icon-folder.gif   Conference Reports for NATAP  
  HIV Glasgow 2020, 5-8 October
Virtual Meeting
Back grey_arrow_rt.gif
Highest Cancer Risk in HIV+ Current
Smokers and Those With Poor HIV Control

  HIV Drug Therapy/Glasgow 2020, October 5-8, 2020
Mark Mascolini
In a study of more than 19,000 Europeans and Australians with HIV infection, current smokers and those with poor HIV control ran the highest risk of cancer diagnosis [1]. But current CD4 count or viral control did not worsen smoking's impact on cancer risk.
Researchers working with the RESPOND International Cohort Consortium of Infectious Diseases [2] reminded colleagues of the high prevalence of smoking and comorbidities in people with HIV infection. Rates of many AIDS cancers and non-AIDS cancers run higher in HIV groups than in the general population, and smoking boosts the risk of many cancers. The RESPOND investigators cited research indicating that smoking spurs HIV replication and heightens inflammation, both of which may contribute to the cancer-causing impact of smoking.
To further explore these issues, the RESPOND team conducted this analysis of their 17 European and Australian cohorts with two goals in mind: (1) to see whether different CD4 count and HIV suppression levels affect smoking's impact on cancer incidence (new diagnoses), and (2) to determine whether any interaction between CD4 count or viral load and cancer incidence holds true for different types of cancer.
RESPOND includes 29,432 Europeans and Australians with HIV for whom standardized data collection began prospectively in 2017 and extends retrospectively back to 2012 [2]. For this analysis, researchers grouped cancers into AIDS-defining cancers, non-AIDS cancers, smoking-related cancers, infection-related cancers, and body mass index (BMI)-related cancers.* They set the study baseline as the time of RESPOND enrollment or the time of known smoking status, whichever came later. The analysis excluded people with no CD4 or viral load data 12 months before or 6 months after the baseline date. Follow-up continued until the first new cancer diagnosis, the last study visit, or December 31, 2018, whichever came first.
The investigators put CD4 and viral load responses into three groups: Good: CD4 at or above 500 and viral load below 200; Poor, CD4 at or below 350 and viral load above 200; and Intermediate, all other CD4 and viral load combinations. They used Poisson regression analysis to assess relationships between current smoking status, and current CD4 count and viral load, on cancer incidence.
The analysis included 19,602 people, 74% men, 77% white, with a median age of 46 (interquartile range [IQR] 38 to 54), and a median 11 years with HIV infection (IQR 4 to 19). While 41.3% of the group never smoked, 44.4% currently smoked, and 14.4% previously smoked. Median age was older in previous smokers (51) than in current smokers (45) or never smokers (46). Previous smokers had a higher proportion with a viral load below 200 copies (79.1%) than did current smokers (66%) or never smokers (68.8%).
Researchers counted 513 new cancer diagnoses in 507 people, including 437 people with non-AIDS cancer, for an overall incidence of 6.9 per 1000 person-years (meaning about 7 of every 1000 people got diagnosed with cancer every year). Proportions of new cancers that were smoking-related were 39.8%, infection-related 35.9%, and BMI-related 23.2%.
Crude cancer incidence proved highest in current smokers (for all cancer types except AIDS cancers) and lowest in never smokers (for all cancer types except AIDS cancers). Crude cancer incidence was generally highest with poor CD4 count and viral load control (for all cancer types except smoking-related cancers). The difference between cancer incidence with poor CD4/viral load control and good CD4/viral load control was greatest for AIDS cancers and infection-related cancers.
But associations between CD4/viral load status (good, intermediate, or poor) and cancer incidence did not differ significantly between never smokers, current smokers, and previous smokers in an analysis adjusted for gender, ethnicity, HIV risk group, hepatitis B or C status, antiretroviral status, undetectable viral load, BMI, hypertension, diabetes, AIDS, cardiovascular disease, end-stage liver disease, chronic kidney disease, prior protease inhibitor use, time with HIV, nadir CD4 count, or baseline age.
RESPOND investigators noted that their analysis is limited by missing data on smoking status and duration, and type or intensity of smoking; missing data on important confounders such as alcohol use, family history, and coinfection with Epstein-Barr virus and human papillomavirus; and insufficient statistical power to analyze the impact of smoking and its interaction with CD4 count and viral load on risk of individual cancers.
The researchers concluded that current smokers and people with a viral load above 200 copies ran the highest risk of a new cancer diagnosis in this large European/Australian population. But the study yielded no evidence that smoking status affected the relationship between current CD4 count/viral load and cancer risk. These findings proved consistent across all cancers, non-AIDS cancers, infection-related cancers, and smoking-related cancers.
*Categories may overlap. Smoking-related cancers are lung, liver, bladder, colon, pancreatic, cervical, kidney, stomach, rectum, esophageal, oropharyngeal, and acute myeloid leukemia. Infection-related cancers are anal, liver, non-Hodgkin lymphoma, Kaposi sarcoma, Hodgkin lymphoma, cervical, stomach, oropharyngeal, and penile. BMI-related cancers are liver, colon, breast, pancreatic, kidney, gall bladder, rectum, esophageal, and thyroid.
1. Mocroft A, Petoumenos K, Wit F, et al. The relationship between smoking, CD4, viral load and cancer risk in HIV-positive adults. HIV Drug Therapy/Glasgow 2020, October 5-8, 2020. Abstract O124.
2. International Cohort Consortium of Infectious Disease (RESPOND). https://chip.dk/Studies/RESPOND