Flow diagram of cutaneous malignancies among participants infected with the human immunodeficiency virus. BCC indicates basal cell carcinoma; KS, Kaposi sarcoma; MM, malignant melanoma; NHL, non-Hodgkin lymphoma; and SCC, squamous cell carcinoma.
Age-adjusted incidence rates of cutaneous malignancies before the highly active antiretroviral therapy (HAART) era (1986-1995) and during the HAART era (1996-2006). The rates for Kaposi sarcoma (KS) apply to men only; the rates for non–AIDS-defining cancers (NADC) apply to white/non-Hispanic men only. The rates are age-adjusted to the appropriate US 2000 standard population. BCC indicates basal cell carcinoma; MM, malignant melanoma; and SCC, squamous cell carcinoma.
Crum-Cianflone N, Hullsiek KH, Satter E, Marconi V, Weintrob A, Ganesan A, Barthel RV, Fraser S, Agan BK. Cutaneous Malignancies Among HIV-Infected Persons. Arch Intern Med. 2009;169(12):1130-1138. doi:10.1001/archinternmed.2009.104
As the life expectancy of persons infected with human immunodeficiency virus (HIV) increases, cancers have become an important cause of morbidity and mortality. Although cutaneous cancers are the most common malignant neoplasms in the general population, little data exist among HIV-positive persons, especially regarding the impact of HIV-specific factors.
We evaluated the incidence rates and factors associated with the development of cutaneous malignancies among HIV-infected persons by examining data that were prospectively collected from a large HIV study that included 4490 participants (1986-2006). Poisson regression and Cox proportional hazards models were performed.
Six percent of HIV-infected persons (n = 254) developed a cutaneous malignancy during 33 760 person-years of follow-up (mean, 7.5 years). Since the advent of highly active antiretroviral therapy (HAART), the incidence rates of cutaneous non–AIDS-defining cancers (NADCs), in particular basal cell carcinoma, have exceeded the rates of cutaneous AIDS-defining cancers such as Kaposi sarcoma. Factors associated with the development of cutaneous NADCs in the multivariate models included increasing age (hazard ratio [HR], 2.1; 95% confidence interval [CI], 1.7-2.6) and race. Compared with the white/non-Hispanic race, African Americans (HR, 0.03; 95% CI, 0.01-0.14) and other races (HR, 0.14; 95% CI, 0.03-0.57) had a lower risk of cutaneous NADCs. There were no significant associations between cutaneous NADCs and time-updated CD4 lymphocyte counts, HIV RNA levels, or receipt of HAART.
At present, the most common cutaneous malignancies among HIV-infected persons are NADCs. Cutaneous NADCs do not appear to be significantly associated with immune function or HAART but rather are related to traditional factors such as aging and skin color.
Cutaneous neoplasms are the most common malignancies in the United States, with 70% to 80% diagnosed as basal cell carcinoma (BCC), followed by squamous cell carcinoma (SCC) and malignant melanoma (MM).1,2 Risk factors include cumulative UV radiation exposure, race, advanced age, reduced DNA repair capability, and immunosuppression.3- 6 Multiple studies have demonstrated an increased risk of cutaneous malignancies in immunosuppressed populations. This risk is most pronounced in solid-organ transplant recipients, who have a 65 to 250 times increased risk as compared with the general population.4- 6 Patients with human immunodeficiency virus (HIV) infection also have an increased risk of some types of cancer.7- 9 Early in the HIV epidemic, Kaposi sarcoma (KS) was the most common malignancy in HIV-positive patients, with a rate that was more than 1000-fold greater than that of the general population.7,10 Because HIV-infected persons have an increased propensity to develop KS as well as non-Hodgkin lymphoma (NHL), these conditions were classified as AIDS-defining cancers (ADCs).11 Since the advent of highly active antiretroviral therapy (HAART), the rate of KS has decreased substantially.12
The incidence and risk factors associated with cutaneous non-ADCs (NADCs) among HIV-infected persons are less defined. Although some studies have shown a higher incidence and/or a more aggressive course of cutaneous NADCs among HIV-infected patients, the number of studies is limited and the results have been mixed.9,13- 17 Important risk factors for the development of skin cancer, regardless of HIV status, include fair skin, family history, and cumulative sun exposure.18,19 However, it is currently unknown whether the CD4 lymphocyte count, HIV RNA level, or receipt of HAART affects the development of cutaneous NADCs among HIV-infected populations. We evaluated a large cohort of HIV-infected persons over the course of the HIV epidemic (1986-2006) to provide data on incidence rates and HIV-specific factors associated with cutaneous malignancies among HIV-infected persons.
We retrospectively examined data collected as part of the US Military HIV Natural History Study, a multicenter, prospective, observational study that enrolled more than 4500 persons at 7 geographic locations in the United States (1986 to 2006). Participants were military beneficiaries who were evaluated on a biannual basis using standardized data collection procedures. All active-duty US military personnel were HIV negative on entry into the service and subsequently underwent HIV screening every 1 to 5 years. Among persons in the US Military HIV Natural History Study with a documented last known HIV-negative date (55%), the mean (SD) seroconversion window was 20.7 (17.0) months. Persons without a documented HIV-positive date, persons with a diagnosis of cancer before 1986, and persons with a diagnosis of cancer more than 6 months before the HIV diagnosis were excluded from the analyses, thereby yielding 4490 participants for our study. Our study was approved by central and local institutional review boards, and all participants provided informed consent.
Cutaneous malignancy cases were identified by database queries using the terms malignancy, cancer, carcinoma, and tumor,20 all of which had a location recorded as skin. Baseline was defined as the time of HIV seroconversion, conservatively estimated as 6 months before the first documented HIV-positive test result. Cutaneous malignancies were included only if they occurred after this point. Cancers were defined as an ADC in cases of cutaneous KS or NHL and as NADC for all other primary skin cancers. The study included only cutaneous SCCs and excluded mucosal SCCs of the mouth, genital, or anal regions.
Data collected included demographics; cutaneous cancer type, location, and date of diagnosis as well as subsequent cutaneous malignancies and their location and date when applicable; last seronegative- and first seropositive-HIV dates; CD4 lymphocyte counts and HIV RNA test results at baseline and sequentially over time; hepatitis B status at HIV diagnosis (defined by at least 2 of the following test results being concurrently positive: core antibody, surface antibody, or surface antigen tests); history of AIDS-defining conditions other than cancer; Walter Reed stage at entry into the US Military HIV Natural History Study, designated as 1 through 6 for ascending degrees of disease based on CD4 lymphocyte counts, opportunistic infections, lymphadenopathy, and delayed-type hypersensitivity21; and use of antiretroviral therapy. Highly active antiretroviral therapy was defined as 2 or more nucleoside reverse transcriptase inhibitors in combination with at least 1 protease inhibitor or 1 non–nucleoside reverse transcriptase inhibitor or an abacavir- or tenofovir-containing regimen of 3 or more nucleoside reverse transcriptase inhibitors.
For the participants with cutaneous cancer, the censoring date was the first cancer diagnosis date for the specific type of cancer considered. For those without cancer, the censoring date was the last study visit or the date of death. Follow-up for this report ended December 31, 2006. Participants were classified for stratification purposes as HIV positive in the pre-HAART era (before 1996) or in the HAART era (1996 or later).
Statistical analyses included descriptive statistics. Values were expressed as mean (SD). χ2 Tests were used to compare proportions; general linear models were used to compare continuous valued measurements. The number of events, person-years (PYs) at risk, and age-adjusted rates of events (per 100 000 PYs) were calculated for 5-year intervals, the pre-HAART era, and the HAART era. Rates for cutaneous ADCs were calculated for men only (because of the predominantly male study population and because most cases occurred among men) and age-adjusted with the US 2000 standard population for males, using the age at the midpoint of the interval. Rates for cutaneous NADC events were calculated for white/non-Hispanic men only (because almost all cases occurred among white men) and age-adjusted with the US 2000 standard population for white/non-Hispanic males. Rates were reported with 95% confidence intervals. Poisson regression analyses were used to test the hypothesis that the cancer rates remained constant over time.
Univariate and multivariate Cox proportional hazard models were used to evaluate the association of specific factors with the type of initial skin cancer. Baseline was the time of HIV diagnosis or January 1, 1986, whichever occurred later. For the CD4 lymphocyte count, HIV RNA level, receipt of HAART, and a history of a noncancer AIDS-defining event, we used both baseline and time-updated covariates, which included all available measurements from baseline through the event or censoring date. All models were stratified by HIV diagnosis era. Hazard ratios were reported with 95% confidence intervals. All analyses were conducted using SAS version 9.1 (SAS Institute Inc, Cary, North Carolina).
The study cohort consisted of 4490 HIV-infected participants who were followed up for a total of 33 760 PYs (mean [SD], 7.5 [4.9] years). The mean (SD) age at HIV diagnosis was 30 (7.9) years; 91% of the participants were male; and the self-reported race was 45% African American, 44% white/non-Hispanic, and 11% other. Of the participants with a documented Walter Reed stage classification at HIV diagnosis (n = 3888), 65% had stage 1 or 2 (early stage) disease. Of 2435 participants with available serologic data, 34% were seropositive for hepatitis B. A baseline CD4 lymphocyte count was available for 3483 participants, with a mean (SD) value of 539 (275) cells per microliter. Many participants were initially infected with HIV before HIV RNA testing was available. A baseline HIV RNA measurement was available for 1317 participants, with a mean (SD) of 4.2 (0.9) log10 copies per milliliter.
A total of 254 of the study cohort (5.7%) developed a cutaneous neoplasm during the study (Figure 1). Eleven participants developed 2 different cutaneous malignancies (5 that were concurrent and 6 that developed during follow-up), for a total of 265 cutaneous cancer events. In total, there were 193 cutaneous ADCs (192 cases of KS and 1 case of NHL) and 72 cutaneous NADCs (51 BCCs, 13 MMs, and 8 SCCs). All cancers occurred among men, except for 1 ADC that occurred in a woman. No Merkel cell carcinoma or other type of cutaneous malignancies were identified during the course of this study.
Age-adjusted rates of KS among men in the cohort varied over the course of the HIV epidemic, as shown in Table 1 and Figure 2. The highest rate was noted in 1991 to 1995, with an age-adjusted rate of 1797 cases per 100 000 PYs. The rates of KS significantly declined during the HAART era (after 1995, P < .001), with rates (per 100 000 PYs) of 205 in 1996 to 2000 and 134 in 2001 to 2006. The age-adjusted rates of cutaneous NADCs did not significantly change over time, but since the advent of HAART, the incidence rates of cutaneous NADCs have exceeded cutaneous ADCs (Table 1 and Figure 2). During 2001 to 2006, the age-adjusted rate of BCCs among white/non-Hispanic men exceeded that of KS (397 cases per 100 000 PYs vs 134 cases per 100 000 PYs, respectively). The rate of MMs (114 per 100 000 PYs) was similar to that of KS, while SCCs occurred at a lower rate (49 per 100 000 PYs).
The most common cutaneous NADC diagnosed was BCC, which occurred in 51 participants; the clinical characteristics are summarized in Table 2. Eight participants (16%) with BCC had more than 1 BCC lesion noted at initial presentation, all of which occurred in the same body location. During a mean (SD) follow-up of 7.8 (6.2) years after the diagnosis of the first BCC, 12 participants (24%) developed new BCC lesions (range, 1-5 lesions) at novel locations, for a total of 83 BCC events. Also, four patients (8%) with an initial BCC subsequently developed a second type of cutaneous malignancy: 3 developed KS and 1 developed MM.
Malignant melanoma was diagnosed in 13 participants; 3 also had BCCs (2 concurrently and 1 subsequent to the diagnosis of MM). Table 2 summarizes the characteristics of participants with MM. During a follow-up of 4.9 (4.3) years after diagnosis of the first MM, 1 participant (8%) developed an additional MM in a different location and 1 developed a BCC. Eight participants had an SCC of the skin (Table 2), 2 also had a BCC, and 1 developed the SCC after having KS. During a follow-up of 2.1 (1.5) years after diagnosis of the first SCC, no patient developed a subsequent cutaneous malignancy.
In summary, of the 65 participants followed up for 5.1 (4.3) years and diagnosed as having a cutaneous NADC, 13 (20%) developed a second cutaneous malignancy of the same type and 5 (8%) developed a second cutaneous malignancy of a novel type. The mean time between the first and second cancers was 3.9 years for those with the same cancer type and 3.7 years for those with a different cancer type. Twenty-two HIV-infected persons (34%) with a cutaneous NADC died during follow-up; however, none of the deaths was attributed to the cutaneous cancer.
We compared the characteristics of participants whose first cancer event was a cutaneous NADC with those whose first cancer event was a cutaneous ADC (Table 3). The 1 participant who was diagnosed with a concurrent cutaneous ADC and a NADC was included in the NADC group. Participants with a cutaneous NADC were older, more likely to be white/non-Hispanic, had a longer mean duration of HIV infection before cancer diagnosis, had a higher CD4 lymphocyte count and a lower HIV RNA level at the time of cancer diagnosis, and were more likely to be receiving HAART than those with a cutaneous ADC. Participants with a cutaneous ADC were more likely to be hepatitis B seropositive than those with a cutaneous NADC.
Using Cox proportional hazard models, we assessed factors associated with cutaneous NADC events (Table 4). Univariate factors associated with time to first cutaneous NADC included increasing age, white/non-Hispanic race, and prior noncancer AIDS-defining event. All cases occurred among men. In the univariate models, there were no significant associations between cutaneous NADC occurrence and Walter Reed stage at entry into the cohort, CD4 lymphocyte count, HIV RNA level, or receipt of HAART. In the final multivariate model, factors associated with cutaneous NADCs included older age and white/non-Hispanic race.
Cutaneous NADCs are now the most frequent cause of cutaneous malignancies among HIV-infected persons. While KS was the most common cutaneous malignancy among HIV-infected persons before the introduction of HAART, BCC has now superseded it as the most frequent cutaneous malignancy. In our study, cutaneous NADCs were not associated with immune status or receipt of HAART but were mostly related to traditional risk factors such as aging and skin color.
Currently, cutaneous malignancies are the most frequently diagnosed cancers in the US general population.1,2 Likewise, cutaneous malignancies likely account for the majority of cancers among HIV-infected persons.20 However, to date, most studies evaluating malignancies among HIV-infected persons have used registry data, which often lack information regarding cutaneous NADCs.22 We evaluated 4490 HIV-infected persons over a 20-year period and found that 6% of the cohort developed a cutaneous malignancy during a mean follow-up of 7.5 years.
In this study, the incidence of cutaneous ADCs significantly decreased after 1995, similar to the findings of other studies and coinciding with the introduction of HAART.12,23,24 On the other hand, the age-adjusted incidence rates of cutaneous NADCs remained stable. As HIV-infected persons are experiencing increasing life expectancies and a reduction of ADCs,25 NADCs have become the most common cutaneous malignancies in this population.
The factors associated with the development of cutaneous NADCs in our study of HIV-infected persons included increased age and white/non-Hispanic race, similar to that seen in other HIV-positive cohorts and in the general population.18,19 On the other hand, CD4 lymphocyte counts were not related to the development of cutaneous NADCs and were relatively high (mean CD4 lymphocyte count, 439/μL) at the time of cancer diagnosis. Other studies have shown that the depth of tumor spread and outcome are typically not associated with the CD4 lymphocyte count.17- 19,26,27 The lack of an association of cutaneous NADC with immune status in HIV-infected patients differs from that in immunosuppressed transplant recipients, whose risk of skin cancer correlates with lower CD4 lymphocyte counts.28
Although a recent study demonstrated similarities in the types of cancers seen among persons with HIV and transplant recipients,29 there are differences among the risk factors and clinical characteristics of cutaneous cancers between these populations. For example, 30% to 70% of solid-organ transplant recipients develop a cutaneous malignancy, whereas only 6% of HIV-infected persons developed either a cutaneous ADC or a cutaneous NADC in our study.30,31 These differences may be attributed to the type or degree of immunosuppression in transplant recipients compared with patients with HIV.30,32 Many transplant recipients receive cyclosporine and/or azathioprine, which not only have immunosuppressive effects that reduce tumor surveillance and increase the risk of acquiring human papillomavirus infection and perhaps other oncogenic viruses but also display intrinsic carcinogenic and photosensitizing effects that may contribute to the development of cutaneous malignancies.6,28
The present study failed to show a protective benefit of HAART against the development of cutaneous NADCs. A recent analysis of the Strategies for Management of Antiretroviral Therapy trial examined the impact of continuous HAART compared with drug conservation on cutaneous NADCs and found a similar rate for both arms (220 per 100 000 PYs; hazard ratio, 1.1; P = .98).33 Because interrupting HAART is associated with rapid CD4 lymphocyte count declines and elevated inflammatory markers, which may affect immune surveillance and cancer incidence, the fact that continuing HAART did not alter cutaneous NADC incidence rates lends further support to the finding that HAART may not reduce rates of cutaneous NADCs. The lack of benefit of HAART in our study may also indicate that any positive effects of HAART may be offset by other factors such as increasing life expectancies and aging, which might increase cancer risk. Although, to our knowledge, HAART has shown no protective effect in the development or prognosis of cutaneous NADCs to date, isolated case reports have shown that BCC and Merkel cell carcinoma may regress with antiretroviral therapy.34,35
It has been hypothesized that the chronic immunosuppressed state seen in HIV-infected persons, which may not be entirely reflected by the CD4 lymphocyte count or improved by HAART, may cause an overall immune dysregulation and lead to cutaneous neoplasms. Other postulated pathogenic mechanisms regarding the risk of cutaneous cancers among patients with HIV include the role of potentially oncogenic HIV proteins (eg, tat and nef), the high incidence of human papillomavirus in this population and its potential role in the pathogenesis of SCC, and/or elevated interleukin levels that result in keratinocyte and melanocyte proliferation.36- 40
It has been suggested that persons infected with HIV have higher rates of some types of cutaneous malignancies, but comparing rates among patients with HIV and the general population is challenging because cancer registries such as the National Cancer Institute Surveillance Epidemiology and End Results Cancer Statistics Review do not provide data for BCC and SCC. Furthermore, age-adjusted rates are only comparable if the same standard population is used. For KS and MM, rates for the general population (age-adjusted to the US 2000 standard population) are available. The rates of KS in the general population before the HIV epidemic (1975-1979) was 4 cases per 100 000 PYs41 compared with our rates of 134 per 100 000 PYs (2001-2006) to 1797 (1991-1995), representing a 34- to 449-fold increased risk among patients with HIV. For MM, the rate in the general population during 2001 to 2005 was 29 per 100 000 PYs41 compared with our rate of 114, suggesting a 3.9-fold higher rate. For rates of BCC and SCC, we recalculated age-adjusted rates using the 1970 US standard population for white males, as there are data in the literature that are standardized to this population.42 The rates for BCC were similar for HIV-infected persons (323, 537, 266, and 354 per 100 000 PYs from the earliest to the most recent 4 periods) compared with those of the general population (360 per 100 000 PYs during 1977-1980).42 Comparison of rates for SCC showed that the recalculated rates for SCC for our 4 periods were 0, 68, 17, and 39 compared with 101 for the general population; however, these data are limited by the small number of SCCs in our cohort.
The current study demonstrated that HIV-infected persons presenting with a cutaneous NADC have a high likelihood of developing subsequent cutaneous malignancies at novel sites. In our study, 24% of the participants who initially presented with a BCC developed a subsequent BCC, and 8% developed a second type of cutaneous cancer. Furthermore, other studies have shown that patients with HIV have high rates of recurrences.16,18,19,43,44 These data suggest that HIV-infected persons with an initial cutaneous NADC should be carefully followed up for both recurrent disease and the development of novel cutaneous malignancies. Similar to other immunosuppressed populations, patients with HIV develop cutaneous malignancies at a younger age than the general population.4,6,16,26,31 In our study, the mean age at the diagnosis of a cutaneous NADC was 43 years, as opposed to a mean age of 70 years in the general population.26
The most frequent cutaneous NADC in our study cohort was BCC, with a 6:1 ratio of BCC to SCC, which concurs with previous studies in HIV-infected populations and parallels the general population but differs from transplant recipients, who disproportionally develop SCC.5,14,16,19,31 In contrast to previous studies that found the trunk to be the most common location of BCC in patients with HIV, the majority of BCC cases in our study occurred in the head and neck region, comparable to the general population.16,19 The current study differed from prior investigations as it included data from the HAART era. The most common locations of SCC (head/neck) and MM (trunk) were similar to those seen in the general population.
As in any study, the current study had limitations. First, 91% of the population consisted of men; hence, the current results may not be generalizable to HIV-infected women. The majority of our HIV cases were diagnosed early during infection and had high initial CD4 lymphocyte counts; therefore, our rates of cancer could be lower than rates for HIV cases that were diagnosed at more advanced stages of disease, especially for ADCs. Our study does provide important data, however, given the recent emphasis on early diagnosis of HIV. Data on an individual's cumulative lifetime of sun exposure were not available, nor could the participants be segregated by geographic location, as military personnel have frequent geographic relocations. These data are important because cumulative UV exposure plays an important role in skin cancer development by altering p53 tumor suppressor genes and reducing the number of Langerhans cells, which are important for immune surveillance.1,3 Other limitations include incomplete data regarding the subtype and aggressiveness of the malignancies, as well as the method of treatment. Because of the observational nature of the study, the ascertainment of cutaneous cancers may have varied over time in our cohort, which may have affected incidence rates. Finally, no information was available regarding human papillomavirus coinfection.36- 38
In conclusion, since the advent of HAART, NADCs are now the most common cutaneous malignancy diagnosed among HIV-infected persons. The present study demonstrates that the development of cutaneous NADCs is associated with increased age and skin color, as in the general population, but is independent of CD4 lymphocyte count or receipt of HAART. These results suggest that regardless of the CD4 lymphocyte level, all HIV-infected persons should be counseled on UV avoidance and that physicians should be aware of the risk of cutaneous malignancies in this population.
Correspondence: Nancy Crum-Cianflone, MD, MPH, c/o Clinical Investigation Department (KCA), Naval Medical Center San Diego, 34800 Bob Wilson Dr, Ste 5, San Diego, CA 92134-1005 (firstname.lastname@example.org).
Accepted for Publication: December 12, 2008.
Author Contributions:Study concept and design: Crum-Cianflone. Acquisition of data: Crum-Cianflone, Marconi, Weintrob, Ganesan, and Agan. Analysis and interpretation of data: Crum-Cianflone, Hullsiek, Satter, Weintrob, Barthel, Fraser, and Agan. Drafting of the manuscript: Crum-Cianflone, Hullsiek, Satter, and Fraser. Critical revision of the manuscript for important intellectual content: Crum-Cianflone, Hullsiek, Satter, Marconi, Weintrob, Ganesan, Barthel, Fraser, and Agan. Statistical analysis: Hullsiek. Obtained funding: Crum-Cianflone and Agan. Administrative, technical, and material support: Crum-Cianflone, Satter, Marconi, Weintrob, Ganesan, Barthel, and Agan. Study supervision: Crum-Cianflone, Ganesan, Barthel, and Agan.
Financial Disclosure: None reported.
Funding/Support: Support for this work was provided by the Infectious Disease Clinical Research Program (IDCRP), Uniformed Services University of the Health Sciences (USUHS), Bethesda, Maryland, of which the TriService AIDS Clinical Consortium (TACC) is a component. The IDCRP is a Department of Defense Tri-service program that is executed through USUHS and the Henry M. Jackson Foundation for the Advancement of Military Medicine in collaboration with the US Department of Health and Human Services, National Institutes of Health, National Institute of Allergy and Infectious Diseases, and Division of Clinical Research through Interagency Agreement HU0001-05-2-0011.
Disclaimer: The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Departments of the Army, Navy, or Air Force or the Department of Defense.
Additional Contributions: Waine MacAllister provided editorial support, and Edith Lederman, MD, provided a critical review of the manuscript.
This article was corrected online for typographical errors on 6/22/2009.