Change in Pattern of Secondary Cancers After Kaposi Sarcoma in the Era of Antiretroviral Therapy

Key Points

Question  Is there a change in the risk of secondary cancers that follow Kaposi sarcoma in the current era of antiretroviral therapy?

Findings  In this longitudinal study of 14 905 cases of Kaposi sarcoma, the incidence of secondary cancers after Kaposi sarcoma has decreased in recent years. Patients with Kaposi sarcoma continue to be at significantly increased risk of developing cancer of the anus and non-Hodgkin lymphoma, while cancers of the tongue and penis and acute lymphocytic leukemia have also become significantly associated with Kaposi sarcoma.

Meaning  The secondary cancers associated with Kaposi sarcoma have changed, and screening for these tumors is important in patients with Kaposi sarcoma.

Importance  Studies performed in the 1980s and early 1990s have shown that people who develop Kaposi sarcoma (KS) are at higher risk of developing other cancers. The demographics of those affected with human immunodeficiency virus (HIV)/AIDS and KS have changed, and individuals with HIV/AIDS and KS now live longer.

Objectives  To test the hypothesis that the secondary cancers developing in patients with KS have changed in recent years and to assess the risk of secondary cancers after KS in different periods.

Design, Setting, and Participants  Longitudinal data from 9 cancer registries in the Surveillance, Epidemiology, and End Results (SEER) database were used to identify cases of KS diagnosed from January 1973 to December 2013. The dates of the analysis were November 2016 to February 2017.

Main Outcomes and Measures  The primary outcome was the development of secondary cancers in individuals with KS. Secondary cancers were considered only if diagnosed 2 months after a diagnosis of KS. Standardized incidence ratios (SIRs) were calculated for the development of new secondary cancers in the pre-AIDS era (1973-1979), pre–highly active antiretroviral therapy (HAART) era (1980-1995), and HAART era (1996-2013). Stratified analysis was then performed on a subset of the cases diagnosed from 1996 to 2013 based on age at diagnosis (<65 and ≥65 years), latency period between KS and the development of secondary cancers (1 year, 2-5 years, >5 to 10 years, and >10 years), and registries with higher vs lower reported rates of HIV/AIDS.

Results  Among 14 905 individuals with diagnosed KS, 13 721 (92.1%) were younger than 65 years at the time of diagnosis, and 14 356 (96.3%) were male. From 1980 to 1995, SIRs were 2.01 (95% CI, 1.00-3.60) for cancer of the rectum, 49.70 (95% CI, 33.53-70.94) for cancer of the anus, 4.98 (95% CI, 2.79-8.22) for cancer of the liver, 13.70 (95% CI, 2.82-40.03) for cancer of the cervix, 6.40 (95% CI, 2.76-12.60) for Hodgkin lymphoma, and 48.97 (95% CI, 44.85-53.36) for non-Hodgkin lymphoma. From 1996 to 2013, cancer of the anus, Hodgkin lymphoma, non-Hodgkin lymphoma, and cancer of the liver remained associated with KS, with the addition of the following significant SIRs: 6.99 (95% CI, 3.20-13.27) for cancer of the tongue, 10.28 (95% CI, 1.24-37.13) for cancer of the penis, and 17.62 (95% CI, 3.63-51.49) for acute lymphocytic leukemia. The SIR of developing any tumor after KS decreased significantly from 3.36 to 1.94 from the pre-HAART era to the HAART era.

Conclusions and Relevance  There has been a significant decline in the overall risk of secondary cancers after KS. Certain cancers, including acute lymphocytic leukemia, cancer of the tongue, and cancer of the penis, are increasingly becoming more common in the HAART era compared with the pre-HAART era. Close monitoring and screening for these secondary cancers is desirable in patients with KS.

Kaposi sarcoma (KS), one of the 3 AIDS-defining cancers, is a tumor that is associated with immunosuppression, and its presence in people with human immunodeficiency virus (HIV) warrants a diagnosis of HIV/AIDS.¹ Kaposi sarcoma manifests as cutaneous purplish, reddish-blue, or dark brown or black macules, plaques, and nodules on the lower extremities,^2-6 which has also been found to spread to the gastrointestinal tract, oral cavity, lungs, and lymph nodes.^7-9 The different forms of identified KS are as follows: (1) classic KS, found commonly in Jewish, Mediterranean, and Middle Eastern HIV-negative elderly men^5,10; (2) African KS, also known as endemic KS; (3) AIDS-associated KS, also known as epidemic KS, which is common in individuals with advanced immunodeficiency²; and (4) iatrogenic KS.⁸ Before the advent of AIDS, KS was found in elderly men.¹¹ However, after the emergence of HIV/AIDS, it has become more common in younger patients with AIDS and weaker immune systems.¹²

There has been a change in the trend and demographics of patients with KS affected by the emergence of the HIV/AIDS epidemic and, subsequently, antiretroviral therapy. Before HIV/AIDS, the incidence of KS was low at 0.3 case per 100 000 but increased to 8.1 cases per 100 000 by 1989 among persons of white race in the United States.¹² In addition, after the emergence of HIV/AIDS, the prevalence of KS rose from 11% to 90% among adults with HIV/AIDS who were aged 20 to 54 years.¹² During the HIV/AIDS epidemic, AIDS-associated cancers (especially KS) peaked in the mid-1990s. An estimated 81.6% of 83 252 KS cases were found during the period from 1980 to 2007 among patients with AIDS.¹³Quiz Ref ID The introduction of highly active antiretroviral therapy (HAART) improved the survival of patients with HIV/AIDS and AIDS-related KS.¹⁴ HAART decreases the suppression of the immune system in HIV-positive patients, with an increased KS tumor response.^2,15 Despite the improvement of the immune system with HAART, there still exists a higher risk of AIDS-associated KS in HIV-positive people compared with the general population.²

Studies^16,17 have shown that individuals with KS are at higher risk of developing other secondary cancers. Biggar et al¹⁶ assessed the risk of secondary cancers in adults younger than 70 years with KS for the period from 1980 to 1990. They found a significantly higher risk of developing non-Hodgkin lymphoma (NHL) in people with AIDS-associated KS both in the first year and in all years after KS.¹⁶ There was no higher risk for secondary NHL among those with classic KS.¹⁶ From 1949 to 1975 at the Memorial Sloan Kettering Cancer Center (New York, New York), 14 of 92 patients with KS developed a secondary cancer after KS; lymphoreticular cancer was found among 9 of the cases.¹⁷ However, the most recent of these studies were performed in the early 1990s, before the introduction of effective antiretroviral therapy.^16,17 Since the time these studies were conducted, several changes have occurred. First, the demographics of people affected with KS have changed, with KS occurring more frequently now in younger adults compared with elderly ones.¹² Second, the demographics of individuals with HIV/AIDS have also changed.¹⁸ Third, antiretroviral agents now provide longer survival for people with HIV/AIDS and consequently those with KS.¹⁴ Therefore, it is likely that the tumors associated with KS have also changed.

Our study hypothesis was that the secondary cancers developing in patients with KS have changed in recent years. We tested this hypothesis by calculating the standardized incidence ratios (SIRs) for the secondary cancers using the Surveillance, Epidemiology, and End Results (SEER) program data from 1973 to 2013.

We used longitudinal data from 9 cancer registries in the SEER database of the National Cancer Institute to identify cases of KS diagnosed from January 1973 to December 2013. The dates of the analysis were November 2016 to February 2017. The SEER registry is a database of 18 cancer registries across the United States, representing approximately 30% of the US population.¹⁹ The SEER data are periodically updated with patient information, including tumor characteristics, cancer treatment, secondary cancers, patient demographics, and cause of death. For the purpose of this analysis, data were restricted to the 9 cancer registries that have been contributing to the SEER data since 1973 onward.¹⁹ The SEER data contain deidentified information and were considered exempt by the institutional review board at the University of South Florida.

The primary outcome was the development of secondary cancers in individuals with KS. Patients with KS were followed up to the date they developed a secondary cancer, date of death, date of last follow-up, or end of the study period. Secondary cancers were considered only if diagnosed 2 months after a diagnosis of KS. We excluded patients with unknown age and cases diagnosed at autopsy or by death certificate and included only cases that were first primary tumors (ie, the patient with KS had no history of cancer). Patient demographic and clinical information was obtained, including age, race, sex, year of diagnosis, cancer registry, and primary site. Because we did not have information on HIV/AIDS diagnosis and KS is associated with HIV/AIDS, we recoded the following variables to attempt to identify cases of KS in individuals with HIV/AIDS compared with those without HIV/AIDS: (1) age was recoded as those younger than 65 years and those 65 years and older, (2) cancer registries were recoded as regions with high vs low reported rates of HIV/AIDS, and (3) year of diagnosis was recoded as 1973 to 1979 (pre-AIDS era), 1980 to 1995 (pre-HAART era), and 1996 to 2013 (HAART era). Race in the SEER data set is recorded as white, black, American Indian or Alaskan native, Asian or Pacific Islander, and unknown.

The SIRs and their respective 95% CIs were calculated for the development of new secondary cancers by comparing the observed with the expected number of cases calculated from the SEER population, adjusted for age, race, sex, and year of diagnosis.²⁰ The analysis was stratified based on year of diagnosis, including the pre-AIDS era (1973-1979), pre-HAART era (1980-1995), and HAART era (1996-2013). Subsequently, further stratified analysis was then performed on a subset of the cases diagnosed from 1996 to 2013. Initially, the SIRs were stratified on the following: (1) age at diagnosis into those younger than 65 years and those 65 years and older to represent individuals with AIDS-related KS and classic KS, respectively; (2) latency period between KS and the development of secondary cancers (1 year, 2-5 years, >5 to 10 years, and >10 years); and (3) registries with higher vs lower reported rates of HIV/AIDS. Owing to the small number of secondary cancers, we categorized all lymphoid tumors and blood tumors as one group, and all other tumors were grouped as solid tumors (eTable 1 in the Supplement). We also present the results for specific secondary cancers that were statistically significant. Analyses were performed using statistical software (SEER*Stat, version 8.3.2; National Cancer Institute). All reported P values are 2-sided, and P ≤ .05 was considered statistically significant.

Table 1 summarizes the demographic characteristics of the study population. A total of 14 905 cases of KS were included in the analysis. There were fewer patients with KS who were 65 years and older (7.9%). Most patients (80.9%) were of white race, followed by black race (14.9%). Most of the cases were from areas with high reported incidences of HIV/AIDS, including the San Francisco-Oakland (California) registry (53.3%), followed by the Atlanta metropolitan (Georgia) registry (14.8%). Up to 73.9% of the cases were diagnosed from 1980 to 1995, 24.6% from 1996 to 2013, and 1.5% from 1973 to 1979.

The Figure shows the incidence of secondary cancers, stratified by year of diagnosis. Only cancers with incidences that were significantly higher than expected are presented. Before 1980, only cancer of the ascending colon was occurring at a rate higher than expected (SIR, 4.34; 95% CI, 1.18-11.12). Quiz Ref IDFrom 1980 to 1995, SIRs were 2.01 (95% CI, 1.00-3.60) for cancer of the rectum, 49.70 (95% CI, 33.53-70.94) for cancer of the anus, 4.98 (95% CI, 2.79-8.22) for cancer of the liver, 13.70 (95% CI, 2.82-40.03) for cancer of the cervix, 6.40 (95% CI, 2.76-12.60) for Hodgkin lymphoma (HL), and 48.97 (95% CI, 44.85-53.36) for NHL. From 1996 to 2013, cancer of the anus, HL, NHL, and cancer of the liver remained associated with KS (Figure and eTable 2 in the Supplement), with the addition of the following significant SIRs: 6.99 (95% CI, 3.20-13.27) for cancer of the tongue, 10.28 (95% CI, 1.24-37.13) for cancer of the penis, and 17.62 (95% CI, 3.63-51.49) for acute lymphocytic leukemia (ALL). Quiz Ref IDThe SIR of developing any tumor after KS decreased significantly from 3.36 to 1.94 from the pre-HAART era to the HAART era (eTable 2 in the Supplement).

Among cases of KS diagnosed from 1996 to 2013, a total of 267 patients developed secondary cancers. People with KS who were younger than 65 years and those from regions with higher reported rates of HIV/AIDS had significantly higher incidences of secondary cancers, including cancer of the tongue, cancer of the anus, cancer of the liver, cancer of the penis, NHL, HL, and ALL. Individuals 65 years and older had a 2.54-fold increased incidence of developing NHL. There were 7 observed cases, and the expected number was 2.75, giving an SIR of 2.5 (7 divided by 2.75), or approximately 3-fold increase risk. Patients from regions with lower reported rates of HIV/AIDS had a 15-fold increased incidence of developing NHL and a 59-fold increased incidence of developing cancer of the anus. When the analysis was stratified by latency period, cancer incidence was highest in the period 1 to 10 years after the diagnosis of KS for all the secondary cancers analyzed except for NHL, which had a higher incidence within the first year after a diagnosis of KS (Table 2).

We believe that our stratification actually represents the people with HIV/AIDS-related KS and classic KS. Similar to the study by Biggar et al,²¹ we took steps to ensure that our stratification represented these population subgroups. Before 1980, only 60 people younger than 65 years had KS, while 162 people with KS were 65 years and older. From 1980 to 2013, the SEER program recorded 1022 cases of KS in those 65 years and older. Had the ratio of KS in those younger than 65 years and those 65 years and older been maintained as 60:162 in the pre-AIDS era, the expected number of KS cases in those younger than 65 years should have been 378 in the AIDS era. However, 13 661 cases of KS were observed in those younger than 65 years in the AIDS era. Consequently, most of the cases of KS (13 283 of 13 661 [97.2%]) in those younger than 65 years in the AIDS era were due to HIV/AIDS and are thus AIDS-related KS. In addition, in the pre-AIDS era, the proportion of men and women with KS in both those younger than 65 years and those 65 years and older were about the same (75.0% and 25.0%, respectively) (eTable 3 in the Supplement). Quiz Ref IDHowever, in the AIDS era, there was a significant change in the male and female proportions only in those who were younger than 65 years (99.3% among men vs 0.7% among women, P < .001), while the proportion of men to women was not significantly different among individuals 65 years and older (P = .35).

In our study, cancer of the ascending colon was the only statistically significant secondary cancer after KS during the pre-AIDS era (1973-1979). A study²² investigating the risk of developing secondary cancers in those with classic KS reported a higher risk of colon cancer among women but not men. However, in our study, only 1 of the 4 patients with cancer of the ascending colon was female. The risk of secondary cancers in the pre-AIDS era was low. Reasons for this low risk could be because of loss of patients to follow-up, poorer diagnostic capabilities, and reduced life expectancy of the population during this era.

In the AIDS era, diagnosis of KS switched from its typical manifestation in elderly men to younger men, possibly due to the rising incidence of HIV/AIDS, coupled with the change in the demographics of people affected with KS. Individuals who developed KS in the HAART era are younger and live longer owing to better treatment. Certain tumors, including ALL, cancer of the tongue, and cancer of the penis, increasingly became more common in the HAART era compared with the pre-HAART era. The higher risk for colon cancer and cervical cancer after KS in the pre-HAART era decreased significantly in the HAART era, perhaps because of improved screening for these cancers and vaccination for human papillomavirus in women.²³

When we stratified our analysis by age at diagnosis and HIV/AIDS region, only NHL occurred significantly more than expected among patients in regions with lower reported rates of HIV/AIDS and among patients who were 65 years and older. All the other secondary cancers, including ALL, HL, and cancers of the tongue, anus, liver, and penis, were more common in patients younger than 65 years and in regions with higher reported rates of HIV/AIDS. All these tumors have been shown to be associated with HIV/AIDS.^1,21,24 The higher incidence of NHL among people with AIDS-related KS compared with classic KS may imply that HIV/AIDS increases the risk, but there is still a substantial risk in those with classic KS as well. A previous study²⁵ has shown that people with NHL are also at increased risk of developing KS. Quiz Ref IDStudies^26-29 have demonstrated a link between KS-associated herpesvirus and certain subtypes of lymphomas, including primary effusion lymphoma, plasmablastic lymphomas, and multicentric Castleman disease. Like KS, primary effusion lymphoma also occurs predominantly in the elderly.²⁷ Therefore, the higher risk of NHL that we found in our study among those younger than 65 years compared with the total risk in both age groups suggests that HIV/AIDS and KS may be interacting to increase the risk of NHL.

The co-occurrence of secondary cancers in individuals with KS could be due to genetic predisposition, similar etiological agents, or effects of treatment. As for treatment, a recent study³⁰ that investigated the effect of therapy with liposomal anthracycline on the development of secondary cancers in people with KS found that such treatment did not increase the risk of secondary cancers. Treatment is unlikely to be responsible for the secondary cancers, as evidenced by the fact that we found secondary cancers occurring more than expected within 1 year of diagnosing KS. If treatment was responsible for the secondary cancers, a certain period would be required for the therapy to exert its carcinogenic effect, and a clustering of cases would be observed during a specific interval. The most likely reason for the co-occurrence of these tumors is the underlying immunosuppression that is common to all these tumors and the possible oncogenic role of human herpesvirus 8. A study³¹ has found a significantly higher prevalence of human herpesvirus 8 among individuals with hematological cancers, including lymphomas, leukemia, and myeloproliferative disorders. Therefore, multiple factors could be responsible for the co-occurrence of these tumors.

There was significant variation in the time it took to develop secondary cancers. The highest risk for NHL was within the first year of developing KS, although the risk remained significantly high in subsequent years. Our findings are similar to the results of studies by Biggar et al¹⁶ and Hjalgrim et al,²² who found the highest risk to be within the first year of developing KS. Higher risk during the first year may be due to detection or surveillance bias because patients with KS may undergo more regular follow-up, leading to increased diagnosis of these tumors. On the other hand, the persistent higher risk beyond 10 years for certain tumors, including NHL, cancer of the anus, and cancer of the liver, indicates that an ongoing chronic process involving inflammation or immunosuppression may be responsible for the higher risk of these tumors in patients with KS.³² Most of the other tumors occurred significantly more than expected after 1 year of developing KS but usually within 10 years.

Our study has some limitations. We did not have data on HIV/AIDS diagnosis and thus were not able to address the influence of HIV/AIDS on an individual level. Having information on HIV/AIDS would have allowed us to assess for confounding and effect modification. However, we were able to show that our age stratification distinguished between KS associated with HIV/AIDS and classic KS. We noted differences in risk when we stratified our analysis by HIV/AIDS region and age at diagnosis of KS to delineate classic and AIDS-related KS. The risk of secondary cancers appears to differ for the 2 types of KS, possibly due to the influence of HIV/AIDS or immunosuppression. We observed an increase in the risk of NHL for those with KS who were younger than 65 years compared with those who were 65 years and older. This difference is most likely due to an interaction between HIV/AIDS and KS because we adjusted for age in our analysis to calculate the SIRs; therefore, the difference is unlikely to be due to age. Further studies exploring the possible interaction between HIV/AIDS and KS in increasing the risk of NHL are required. Furthermore, multiple secondary cancers are rare in general; hence, the numbers of some of these cases are small, which creates wide point estimates that reduce the precision of the risk of these tumors. Larger samples will result in better estimates of the higher risk of these secondary cancers. We did not have data on other risk factors, such as sexual behavior, oncogenic viruses, therapy, history of transplant or other immunosuppressive therapy, and exposure to other environmental carcinogens, which could act as potential confounders. Individuals with risky sexual behaviors are more likely to have more exposure to other carcinogens like human papillomavirus and are more likely to smoke.³³ Men who have sex with men have higher risk of HIV/AIDS and are more likely to develop anal cancers.³³ In addition, there is a possibility of surveillance bias because people with HIV/AIDS and KS may have more regular follow-up compared with the general population.

In summary, we found a decline in the occurrence of secondary cancers after KS in the AIDS era and observed the emergence of new tumors that are becoming more common among HIV/AIDS populations. It may be that HIV/AIDS and KS are interacting to increase the risk of secondary cancers. Vigilant monitoring and screening for the occurrence of these cancers is important in individuals who are diagnosed as having KS because subsequent cancers may develop several years later.

Accepted for Publication: June 7, 2017.

Corresponding Author: Fahad Mukhtar, MD, MPH, Department of Epidemiology and Biostatistics, College of Public Health, University of South Florida, 13201 Bruce B. Downs Blvd, Bldg MDC56, Tampa, FL 33612 (fahadmukhtar@health.usf.edu).

Published Online: August 24, 2017. doi:10.1001/jamaoncol.2017.2395

Author Contributions: Dr Mukhtar had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Mukhtar, Ilozumba, Cimenler.

Acquisition, analysis, or interpretation of data: Mukhtar, Utuama.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: Mukhtar, Ilozumba.

Statistical analysis: Mukhtar, Cimenler, Utuama.

Administrative, technical, or material support: Ilozumba.

Study supervision: Mukhtar.

Conflict of Interest Disclosures: None reported.

Additional Information: The Surveillance, Epidemiology, and End Results (SEER) program (http://www.seer.cancer.gov) research data (1973-2013), National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Surveillance Systems Branch, released April 2016, based on the November 2015 submission of SEER data.