Incidence and risk factors for the occurrence of non-AIDS-defining cancers among human immunodeficiency virus-infected individuals

Authors

  • Alina Burgi M.P.H.,

    1. Graduate School of Public Health, San Diego State University, San Diego, California
    Search for more papers by this author
  • Stephanie Brodine M.D.,

    1. Graduate School of Public Health, San Diego State University, San Diego, California
    Search for more papers by this author
  • Scott Wegner M.D.,

    1. United States Military HIV Research Program, Rockville, Maryland
    2. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    Search for more papers by this author
  • Mark Milazzo,

    1. United States Military HIV Research Program, Rockville, Maryland
    2. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    Search for more papers by this author
  • Mark R. Wallace M.D.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Naval Medical Center, San Diego, California
    Search for more papers by this author
  • Katherine Spooner M.D.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Walter Reed Army Medical Center, Washington, DC
    Search for more papers by this author
  • David L. Blazes M.D., M.P.H.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Brian K. Agan M.D.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Wilford Hall United States Air Force Medical Center, Lackland Air Force Base, Texas
    Search for more papers by this author
  • Adam Armstrong D.O.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Naval Medical Center Portsmouth, Portsmouth, Virginia
    Search for more papers by this author
  • Susan Fraser M.D.,

    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Tripler Army Medical Center, Honolulu, Hawaii
    Search for more papers by this author
  • Nancy F. Crum M.D., M.P.H.

    Corresponding author
    1. Tri-Service AIDS Clinical Consortium, Rockville, Maryland
    2. Infectious Diseases Division, Naval Medical Center, San Diego, California
    • c/o Clinical Investigation Department (KCA), Naval Medical Center San Diego, 34800 Bob Wilson Drive, Suite 5, San Diego, CA 92134-1005
    Search for more papers by this author
    • Fax: (619) 532-8137


  • This is a U.S. government work and, as such, is in the public domain in the United States of America.

  • Presented in condensed form as abstract 886 at the 42nd Annual Meeting of the Infectious Diseases Society of America, Boston, Massachusetts, September 30–October 3, 2004.

Abstract

BACKGROUND

The objective of this study was to determine the rates and predictors of non-AIDS-defining cancers (NADCs) among a cohort of human immunodeficiency virus (HIV)-infected individuals.

METHODS

The authors conducted a retrospective study of 4144 HIV-infected individuals who had 26,916 person-years of follow-up and who had open access to medical care at 1 of the United States military HIV clinics during the years 1988–2003. Cancer incidence rates were race specific and were adjusted for age; these were compared with national rates using logistic regression to assess predictors of NADC development.

RESULTS

One hundred thirty-three NADCs were diagnosed with a rate of 980 diagnoses per 100,000 person-years. The most frequent NADCs were skin carcinomas (basal cell and squamous cell), Hodgkin disease, and anal carcinoma. The results showed that there were higher rates of melanoma, basal and squamous cell skin carcinomas, anal carcinoma, prostate carcinoma, and Hodgkin disease among the HIV-infected cohort compared with age-adjusted rates for the general United States population. Predictors of NADCs included age older than 40 years (odds ratio [OR], 12.2; P < 0.001), Caucasian/non-Hispanic race (OR, 2.1; P < 0.001), longer duration of HIV infection (OR, 1.2; P < 0.001), and a history of opportunistic infection (OR, 2.5; P < 0.001). The use of highly active antiretroviral therapy (HAART) was associated with lower rates of NADCs (OR, 0.21; P < 0.001). A low CD4 nadir or CD4 count at diagnosis (< 200 cells/mL) was not predictive of NADCs.

CONCLUSIONS

The most frequent NADCs were primary skin malignancies. Melanoma, basal and squamous cell skin carcinomas, anal carcinoma, prostate carcinoma, and Hodgkin disease occurred at higher rates among HIV-infected individuals. The implementation of screening programs for these malignancies should be considered. Most risk factors for the development of NADCs are nonmodifiable; however, the use of HAART appeared to be beneficial in protecting against the development of malignant disease. Cancer 2005. Published 2005 by the American Cancer Society.

The incidence of malignancies is increased among individuals who are infected with human immunodeficiency virus (HIV), and an estimated 40% of all patients with acquired immunodeficiency syndrome (AIDS) develop cancer during the course of their disease.1, 2 There are three AIDS-defining malignancies, which include Kaposi sarcoma, non-Hodgkin lymphoma, and cervical carcinoma.3 In addition, many non-AIDS-defining cancers (NADCs) have increasingly been reported among HIV-infected patients, including carcinoma of the anus, head and neck, testis, lung, colon, skin (basal cell skin carcinoma, squamous cell skin carcinoma, and melanoma), and Hodgkin disease.4 Although several risk factors have been associated with the development of AIDS-defining cancers, predictors for the occurrence of NADCs have not been defined so well. In our search of the literature, we found that some studies suggested an association between immunosuppression, as measured by CD4 count, and the development of NADCs, whereas other studies reported no such association.5, 6 Identifying predictors for the development of NADCs among HIV-infected patients may facilitate earlier diagnosis and therapy of these conditions. We performed a retrospective study to determine the rates and predictors of NADCs among HIV-infected United States Department of Defense (DoD) beneficiaries.

MATERIALS AND METHODS

We examined data that were collected during a multicenter, HIV natural history study, which prospectively had evaluated DoD beneficiaries on a biannual basis utilizing standardized data collection procedures. Details of this cohort have been published previously.7, 8 Research study coordinators and nurses obtained data from patients' medical records, laboratory reports, hospital information systems, and patient/coordinator interviews. All HIV-infected beneficiaries were invited to join this observational study at one of seven United States military Tri-Service HIV Evaluation Clinics, including Walter Reed Army Medical Center (Washington, DC), Wilford Hall Medical Center and Brooke Army Medical Center (San Antonio, TX), National Naval Medical Center (Bethesda, MD), Naval Medical Center San Diego (San Diego, CA), Naval Medical Center Portsmouth (Portsmouth, VA), and Tripler Army Medical Center (Honolulu, HI). The protocols and consent forms for the study were approved by the Institutional Review Boards of each participating institution and by the Tri-Service Human Subjects Research Review Board.

In total, 4144 participants had available data for the years 1988–2003. Patients with cancer were identified by searching the data base for coding diagnoses and by using the terms “cancer”, “malignancy”, “malignant”, “melanoma”, “tumor”, “neoplasm”, “carcinoma”, “adenocarcinoma”, “carcinoid”, “Hodgkin disease”, “myeloma”, “sarcoma”, “angiosarcoma”, “fibrosarcoma”, “lymphoma”, “leukemia”, “squamous”, “basal”, and ”ca“ (abbreviation for cancer). The diagnosis of a malignancy was verified by laboratory, radiology, and/or histopathology results. Patients who had benign tumors or who had cancer diagnosis dates that preceded their HIV diagnosis were excluded. Data collected on patients (cases) (diagnosis of NADC) and controls included demographics, CD4 cell count (at the time of cancer diagnosis and nadir), HIV viral load (at the time of cancer diagnosis), history of opportunistic infection, HIV stage according to the Walter Reed classification system (WRCS), and duration of HIV infection. CD4 measurements, the use of highly active antiretroviral therapy (HAART), a history of opportunistic infection, and WRCS HIV stage were obtained at the time of cancer diagnosis for patients with cancer and at the last available date for the group without cancer (controls). HAART was defined as the concurrent use of three or more antiretroviral medications. HIV viral load was not included in the analyses of this study due to its unavailability before 1996.

Statistical analyses included descriptive statistics and univariate analyses of all variables. Multivariate analyses were performed using a backward, stepwise, logistic regression model to assess the predictors of NADC development; models for individual cancer types were not performed due to the small sample size for each type. A variable that changed the measure of association by ≥ 10% between 1 of the exposure variables and the outcome was considered a confounder and remained in the model. Determination of the goodness of fit utilized the Hosmer–Lemeshow statistic, with a good fit defined as a P value > 0.10 (SAS software; version 8e, SAS, Inc., Cary, NC).

Cancer incidence rates (per 100,000 person-years) were determined using the number of individuals at risk by year and were age-adjusted. Our data were compared with the Surveillance, Epidemiology, and End Results (SEER) Program cancer rates.9 The 95% confidence intervals (95%CI) for incidence rates obtained from the RV122 study were calculated based on Poisson distribution. SEER data were unavailable for basal and squamous cell skin carcinoma rates; annual rates were obtained from other sources.10, 11

RESULTS

Among 4144 participants in this study population with a total of 26,916 person-years of follow-up, 133 NADCs were diagnosed, yielding an age-adjusted rate of 980 per 100,000 person-years. In addition, 318 patients had an AIDS-defining cancer, including Kaposi sarcoma (n = 224 patients), non-Hodgkin lymphoma (n = 91 patients), and cervical carcinoma (n = 3 patients).

The primary sites and frequency of each NADC are shown in Table 1. The most frequent non-AIDS-defining malignancies reported were basal cell skin carcinoma (32.3%), squamous cell skin carcinoma (12.0%), Hodgkin disease (9.8%), and anal carcinoma (7.5%). The age-adjusted rate in Caucasian males (n = 38 participants) for basal cell skin carcinoma was 795 per 100,000 person-years; this rate was significantly higher than the United States annual age-adjusted incidence of basal cell skin carcinoma for Caucasian males of 475 per 100,000 (binomial test of proportions; P = 0.005).10 In our HIV-positive cohort, the age-adjusted rate for squamous cell skin carcinoma was higher for Caucasians (159 per 100,000 person-years) compared with African Americans (39 per 100,000 person-years) and mirrored the higher rate of these cancers among the general Caucasian United States population. In addition, the rate of squamous skin carcinoma was elevated in our cohort compared with general United States population rates (P = 0.001).11

Table 1. Types of Non-AIDS-Defining Malignancies (N = 133 patients)
Primary type of malignancyNo. of malignancies (%)a
  • AIDS: acquired immunodeficiency syndrome.

  • a

    Single occurrences of breast adenocarcinoma, breast carcinoma, endocrine system carcinoma, esophageal carcinoma, hepatic carcinoma, oral cavity, pancreatic carcinoma, papillary cell (locus unknown) carcinoma, skin carcinoma (excluding basal and squamous cell carcinoma), gastric carcinoma, testicular carcinoma, throat carcinoma, tongue carcinoma, and uterine carcinoma were reported.

Basal cell skin carcinoma43 (32.3)
Squamous cell skin carcinoma16 (12)
Hodgkin disease13 (9.8)
Anal carcinoma10 (7.5)
Melanoma9 (6.8)
Colorectal carcinoma8 (6)
Prostate carcinoma6 (4.5)
Renal carcinoma4 (3)
Lung carcinoma3 (2.3)
Salivary gland carcinoma3 (2.3)
Brain carcinoma2 (1.5)
Bone carcinoma2 (1.5)
Male genitalia carcinoma2 (1.5)

For Caucasians, NADC rates exceeded national rates in the United States after age adjustment for the following malignancies: prostate carcinoma, melanoma, anal carcinoma, and Hodgkin disease. For African Americans, NADC rates exceeded United States national rates after age adjustment for colorectal carcinoma, anal carcinoma, and Hodgkin disease (Table 2). In the current study, only Caucasians developed melanoma.

Table 2. Age-Adjusted Cancer Incidence Rates by Race and Primary Site, per 100,000 Person-Years
MalignancyIncidence rate (95%CI)
CaucasiansAfrican Americans
HIV-infected patientsGeneral US populationaChi-square testHIV-infected patientsGeneral US populationaChi-square test
  • 95%CI: 95% confidence interval; US: United States; HIV: human immunodeficiency virus.

  • a

    Data were based on National Program of Cancer Registries and Surveillance, Epidemiology, and End Results Program's population-based registries. Cancer incidence rates were based on the 2000 United States standard population.

  • b

    A gender-specific denominator was used to calculate the rate for prostate carcinoma.

  • c

    Statistically significant (P value < 0.05).

Prostate carcinomab258.8 (227.3–290.4)150.5 (149.7–151.3)78.0c221.7 (192.5–250.9)233.8 (230.4–237.2)0.63
Melanoma175.7 (149.7–201.7)16.7 (16.6–16.9)1513.8c00.9 (0.7–1.0)
Colorectal carcinoma13.7 (6.7–20.9)54.1 (53.8–54.4)30.2149.1 (125.2–173.1)58.0 (57.0–59.1)143.2c
Anal carcinoma54.5 (40.0–69.0)1.3 (1.3–1.4)2177.9c34.9 (23.3–46.5)1.3 (1.1–1.4)868.4c
Hodgkin disease74.7 (57.8–91.6)2.9 (2.8–2.9)1777.7c26.4 (16.3–36.4)2.3 (2.1–2.5)251.9c

Descriptive and univariate analyses of the groups with and without cancers are shown in Table 3. Age was correlated with NADCs, as participants age 40 years or older had significantly higher rates of NADC with compared participants age 24 years or younger (odds ratio [OR], 10.49; 95%CI, 5.79–19.0). Caucasian/non-Hispanic participants had significantly higher rates of NADCs compared with African-American participants (OR, 2.45; 95%CI, 1.67–3.62); this likely was related to our high rate of skin carcinomas, which occurred predominantly among Caucasian/non-Hispanics. We determined that age and race were confounders, whereas gender was neither a confounder nor a significant predictor of cancer. Duration of HIV infection (OR, 1.13; 95%CI, 1.09–1.17), CD4 counts at cancer diagnosis of 200–499 cells/mm3 (OR, 2.25; 95%CI, 1.49–3.40), a history of opportunistic infection (OR, 2.49; 95%CI, 1.75–3.54), and HAART (OR, 0.56; 95%CI, 0.38–0.82) were associated strongly with the development of NADC in univariate analyses.

Table 3. Univariate Associations of Non-AIDS-Defining Malignancy Occurrences and Demographics, Laboratory Values, and Clinical Markers
VariableNo. of patients (%)OR95%CIP value
With malignancy (n = 133)Without malignancy (n = 4011)
  • AIDS: acquired immunodeficiency syndrome; OR: odds ratio; 95%CI: 95% confidence interval; HIV: human immunodeficiency virus; +: positive; HAART: highly active antiretroviral therapy.

  • a

    The category “Other”: included patients who were Hispanic/Puerto Rican/Mexican, Asian/Pacific Islander, Native American/Alaskan Native, Asian, or other.

Gender     
 Male123 (92.5)3619 (90.2)1.00  
 Female10 (7.5)392 (9.8)0.750.39–1.440.390
Age    < 0.001
 ≤ 24 yrs15 (11.3)1248 (31.2)1.00  
 25–29 yrs33 (24.8)1174 (29.4)2.341.26–4.330.007
 30–34 yrs23 (17.3)759 (19.0)2.521.31–4.860.006
 35–39 yrs16 (12.0)449 (11.2)2.971.45–6.050.003
 ≥ 40 yrs46 (34.6)365 (9.2)10.495.79–19.0< 0.001
Race    < 0.001
 African American38 (28.6)1850 (46.2)1.00  
 Caucasian86 (64.7)1706 (42.6)2.451.67–3.62< 0.001
 Othera9 (6.8)452 (11.3)0.970.47–2.020.934
Walter Reed disease Stage     
 Stage 1–249 (38.9)1483 (40.6)1.00  
 Stage 3–677 (61.1)2166 (59.4)1.080.75–1.550.694
Mean duration of HIV infection (yrs)9.026.371.131.09–1.17< 0.001
CD4+ cell count at diagnosis    < 0.001
 ≥ 500 cells/mm336 (27.5)1619 (40.6)1.00  
 200–499 cells/mm366 (50.4)1319 (33.1)2.251.49–3.40< 0.001
 < 200 cells/mm329 (22.1)1050 (26.3)1.240.76–2.040.391
Nadir CD4+ cell count    0.747
 ≥ 500 cells/mm317 (15.9)738 (18.5)1.00  
 200–499 cells/mm347 (43.9)1754 (44.0)1.160.66–2.040.598
 < 200 cells/mm343 (40.2)1496 (37.5)1.250.71–2.200.445
HIV viral load at time of diagnosis (copies/mL)    0.013
 < 40014 (19.2)796 (35.5)1.00  
 400–75,00045 (61.6)1167 (52.1)2.191.20–4.020.011
 >75,00014 (19.2)279 (12.4)2.851.34–6.060.006
Opportunistic infection(s)     
 No53 (39.9)2497 (62.3)1.00  
 Yes80 (60.1)1514 (37.8)2.491.75–3.54< 0.001
Use of antiretroviral drugs     
 0, 1, 2 drugs95 (71.4)2342 (58.4)1.00  
 HAART38 (28.6)1669 (41.6)0.560.38–0.820.003

In the multivariate model (Table 4), significant predictors (P < 0.001) for the occurrence of NADCs included age 40 years or older (OR, 12.22; 95%CI, 6.51–22.95), Caucasian/non-Hispanic race (OR, 2.14; 95%CI, 1.43–3.22), duration of HIV infection (OR, 1.20; 95%CI, 1.15–1.25), and a history of opportunistic infection (OR, 2.53; 95%CI, 1.63–3.94). HAART use (OR, 0.21; 95%CI, 0.14–0.34) was protective against the development of NADC.

Table 4. Final Multivariate Model for the Occurrence of Non-AIDS-Defining Malignancy
VariableOR95%CIP valuea
  • AIDS: acquired immunodeficiency syndrome; OR: odds ratio; 95%CI: 95% confidence interval; HIV: human immunodeficiency virus; +: positive; HAART: highly active antiretroviral therapy.

  • a

    Goodness of fit chi-square = 9.19; P value = 0.326.

  • b

    The category “Other” included patients who were Hispanic/Puerto Rican/Mexican, Asian/Pacific Islander, Native American/Alaskan Native, Asian, or other.

Age  < 0.001
 ≤ 24 yrs1.00  
 25–29 yrs2.251.20–4.220.011
 30–34 yrs2.631.34–5.160.005
 35–39 yrs3.231.55–6.720.002
 ≤ 40 yrs12.226.51–22.95< 0.001
Race  < 0.001
 African American1.00  
 Caucasian, not Hispanic2.141.43–3.22< 0.001
 Otherb1.000.46–2.200.999
Duration of HIV infection (yrs)1.201.15–1.25< 0.001
CD4+ cell count at diagnosis  < 0.001
 ≥ 500 cells/mm31.00  
 200–499 cells/mm31.631.04–2.550.032
 < 200 cells/mm30.430.23–0.800.007
History of opportunistic infection   
 No1.00  
 Yes2.531.63–3.94< 0.001
Use of antiretroviral drugs   
 No use, 1 drug, 2 drugs1.00  
 HAART0.210.14–0.34< 0.001

DISCUSSION

The most frequent non-AIDS-defining malignancies among HIV-infected patients in our cohort were primary carcinomas of the skin. We also noted a higher rate of prostate carcinoma, melanoma, anal carcinoma, and Hodgkin disease among HIV-infected patients compared with age-adjusted, race-specific individuals residing in the United States. Predictors of NADCs included increasing age, Caucasian/non-Hispanic race, a history of opportunistic infections, and longer duration of HIV infection. HAART was associated with a lower risk of cancer in our cohort.

Several studies have demonstrated a progressive decline in AIDS-defining cancers in the post-HAART era. This has been noted particularly for Kaposi sarcoma and cervical carcinoma; non-Hodgkin lymphoma rates have been affected less.12–15 Simultaneously, the rates of NADCs are rising,16, 17 with 1 study showing a significant increase during the post-HAART era versus the pre-HAART era (rate ratio, 10.9; P < 0.0002).18 Before the availability of HAART, NADCs accounted for < 1% of deaths.19 In a recent study, it was demonstrated that 13% of deaths among HIV-infected patients were due to NADCs.20 The etiology of this trend is unknown; however, our data suggest that increasing age and duration of HIV infection among this population may be important factors.

NADCs that reportedly occur at an increased rate among HIV-infected patients compared with the general population include anorectal carcinoma(80-fold increase), Hodgkin disease (17-fold), melanoma (6-fold), and lung carcinoma (3-fold).21 Other studies also noted increased rates for brain tumors, leukemia, myeloma, as well as malignancies of the head and neck, lip, and testis.4, 5, 22 Hepatocellular tumors have accounted for several NADCs and may become increasingly important among hepatitis B or C coinfected patients.20, 22 Some studies also have noted an increase in the aggressiveness of certain NADCs (e.g., melanoma, breast) in the setting of HIV infection.23, 24 The most common NADCs in our study were primary skin carcinoma (most commonly, basal cell skin carcinoma, followed by squamous cell skin carcinoma). Other studies also reported the propensity of HIV-infected patients to develop nonmelanoma skin malignancies.18, 25, 26 We also noted both in the current study and in a prior investigation that the rate of prostate carcinoma was elevated among HIV-infected men.27

The precise cause for the increased rate of certain NADCs remains unclear. Some studies have suggested lifestyle factors6; certainly, modification of tobacco and alcohol use is recommended. HIV-infected patients frequently are coinfected with oncogenic viruses, such as hepatitis B and C, human herpesvirus 8, and Epstein–Barr virus. Anorectal carcinoma has been associated with the human papillomavirus and has a marked increased rate among men who have sex with men.28 Others have focused on the oncogenic potential of HIV infection itself, including by induction of specific protooncogenes.29–32

We examined risk factors for NADC using an epidemiologic approach and found that the duration of HIV infection was predictive independent of the patient's age. A possible explanation of this finding is that ongoing loss of the immune repertoire and adequate tumor surveillance may be important in cancer pathogenesis. Additional time is provided for patients to develop cancer as HIV-infected patients live longer and as deaths due to opportunistic infections continue to decline.

We did not find an association between cancer development and a low CD4 nadir or CD4 count at cancer diagnosis. The latter finding may be due to the fact that the controls were not matched to the cases, and that the latest CD4 counts were utilized for this group, which may have biased our conclusions toward the null. Another possible explanation is that, we did not investigate immunosuppression as a risk factor for individual NADCs. Because skin cancers were the most common malignancy in our cohort, these data suggest that CD4 counts do not predict skin cancer development.25, 26 The results from other investigations concur on finding the lack of an association with the CD4 count, with the possible exceptions of Hodgkin disease and anal carcinoma.13, 15, 18, 33 For instance, anal carcinoma was correlated with low CD4-positive cell counts, high HIV plasma viral loads, and a previous AIDS-defining event.28, 34 These findings suggest that immunosuppression may be important for the development of certain cancers but not others.

Greater than 75% the patients with NADCs in our cohort had a CD4 count > 200 cells/mm3 at the time of their cancer diagnosis. Others also have noted that NADCs frequently occur in the setting of controlled HIV infection.18, 20, 35 In a study that examined all NADCs, CD4 counts at cancer diagnosis were significantly lower in the pre-HAART era versus the post-HAART era (35 cells/mm3 vs. 123 cells/mm3); this may be related to the increasing survival with more robust CD4 counts after the introduction of HAART.17

With regard to the use of antiretroviral drugs, patients who did not receive HAART were more likely to develop cancer. HAART maintains immune function and may enhance tumor surveillance. Conversely, some investigators have stated concerns for the possible mutagenic, long-term effects of these medications.35, 36 In a study that compared the incidence of NADCs in the pre-HAART era and the post-HAART era, no significant differences were shown in the incidence rate; however, specific medication use was not examined.35 Further investigation into the role of HAART on cancer incidence, specifically for non-Hodgkin lymphoma and NADCs, is warranted.

Our finding that Caucasian/non-Hispanics had a two-fold greater risk of NADCs was noteworthy and likely is related to the high rate of skin cancers within our study cohort. Among the general population, overall, African Americans are more likely to develop cancer than any other racial/ethnic group37; however, Caucasians have higher rates of certain cancers, such as skin, bone, breast, testicular, and brain malignancies as well as Hodgkin disease. The rate for anal carcinoma is the same, whereas prostate carcinoma risk is highest in African Americans. In our study, we noted that HIV-infected Caucasian/non-Hispanic patients had a higher rate of prostate carcinoma, melanoma, squamous cell and basal cell skin carcinomas, anal carcinoma, and Hodgkin disease compared with the general United States population, whereas HIV-positive African Americans had higher rates of colorectal carcinoma, anal carcinoma, and Hodgkin disease when race-specific data were analyzed.

The strengths of the current study include a well defined cohort with ongoing, long-term follow-up spanning over a decade. Given open access medical care through the military healthcare system, it is unlikely that we misdiagnosed cancer events. Limitations of this study include the lack of women and the lack of data on smoking history in the cohort. Due to the small numbers of patients with each NADC, for the current risk factor analysis, we used combined data; thus, we were not able to examine predictors for individual cancer types or the potential role of viral coinfections. Because our population receives frequent medical evaluations, there may be a screening bias toward an increased diagnosis rate of cancer compared with the rate among the general United States population.

Based on the findings of the current study, we recommend considering the placement of cancer surveillance programs within HIV clinics. Screening for skin malignancies as well as prostate and anal carcinomas among HIV-infected patients appears to be the highest priority; prospective studies examining the efficacy and cost-effectiveness of this approach are needed. Age older than 40 years, Caucasian/non-Hispanic ethnicity, longstanding HIV infection, and a lack of HAART are risk factors for NADCs. Early diagnosis and therapy for cancer likely will become more important for improved quality of life and survival among HIV-infected individuals as more patients suffer from non-AIDS-related diagnoses.

Acknowledgements

The Chief, Bureau of Medicine and Surgery, Navy Department, Washington, DC, Clinical Investigation Program, sponsored this report (no. S93-058), as required by Naval School of Health Sciences, Bethesda, Instruction (NSHSBETHINST) 6000.41B. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of Defense, or the United States Government. This statement is required by Air Force Instruction (AFI) 40-402 in reports of research involving humans: “The voluntary fully informed consent of the subjects used in this research was obtained as required by 32 Code of Federal Regulations (CFR) 219 and AFI 40-402.”

Ancillary