doi.org/10.2105/AJPH.2016.303472. Niccolai LM, Meek JI, Brackney M, Hadler JL, Sosa LE, Weinberger DM. Declines in human papillomavirus (HPV)-associated high-grade cervical lesions after introduction of HPV vaccines in Connecticut, United States, 2008−2015. Clin Infect Dis. 2017;65(6):884–889. https://doi.org/10.1093/cid/cix455. Hariri S, Unger ER, Powell SE, et al. Human papillomavirus genotypes in high-grade cervical lesions in the United States. J Infect Dis. 2012;206(12):1878–1886. https://doi.org/10.1093/infdis/jis627. Hariri S, Unger ER, Schafer S, et al. HPV type attribution in highgrade cervical lesions: assessing the potential benefits of vaccines in a population-based evaluation in the United States. Cancer Epidemiol Biomarkers Prev. 2015;24(2):393–399. https://doi.org/10.1158/10559965.EPI-14-0649. McClung NM, Gargano JW, Bennett NM, et al. Trends in human papillomavirus vaccine types 16 and 18 in cervical precancers, 2008−2014. Cancer Epidemiol Biomarkers Prev. 2019;28(3):602–609. https://doi.org/10.1158/1055-9965.EPI-18-0885. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Painting a truer picture of US socioeconomic and racial/ethnic health inequalities: the Public Health Disparities Geocoding Project. Am J Public Health. 2005;95(2):312–323. https://doi.org/10.2105/AJPH.2003.032482. Krieger N, Waterman PD, Chen JT, Soobader MJ, Subramanian SV. Monitoring socioeconomic inequalities in sexually transmitted infections, tuberculosis, and violence: geocoding and choice of area-based socioeconomic measures−the Public Health Disparities Geocoding Project (US). Public Health Rep. 2003;118(3):240–260. https://doi.org/ 10.1093/phr/118.3.240. Yousey-Hindes KM, Hadler JL. Neighborhood socioeconomic status and influenza hospitalizations among children: New Haven County, Connecticut, 2003-2010. Am J Public Health. 2011;101(9):1785–1789. https://doi.org/10.2105/AJPH.2011.300224. Singh GK, Miller BA, Hankey BF, Edwards BK. Persistent area socioeconomic disparities in U.S. incidence of cervical cancer, mortality, stage, and survival, 1975−2000. Cancer. 2004;101(5):1051–1057. https://doi.org/10.1002/cncr.20467. Gomez N, Guendelman S, Harley KG, Gomez SL. Nativity and neighborhood characteristics and cervical cancer stage at diagnosis and survival outcomes among Hispanic women in California. Am J Public Health. 2015;105(3):538–545. https://doi.org/10.2105/AJPH.2014.302261. Niccolai LM, Russ C, Julian PJ, et al. Individual and geographic disparities in human papillomavirus types 16/18 in high-grade cervical lesions: associations with race, ethnicity, and poverty. Cancer. 2013;119(16):3052–3058. https://doi.org/10.1002/cncr.28038. Niccolai LM, Julian PJ, Meek JI, McBride V, Hadler JL, Sosa LE. Declining rates of high-grade cervical lesions in young women in Connecticut, 2008−2011. Cancer Epidemiol Biomarkers Prev. 2013;22 (8):1446–1450. https://doi.org/10.1158/1055-9965.EPI-13-0272. Hariri S, Unger ER, Powell SE, et al. The HPV Vaccine Impact Monitoring Project (HPV-IMPACT): assessing early evidence of vaccination impact on HPV-associated cervical cancer precursor lesions. Cancer Causes Control. 2012;23(2):281–288. https://doi.org/10.1007/ s10552-011-9877-6. Quick Facts. New Haven County, Connecticut: U.S. Bureau of the Census. www.census.gov/quickfacts/newhavencountyconnecticut. Published 2018. Accessed October 23, 2019. 2010−2014. American Community Survey 5-year estimates, demographic and housing estimates. U.S. Bureau of the Census. https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml? pid=ACS_15_5YR_DP05&prodType=table. Published 2014. Accessed May 24, 2018. 2010−2014. American Community Survey 5-year estimates, poverty status in the past 12 months. U.S. Bureau of the Census. https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?
www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9 pid=ACS_13_5YR_S1701&prodType=table. Published 2014. Accessed May 24, 2018. 24. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Race/ethnicity, gender, and monitoring socioeconomic gradients in health: a comparison of area-based socioeconomic measures−the Public Health Disparities Geocoding Project. Am J Public Health. 2003;93(10):1655–1671. https://doi.org/10.2105/ajph.93.10.1655. 25. Waggaman C, Julian P, Niccolai LM. Interactive effects of individual and neighborhood race and ethnicity on rates of high-grade cervical lesions. Cancer Epidemiol. 2014;38(3):248–252. https://doi.org/ 10.1016/j.canep.2014.03.004. 26. Hurley LP, Lindley MC, Allison MA, et al. Primary care physicians’ perspective on financial issues and adult immunization in the era of
& 2020
9
the Affordable Care Act. Vaccine. 2017;35(4):647–654. https://doi.org/ 10.1016/j.vaccine.2016.12.007. 27. Saslow D, Solomon D, Lawson HW, et al. American Cancer Society, American Society for Colposcopy and Cervical Pathology, and American Society for Clinical Pathology screening guidelines for the prevention and early detection of cervical cancer. CA Cancer J Clin. 2012;62(3):147–172. https://doi.org/10.3322/caac. 21139. 28. Greene SK, Levin-Rector A, Hadler JL, Fine AD. Disparities in reportable communicable disease incidence by census tractlevel poverty, New York City, 2006-2013. Am J Public Health. 2015;105(9):e27–e34. https://doi.org/10.2105/AJPH.2015. 302741.
RESEARCH ARTICLE
Human Papillomavirus 16/18−Associated Cervical Lesions: Differences by Area-Based Measures of Race and Poverty Monica M. Brackney, MS,1 Julia W. Gargano, PhD,2 Susan E. Hannagan, MPH,1 James Meek, MPH,1 Troy D. Querec, PhD,3 Linda M. Niccolai, PhD1
Introduction: This analysis evaluates trends in cervical lesions with human papillomavirus 16/18 detected by area-based measures of race, ethnicity, and poverty during 2008−2015.
Methods: Trends in the proportion of lesions with human papillomavirus 16/18 detected among residents of New Haven County, Connecticut were examined by area-based measures of race, ethnicity, and poverty. Area-based measures are aggregate descriptors of census tract characteristics useful for measuring differences in health outcomes in the context of where people live. Multivariable logistic regression modeling was conducted, adjusted for individual-level race, ethnicity, and insurance status to assess the independent effects of area-based measures. Data were analyzed in 2018−2019. Results: Among women aged 21−24 years and 25−29 years, significant declines in the proportion of lesions with human papillomavirus 16/18 were observed. Among women aged 21−24 years, declines began earlier and were greater in magnitude in areas of lower poverty (OR=0.55, 95% CI=0.36, 0.85 for 2010−2012 vs 2008−2009 and OR=0.30, 95% CI=0.18, 0.51 for 2013−2015 vs 2008−2009) compared with higher poverty (OR=1.66, 95% CI=0.86, 3.21 and OR=0.48, 95% CI=0.19, 1.20). Similar patterns were observed for women aged 25−29 years, and for area-based measures of race and ethnicity. Conclusions: Differences were observed in declines in the proportion of human papillomavirus 16/18 lesions by area-based measures since the introduction of human papillomavirus vaccines, with greater and earlier declines in areas with fewer residents living in poverty and racial minorities. Ongoing human papillomavirus vaccine impact monitoring is necessary to track differences by sociodemographic characteristics. Am J Prev Med 2020;000(000):1−9. © 2020 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
H
uman papillomavirus (HPV) is the primary cause of cervical and other anogenital cancers, oropharyngeal cancer, and genital warts.1 In 2006, a safe and efficacious quadrivalent HPV vaccine was licensed for use in the U.S. by the Food and Drug Administration.2 In 2007, the Advisory Committee on Immunization Practices recommended routine vaccination for female youth aged 11−12 years with catch-up vaccination through age 26 years.3 In 2011, the Advisory Committee on Immunization Practices extended the routine recommendation to male youth aged 11−12 years with catch-up vaccination up to age
21 years.3 All 3 HPV vaccines that are approved in the U.S. target HPV 16 and 18, which are responsible for 70% of cervical cancer cases.4 From the 1Connecticut Emerging Infections Program, Yale School of Public Health, New Haven, Connecticut; 2Division of Viral Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia; and 3Division of High Consequence Pathogens and Pathology, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia Address correspondence to: Monica M. Brackney, MS, Connecticut Emerging Infections Program, Yale School of Public Health, 1 Church Street, 7th Floor, New Haven CT 06520. E-mail: [email protected]. 0749-3797/$36.00 https://doi.org/10.1016/j.amepre.2019.12.003
© 2020 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.
Am J Prev Med 2020;000(000):1−9
1
ARTICLE IN PRESS 2
Brackney et al / Am J Prev Med 2020;000(000):1−9
It is important to monitor HPV vaccine impact at the population level, especially when vaccination coverage is suboptimal, as it is in the U.S. Although HPV vaccination coverage has increased over time, the results from the 2018 National Immunization Survey−Teen showed that coverage for at least 1 dose was 69.9% in female youth and 66.3% in male youth.5 This coverage is substantially lower than coverage for other adolescent vaccines, including tetanus, diphtheria, and pertussis vaccine at 88.9% and meningococcal vaccine at 86.6%.5 In Connecticut, coverage increased from 45.0% for at least 1 dose for adolescent girls aged 13−17 years in 2008 to 70.5% for at least 1 dose for adolescent girls and boys aged 13−17 years in 2018.5,6 Despite this moderate coverage, declines in precancerous cervical lesions caused by HPV have been observed in the U.S.7−9 Monitoring trends in precancerous cervical lesions by race, ethnicity, and poverty is important to assess whether vaccine impact is occurring similarly by these sociodemographic and economic factors. Before vaccine impact, differences in HPV type distribution by race and ethnicity had been observed, with black and Hispanic women having a lower prevalence of HPV 16/18 in cervical lesions.10,11 Since the HPV vaccination has been introduced in the U. S., declines in HPV 16/18-associated cervical lesions have also differed by race and ethnicity with declines being observed in non-Hispanic white and non-Hispanic black women, but not in Asian or Hispanic women.12 Analysis of census tract−level area-based measures of sociodemographic characteristics is another way to monitor health outcomes by characteristics of communities.13 Area-based measures can reflect the immediate surroundings in which people live, such as access to health services, resources, socioeconomic advantage/disadvantage, and other living conditions. These measures have been demonstrated to be strong predictors of health outcomes.14,15 For example, neighborhood poverty has been demonstrated to be associated with poorer health outcomes, including cervical cancer.16,17 Early data from the HPV-IMPACT surveillance program has shown differences in HPV type distribution by area-based measures of poverty with women living in areas of high poverty being less likely to have HPV 16/18−associated lesions after controlling for individual-level race and ethnicity.18 Furthermore, since the introduction of HPV vaccines, greater declines in the incidence of cervical lesions have been observed among women living in areas with lower proportions of black, Hispanic, and people living in poverty.19 To the best of our knowledge, no previous studies have examined the extent to which HPV 16/18−associated cervical lesions have declined by area-based measures of race, ethnicity, and poverty since the introduction of HPV vaccines. Therefore, the objective of this study is to
determine if the proportion of HPV 16/18−associated lesions declined similarly across census tract area-based measures of race, ethnicity, and poverty while adjusting for individual measures of race, ethnicity, and insurance (a proxy for poverty).
METHODS The HPV-IMPACT program, established in 2008, is a collaboration of the Centers for Disease Control and Prevention (CDC) and 5 Emerging Infections Program sites to monitor the population-level impact of the HPV vaccine on precancerous cervical lesions. Methods of HPV-IMPACT surveillance have been described previously.20 In Connecticut, cervical intraepithelial neoplasia (CIN) with grades 2 or higher and adenocarcinoma in situ (collectively known as CIN2 +) has been a reportable condition since January 1, 2008. This program has since collected data from >30 laboratories statewide. Enhanced surveillance, conducted for women aged 18−39 years residing in New Haven County, includes the collection of data from medical chart reviews, patient interviews, and residual specimen tissue that is sent to CDC for HPV DNA typing. This work has been deemed public health surveillance by the university, state, and federal IRBs and thus exempt from the requirement of human subjects approval. For this study, HPV typing was performed on 1 archived block from each case using previously described methods established for this surveillance system.10 Briefly, after histologic verification of the lesion in the formalin-fixed, paraffin-embedded tissue sections, DNA was extracted and tested using the Linear Array HPV Genotyping Assay (Roche Diagnostics). Samples with inadequate or negative results were retested with the INNO-LiPA HPV Genotyping Extra Assay (Innogenetics), with a smaller amplicon.
Study Population This analysis included CIN2+ women residing in New Haven County aged 21−39 years (n=2,779). According to the 2018 U.S. Census Bureau, New Haven County has approximately 857,620 people: 77.6% white, non-Hispanic; 15.0% black; 18.6% Hispanic; and <5% Asian or American Indian.21 Women aged 18−20 years were excluded because updated guidelines beginning in 2009 no longer recommended cervical cancer screening in this age group. Women who did not have a lesion typed (n=1,420) or could not be geocoded (n=23) were excluded. Reasons lesions were not typed included: specimen tissue could not be retrieved from the pathology laboratory, no residual tissue was available, or pathology review at CDC did not confirm the presence of a lesion.
Measures Each woman diagnosed with CIN2+ and reported to the surveillance system was geocoded to the census tract level using the residential address and the Federal Financial Institutions Examination Council Geocoding/Mapping System database. Census tract−level area-based measures of race, ethnicity, and poverty were obtained from the 2010−2014 American Community Survey provided by the U.S. Census Bureau.22,23 Geocoded cases were matched to the corresponding area-based measures from the American Community Survey. Areabased measures were examined by dividing them into 2 groups: high, defined as ≥20% of the census tract, and low, defined as <20% of the www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9 census tract. These categories were adapted from the cut points used in the Public Health Disparities Geocoding Project with the 3 lower levels combined into 1 category to increase the statistical power for analyses.24 The surveillance period was divided into 3 periods: 2008−2009, 2010−2012, and 2013−2015. Multiple years were combined for statistical power to permit a more robust analysis of trends. The first period, 2008−2009, was selected because declines in precancerous cervical lesions had not yet occurred in Connecticut.19 The second and third periods were divided evenly with 3 years in each group. Individual race and ethnicity variables were combined into the following categories: white, non-Hispanic; Hispanic; black, non-Hispanic; and other race, non-Hispanic. Categories that were too small to evaluate independently (Asian, American Indian, Pacific Islander, other race, multiple races) were combined into a single category, “other.” Information on individuallevel race, ethnicity, and insurance status were collected from laboratory reports, medical chart reviews, and case interviews. Insurance status was categorized into 2 groups: private insurance and no private insurance.
Statistical Analysis The primary outcome of interest was the proportion of CIN2+ lesions with HPV 16/18 detected, the oncogenic types included in all vaccines used in the U.S. since 2006. Analyses were stratified by 4 age groups: 21−24, 25−29, 30−34, and 35−39 years. Unadjusted logistic regression analysis was conducted to examine changes in the proportion of lesions with HPV 16/18 detected with the study period as the primary exposure variable of interest. Only age groups that demonstrated significant changes over time were analyzed further by area-based measures. Unadjusted and adjusted logistic regression models were created to examine the proportion of CIN2+ with HPV 16/18 detected over time. To determine whether declines were evident in all areas, separate models were developed for each level (≥20%, high and <20%, low) of the 3 area-based measures (race, ethnicity, and poverty), in the 2 age groups (21−24 years and 25−29 years) that showed overall declines in the proportion of cases with HPV 16/18. A previous Connecticut HPV-IMPACT analysis demonstrated that both area-based measures and individual demographic characteristics were associated with the rate of CIN2+.25 Therefore, all models also included individual-level race/ethnicity and insurance status. Data analysis was completed in 2019 using SAS, version 9.4. The declines of each level, high (≥20%) or low (<20%), for each of the area-based measures were compared by examining when/if significant declines began and the magnitude of these declines. Earlier declines were defined as declines that began in the second period (2010−2012); later declines were defined as declines that began in the third period (2013−2015). Larger declines were defined as a difference in AORs over time that were greater in magnitude.
RESULTS A total of 4,222 women aged 21−39 years were reported to have CIN2+ in New Haven County from 2008 to 2015. Included in the analysis were 2,779 women (65.8%) who
& 2020
3
were geocoded and had cervical biopsy tissue that was typed successfully. Women who were included in the analysis were significantly more likely to be white, have private health insurance, and live in areas with low proportions of black, Hispanic, and poor residents compared with those who were excluded (p<0.05 for each). During the study period, HPV 16/18 was detected in 1,075 (38.7%) lesions. The unadjusted logistic regression analysis demonstrated a significant overall decline in the proportion of CIN2+ lesions with HPV 16/18 detected when comparing 2013−2015 with 2008−2009 (OR=0.73, 95% CI=0.60, 0.89) (Table 1). The stratified analysis showed that changes in the HPV 16/18 proportion over time differed by age groups, with women in the younger age groups, 21−24 and 25−29 years, demonstrating significant decreases in the later periods but women aged 30−34 and 35−39 years showing no significant decreases (Table 2). When examining the effect of individual unadjusted characteristics of race/ethnicity, declines in white, non-Hispanic women aged 21−24 years began in the second period, 2010−2012 (OR=0.62, 95% CI=0.39, 0.98), and continued through the third period, 2013−2015 (OR=0.24, 95% CI=0.13, 0.43). Black women aged 21−24 years demonstrated significant declines beginning in the third period (OR=0.26, 95% CI=0.08, 0.92). Declines in white, non-Hispanic women aged 25−29 years began in the second period (OR=0.62, 95% CI=0.41, 0.94) but were not significant in the third period. No declines were seen in black or Hispanic women aged 25−29 years in either time period (data not shown). The subsequent analysis focused on area-based measures to address the primary research question. Within the 2 younger age groups with significant declines in the proportion of HPV 16/18 overall, declines in the proportion with HPV 16/18 began at different times and to different degrees of magnitude in areas with low versus high proportions of each area-based measure (Figure 1A and B). To quantify these changes in the proportion of HPV 16/18 over time in the context of area-based measures, ORs were estimated for the latter 2 periods (2010−2012, 2013−2015) compared with the first period (2008−2009), stratified by area-based measures and adjusted for individual-level characteristics (Table 3). Among women living in areas with lower poverty, significant declines in the proportion of HPV 16/18−associated lesions were observed in women aged 21−24 years in 2010−2012 (AOR=0.55, 95% CI=0.36, 0.85) and 2013 −2015 (AOR=0.30, 95% CI=0.18, 0.52) when compared with 2008−2009 (Table 3). A similar pattern was seen in women aged 25−29 years. In high-poverty areas, no significant declines were observed in any of the age groups.
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
4
Table 1. Proportion of HPV 16/18 Detected in CIN2+ Lesions Among Women Aged 21‒39 Years, 2008‒2015 Variable Period of diagnosis 2008‒2009 2010‒2012 2013‒2015 Age groups, years 21‒24 25‒29 30‒34 35‒39 Diagnosis CIN2/CIN2/3 CIN3/AIS Race/ethnicitya White Black Hispanic Other race Insurancea Private Not private Area-based measures Percent black <20% ≥20% Percent Hispanic <20% ≥20% Percent poverty <20% ≥20%
Total N=2,779
Number 16/18-positive (%) n=1,075
Number 16/18-negative (%) n=1,704 Unadjusted OR (95% CI)
783 1,138 858
333 (42.5) 442 (38.8) 300 (35.0)
450 (57.5) 696 (61.2) 558 (65.0)
1.00 0.86 (0.71, 1.03) 0.73 (0.60, 0.89)
757 983 678 361
273 (36.1) 413 (42.0) 265 (39.1) 124 (34.4)
484 (63.9) 570 (58.0) 413 (60.9) 237 (65.7)
1.08 (0.83, 1.40) 1.39 (1.08, 1.78) 1.23 (0.94, 1.60) 1.00
1,968 811
615 (31.3) 460 (56.7)
1,353 (68.8) 351 (43.3)
1.00 2.88 (2.44, 3.41)
1,575 384 523 126
635 (40.3) 120 (31.3) 202 (38.6) 47 (37.3)
940 (59.7) 264 (68.8) 321 (61.4) 79 (62.7)
1.00 0.67 (0.53, 0.85) 0.93 (0.76, 1.14) 0.88 (0.61, 1.28)
1,694 1,031
655 (38.7) 396 (38.4)
1,039 (61.3) 635 (61.6)
1.00 0.99 (0.84, 1.16)
1,946 833
787 (40.4) 288 (34.6)
1,159 (59.6) 545 (65.4)
1.00 0.78 (0.66, 0.92)
1,759 1,020
679 (38.6) 396 (38.8)
1,080 (61.4) 624 (61.2)
1.00 1.01 (0.86, 1.18)
2,017 762
782 (38.8) 293 (38.5)
1,235 (61.2) 469 (61.6)
1.00 0.99 (0.83, 1.17)
Note: Boldface indicates statistical significance at p<0.05. There were 1,420 women excluded because no lesion was typed, and 23 women excluded because they could not be geocoded. a Total numbers may not sum because of missing data; missing race/ethnicity n=171 (6.3%); missing insurance n=54 (2.0%). AIS, adenocarcinoma in situ; CIN, cervical intraepithelial neoplasia; CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
Among women living in areas with low proportions of black residents, significant declines in the proportion of HPV16/18−associated lesions were observed in women aged 21−24 years in 2010−2012 (AOR=0.61, 95% CI=0.40, 0.95) and 2013−2015 (AOR=0.37, 95% CI=0.22, 0.63) when compared with 2008−2009 (Table 3). A similar pattern was seen in women aged 25−29 years. However, in areas with high proportions of black residents, significant declines in the HPV 16/ 18−positive proportion did not begin until 2013−2015 for those aged 21−24 years, and not observed at all for those aged 25−29 years. Finally, in analyses stratified by proportion Hispanic, significant declines began in 2013−2015 for women
aged 21−24 years in both low- and high-Hispanic areas; declines were similar but only significant in the low-Hispanic areas among those aged 25−29 years.
DISCUSSION The results from this study contribute to the growing body of literature demonstrating HPV vaccine impact. These results demonstrated significant declines in the proportion of HPV 16/18−associated lesions in young women, aged 21−24 and 25−29 years, whereas no declines were seen in women aged 30−34 and 35−39 years. Women who were in the younger age groups were more likely to be vaccinated, which strengthens evidence of vaccine impact. www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
5
Table 2. Proportion of HPV 16/18 Detected in CIN2+ Lesions and Unadjusted ORs, 2008‒2015 Variable Age 21‒24 years (n=757) 2008‒2009 2010‒2012 2013‒2015 Age 25‒29 years (n=983) 2008‒2009 2010‒2012 2013‒2015 Age 30‒34 years (n=678) 2008‒2009 2010‒2012 2013‒2015 Age 35‒39 years (n=361) 2008‒2009 2010‒2012 2013‒2015
Totals N=2,779
Number 16/18 positive (%)
Number 16/18 negative (%)
Unadjusted OR (95% CI)
252 318 187
111 (44.1) 124 (39.0) 38 (20.3)
141 (56.0) 194 (61.0) 149 (79.7)
1.00 0.81 (0.58, 1.14) 0.32 (0.21, 0.50)
259 395 329
128 (49.4) 160 (40.5) 125 (38.0)
131 (50.6) 235 (59.5) 204 (62.0)
1.00 0.70 (0.51, 0.96) 0.63 (0.45, 0.87)
167 284 227
57 (34.1) 114 (40.1) 94 (41.4)
110 (65.9) 170 (59.9) 133 (58.6)
1.00 1.29 (0.87, 1.93) 1.36 (0.90, 2.07)
105 141 115
37 (35.2) 44 (31.2) 43 (37.4)
68 (64.8) 97 (68.8) 72 (62.6)
1.00 0.83 (0.49, 1.43) 1.10 (0.63, 1.90)
Note: Boldface indicates statistical significance (p<0.05). There were 1,420 women excluded because no lesion was typed, and 23 women excluded because they could not be geocoded. CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
Vaccines can reduce disease morbidity, mortality, and disparities. However, unequally distributed vaccine uptake can increase disparities, especially when vaccination coverage has not yet reached levels high enough to produce herd immunity. Building on previous findings of area-based differences in the prevalence of HPV 16/ 18 in cervical lesions and area-based differences in the declines in rates of CIN2+ overall in young women in Connecticut,18,19 this analysis further demonstrates differences in declines in the proportion of cervical lesions with HPV 16/18 detected. After adjusting for individuallevel race, ethnicity, and SES, earlier and greater declines in the proportion of lesions with HPV 16/18 were detected in areas of lower poverty and, to a lesser extent, in areas of lower proportions black and Hispanic residents. A recent analysis of data from the 5 HPV-IMPACT sites through 2014 demonstrated significant declines in the proportion of CIN2+ lesions with HPV 16/18 among white, non-Hispanic and black, non-Hispanic women aged 18−39 years, but no significant declines in Hispanic or Asian women.12 The present analysis of data from the Connecticut site of HPV-IMPACT demonstrated that similar results by individual-level race/ethnicity, but in addition to this, declines differed across area-based measures of poverty and to a lesser extent, race, even when controlling for individual-level race and insurance status, a proxy for income. Although vaccination coverage by area-based measures was not available, differences in patterns of vaccine & 2020
uptake such as overall coverage and age at vaccination initiation may be responsible for these differences. Furthermore, high cost may have been a contributing factor for uninsured young adult women. The HPV vaccine has been covered by the Vaccines for Children Program, which has covered the cost of vaccines for children aged ≤18 years who are uninsured or underinsured since 2006. Clinicians in low-income communities may be less likely to offer vaccines to women aged >18 years if they perceive them to be unaffordable.26 Additional possible reasons include lower access to preventive healthcare services and differences in sexual networks.
Limitations This analysis has some limitations worth noting. First, individual-level vaccine status was not assessed. Limited statistical power may have affected significance testing for some comparisons owing to stratification by both area-based measures and age groups, resulting in small numbers in some groups. In particular, the strata with higher proportions black, Hispanic, and living in poverty are smaller than the areas with lower proportions, and differences in observed timing of declines may be attributable to differences in precision. Second, this analysis does not account for changes in cervical cancer screening recommendations over the study period, which included increasing the age of screening initiation, longer screening intervals, and adding HPV co-testing.27 However, even if screening changes were adopted
ARTICLE IN PRESS 6
Brackney et al / Am J Prev Med 2020;000(000):1−9
Figure 1. Proportion of HPV 16/18 detected in CIN2+ lesions, among women aged 21‒24 years (A) and 25‒29 years (B), stratified census track area-based measures and year of diagnosis, 2008‒2015. HPV, human papillomavirus; CT, census tract; CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ.
www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9
7
Table 3. Adjusted Associations Between Years of Diagnosis and Proportion HPV 16/18 Detected in CIN2+ Lesions, 2008‒ 2015 Year of diagnosis 2008−2009 Variable Age group 21‒24 years Area-based poverty <20% (n=498) ≥20% (n=198) Area-based black <20% (n=473) ≥20% (n=223) Area-based Hispanic <20% (n=433) ≥20% (n=263) Age group 25‒29 years Area-based poverty <20% (n=665) ≥20% (n=254) Area-based black <20% (n=636) ≥20% (n=283) Area-based Hispanic <20% (n=595) ≥20% (n=324)
2010−2012
2013−2015
AOR (95% CI)
1.00 1.00
0.55 (0.36, 0.85) 1.66 (0.86, 3.21)
0.30 (0.18, 0.52) 0.48 (0.19, 1.20)
1.00 1.00
0.61 (0.40, 0.95) 1.22 (0.64, 2.31)
0.37 (0.22, 0.63) 0.24 (0.09, 0.65)
1.00 1.00
0.65 (0.41, 1.03) 0.98 (0.55, 1.75)
0.31 (0.17, 0.56) 0.42 (0.19, 0.87)
1.00 1.00
0.64 (0.43, 0.95) 0.84 (0.46, 1.53)
0.62 (0.41, 0.93) 0.74 (0.38, 1.41)
1.00 1.00
0.61 (0.41, 0.91) 0.99 (0.54, 1.83)
0.61 (0.40, 0.91) 0.78 (0.40, 1.50)
1.00 1.00
0.73 (0.48, 1.10) 0.62 (0.36, 1.09)
0.64 (0.41, 0.98) 0.64 (0.35, 1.17)
Note: Boldface indicates statistical significance (p<0.05). All analyses adjusted for individual-level race/ethnicity and insurance. There were 849 women excluded because no lesion was typed and 18 women excluded because they could not be geocoded. CIN2+, cervical intraepithelial neoplasia grades 2 or greater and adenocarcinoma in situ; HPV, human papillomavirus.
differently for residents of relatively advantaged and disadvantaged neighborhoods, these would have affected the number of detected lesions but it seems unlikely that they would have impacted the proportion testing positive for HPV 16/18. Finally, the results may not be generalizable beyond New Haven County or Connecticut. However, the demographics of sex, race/ethnicity, and age are similar to the U.S. as a whole.
CONCLUSIONS Results from this analysis extend previous research showing declines in CIN2+ lesions and the proportion of lesions with HPV 16/18 in the era of HPV vaccination.7,12 Overall declines are encouraging; however, the delayed impact seen in neighborhoods with large proportions of black and poor residents indicates a need for continued monitoring of these trends and continuing efforts for optimal vaccine uptake. Residing in areas of concentrated poverty is associated with several poor health outcomes, including cervical cancer.14−16,28 Previous studies demonstrating higher late-stage cervical cancer diagnoses and higher rates of death from cervical & 2020
cancer in areas of high poverty indicate that cervical cancer prevention efforts, including vaccination and screening, need to be stronger in these communities to eliminate disparities.16,17 In addition, efforts to vaccinate all eligible adolescents and young adults would also contribute to driving down these disparities by increasing herd immunity. Efforts to increase vaccine uptake should include both a focus on the administration of vaccine at the recommended ages of 11−12 years for optimal impact, as well as catch-up vaccine for those aged 13−26 years who were not previously vaccinated. Furthermore, this analysis demonstrates the utility of census tract area-based measures that allows public health officials, advocates, and community partners to guide efforts that effectively target interventions at the community level. Future research should consist of continued population-based monitoring of HPV-related outcomes, vaccination uptake, and screening. In addition, in-depth studies at the individual level of missed opportunities would provide valuable insight that may help explain the underlying cause of socioeconomic disparities identified herein.
ARTICLE IN PRESS 8
Brackney et al / Am J Prev Med 2020;000(000):1−9
ACKNOWLEDGMENTS The Connecticut Emerging Infections Program would like to thank James Hadler, Elizabeth Unger, and Lauri Markowitz for their consultation on this manuscript. The Connecticut Emerging Infections Program would also like to thank the data collection efforts of Kyle Higgins and the many students in the Yale School of Public Health over the years that have contributed to the surveillance system that provided data for this analysis. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. This work was supported by a Centers for Disease Control and Prevention cooperative agreement (5 NU50CK000488-0200) with the Connecticut Emerging Infections Program. The study sponsor had a role in the study design, collection, analysis, and interpretation of data, or writing the report. MMB contributed to the conception and design, acquisition of data, analysis, and interpretation of data, and writing and revision of the manuscript. LMN, JWG, SEH, and JM contributed to the conception and design, analysis and interpretation of data, and review and revision of the manuscript. TDQ contributed to the acquisition of data, analysis and interpretation of data, and the review and revision of the manuscript. Linda M. Niccolai has served as a scientific advisor for Merck. No other financial disclosures were reported by the authors of this paper.
9.
10.
11.
12.
13.
14.
15.
REFERENCES 1. Van Dyne EA, Henley SJ, Saraiya M, Thomas CC, Markowitz LE, Benard VB. Trends in human papillomavirus-associated cancers United States, 1999‒2015. MMWR Morb Mortal Wkly Rep. 2018;67 (33):918–924. https://doi.org/10.15585/mmwr.mm6733a2. 2. Centers for Disease Control and Prevention (CDC). FDA licensure of quadrivalent human papillomavirus vaccine (HPV4, Gardasil) for use in males and guidance from the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep. 2010;59(20):630–632. https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5920a5.htm. Accessed October 9, 2018. 3. Markowitz LE, Dunne EF, Saraiya M, et al. Human papillomavirus vaccination: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep. 2014;63(RR-05):1–30. https:// www.cdc.gov/mmwr/preview/mmwrhtml/rr6305a1.htm. Accessed February 7, 2019. 4. de Martel C, Plummer M, Vignat J, Franceschi S. Worldwide burden of cancer attributable to HPV by site, country and HPV type. Int J Cancer. 2017;141(4):664–670. https://doi.org/10.1002/ijc.30716. 5. Walker TY, Elam-Evans LD, Yankey D, et al. National, regional, state, and selected local area vaccination coverage among adolescents aged 13−17 years - United States, 2018. MMWR Morb Mortal Wkly Rep. 2019;68(33):718–723. https://doi.org/10.15585/mmwr.mm6833a2. 6. Centers for Disease Control and Prevention (CDC). National, state, and local area vaccination coverage among adolescents aged 13−17 years—United States, 2008. MMWR Morb Mortal Wkly Rep. 2009;58 (36):997–1001. https://www.ncbi.nlm.nih.gov/pubmed/19763075. Accessed October 9, 2018. 7. Gargano JW, Park IU, Griffin MR, et al. Trends in high-grade cervical lesions and cervical cancer screening in 5 states, 2008−2015. Clin Infect Dis. 2019;68(8):1282–1291. https://doi.org/10.1093/cid/ciy707. 8. Flagg EW, Torrone EA, Weinstock H. Ecological association of human papillomavirus vaccination with cervical dysplasia prevalence
16.
17.
18.
19.
20.
21.
22.
23.
in the United States, 2007-2014. Am J Public Health. 2016;106 (12):2211–2218. https://doi.org/10.2105/AJPH.2016.303472. Niccolai LM, Meek JI, Brackney M, Hadler JL, Sosa LE, Weinberger DM. Declines in human papillomavirus (HPV)-associated high-grade cervical lesions after introduction of HPV vaccines in Connecticut, United States, 2008−2015. Clin Infect Dis. 2017;65(6):884–889. https://doi.org/10.1093/cid/cix455. Hariri S, Unger ER, Powell SE, et al. Human papillomavirus genotypes in high-grade cervical lesions in the United States. J Infect Dis. 2012;206(12):1878–1886. https://doi.org/10.1093/infdis/jis627. Hariri S, Unger ER, Schafer S, et al. HPV type attribution in highgrade cervical lesions: assessing the potential benefits of vaccines in a population-based evaluation in the United States. Cancer Epidemiol Biomarkers Prev. 2015;24(2):393–399. https://doi.org/10.1158/10559965.EPI-14-0649. McClung NM, Gargano JW, Bennett NM, et al. Trends in human papillomavirus vaccine types 16 and 18 in cervical precancers, 2008−2014. Cancer Epidemiol Biomarkers Prev. 2019;28(3):602–609. https://doi.org/10.1158/1055-9965.EPI-18-0885. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Painting a truer picture of US socioeconomic and racial/ethnic health inequalities: the Public Health Disparities Geocoding Project. Am J Public Health. 2005;95(2):312–323. https://doi.org/10.2105/AJPH.2003.032482. Krieger N, Waterman PD, Chen JT, Soobader MJ, Subramanian SV. Monitoring socioeconomic inequalities in sexually transmitted infections, tuberculosis, and violence: geocoding and choice of area-based socioeconomic measures−the Public Health Disparities Geocoding Project (US). Public Health Rep. 2003;118(3):240–260. https://doi.org/ 10.1093/phr/118.3.240. Yousey-Hindes KM, Hadler JL. Neighborhood socioeconomic status and influenza hospitalizations among children: New Haven County, Connecticut, 2003-2010. Am J Public Health. 2011;101(9):1785–1789. https://doi.org/10.2105/AJPH.2011.300224. Singh GK, Miller BA, Hankey BF, Edwards BK. Persistent area socioeconomic disparities in U.S. incidence of cervical cancer, mortality, stage, and survival, 1975−2000. Cancer. 2004;101(5):1051–1057. https://doi.org/10.1002/cncr.20467. Gomez N, Guendelman S, Harley KG, Gomez SL. Nativity and neighborhood characteristics and cervical cancer stage at diagnosis and survival outcomes among Hispanic women in California. Am J Public Health. 2015;105(3):538–545. https://doi.org/10.2105/AJPH.2014.302261. Niccolai LM, Russ C, Julian PJ, et al. Individual and geographic disparities in human papillomavirus types 16/18 in high-grade cervical lesions: associations with race, ethnicity, and poverty. Cancer. 2013;119(16):3052–3058. https://doi.org/10.1002/cncr.28038. Niccolai LM, Julian PJ, Meek JI, McBride V, Hadler JL, Sosa LE. Declining rates of high-grade cervical lesions in young women in Connecticut, 2008−2011. Cancer Epidemiol Biomarkers Prev. 2013;22 (8):1446–1450. https://doi.org/10.1158/1055-9965.EPI-13-0272. Hariri S, Unger ER, Powell SE, et al. The HPV Vaccine Impact Monitoring Project (HPV-IMPACT): assessing early evidence of vaccination impact on HPV-associated cervical cancer precursor lesions. Cancer Causes Control. 2012;23(2):281–288. https://doi.org/10.1007/ s10552-011-9877-6. Quick Facts. New Haven County, Connecticut: U.S. Bureau of the Census. www.census.gov/quickfacts/newhavencountyconnecticut. Published 2018. Accessed October 23, 2019. 2010−2014. American Community Survey 5-year estimates, demographic and housing estimates. U.S. Bureau of the Census. https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml? pid=ACS_15_5YR_DP05&prodType=table. Published 2014. Accessed May 24, 2018. 2010−2014. American Community Survey 5-year estimates, poverty status in the past 12 months. U.S. Bureau of the Census. https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?
www.ajpmonline.org
ARTICLE IN PRESS Brackney et al / Am J Prev Med 2020;000(000):1−9 pid=ACS_13_5YR_S1701&prodType=table. Published 2014. Accessed May 24, 2018. 24. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Race/ethnicity, gender, and monitoring socioeconomic gradients in health: a comparison of area-based socioeconomic measures−the Public Health Disparities Geocoding Project. Am J Public Health. 2003;93(10):1655–1671. https://doi.org/10.2105/ajph.93.10.1655. 25. Waggaman C, Julian P, Niccolai LM. Interactive effects of individual and neighborhood race and ethnicity on rates of high-grade cervical lesions. Cancer Epidemiol. 2014;38(3):248–252. https://doi.org/ 10.1016/j.canep.2014.03.004. 26. Hurley LP, Lindley MC, Allison MA, et al. Primary care physicians’ perspective on financial issues and adult immunization in the era of
& 2020
9
the Affordable Care Act. Vaccine. 2017;35(4):647–654. https://doi.org/ 10.1016/j.vaccine.2016.12.007. 27. Saslow D, Solomon D, Lawson HW, et al. American Cancer Society, American Society for Colposcopy and Cervical Pathology, and American Society for Clinical Pathology screening guidelines for the prevention and early detection of cervical cancer. CA Cancer J Clin. 2012;62(3):147–172. https://doi.org/10.3322/caac. 21139. 28. Greene SK, Levin-Rector A, Hadler JL, Fine AD. Disparities in reportable communicable disease incidence by census tractlevel poverty, New York City, 2006-2013. Am J Public Health. 2015;105(9):e27–e34. https://doi.org/10.2105/AJPH.2015. 302741.