Blood Transfusions and the Risk of Intermediate- or High-Grade Non-Hodgkin's Lymphoma

The incidence of non-Hodgkin's lymphomas (NHLs) has increased for the past 30 years among men and women of all ages ( 1 ) . A number of epidemiologic studies have been undertaken to determine the causes of this increase. In several studies ( 2–5 ) , prior history of blood transfusion has been associated with an increased risk of NHL. This increased risk has been attributed to the immunosuppressive effects of allogeneic blood transfusions as well as the increased susceptibility of those persons who receive transfusions to infections caused by blood-borne organisms ( 3 ) . A recent study ( 6 ) conducted in Sweden found no association between blood transfusion and risk of NHL. To address this issue, along with other NHL-related issues, we conducted a population-based, case-control study of NHL in two groups of subjects—those who were infected with the human immunodeficiency virus (HIV) and those who were not. Patients who were aged 18-75 years when they were newly diagnosed with high- or intermediate-grade NHL [classified according to the Working Formulation ( 7 ) ] and who lived in Los Angeles County were identified by the Cancer Surveillance Program, the populationbased cancer registry for Los Angeles County. Because the study's major focus was acquired immunodeficiency syndrome (AIDS)-NHL, we limited case eligibility to patients with high- or intermediate-grade lymphoma. Our study methods have been described previously ( 8 , 9 ) . Briefly, we identified 1431 potentially eligible patients diagnosed between April 1989 and November 1992 and completed interviews with 525. We were unable to interview (and, hence, excluded) the remaining patients for the following reasons: 658 as a result of death, 44 for being too ill, two for being mentally incapacitated, 145 because of patient refusals, and 57 because of physician refusals for patient contact. Twenty-seven of the interviewed patients were ineligible based on pathology review and were excluded. Of the remaining 498 patients, 378 patients with NHL (193 females and 185 males) were confirmed to be HIV seronegative by standard HIV blood testing procedures. These patients were compared with neighborhood control subjects individually matched by age, race, and sex. Among the 120 seropositive patients with high- or intermediate-grade NHL, 113 were homosexual or bisexual men. Men who were patients diagnosed with AIDS on the basis of factors other than NHL (and who were identified by the Los Angeles County AIDS Program and were treated at Los Angeles County- University of Southern California Medical Center) were selected as control subjects for the first 50 of the 113 patients with AIDS-NHL. These AIDS control subjects were matched to patients with AIDS-NHL by age, race, and method of HIV acquisition. Because of limited resources, we could not match additional AIDS control subjects to our patients with AIDS-NHL. The 50 matched AIDS case-control pairs form the basis of the AIDS-NHL analyses. Signed, informed consent was obtained from each subject, and study procedures were approved by the Los Angeles County-University of Southern California Institutional Review Board in accordance with assurances approved by the U.S. Department of Health and Human Services. We estimated odds ratios (ORs) and their 95% confidence intervals (CIs) using conditional logistic regression methods. Because previous analyses of the HIV-negative case patients and control subjects showed an association between NHL risk and history of recreational drug use ( 9 ) , we conducted analyses that adjusted for use of these drugs. This study provides no statistically significant evidence to support the hypothesis that past history of blood transfusion is associated with an increased risk of high- or intermediate-grade NHL ( Table 1 ), although—based on the upper 95% confidence limit for the adjusted OR—we cannot exclude the possibility of a relatively small (72%) increase in risk of NHL among HIV-negative individuals or a nearly fivefold increase in risk among HIV-infected patients with AIDS. Following exclusion of subjects with a recent transfusion (any that occurred within 3 years of the date of the case patient's NHL diagnosis date, for each case-control pair), a possible increased risk for AIDS-NHL remained among HIV-positive subjects who underwent a transfusion (five case patients and one control subject exposed), but this increased risk of NHL was not confirmed among the HIV-negative subjects. Multiple transfusion episodes were not associated with greater risk of NHL. Among HIV-negative subjects, results were similar when males and females were examined separately (data not shown). When all 113 AIDS-NHL case patients were compared with the 50 AIDS control subjects, the results were similar to those for the 50 matched pairs shown in the table (data not shown). As discussed by others ( 10 ) , all such studies are potentially confounded by the underlying illnesses for which transfusions are given. It is also possible that receiving a blood transfusion, especially in the late 1970s to early 1980s, may have increased a patient's likelihood of exposure to HIV. Thus, in the study of Blomberg and colleagues ( 5 ) , in which blood transfusions were given during 1981-1982, of the seven patients who developed NHL, three developed extranodal high-grade NHL, a disease type that is particularly associated with AIDS ( 11 , 12 ) . This is also relevant when considering the elevated risk of high-grade extranodal lymphoma reported by Brandt and colleagues ( 2 ) . In the current study, case patients with AIDS-NHL were restricted to homosexual and bisexual men. For this population, exposure to HIV was through sexual contact and not a blood transfusion. We excluded recent transfusions (within 3 years of the case patient's diagnosis) to rule out transfusions that might have been received because of the subject's underlying illness, and we observed no association. It is possible that low-grade NHL is particularly associated with blood transfusions ( 10 ) . In the study conducted by Brandt and colleagues ( 2 ) , the only statistically significant associations with blood transfusions were for low-grade NHL or with extranodal high-grade NHL. Neither Memon and Doll ( 4 ) nor Cerhan and colleagues ( 3 ) evaluated risk by grade of NHL. In the former study, risk estimates were nonsignificant and based on only four cases. The latter study of older women included four patients with Hodgkin's disease and reported a statistically significant increased risk of lymphoma associated with a history of transfusion. The breakdown of patients by grade in the study by Cerhan and colleagues suggests an association with low-grade—as opposed to intermediate- or high-grade—disease. Thus, the reported relationship between NHL and blood transfusion in previously published studies appears to be restricted to patients with low-grade or extranodal NHL. Our results may differ from those previously reported because we limited our study to patients with high- or intermediate-grade nodal and extranodal NHL and because we distinguished between AIDS-associated NHL and NHL among patients not infected by HIV. Future studies will need to examine carefully whether transfusion recipients are at risk for specific subtypes of NHL.