The Incidence and Risk Factors For Hematological Malignancies and Premalignant Hematological Disorders In HIV

An increasing incidence of hematological malignancies (HM) has been noted in HIV+ patients, however the risk factors remain unclear. We retrospectively explored our large database of HIV+ patients to assess the incidence and distribution of different HM and premalignant hematological disorders (PMD) in HIV+ patients and the HIV and non HIV related risk factors for development.

Patients over the age of 18 with a positive HIV test (western blot or detectable viral load) between Jan 1^st 2001 and Dec 31^st 2011 were identified using the medical center database Clinical Looking Glass. Data were collected regarding demographics, Hepatitis B, C status, paraproteinemia and HIV characteristics. Results of all biopsies performed for each patient, ICD9 and cancer registry diagnoses were reviewed for a diagnosis of hematological malignancy (HM). Data collection was censored as of April 1^st 2013.

10,293 patients with a positive HIV test were identified with a median duration of follow up of 60 months (0-146 months) from the test date. 747 of these patients underwent a total of 948 biopsies for diagnosis of hematological disorders. 46 patients were excluded for an insufficient biopsy specimen. 255 of 10,293 patients (2.5%) were diagnosed with a HM or PMD. 241 patients had biopsy confirmation and 14 patients had ICD9 code/cancer registry data which were confirmed by chart review. Ten patients were diagnosed with HM when their HIV status was unknown and were excluded. 2 of the remaining 245 HM positive patients had two HM each, myeloma followed by extranodal NK/ T cell lymphoma and Castleman followed by DLBCL.

Mean age at diagnosis of the first HM/PHD was 46 years (SD:9.9). HM positive cases (n=245) in comparison to HM negative (n=9992) controls were more likely to be male (69% vs 57.6%, p<0.0001). More than 85% of patients in both groups were tested for hepatitis C and active hepatitis B with no difference in prevalence. HM positive cases were more often tested for paraprotein (OR:3.63, 95% CI: 2.78-4.75). Of those tested (n=1381), HM positive cases were more likely to have a discrete paraprotein. (OR: 2.79, 1.05-7.43) but no significant difference in prevalence of faint, oligoclonal or multiple paraprotein bands was detected. Binary logistic regression analysis of the paraproteinemia tested group showed increased odds of developing HM with male gender (OR:2.01, 1.27-3.19) and discrete (OR:3.25, 1.2-8.79), but not faint (OR: 1.4, 0.8-2.46) paraproteins.

The distribution of HM is noted in Table 1. High grade lymphomas accounted for 80.6% of HM, with DLBCL being the most frequent category. Of all HM, 43.5% occurred as non-nodal primaries The closest CD4 count within 6 months prior to HM diagnosis was obtained and median CD4 count compared. High grade lymphomas had a significantly lower median CD4 count close to HM diagnosis compare to low grade lymphomas and plasma cell dyscrasias. (150 vs 314 vs 281, p: 0.012)

Our study suggests that in HIV related HM, male gender and discrete paraproteins are likely to be significant risk factors but oligoclonal or multiple paraproteins are not. No additional risk appeared to be conferred by concurrent viral illnesses. The relationship of CD4 count to HM development suggests the role of immune mechanisms in pathogenesis.

No relevant conflicts of interest to declare.