Association of Multiple Primary Skin Cancers With Human Immunodeficiency Virus Infection, CD4 Count, and Viral Load

Key Points

Question

Among individuals with human immunodeficiency virus (HIV), how do biologic markers of immune function (CD4 count and viral load) correlate with risk of subsequent nonmelanoma skin cancer (NMSC)?

Findings

This cohort study analyzed 455 HIV-infected and 1945 HIV-uninfected age- and sex-matched non-Hispanic white participants diagnosed with at least 1 NMSC for subsequent NMSC development. Risk of subsequent squamous cell carcinoma (SCC), but not basal cell carcinoma, was associated with lower CD4 counts and higher viral load.

Meaning

Increased patient and clinician awareness of risk of multiple SCCs associated with low CD4 counts or high viral loads among HIV-infected individuals may be warranted.

Abstract

Importance

Persons with human immunodeficiency virus (HIV) have a 2.8-fold higher risk than HIV-uninfected persons of nonmelanoma skin cancer (NMSC), defined as basal cell carcinoma (BCC) and squamous cell carcinoma (SCC). Individuals with a prior NMSC history are at increased risk for subsequent NMSC, but the magnitude of risk and its relation to HIV disease-related factors, including CD4 count and viral load (VL), are unknown.

Objective

To better understand how laboratory markers currently used to evaluate HIV disease progression may be associated with subsequent NMSC risk.

Design, Setting, and Participants

This cohort study analyzed 455 HIV-infected and 1945 HIV-uninfected patients, all of them members of the Kaiser Permanente Northern California (KPNC) health care plan, diagnosed with at least 1 NMSC from 1996-2008 to determine risk of subsequent NMSCs in relation to CD4 count and VL. All participants were white, non-Hispanic persons 18 years or older who had had at least 1 NMSC during the 1996-2008 period. Participants entered the cohort at their first NMSC diagnosis and were observed through 2008. Incidence rates were calculated and adjusted hazard ratios were estimated using extended Cox regression models with recent CD4 count and VL analyzed as time-changing covariates.

Main Outcomes and Measures

Measured CD4 count, VL, and subsequent NMSC (BCC and SCC).

Results

The cohort comprised 455 HIV-infected participants (13 [3%] women) and 1952 HIV-uninfected participants (154 [8%] women). Median duration of observation was 4.6 years, and 16.5% (n = 390) either died (n = 35) or lost KPNC membership status (n = 355) without having a subsequent primary NMSC. Compared with HIV-uninfected persons, HIV-infected individuals were slightly younger (mean age, 52.5 vs 55.5 years), more likely men (97% vs 92%), more likely to have smoked (57% vs 45%), and less likely to be overweight/obese (50% vs 61%). The small observed differences by HIV status in matching characteristics (ie, age and sex) resulted from the restriction of the original cohort to those with at least 1 NMSC. Compared with uninfected individuals, those with HIV infection with a recent biomarker of more severe immune deficiency (CD4 count <200 cells/mL) had a 44% increased risk of subsequent NMSC overall and a 222% increase risk of SCC in particular, suggesting that subsequent SCC risk is associated with immune dysfunction.

Conclusions and Relevance

HIV-infected persons compared with HIV-uninfected persons were are at higher risk for subsequent new SCC but not BCC, with a dose-response relationship between risk and lower CD4 counts and higher VLs. Subsequent new primary SCCs had a strong association with lower CD4 and higher VL among HIV-infected persons, suggesting that immune dysfunction might contribute to increased SCC risk. Clinical implications include targeted monitoring for SCC among HIV-infected individuals, particularly those with low CD4 counts or high VLs.

This cohort study of patients with at least 1 prior nonmelanoma skin cancer (NMSC) evaluates the association of subsequent NMSCs with HIV infection, CD4 count, and viral load.

Introduction

Nonmelanoma skin cancer (NMSC), defined as basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), is the most common malignant condition in the United States, affecting over 2 million Americans per year, exceeding the incidence of all other cancers combined. Among immunocompetent individuals, BCCs are more common than SCCs. However, among immunocompromised persons, including organ transplant recipients and persons with immune dysfunction, SCCs predominate over BCCs. Our research group previously demonstrated a 2.8-fold higher risk of first NMSCs in persons infected with human immunodeficiency virus (HIV) vs HIV-uninfected individuals, largely driven by the influence of immunosuppression on SCC risk. Poor control of HIV disease may suggest innate and cell-mediated immune dysfunction, which could modify risk of multiple primary SCCs, although this has not been evaluated. We investigated the risk of multiple primary NMSCs among HIV-infected individuals in relation to CD4 count and HIV viral load (VL) using a retrospective cohort of white, non-Hispanic patients 18 years or older who had had at least 1 NMSC during the 1996-2008 period and were members of the Kaiser Permanente Northern California (KPNC) health care plan (hereinafter “members”).

Methods

Our group has described previously the identification of a cohort of adult HIV-infected and matched HIV-uninfected individuals within KPNC, an integrated health system caring for over 3.4 million Californians. Briefly, we identified all adult HIV-positive non-Hispanic white members (n = 6567) and matched them with HIV-negative members (n = 36 887) for age, sex, and race. The HIV-infected individuals were identified from an HIV registry that includes all known cases among KPNC members since 1980. Confirmation of HIV infection was determined by medical chart review and comparisons of case lists with KPNC HIV clinics.

The present study was approved by the KPNC institutional review board, including a waiver of patient written informed consent. For purposes of the present study, we restricted the previously described cohort to participants who had at least 1 NMSC. The index date was defined as the biopsy date of the first NMSC (BCC or SCC) after study entry. Participants were observed until a subsequent NMSC outcome, termination of health-plan membership, death, or December 31, 2008. The NMSC outcomes were defined as nonrecurrent, nongenital, and nonoral BCCs or SCCs. We identified potential NMSCs by electronically searching pathology records for all KPNC specimens during the 1996-2008 period, as previously described, and only included cases confirmed after review by a dermatologist (M.M.A.). We used other KPNC databases to obtain laboratory, sociodemographic, clinical, and behavioral data (Table 1), including laboratory test results (CD4 and VL), patient demographics (age, sex), health plan enrollment periods, and clinical diagnoses, including overweight or obesity diagnoses (International Classification of Diseases, Ninth Revision [ICD-9], codes 278, 259.9, and V85; and internal weight/height codes) and tobacco use (ICD-9 codes 305.1, V15, V65, and 649; and internal social history codes).

Table 1. Characteristics of the Study Cohort^a.

Characteristic	HIV-Infected (n = 455)b	HIV-Uninfected (n = 1952)
Sex
Female	13 (3)	154 (8)
Male	442 (97)	1798 (92)
Age, mean (median), y	52.6 (51.9)	55.5 (55.0)
Age, categorical, y
18-49	184 (40)	575 (29)
50-64	224 (49)	1032 (53)
≥65	47 (10)	345 (18
Ever smoker
Yes	260 (57)	887 (45)
No	195 (43)	1065 (55)
Ever overweight/obesec
Yes	228 (50)	1186 (61)
No	227 (50)	766 (39)
Duration of follow-up, mean (median), y	5.2 (4.8)	5.1 (4.6)
New NMSCs, mean No.	1.00	0.73
New NMSCs, categorical
None	276 (61)	1276 (65)
1	98 (22)	380 (19)
2	33 (7)	144 (7)
3-4	26 (6)	99 (5)
≥5	22 (5)	53 (3)

Abbreviations: HIV, human immunodeficiency virus; NMSC, nonmelanoma skin cancer.

^aUnless otherwise indicated, data are reported as number (percentage) of participants.

^bAmong HIV-infected individuals, 358 (79%) were men who had sex with men, 26 (6%) were injection drug users, 26 (6%) were heterosexual, and 45 (10%) had unknown HIV risk factors.

^cDefined as having had at least 1 of the following International Classification of Diseases, Ninth Revision (ICD-9) codes: 278, 259.9, and/or V85.

We computed incidence rates per 1000 person-years for subsequent NMSC, BCC, and SCC by HIV status, and by categories defined by recent CD4 counts and HIV VL. We estimated unadjusted and adjusted hazard ratios (HRs) using extended Cox regression models for multiple events. We used the conditional model approach, which assumes that incurring risk of a subsequent event can occur only after experiencing a previous event, with time since index date as the time scale. We used robust sandwich variance estimators to account for correlation of multiple events for the same person. We adjusted models for the following variables: baseline sex, age at index, obesity, and smoking status. Our reference group was HIV-uninfected individuals. To test for statistically significant trends in the relationships between the outcomes and CD4 count and VL, we ran separate Cox models in which we treated the CD4 and VL categories as ordered numeric variables (Table 2). All analyses were performed with SAS software (version 9.3; SAS Institute Inc).

Table 2. Risk of Subsequent Primary NMSC, BCC, and SCC by HIV and Biomarker Status^a.

Model	NMSC				BCC				SCC
	No.	Rateb	uHR (95% CI)	aHR (95% CI)c	No.	Rateb	uHR (95% CI)	aHR (95% CI)c	No.	Rateb	uHR (95% CI)	aHR (95% CI)c
Overall
HIV−	1418	143	1 [Reference]	1 [Reference]	1134	114	1 [Reference]	1 [Reference]	284	29	1 [Reference]	1 [Reference]
HIV+	456	193	1.14 (1.01-1.29)	1.15 (1.02-1.31)	339	143	1.12 (0.98-1.28)	1.13 (0.98-1.29)	117	49	1.18 (0.91-1.53)	1.23 (0.95-1.61)
CD4,d cells/µL
HIV−	1418	143	1 [Reference]	1 [Reference]	1134	114	1 [Reference]	1 [Reference]	284	29	1 [Reference]	1 [Reference]
≥500	130	155	1.04 (0.86-1.26)	1.06 (0.87-1.29)	111	132	1.05 (0.85-1.29)	1.06 (0.86-1.31)	19	23	0.84 (0.52-1.36)	0.89 (0.55-1.44)
200 to <500	237	193	1.13 (0.97-1.32)	1.14 (0.97-1.33)	180	147	1.16 (0.98-1.37)	1.16 (0.98-1.37)	57	46	1.08 (0.77-1.52)	1.13 (0.79-1.60)
<200	88	294	1.44 (1.10-1.88)	1.44 (1.10-1.88)	48	160	1.19 (0.87-1.63)	1.19 (0.86-1.64)	40	134	2.15 (1.39-3.35)	2.22 (1.42-3.48)
P for trend	NA	NA	NA	.25	NA	NA	NA	.75	NA	NA	NA	.04
VL,d copies/mL
HIV−	1418	143	1 [Reference]	1 [Reference]	1134	114	1 [Reference]	1 [Reference]	284	29	1 [Reference]	1 [Reference]
<500	308	179	1.09 (0.95-1.26)	1.10 (0.95-1.26)	233	135	1.10 (0.94-1.28)	1.10 (0.94-1.28)	75	44	0.99 (0.72-1.35)	1.01 (0.73-1.39)
500 to <10 000	57	183	1.17 (0.88-1.54)	1.21 (0.91-1.60)	39	125	1.04 (0.75-1.45)	1.08 (0.77-1.50)	18	58	1.39 (0.81-2.39)	1.53 (0.88-2.67)
≥10 000	83	265	1.27 (0.98-1.65)	1.31 (1.00-1.72)	60	191	1.20 (0.90-1.60)	1.21 (0.91-1.62)	23	73	1.97 (1.21-3.20)	2.28 (1.39-3.75)
P for trend	NA	NA	NA	.66	NA	NA	NA	.70	NA	NA	NA	.01

Abbreviations: aHR, adjusted hazard ratio; BCC, basal cell carcinoma; HIV, human immunodeficiency virus; NMSC, nonmelanoma skin cancer; SCC, squamous cell carcinoma; uHR, unadjusted hazard ratio; VL, viral load.

^aPatients could have multiple subsequent primary NMSCs during follow-up.

^bIncidence rate is per 1000 person-years; persons could contribute multiple events to the numerator in the calculation of the incidence rate.

^cAdjusted for age, sex, smoking status, and obesity.

^dHIV+ patients could contribute at-risk person-time and events to more than 1 category. The models exclude HIV+ persons without any measure of CD4 count or VL (1 outcome was associated with persons missing CD4 count, and 8 outcomes were associated with persons missing VL measures).

Results

The cohort comprised 455 HIV-infected and 1952 HIV-uninfected persons. Median duration of observation was 4.6 years, and 16.5% (n = 390) either died (n = 35) or lost membership status (n = 355) without having a subsequent primary NMSC. Cohort characteristics are summarized in Table 1. Compared with HIV-uninfected persons, HIV-infected individuals were slightly younger (mean age, 52.5 vs 55.5 years), more likely to be men (97% vs 92%), more likely to have smoked (57% vs 45%), and less likely to be overweight or obese (50% vs 61%). The small observed differences by HIV status in matching characteristics (ie, age and sex) resulted from the restriction of the original cohort to individuals with at least 1 NMSC.

Subsequent incidence rates, and unadjusted HRs and adjusted HRs (aHRs) for subsequent NMSC, BCC, and SCC are listed in Table 2. Overall, HRs and aHRs did not differ substantially in any analyses. The CD4 counts and VLs were not completely collinear: during the approximately 300 patient-years of observation during which patients had CD4 counts lower than 200 cells/µL, they had VLs greater than 10 000 cells/µL for only 85 of these patient-years. The association between NMSC and HIV status was more pronounced for those with lower CD4 counts (aHR 1.44; 95% CI, 1.10-1.88, comparing persons with CD4 count <200 cells/µL vs uninfected individuals) and higher VLs (aHR 1.31; 95% CI, 1.00-1.72, comparing persons with VL >10 000 copies/mL vs uninfected individuals). Comparing HIV-infected with HIV-uninfected persons, we found a trend of higher aHRs with lower recent CD4 counts and with higher viral loads for SCCs. The aHR for subsequent primary SCC for a CD4 count below 200 cells/µL was 2.22 (95% CI, 1.42-3.48; P = .04) compared with uninfected individuals. Similarly, the aHR for subsequent primary SCC for a VL greater than 10 000 copies/mL was 2.28 (95% CI, 1.39-3.75; P = .01). For BCCs, although the point estimates for the HR were higher for lower CD4 and higher VL categories, the differences were not significant. Additional models in which HIV-infected individuals with CD4 counts of 500 cells/µL or higher and VL below 500 copies/mL were the reference group did not substantially alter the point estimates (data not show), suggesting that risk of subsequent primary NMSC is very similar among HIV-infected individuals with CD4 counts of 500 cells/µL or higher and VL below 500 copies/mL and HIV-uninfected individuals.

Discussion

Among non-Hispanic white individuals with HIV infection and a history of skin cancer, we found a 15% increased risk of subsequent NMSC over an average follow-up period of about 5 years compared with HIV-uninfected individuals receiving care in the same integrated health care delivery system. Compared with uninfected individuals, those with HIV infection and a recent biomarker of more severe immune deficiency (CD4 count <200 cells/mL) had a 44% increased risk of subsequent NMSC overall and a 222% increased risk of SCC in particular. Our findings suggest that HIV infection and its associated immunodeficiency contribute to the higher risk of NMSC overall and SCC in particular. The HIV VL, which is often associated with antiretroviral treatment adherence, was associated with subsequent primary SCC (aHR for a VL >10 000 copies/mL, 2.28; 95% CI, 1.39-3.75; P = .01) but not for BCC.

This study addresses key questions regarding subsequent NMSC risk among a high-risk subgroup of HIV-infected population who, by virtue of having had a pathologically validated skin cancer, is at increased risk of subsequent NMSCs. Specifically, it was previously not known precisely which NMSC subtype is increased in high-risk persons with HIV, and whether biomarkers of HIV infections, such as degree of immune dysfunction, are associated with subsequent skin cancer risk. Prior studies examining NMSC risk have not included HIV-uninfected control groups, were unable to differentiate among BCC and SCC (they share the same ICD-9 code), and did not examine biomarkers such as CD4 count and VL as time-varying covariates that could affect NMSC risk.

Strengths of the present study include a large cohort with confirmed HIV status; inclusion of time-varying laboratory data capturing immunologic and virology markers; and availability of all pathology reports at KPNC to capture NMSC, a cancer that is otherwise difficult to study because it is not generally reported to cancer registries.

Limitations

Given that this study was based on electronic medical record data, a limitation is lack of measurement of known skin cancer risk factors, such as skin type and sun exposure history. However, limiting the cohort to non-Hispanic white individuals with at least 1 NMSC selected for individuals with fair-skin phenotype and some sun-exposure history, thereby reducing the impact of potential confounding. Additional limitations include potential lack of generalizability to groups besides non-Hispanic whites, although other groups have markedly lower NMSC risk.

It is also possible that some individuals in the study sought care outside of health plan facilities and that some variables or outcomes of interest were not captured in the electronic medical record. However, since cohort members already had at least 1 NMSC treated at health plan facilities, it is unlikely that subsequent NMSC would be treated outside the health plan. Also, given that KPNC is a prepaid comprehensive health care system and members would have had to pay out of pocket for services received outside the health plan, seeking care outside the plan seems unlikely.

Finally, although our data may have been underpowered to detect a potential association between HIV biomarkers and subsequent NMSCs, the confidence intervals surrounding our HRs are not wide, suggesting that our findings were sufficiently powered. Our findings would benefit from replication in a different study population with more female representation and a larger sample size.

Conclusions

In conclusion, among non-Hispanic whites with a pathologically confirmed history of NMSC, HIV-infected persons are at higher risk for subsequent primary NMSC compared with HIV-uninfected persons. There was a dose-response relationship between increased NMSC risk and lower CD4 counts and higher VLs, suggesting that NMSC risk was associated with biomarkers of immunodeficiency in HIV-infected patients. Compared with BCC risk, subsequent new primary SCC risk had a strong association with lower CD4 and higher VL among HIV-infected persons, suggesting that immune function might contribute to increased SCC risk. Previous studies have shown that HIV-infected persons have higher rates of SCC recurrence after treatment, even among younger individuals with well-controlled HIV. Our findings support immune dysfunction as a risk factor for SCCs and dovetail with SCC risk data from iatrogenic immunodeficiency states such as organ transplantatin, and autoimmune diseases. Future studies are needed to validate our findings.

Our study results provide a more precise risk estimate for subsequent primary NMSC among non-Hispanic white HIV-infected patients with a prior NMSC and could have implications for screening for the broader HIV-infected community. Results could inform ongoing skin cancer screening practices, which do not currently consider HIV infection status or biomarkers such as CD4 count and VL. Clinical implications of our findings suggest a potential benefit from targeted monitoring for SCC among HIV-infected individuals, particularly those with low CD4 counts or high VLs.