National Analysis of Over 48,000 Veterans With HIV Demonstrates CD4/CD8 Ratio as a Risk Marker for Anal Intraepithelial Lesions and Anal Cancer

Abstract

BACKGROUND:

Anal squamous intraepithelial lesions are identifiable and treatable precancerous lesions that lack defined risk factors determining screening necessity.

OBJECTIVE:

Assess the prevalence and risk factors associated with low- and high-grade anal squamous intraepithelial lesions and anal squamous cell carcinoma.

DESIGN:

Retrospective cohort analysis of veterans with HIV between 1999–2023.

SETTINGS:

National multicenter study of the Department of Veterans Affairs.

PATIENTS:

Veterans with HIV who had >1 year of follow-up and no anal squamous intraepithelial lesions or anal cancer diagnosis prior to the study period.

MAIN OUTCOMES AND MEASURES:

Primary outcomes include the prevalence, disease-free survival rates, and hazard ratios associated with risk factors for developing anal squamous intraepithelial lesions and/or anal cancer.

RESULTS:

48,368 patients were analyzed. The average age of patients at study initiation was 47.8 years with a mean follow-up of 12.3 years. 7,572 (16%) patients had at least one anal cytopathology or histopathology result. Prevalence of anal disease was recorded for low-grade disease (n = 1,513, 3.1%), high-grade disease (n = 1,484, 3.1%), and cancer (n = 664, 1.4%). Mean times to first incident low-grade disease, high-grade disease, and cancer were 8.5 (SD = 6.0), 9.1 (SD = 6.0), and 9.7 (SD = 6.2) years, respectively. 5-year, 10-year, and 20-year disease-free survival rates for development of low-grade disease, high-grade disease, or cancer were 97.5%, 94.5%, and 88.4%, respectively. Cox regression modeling demonstrated CD4/CD8 ratios <0.5 were associated with increased risk of anal cancer (HR: 3.93, 95% CI: 3.33–4.63, p < 0.001).

LIMITATIONS:

Retrospective study that focuses almost exclusively on male U.S. veterans. Results might not apply to non-male, non-U.S. populations.

CONCLUSIONS:

National analysis of over 48,000 veterans with HIV demonstrates 16% had anal cytopathology or histopathology results with an anal cancer prevalence of 1.4%. CD4/CD8 ratios <0.5 correlate strongly with severity of anal disease and can help identify patients at highest risk for anal cancer to prioritize screening efforts. See Video Abstract.

Video Abstract

Download video file ^{(22.4MB, mp4)}

INTRODUCTION

Anal squamous cell carcinoma (ASCC) and its treatable precursor lesions, known as anal squamous intraepithelial lesions (ASIL), are primarily caused by chronic human papillomavirus (HPV) infections.^1,2 These infections are more common in immunocompromised individuals. As a result, specific populations, such as people with HIV (PWH), are disproportionately affected by ASIL and its attendant morbidity, as well as ASCC and its attendant morbidity and mortality.^2,3 Results from the ANCHOR study demonstrate that treating high-grade ASIL (HSIL) in PWH can effectively prevent progression to ASCC. These findings provide evidence to support screening for ASIL in PWH.⁴ However, numerous barriers to implementing successful anal disease screening programs exist and have contributed to the underutilization of screenings in this high-risk population.^5–8 Given the barriers to screening, it is necessary to identify non-invasive adjuncts that can aid in risk stratifying anal disease development in PWH. Such risk stratification will aid in individualizing screening protocols for ASIL and ASCC by prioritizing the utilization of limited resources in individuals at highest risk, thereby decreasing the burden of screening across health systems.

One promising adjunct to traditional screening measures is the CD4/CD8 ratio.⁹ Low CD4/CD8 ratios have been shown to correlate with risk of developing solid organ cancers in PWH, independent of CD4 T-cell counts, and the labs that comprise this ratio are routinely drawn in PWH though the ratio is typically not utilized clinically.^9–13 CD4/CD8 ratio inversion (<1) is common in PWH and is associated with several immunological abnormalities, such as increased innate and adaptive immune activation that can lead to the expansion of defective CD8 T-cells.^9,10,14 Notably, this immune dysfunction has been shown to persist despite successful virologic suppression on antiretroviral therapy and normalization of CD4 T-cell counts.^9,10 While we have previously demonstrated low CD4/CD8 ratios correlate with risk of anal HSIL and ASCC in a small, regional cohort of veterans with HIV, it remains unknown if these findings are applicable to a larger, national population.¹⁵

The prevalence of and risk factors for developing ASIL represent additional unknowns within the understudied population of PWH due to a lack of large-scale studies.^2,16,17 The majority of existing analyses related to risk of ASIL and ASCC development in PWH have involved small, specialized centers, included relatively short observation periods, and/or sourced data from large cancer registries, which generally lack the granular data necessary for evaluating ASIL.^17,18 We have sought to overcome these limitations by applying rule-based text analysis (RBTA) to free-text anal pathology reports, which for the purposes of this manuscripts refers to both cytopathology and histopathology, among a national cohort of veterans with HIV over a greater than 20-year timeframe. The Department of Veterans Affairs (VA) is well-suited for this investigation as it is the largest single healthcare system in the U.S. that cares for PWH, and it recommends annual screening for ASIL and ASCC in all veterans with HIV.¹⁹ The VA also possesses a centralized electronic medical record system that contains comprehensive, longitudinal clinical data on veterans in the U.S. and a robust research apparatus for acquiring this data.^20–23 In this context, we aimed to describe the prevalence of ASIL and ASCC among a national population of PWH and to evaluate risk factors, including the CD4/CD8 ratio, for the development of ASIL and ASCC.

MATERIALS AND METHODS

Data Source and Patient Population

After obtaining approval from the University of Wisconsin institutional review board and VA Research and Development Committee (ID 2019–0276), we queried the VA Corporate Data Warehouse (CDW) to generate a retrospective national cohort of veterans with HIV. Patients were identified via a previously validated method using two diagnostic criteria: 1) ICD-9/10 codes for HIV and 2) HIV lab test results (HIV viral load, CD4 T-cell count, or CD4/CD8 ratio).²⁴ Patients were included if they had at least one positive HIV lab test result within the VA between 1999–2023. Individual observation periods were initiated at the time of a patient’s first positive HIV lab result. Patients with less than one year of follow-up and those with anal pathology (i.e., cytopathology or histopathology) results of LSIL, HSIL, or ASCC prior to study initiation were excluded.

Variables and Outcomes

The primary outcomes in this study are prevalence of ASIL and ASCC (based on a patient’s most severe anal disease diagnosis), disease-free survival rates for ASIL and ASCC, and hazard ratios for ASIL and ASCC development that are linked to specific risk factors like the CD4/CD8 ratio. Demographic characteristics collected included age (at time of study initiation), sex, race, and VA location (region with the majority of an individual’s HIV lab results). Follow-up for each patient was defined as time from first positive HIV lab result (starting from the year 1999) to their most recent VA encounter or date of death before the end of 2023. Clinical characteristics with associated occurrence dates were collected including HIV lab results (HIV viral loads, CD4 T-cell counts, and CD4/CD8 ratios), anal cytopathology and histopathology results, and a history of the following factors: non-HPV-related sexually transmitted infections (STIs: gonorrhea, chlamydia, hepatitis B, hepatitis C, syphilis, or genital herpes simplex), smoking, and anogenital condyloma. Patients’ lowest ever CD4 count and CD4/CD8 ratio were recorded and labeled nadir. Nearest CD4 count and CD4/CD8 ratio were recorded as the closest lab result that preceded an anal pathology result. The median and mean number of days between pathology results and nearest CD4 counts and CD4/CD8 ratios were 49 days and 212 days, respectively.

We evaluated ASIL and ASCC by performing RBTA of free-text anal cytopathology and histopathology reports. This process was previously implemented in a regional study of anal disease in veterans with HIV but was modified for use in the current study to address variations in data on a national level.¹⁵ Anal pathology reports were analyzed for any of the following six results: 1) ASCC, 2) HSIL, 3) low-grade squamous intraepithelial lesions (LSIL), 4) atypical squamous cells cannot exclude HSIL (ASC-H), 5) atypical squamous cells of undetermined significance (ASCUS), and 6) negative for intraepithelial lesions or malignancy (NILM).²⁵ Non-contemporary terminology for anal cytopathology and histopathology reports was observed and translated to align with contemporary terminology. For example, the terms Bowen’s disease and anal squamous carcinoma in situ were both labeled as HSIL. We performed manual data validation on a randomized sample (10%) of each pathology result type obtained via our RBTA to evaluate for accuracy in labeling pathology results. ICD-9/10 diagnostic codes were used as a secondary means to identify anal pathology results when no cytopathology or histopathology reports were directly retrieved via querying the VA CDW (e.g., scanned reports or pathology results performed outside the VA). In these cases, manual chart review was performed to evaluate the presence and timing of anal disease.

Statistical Analysis

Individuals without anal disease prior to study initiation were followed for subsequent development of first incident ASIL or ASCC. Three mutually inclusive diagnostic groupings of anal disease(s) were evaluated: 1) LSIL, HSIL, or ASCC; 2) HSIL or ASCC; or 3) ASCC alone. Rates of development for each of these events were estimated using Kaplan-Meier methods. Disease-free survival rates from time of study initiation were determined at 5 years, 10 years, and 20 years for each diagnostic grouping. Factors hypothesized to be associated with development of anal disease were evaluated using Cox proportional hazards models. Univariable analyses were performed to assess for interactions between certain risk factors and development of anal disease. STI history, smoking history, and anogenital condyloma history were treated as time-varying covariates. Anogenital condyloma history was excluded as a variable in the analysis aimed at evaluating the diagnostic grouping of LSIL, HSIL, or ASCC, as anogenital condyloma is considered a form of LSIL. Because care for PWH changed dramatically during our study period, we performed a time-based sensitivity analysis using univariable Cox regression during the periods of 2004–2010 and 2016–2022. These two time periods were chosen on the basis that they roughly represent different eras in HIV care. For example, the latter time period includes the introduction of single-tablet anti-retroviral therapy (ART) regimens with potent integrase inhibitors and an increasing focus on immediate ART initiation regardless of CD4 counts.

To evaluate the impact of immunologic markers on the development of anal disease, we generated multivariable Cox regression models adjusting for clinically relevant factors among each diagnostic grouping of anal disease for nadir CD4 count, nadir CD4/CD8 ratio, nearest CD4 count, and nearest CD4/CD8 ratio (12 total models). We removed time-varying covariates from the multivariable models. CD4 counts and CD4/CD8 ratios were evaluated as binary variables (CD4 count ≥200 cells/mm³ or <200 cells/mm³ and CD4/CD8 ratio ≥0.5 or <0.5). For both univariable and multivariable analyses, a unit change of 100 was used for CD4 counts and a unit change of 1.0 was used for CD4/CD8 ratios. Statistical analyses were performed at a significance level of 0.05. All analyses were performed using SAS statistical software version 9.4.

RESULTS

A total of 62,750 veterans were identified as having a history of HIV via ICD codes. We excluded 11,228 veterans who did not have at least one positive HIV lab result identified within the VA. We also excluded 2,971 veterans who had less than one year of follow-up and an additional 183 veterans who had documented LSIL, HSIL, or ASCC prior to their first HIV lab result. After applying these exclusions, 48,368 veterans with HIV remained for analysis (Fig. 1).

Figure 1.

Flow chart of inclusion and exclusion criteria. Abbreviations: ASCC, anal squamous cell carcinoma; HSIL, high-grade squamous intraepithelial lesions; ICD, International Classification of Diseases; LSIL, low-grade squamous intraepithelial lesions; PWH, people with HIV.

The average age of patients at the beginning of the observation period was 47.8 years with a mean follow-up time of 12.3 years (Table 1). Patients were predominantly male (96.8%) and the majority self-identified racially as either Black (47.0%) or White (39.2%). While veterans from all major U.S. geographical regions were represented, the majority of representation came from the North Atlantic (27.5%) and Southeast (26.4%) regions. One in six patients (16.6%) had a history of a prior non-HPV-related STI and nearly half (49.0%) of patients were current smokers. A small proportion (0.9%) of patients had a prior history of anogenital condyloma. None of these individuals were included in our analyses aimed at assessing the diagnostic grouping LSIL, HSIL, or ASCC.

Table 1.

Characteristics of study population at the time of study initiation

Characteristics	Patients (n = 48368)
Mean age, years (SD)	47.8 (11.2)
Sex, n (%)
Female	1549 (3.2)
Male	46819 (96.8)
Race, n (%)
American Indian or Alaska Native	268 (0.6)
Asian	207 (0.4)
Black	22733 (47.0)
Multiple	573 (1.2)
Native Hawaiian or other Pacific Islander	303 (0.6)
Unknown	5326 (11.0)
White	18958 (39.2)
VA Region, n (%)
Continental	8023 (16.6)
Midwest	5699 (11.8)
North Atlantic	13318 (27.5)
Pacific	8545 (17.7)
Southeast	12783 (26.4)
STI history, n (%)	8071 (16.7)
Smoking history, n (%)
Current user	23721 (49.0)
Former user	4069 (8.4)
Non-smoker/Unknown	20578 (42.6)
Anogenital condyloma history, n (%)	437 (0.9)
Mean time of follow-up, years (SD)	12.3 (7.2)

Abbreviations: SD, standard deviation; STI, sexually transmitted infection; VA, Veterans Affairs.

Manual review of anal cytopathology and histopathology results (10% sampling) labeled by RBTA yielded an overall accuracy rate of 96% (694/725). Using ICD-codes (with manual chart review), we identified 445 additional patients with anal pathology results who were not originally captured by the RBTA. A more severe anal pathology result was identified in 57 (0.8%) patients. By combining the results of the RBTA and ICD-code approaches, the total number of patients with at least one anal cytopathology or histopathology result was 7,572 (16% of included patients) and the total with at least one abnormal result was 4,742 (9.8% of included patients or 62.6% of patients with at least one anal pathology result). Among the entire cohort of included patients, the following prevalence rates were determined: 664 ASCC (1.4%), 1484 HSIL (3.1%), 1,513 LSIL (3.1%), 47 ASC-H (0.1%), 1034 ASCUS (2.1%), 2,830 NILM (5.9%), and 40,796 with no anal pathology result (84.3%). Mean times to first incident LSIL, HSIL, and ASCC were 8.5 (SD = 6.0), 9.1 (SD = 6.0), and 9.7 (SD = 6.2) years, respectively.

5-year, 10-year, and 20-year disease-free survival rates from time of study initiation for each anal disease diagnostic grouping were as follows: LSIL, HSIL, or ASCC (97.5%, 94.5%, and 88.4%, respectively), HSIL or ASCC (98.7%, 97.0%, and 92.9%, respectively), and ASCC alone (99.5%, 98.9%, and 97.5%, respectively) (Fig. 2). Results for the univariable Cox regression analyses demonstrate Black race was associated with a lower risk of anal disease as compared to White race (Table 2). Conversely, receiving the majority of one’s care in either the Midwest or Pacific regions was associated with an increased risk of anal disease compared to the Southeast region. History of a non-HPV-related STI and history of smoking were significantly associated with each of the three anal disease diagnostic groupings. History of anogenital condyloma was significantly associated with the development of HSIL or ASCC and ASCC alone. Nadir CD4 counts <200 cells/mm³, nearest CD4 counts <200 cells/mm³, nadir CD4/CD8 ratios <0.5, and nearest CD4/CD8 ratios <0.5 were significantly associated with each anal disease diagnostic grouping, and these associations were stronger with increasing disease severity.

Figure 2.

Kaplan-Meier curve of time to development of anal cancer for all patients (n = 48,368). Data censored based on patients lost to follow-up or death. Date of first positive HIV lab result is the starting timepoint for all patients.

Table 2.

Univariable Cox regression models with unadjusted hazard ratios for development of anal LSIL, HSIL, and/or ASCC in PWH

Anal Disease	LSIL, HSIL, or ASCC				HSIL or ASCC				ASCC Alone
Characteristics	N	HR	(95% CI)	p value	N	HR	(95% CI)	p value	N	HR	(95% CI)	p value
Age	3323	0.97	(0.97–0.97)	<0.001	1960	0.98	(0.98–0.98)	<0.001	664	1.01	(0.99–1.01)	0.157
Race	3323				1960				664
White		Ref.	–	–		Ref.	–	–		Ref.	–	–
Black		0.69	(0.64–0.74)	<0.001		0.61	(0.56–0.67)	<0.001		0.56	(0.47–0.66)	<0.001
Other		0.72	(0.63–0.82)	<0.001		0.69	(0.59–0.82)	<0.001		0.85	(0.65–1.10)	0.217
VA Region	3323				1960				664
Southeast		Ref.	–	–		Ref.	–	–		Ref.	–	–
Continental		0.93	(0.83–1.05)	0.257		0.90	(0.77–1.05)	0.175		1.09	(0.85–1.40)	0.493
Midwest		2.04	(1.84–2.27)	<0.001		2.11	(1.85–2.42)	<0.001		1.62	(1.27–2.06)	<0.001
North Atlantic		0.94	(0.84–1.04)	0.194		0.93	(0.82–1.07)	0.307		0.92	(0.74–1.15)	0.466
Pacific		1.47	(1.33–1.63)	<0.001		1.35	(1.18–1.54)	<0.001		1.36	(1.08–1.70)	0.009
STI history	3323	57.18	(53.34–61.30)	<0.001	1960	96.81	(88.41–106.01)	<0.001	664	251.15	(215.18–293.15)	<0.001
Smoking history	3323	38.02	(35.48–40.75)	<0.001	1960	61.87	(61.08–73.21)	<0.001	664	210.39	(179.88–246.08)	<0.001
Anogenital condyloma history	–	–	–	–	1960	183.41	(167.52–200.80)	<0.001	664	189.15	(159.40–224.45)	<0.001
Nadir CD4 <200 cells/mm3	3070	1.08	(1.01–1.16)	0.026	1792	1.26	(1.15–1.39)	<0.001	614	2.19	(1.85–2.59)	<0.001
Nadir CD4/CD8 <0.5	3070	1.18	(1.08–1.28)	<0.001	1792	1.33	(1.18–1.49)	<0.001	614	2.88	(2.22–3.74)	<0.001
Nearest CD4 <200 cells/mm3	3070	1.31	(1.19–1.44)	<0.001	1792	1.50	(1.32–1.70)	<0.001	614	2.37	(1.97–2.85)	<0.001
Nearest CD4/CD8 <0.5	3070	2.00	(1.87–2.15)	<0.001	1792	2.33	(2.13–2.56)	<0.001	614	3.84	(3.26–4.52)	<0.001

Abbreviations: ASCC, anal squamous cell carcinoma; CD, cluster of differentiation; CI, confidence interval; HR, hazard ratio; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion; STI, sexually transmitted infection; VA, Veterans Affairs; PWH, people with HIV.

For our sensitivity analysis involving the time periods 2004–2010 and 2016–2022, similar trends were observed during these two time periods as compared to the entire study period with respect to the studied immune markers and their increasingly strong associations with anal disease severity. Nearest CD4/CD8 ratios <0.5 were significantly associated with ASCC from 2004–2010 (HR: 4.66, 95% CI: 3.18–6.82, p < 0.001) and 2016–2022 (HR: 5.17, 95% CI: 2.23–11.96, p < 0.001).

In all multivariable Cox regression models, Black race was associated with a decreased risk of anal disease while the Midwest region was associated with an increased risk (Table 3). Across all models, lower nadir and nearest CD4 counts (<200 cells/mm³) and CD4/CD8 ratios (<0.5) were associated with an increased risk of developing anal disease, and these risks are increasingly strong with worsening severity of anal disease. Among the immune markers, nearest CD4/CD8 ratios demonstrated the strongest risk association across each anal diagnostic grouping.

Table 3.

Multivariable Cox regression models with adjusted hazard ratios for development of anal disease in PWH

LSIL, HSIL, or ASCC (n = 3070)
Immune Markers	Nadir CD4 <200			Nadir CD4/CD8 <0.5			Nearest CD4 <200			Nearest CD4/CD8 <0.5
Characteristics	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value
Age	0.97	(0.97–0.97)	<0.001	0.97	(0.97–0.97)	<0.001	0.97	(0.97–0.97)	<0.001	0.97	(0.96–0.97)	<0.001
Race
White	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Black	0.70	(0.65–0.76)	<0.001	0.71	(0.65–0.76)	<0.001	0.71	(0.65–0.76)	<0.001	0.69	(0.64–0.75)	<0.001
Other	0.70	(0.62–0.80)	<0.001	0.71	(0.62–0.81)	<0.001	0.70	(0.61–0.79)	<0.001	0.67	(0.59–0.77)	<0.001
VA Region
Southeast	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Continental	0.86	(0.77–0.97)	0.02	0.87	(0.77–0.98)	0.02	0.87	(0.77–0.98)	0.02	0.87	(0.77–0.98)	0.02
Midwest	2.05	(1.84–2.28)	<0.001	2.04	(1.84–2.27)	<0.001	2.04	(1.84–2.27)	<0.001	2.03	(1.82–2.26)	<0.001
North Atlantic	0.85	(0.76–0.95)	0.004	0.85	(0.76–0.95)	0.005	0.85	(0.76–0.95)	0.003	0.84	(0.75–0.93)	0.002
Pacific	1.40	(1.26–1.55)	<0.001	1.40	(1.26–1.55)	<0.001	1.40	(1.26–1.55)	<0.001	1.39	(1.25–1.54)	<0.001
Nadir or nearest count or ratio	1.16	(1.08–1.25)	<0.001	1.20	(1.10–1.31)	<0.001	1.41	(1.27–1.55)	<0.001	2.12	(1.97–2.27)	<0.001
HSIL or ASCC (n = 1792)
Immune Markers	Nadir CD4 <200			Nadir CD4/CD8 <0.5			Nearest CD4 <200			Nearest CD4/CD8 <0.5
Characteristics	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value
Age	0.98	(0.98–0.98)	<0.001	0.98	(0.98–0.99)	<0.001	0.98	(0.98–0.98)	<0.001	0.98	(0.97–0.98)	<0.001
Race
White	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Black	0.61	(0.55–0.68)	<0.001	0.62	(0.56–0.68)	<0.001	0.62	(0.56–0.68)	<0.001	0.61	(0.55–0.67)	<0.001
Other	0.67	(0.56–0.80)	<0.001	0.67	(0.56–0.80)	<0.001	0.66	(0.55–0.79)	<0.001	0.64	(0.54–0.76)	<0.001
VA Region
Southeast	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Continental	0.84	(0.72–0.99)	0.03	0.85	(0.73–0.99)	0.04	0.85	(0.73–1.00)	0.04	0.85	(0.73–0.99)	0.04
Midwest	2.09	(1.83–2.40)	<0.001	2.08	(1.82–2.39)	<0.001	2.09	(1.82–2.40)	<0.001	2.07	(1.80–2.37)	<0.001
North Atlantic	0.84	(0.73–0.98)	0.02	0.85	(0.73–0.98)	0.03	0.84	(0.72–0.97)	0.02	0.82	(0.71–0.95)	0.009
Pacific	1.22	(1.07–1.41)	0.005	1.22	(1.06–1.40)	0.006	1.22	(1.06–1.40)	0.006	1.20	(1.04–1.38)	0.01
Nadir or nearest count or ratio	1.35	(1.23–1.48)	<0.001	1.36	(1.21–1.52)	<0.001	1.60	(1.41–1.81)	<0.001	2.44	(2.22–2.67)	<0.001
ASCC Alone (n = 614)
Immune Markers	Nadir CD4 <200			Nadir CD4/CD8 <0.5			Nearest CD4 <200			Nearest CD4/CD8 <0.5
Characteristics	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value	HR	(95% CI)	p value
Age	1.01	(1.00–1.01)	0.24	1.01	(1.00–1.01)	0.14	1.00	(1.00–1.01)	0.27	1.00	(1.00–1.01)	0.50
Race
White	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Black	0.55	(0.46–0.65)	<0.001	0.56	(0.47–0.67)	<0.001	0.57	(0.48–0.68)	<0.001	0.56	(0.47–0.66)	<0.001
Other	0.77	(0.58–1.02)	0.07	0.78	(0.59–1.04)	0.09	0.76	(0.57–1.00)	0.05	0.73	(0.55–0.97)	0.03
VA Region
Southeast	Ref.	-	-	Ref.	-	-	Ref.	-	-	Ref.	-	-
Continental	0.99	(0.77–1.28)	0.95	1.02	(0.79–1.31)	0.91	1.03	(0.80–1.32)	0.83	1.02	(0.80–1.32)	0.86
Midwest	1.50	(1.17–1.94)	0.002	1.48	(1.15–1.91)	0.002	1.50	(1.16–1.93)	0.002	1.47	(1.14–1.90)	0.003
North Atlantic	0.95	(0.75–1.20)	0.66	0.96	(0.76–1.22)	0.74	0.93	(0.74–1.18)	0.57	0.92	(0.73–1.16)	0.48
Pacific	1.18	(0.93–1.50)	0.17	1.16	(0.91–1.47)	0.23	1.17	(0.92–1.48)	0.20	1.13	(0.89–1.44)	0.30
Nadir or nearest count or ratio	2.32	(1.96–2.74)	<0.001	3.00	(2.31–3.89)	<0.001	2.47	(2.05–2.98)	<0.001	3.93	(3.33–4.63)	<0.001

Abbreviations: ASCC, anal squamous cell carcinoma; CD, cluster of differentiation; CI, confidence interval; HR, hazard ratio; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion; VA, Veterans Affairs; PWH, people with HIV.

DISCUSSION

This is the largest study of ASIL and ASCC in PWH in the U.S. over the longest timeframe. Our results provide a detailed analysis of unselected patients from an entire healthcare system; whereas most studies in the literature analyze patients who have self-selected into high-volume centers.^17,18 In 48,368 veterans over a 24-year period, we identified 16% of patients had at least one anal cytopathology or histopathology result collected during the study period. Reasons for the underutilization of pathologic evaluations for anal cancer screenings are multifactorial and not unique to the VA.^26–28 Many centers do not have the knowledge, skills, or resources to perform regular anal cancer screening tests.⁸ It is also possible that veterans with HIV had competing health priorities that did not make them good candidates for cancer screenings. We showed that when anal pathology was collected, 63% of cases yielded abnormal results and 45% of these abnormal results were either HSIL or ASCC.

Within this large cohort of PWH, we have shown that the 20-year disease-free survival for ASCC was 97.5%. This finding suggests that while PWH are at an elevated risk of developing ASCC, the overall risk of developing ASCC 20 years after HIV diagnosis is relatively low, even in the absence of standardized screening. On average, LSIL, HSIL, and ASCC were diagnosed 8.5, 9.1, and 9.7 years after a patient was diagnosed with HIV, respectively. Taken together, these findings reinforce the need for careful, long-term follow-up for evaluation of anal disease in this high-risk population and that screenings are more important after multiple years of HIV care rather than at the time of first HIV diagnosis. These diagnosis timelines on a large population can help guide us to focus testing for those with longer follow-up times in centers where screening resources are limited.

The results of this study demonstrate numerous risk factors for the development of ASIL and ASCC. History of STIs and smoking have strong correlations to the development of ASIL and ASCC, and history of anogenital condyloma has a strong correlation to the development of HSIL and ASCC. We utilized a time-varying covariate analysis, which yields larger hazard ratios when risk factors are more closely related chronologically to events and allows us to assess the impact of these risk predictors on disease status more accurately. These risk predictors have been previously described by other, smaller studies.^12,15,29 Validation in this large cohort is powerful to help clinicians determine higher risk patients for more vigilant anal cancer screenings. Interestingly, we identified variation in anal disease by race and geographic location but it is unclear if these differences are truly reflective of different disease status rates versus different screening rates among these groups.^30,31

The CD4/CD8 ratio is an objective marker that does not rely on patient recall of other health conditions like STIs, which might not be readily apparent in the medical record if diagnosed in the distant past or in an outside health system. Low CD4/CD8 ratios have are associated with a variety of health conditions including cancer in PWH.^9,12 In large cohorts of PWH from North America, Europe, and Australia, low CD4/CD8 ratios precede cancer diagnoses by at least 6–36 months, suggesting cancer itself does not cause a low ratio.^9,12 While we have previously demonstrated a strong correlation between low CD4/CD8 ratios and ASIL and ASCC in small populations of PWH, the CD4/CD8 ratio has not been evaluated as a risk factor for ASIL and ASCC on a large, national scale prior to this study.^11,15 Our results demonstrate a strong relationship between low CD4/CD8 ratios and severity of anal disease. A ratio <0.5 nearest to disease was the most predictive, suggesting the CD4/CD8 ratio is a clinically useful marker to risk stratify patients for screening and can help with shared decision making.³²

This study is limited in several ways. It is retrospective in nature, and as such it is not feasible to evaluate for important sexual risk factors like history of anoreceptive intercourse or number of sexual partners. Because this study focuses nearly exclusively on U.S. male patients, the results might not apply to non-male or non-U.S. populations. In this study, we evaluated anal specimens by analyzing both cytopathology and histopathology reports as part of a combined assessment. It is unclear what impact this might have on our results, given the known limitations in sensitivity and specificity of anal Pap smears.³³ Additionally, this combined analysis introduces a potential source of selection bias, as we did not distinguish between patients whose pathology was obtained as part of routine anal cancer screening versus those who were symptomatic enough to warrant a evaluation. Anal pathology results were likely not captured for patients who had anal evaluations performed outside the VA that were not documented with an ICD code within the VA.

In this national study of over 48,000 veterans with HIV, 16% had anal cytopathology or histopathology results available for analysis. On average, anal disease was diagnosed 8.5–9.7 years after HIV care was initiated. The CD4/CD8 ratio correlates strongly with severity of anal disease and can help clinicians identify veterans with HIV at highest risk for anal cancer to prioritize screening efforts.