Nearly 52,700 persons were expected to be diagnosed with Non-Hodgkin's lymphoma (NHL) in the United States during 1996, with 23,300 deaths expected to be attributable to this cancer(1). The Surveillance, Epidemiology and End Results (SEER) program at the National Cancer Institute reported a 50% increase in the incidence of NHL between 1973 and 1987, from 9 cases per 100,000 to 13 cases per 100,000, with an increase of nearly 4% per year among men and 3% per year among women. These increases exceeded those of all other cancers except melanoma of the skin and lung cancer among women(2). The changes in age-adjusted incidence and mortality rates for NHL over time have occurred worldwide, and although some of the increase may result from better diagnostic capabilities, most of it remains unexplained(3).
Individuals infected with the human immunodeficiency virus (HIV) are at increased risk for the development of high-grade NHL(4). Most investigators have reported a 5% to 10% incidence of NHL among persons infected with HIV, as reviewed by Obrams and Grufferman in 1991(5). Between 1980 and 1986, a fivefold increase in the incidence of NHL was reported among never-married men who lived in census tracts with a high rate of acquired immunodeficiency syndrome (AIDS)(6,7). In 1990, about 2000 HIV-associated lymphomas occurred in the United States(8), and a report in 1992 estimated that 10% of all lymphomas would be HIV related by 1994 and that this proportion would increase over time(9). An analysis of the data collected by the SEER program in San Francisco in 1985 demonstrated a 10-fold increase in the annual incidence of NHL among never-married men in the San Francisco Bay Area, with a more than 100-fold increase in high-grade NHL among these men(10). An excess risk of 225 cases of NHL per 100,000 person years was attributable to HIV infection among homosexual men in the San Francisco City Clinic cohort study, and an age-adjusted morbidity ratio of 38 for NHL also was reported(11).
To better understand the magnitude of this excess incidence in the entire population and that attributable to the HIV epidemic, we conducted a population-based case-control study of NHL in the San Francisco Bay Area. This report describes the medical history, sexual practices, and demographic risk factors for NHL among homosexual men who were diagnosed with NHL and for their homosexual control subjects. Results for occupational and other exposures are presented in a companion report in this issue(12).
MATERIALS AND METHODS
A rapid case-finding system operated by the Northern California Cancer Center (NCCC) was used to identify NHL patients within 1 month of diagnosis in hospitals in the six counties of Alameda, Contra Costa, Marin, San Francisco, Santa Clara, and San Mateo in the San Francisco Bay Area. This method ensures that, relative to their diagnosis date, patients will be identified and interviewed rapidly, removing effects that may be associated with enrolling long-term NHL survivors. The study was conducted between 1988 and 1995. To be eligible for the study, subjects must have been between 21 and 74 years of age and resided in one of the six counties listed at the time of diagnosis. All study cases were diagnosed between January 1988 and July 1994. Information provided by the NCCC included the patient's name, age, race, address, and telephone number: histologic confirmation for diagnosis of NHL (ICDO codes: 9590 to 9594, 9670 to 9704, 9723, and 9750 from the Field Trial Edition of the ICDO, March 1988); diagnosis date and place of diagnosis; and names of physicians who had clinical responsibility for the subjects.
Physicians for the identified persons with NHL were contacted by telephone to determine whether there were medical contraindications to contacting their patients. If there were no contraindications, letters were sent to potential subjects explaining the study and asking them to participate. One to two weeks after the letter was mailed, the subjects were telephoned and asked whether they had questions regarding the study and whether they would like to participate. If the subjects chose to be interviewed, a convenient time was arranged.
Using these procedures, 1593 individuals who had been diagnosed recently with NHL were identified and interviewed. Of the 2812 patients eligible for interview, 593 (21%) had died before they could be contacted. Medical contraindications were reported by subjects' physicians for 78 (2.8%), 275 (9.8%) were too ill to be interviewed, 74(2.6%) could not be located, and 182 (6.5%) patients refused to participate in the study. No proxy interviews were conducted. Of the total number of subjects, 312 homosexual and bisexual men (hereafter called homosexual) with NHL were eligible and completed interviews.
Control subjects were identified using random-digit dial(13-16) and were frequency matched to the patients by sex, county of residence, and by age in 5-year age groups. Randomly generated telephone numbers were dialed a minimum of 10 times on weekdays and weekends at different hours of the day and evening as needed to reach the residents. Seventy-five percent of all control subjects who were contacted completed interviews. A total of 420 homosexual and bisexual men (here-after called homosexual) who were free of NHL were among the 2515 control subjects who completed interviews, and these 420 men formed the control group for this report. If more than one subject was eligible in the household, a random-number procedure was followed to choose among those eligible so that only one subject was interviewed in each household.
All subjects were asked about their HIV status in the questionnaire. At the end of the interview, men were asked if they wished to participate in an optional laboratory portion of the study that included having their blood drawn to test for HIV status. If they chose to participate, a phlebotomist obtained blood specimens within the next few days. One man did not know his HIV status (patient), and two men (controls) did not wish to participate in the blood draw. These three men were omitted from these analyses. For purposes of this analysis, all men are categorized by HIV status according to their response to questions on the questionnaire, although there was 92% agreement between subject response and blood tests for all men for whom blood was drawn. Not all men had blood drawn, because men on chemotherapy were not eligible, and some men chose not to participate. Patients and controls had their blood tested for HIV using the autoantigen enzyme-linked immunosorbent assay (ELISA) with a Western blot assay confirmation.
Structured personal interviews were conducted by trained interviewers in study subjects' homes or at a place convenient to the subject. Interview topics covered family medical history, complete medical history of the subject (including common allergies and other disturbances of the immune system), use of therapeutic drugs, vaccinations and viral infections, sexual history, use of tobacco, alcohol and illicit drugs, demographic characteristics, and other factors. Men were categorized as homosexual based on self-report.
Unconditional logistic regression was used to obtain odds ratios as estimates of relative risks that first were adjusted for age alone and later were adjusted for potential confounders in multivariate models. The 95% confidence intervals for odds ratios and two-sided p values were calculated. Effects are referred to as being somewhat elevated or borderline when p values lie between 0.05 and 0.10.
Possible associations and interactions between predictors were assessed. Necessary precautions were taken to avoid collinearity in the multivariate models. Whenever two or more highly correlated variables that measured similar biologic phenomena were candidates for the multivariate model, only the variable that had greater biologic relevance was retained for entry into the multivariate model. Variables with spurious correlations also were not simultaneously allowed into the multivariate models. When the heterogeneity test for odds ratios applied to stratified data indicated a possible interaction, a logistic model was fit and the interaction term was removed from the model if it failed to attain a 10% level of significance(17).
Because HIV status was associated with increased risk for NHL, all analyses were stratified by HIV status and are presented separately. Age was categorized into eight groups, and all odds ratios were adjusted for age group. Some covariates, especially sexual history characteristics, were correlated with HIV status, and odds ratios for some factors were heterogeneous between the HIV groups. Covariates not inherently categorical were divided into quartiles or tertiles using all control subjects unless a paucity of measurements or meaningfulness of categories forced a slightly different classification. Only subjects' medical conditions that occurred more than 5 years before the date of interview were considered in the analysis. This was done to avoid confounding with NHL diagnoses because of symptoms that may have been related to the onset of NHL and HIV.
The best-fitting model was obtained from SAS software using the forward stepwise procedure, and then backward pruning was done through manual intervention to construct parsimonious models containing biologically relevant covariates. The two-tailed statistical tests that were used included chi-square, Fisher's exact. Wilcoxon rank-sum, the z-statistic tests and the test for homogeneity of odds ratios. Multivariate models were fit separately for each HIV group. A new variable was added to or retained in the model if its coefficient was significant at the 10% level or the factor was thought to be biologically relevant to NHL.
Table 1 gives the distribution by age, by race and ethnic groups, and by HIV status of the 732 homosexual men in this report, 312 with NHL and 420 who formed the control group. About one half of the men were HIV positive (49%) and one half (51%) HIV negative. The median age was 40 years for patients and 39 years for the control subjects. The distribution of NHL and control subjects by the age groups given in Table 1 was not different using the chi-square test (p = 0.14). Time since first laboratory test that indicated HIV positivity was different for men with NHL (median 2 years) and those without NHL (median 4 years, p = 0.0001). The median age of 40 years in the HIV-positive patients with NHL was lower than the median age of 48 years for the HIV-negative patients with NHL (Wilcoxon two-sample test,p = 0.0004). As seen in Table 1, NHL patients were much more likely to be HIV positive than were control subjects, with a 20-fold increased risk for NHL compared with HIV-negative men. Because of this strong association of HIV status with risk for NHL, analyses for most factors are presented separately by HIV status. One NHL patient and two control subjects did not know their HIV status and were ineligible or chose not to have their blood drawn for HIV testing. The analyses by HIV status therefore are based on 311 patients and 418 control subjects.
The proportion of cases among homosexual men attributable to HIV infection can be estimated to be 80% with a 95% confidence interval of 74% to 84%(18). This means that, for all subtypes combined, 80% of the NHL incidence among homosexual men is attributable to HIV infection, assuming that the group of patients in the study is representative of the population of homosexual men with NHL from the San Francisco Bay Area.
Subjects are presented in Table 2 according to their HIV and NHL status and demographic factors. Higher education of subjects and higher family income were associated with a trend toward a decreased risk for NHL among the HIV-negative men, but these factors were not consistently related to NHL risk among the HIV-positive men. A greater proportion of HIV-negative control subjects never had been married compared with the HIV-negative men with NHL. A greater than threefold risk for NHL was detected among HIV-negative homosexual men with five or more siblings, when one or no sibling was the baseline category. When numbers of brothers and sisters of HIV-negative men were examined separately, there was a trend with increasing risk for an increase in number of brothers or sisters. Among the HIV-negative men, a greater than fourfold risk was associated with having four or more brothers relative to one or no brother, with a greater than threefold risk for sisters.
A detailed sexual history was asked in an attempt to determine whether the risk for NHL among homosexual men was related to sexually transmitted agents, and results for these factors are presented in Table 3. There was an increasing trend in risk for NHL with an increase in the total number of sexual partners of the opposite that was of borderline significance among the HIV-negative men, with the suggestion of an increased risk among the HIV-positive men. In contrast, there was a decreasing trend in risk for NHL with increasing total number of partners of the same sex among the HIV-negative men, with reduced odds ratios for more partners of the same sex also among the HIV-positive men. When number of partners of the same sex was examined by age interval, HIV-positive and HIV-negative NHL patients had had fewer partners between the ages of 30 and 49 than their respective control subjects. When all men were combined, marital status was related to the number of sexual partners in that men who never had been married to women had had more partners of the same sex (p < 0.0001) and fewer partners of the opposite sex (p < 0.0001). Ever having had receptive anal intercourse was related to a reduced risk for NHL among HIV-negative subjects with a somewhat decreased risk among HIV-positive men, although there were few HIV-positive men who had had no anal intercourse. Frequency of anal intercourse was associated with a reduced risk for NHL among men between the ages of 20 and 29 and of 30 and 39 in both HIV categories, as was ever having engaged in rimming (i.e., anal-oral contacts).
Results of analyses of tobacco and illicit drug use are presented in Table 4. Although risk ratios were slightly elevated for the number of cigarette smokers in the household during childhood among HIV-positive men, this result could have resulted from chance. A fourfold risk for NHL was observed among the HIV-negative men who were nonsmokers, with a significant trend of increasing risk with increasing number of smokers in the household 5 years before the interview. There were no differences in the history of cigarette smoking between patients and control subjects (data not shown). Use of other tobacco products could not be analyzed because of the small numbers. When lifetime consumption of alcoholic beverages was considered in tertiles, the HIV-negative men had a decreasing trend in risk for NHL with an increase in wine consumption (lower third: OR = 0.49; medium: OR = 0.19; highest: OR = 0.32; trend: p= 0.001). There was no relation between NHL and any other type of alcohol use.
Other substance use included marijuana, cocaine, crack, amyl or butyl nitrate (i.e., poppers), speed, crank or amphetamines, LSD (i.e., acid), and other social drugs(Table 4). Substance abuse was more common among the younger than the older men, and there was some statistical interaction between age and illicit drug use for most drugs, especially among the HIV-positive subjects. However, when odds ratios and confidence intervals were calculated separately for subjects younger than 40 years of age and those 40 years or older, small numbers of nonusers resulted in wide and overlapping confidence intervals for the two age groups for all drugs except cocaine or crack. Only common age-adjusted odds ratios and confidence intervals are presented in Table 4. Cocaine or crack use was associated with a somewhat increased risk for NHL among the younger HIV-positive men (1 to 19 times: OR = 1.9, CI = 0.81 to 4.4; 20+ times: OR = 1.7, CI = 0.78 to 3.9), but was associated with a decreased risk for NHL among the older HIV-positive men (1 to 19 times: OR = 0.17, CI = 0.06 to 0.60; 20+ times: OR = 0.16, CI = 0.06 to 0.49; trend: p= 0.002). Decreased risk for NHL also was associated with an increase in the lifetime frequency of speed, crank, or amphetamine use in HIV-positive subjects. Among the HIV-negative subjects, lifetime use of nitrates was associated with a lower incidence of NHL, and lifetime use of speed, crank or amphetamines showed a borderline effect. Other social drugs used by the subjects included angel dust (i.e., PCP), magic mushrooms (i.e., psilocybin), MDM, MDA or Ecstasy powder, and Quaalude, or Mandrax. These drugs were not used as frequently as those presented in Table 4, and their use had a similar effect on NHL status among HIV-positive and HIV-negative subjects. When pooled over HIV status, ever use of Quaalude, or Mandrax, was associated with an increased risk for NHL (OR = 2.0, CI = 1.0 to 3.7, p = 0.04).
Odds ratios associated with allergies, bee or wasp stings, therapeutic drug use, and vaccinations are presented in Table 5. HIV-positive and HIV-negative patients were less likely to have had allergies to anything other than medications compared with the control subjects. Odds ratios for individual allergies were consistently below 1.0 for HIV-positive and HIV-negative men. Allergies to grass, hay, trees, leaves, other plants, or pollen and to dust, chalk, or mold were associated with a reduced risk for NHL among HIV-positive and HIV-negative men. Allergies to foods were associated with a reduced risk for NHL among HIV-negative men. Detailed analysis of insect stings or bites showed HIV-positive NHL patients to have been less likely than HIV-positive control subjects ever to have been bitten by animals or insects other than fleas and mosquitoes (OR = 0.36, CI = 0.18 to 0.72). In particular, these patients were less likely to report having been stung by bees or wasps.
When therapeutic drug use for 4 consecutive weeks or longer was examined for its relation to NHL, control subjects were found to have been more likely than NHL patients to have used antihistamines or other allergy medications. Among HIV-positive men, control subjects were more likely to have used cimetidine. Results for history of vaccinations showed similar risk ratios for HIV-positive and HIV-negative men for nearly all immunizations. HIV-positive NHL patients were more likely to have been vaccinated against measles and somewhat more likely to have been vaccinated against rubella than were HIV-positive control subjects, but they were less likely to have been vaccinated against influenza, to have received hepatitis vaccine, and to have been vaccinated against poliomyelitis before 10 years of age. Analysis among HIV-negative men showed that NHL patients were less likely to have been vaccinated against influenza, tetanus, smallpox, and hepatitis and also were less likely to have been vaccinated against poliomyelitis before 10 years of age.
There was no association between risk for NHL and any of the following factors among HIV-positive or HIV-negative homosexual men: allergic reactions to medications: use of Valium: having had vaccinations for diphtheria-pertussis-typhoid, cholera, or yellow fever; having had a tonsillectomy or appendectomy; and having had blood transfusions or numerous diagnostic radiographs. There also was no association with ever having had the following conditions more than 5 years before diagnosis with NHL: hepatitis; shingles; mononucleosis; chickenpox; mumps; measles; pneumonia; whooping cough; scarlet fever; parasitic disease; heart, kidney, or thyroid disorders; cold sores; eczema; gonorrhea; syphilis; herpes; anal or genital warts; or other cancers. There were no consistent associations by group with usual weight as an adult, ever having been overweight by 20 pounds or more, or religion as a child. Nor did diseases and conditions among family members of subjects with NHL differ from those found in control subjects' families.
Multiple logistic models for HIV-positive subjects and for HIV-negative subjects are presented in Table 6. For each of the two models, each factor presented is mutually adjusted for each other and for age; family income is in the model for the HIV-negative men. Among the HIV-positive men, most of the difference between NHL patients and control subjects was explained by frequency of anal intercourse between the ages of 20 and 29; allergy to grass, hay, trees, leaves, other plants, and pollen (p = 0.0007); number of bee or wasp stings; use of cimetidine for 4 consecutive weeks or longer (p = 0.03); vaccination against influenza (p = 0.003); and lifetime use of speed, crank, or amphetamine. Table 6 also shows that several factors were associated with NHL among HIV-negative subjects. Frequency of anal intercourse between the ages of 20 and 29, nonmedication allergies (p = 0.02), early vaccination against poliomyelitis (p = 0.05), and number of siblings (p = 0.001) showed an association with NHL status.
Individuals infected with HIV are at increased risk for NHL(5,10), and the risk for NHL in individuals with AIDS has been computed to be 60 times that of the general population(19). The data from this population-based study showed that HIV infection resulted in a 20-fold increase in the risk for NHL among homosexual men. Others have reported that subjects with AIDS-related lymphomas are considerably younger than those with NHL not related to HIV infection(20), and our data supported these findings. The HIV-positive NHL patients were an average of 8 years younger than the HIV-negative NHL patients. Because men with NHL often have a rapidly progressive course of disease, results related to HIV are likely to be underrepresented and should be interpreted with caution because of the possibility of selected survival of the men in our study.
This study was designed to examine several a priori hypotheses about potential risk factors for NHL suggested by others, to identify and examine new potential risk factors in the general population, and to study whether the magnitude of their effect varied according to HIV status in the high-risk subpopulation of homosexual men. Some of the risk factors of interest were common disorders of the immune system, such as common allergies, poliomyelitis and other vaccinations, use of cimetidine, number of siblings, history of sexually transmitted diseases, and homosexual experiences among men. Some of these risk factors were confirmed in the subgroup of homosexual men, and the effect of many factors differed between HIV-positive and HIV-negative patients. Results from the multivariate models provided more precise estimates of the effects of various risk factors after control for other relevant factors.
The HIV-negative homosexual NHL patients from our study were more likely to have been married compared with the HIV-negative homosexual control subjects. A previous study had found no association between marital status and risk for NHL in the general population(21), and to our knowledge, marital status in relation to NHL has not been explored in the homosexual population. Our data also showed a decreased risk for NHL among men with more same-sex sexual activity, which was more common among the never-married men. Other factors such as rimming that are related to increased sexual activity were associated with a decreased risk for NHL among HIV-positive and HIV-negative men. Both groups had a reduced risk for NHL with an increase in the frequency of same-sex partners and, in particular, of receptive anal intercourse, most notably between the ages of 20 and 29 and of 30 and 39. These results are consistent with increased immunosuppressive effects of seminal fluid(22-25,25a,25b,25c). Early studies suggested that exposure to semen through receptive anal intercourse affected T cell subsets in homosexual men(25a) and resulted in higher levels of semen-related antigen and antibodies in serum when compared with heterosexual men(25b,25c). The results of animal studies are consistent with our results. Pooled rabbit semen placed into the rectum of healthy male rabbits resulted in the appearance of antibodies to sperm and to peripheral blood lymphocyte antigens, suggesting that repeated rectal deposition of semen may compromise immune system functions(26). Our results also are consistent with the somewhat elevated risk for total number of partners of the opposite sex, because men who were married had fewer same-sex partners and would not be exposed as often to the immunosuppressive effects of semen on sensitive rectal tissue. Earlier studies have suggested that HIV-related lymphomas may be outgrowths of antigen-driven B cells(27), and this topic is explored in detail in another work on AIDS retroviruses(28). Subjects who have had more sexual activity would have a source of immunosuppressive factors, such as semen, viruses, and other infectious organisms, and suppression of the immune system may prevent or decrease the frequency of antigen-driven lymphomas. If these associations have a biologic basis, they may be related to the occurrence of an antigen-driven process early in lymphomagenesis(27,28).
Data from a college alumni cohort study showed no association between NHL and only-child status or with number of siblings(29), whereas a report from Yorkshire found an elevated risk for NHL in single-child families(30). In contrast, our study indicated an increased risk for NHL associated with five or more siblings among HIV-negative homosexual men. Children with many siblings are likely to have numerous viral infections when young, and it has been postulated that more viral infections as a child produce fewer allergies as an adult(31). Although allergies were found to be highly protective against NHL in our study, an increased risk for NHL was associated only with a large number of siblings.
We found allergies to anything other than medications to be associated with a decreased risk for NHL among HIV-positive and HIV-negative men. Similar to our results were those of a study conducted in Los Angeles that reported that allergy to nuts and berries among men and allergy to insect bites or stings among women were associated with lower risk for NHL(32); however, an Italian study provided no such associations(33). Others have reported a reduced risk for NHL associated with allergic disorders(34) and with hay fever (35). We also found a decreased risk for NHL associated with bee or wasp stings among HIV-positive homosexual men. A study from Denmark reported that stings from honey bees and wasps are of clinical significance and that these insects contain separate allergenic components(36).
With respect to medical history, vaccination against influenza, hepatitis, and early vaccination against poliomyelitis was protective for NHL among the HIV-positive men, but vaccination against rubella and measles were associated with increased risk ratios in the same group. Among the HIV-negative men, vaccination against poliomyelitis before the age of 10 was related to a reduced risk for NHL, as were vaccinations against influenza, tetanus, smallpox, and hepatitis in the univariate results. Vaccination against cholera and yellow fever was related to reduced risk for NHL in the Los Angeles study(32). Similarly, in our study, risk ratios were below 1.0, although confidence limits overlapped unity.
Univariate analyses of factors corresponding to medical history and past therapeutic drug use reported from earlier studies and based on the general population also have had varying results. Among viral diseases, an elevated risk for NHL was associated with previous infections with herpes zoster(33) and with chickenpox (29). Herpes simplex virus, most likely to have been a reflection of immunosuppression, also was reported to have been associated with development of NHL among patients with advanced HIV infection(37). A report from northeastern Italy indicated an association between a history of various chronic infectious diseases (e.g., tuberculosis, pyelonephritis, chronic bronchitis, malaria) and increased risk for NHL(38). Two other reports from Italy and England found that scarlet fever and shingles were associated with an increased risk for NHL(30,33). The English results showed an elevated risk for NHL associated with a previous history of diabetes, pneumonia, other malignancy, eczema, and other miscellaneous skin conditions(30), but a report from Los Angeles county showed history of eczema to be associated with a reduced risk for NHL(32). Autoimmune disorders and tonsillectomy also have provided mixed results, with some studies showing positive associations with NHL(33) and others reporting no associations (32). Our results showed no association between these factors and the risk for NHL among HIV-positive or HIV-negative homosexual men.
An elevated risk for NHL was associated with the use of aspirin, penicillin or other antibiotics, and steroids in the Los Angeles study(32). Among HIV-positive men, we found a protective association for the intake of cimetidine, a histamine antagonist, for 4 consecutive weeks or longer. Because NHL occurs at much higher rates in those who are immunosuppressed, the results regarding allergies and antihistamines bear more relevance in this context. Histamine is one of several mediators that helps to regulate immune response, and it has been reported that cimetidine can block histamine release and inhibit suppressor cell function(39). Others have reported that patients who received cimetidine exhibited enhanced cell-mediated immunity(40,41).
Analyses concerning substance abuse in our study indicated that the use of speed, crank, or amphetamines was associated with reduced risk ratios for NHL among both groups, although these results were only somewhat related in HIV-negative men. Use of other drugs reported in this study also was associated with a nonstatistically significant but consistently reduced risk for NHL. Studies using rats and monkeys have shown an immunosuppressive effect of drugs(42-44) on helper T-cell function and on the absolute number of circulating lymphocytes. When antigen drive is involved, mild immunosuppression such as that caused by drug use may slow the lymphogenic process, because fewer B cells are at risk for division.
This study provides strong support for a role of HIV infection in the incidence of NHL, and we estimate that 80% of the incidence of NHL among homosexual men was attributable to HIV infection. Although we do not have precise figures, we think this estimate is conservative because of the high death rate among the HIV-positive men with NHL. HIV infection was strongly related to many sexual practices measured in this study and to other determinants that appeared to be associated with the incidence of NHL before stratification by HIV status. The findings for a history of more sexual partners and of receptive anal intercourse while young, allergies to plant materials, bee or wasp stings, therapeutic use of cimetidine for 4 weeks or more, and vaccination against influenza may provide clues for further exploration of the cause and immunology of NHL among those infected with HIV. Among HIV-negative homosexual men, some clues to the etiology of NHL may be provided by history of more sexual partners and of receptive anal intercourse when young, nonmedication allergies, and poliomyelitis vaccination at an early age. The reduced risk associated with an increased number of allergies involves complex immunologic processes. To better understand these associations requires multidisciplinary efforts in large, carefully designed epidemiologic studies. If found in other populations, these results may provide insights into the mechanisms of NHL pathogenesis.
Acknowledgments: The authors thank Jennifer Kristiansen for her assistance in this study and Dr. Michael McGrath for his thoughts regarding immunologic processes. This work was supported in part by grant numbers R01-CA45614 and U01-CA66529 from the National Cancer Institute, National Institutes of Health.
1. Parker SL, Tong T. Bolden S, Wingo PA. Cancer statistics, 1996. CA Cancer J Clin
2. Devea SS, Fears T. Non-Hodgkin's lymphoma time trends: United States and international data. Cancer Res
3. Hartge P, Devesa SS, Fraumeni Jr JF. Hodgkin's and non-Hodgkin's lymphomas. Cancer Surv
4. Ziegler JL, Beckstead JA, Volberding PA, et al. Non-Hodgkin's lymphoma in 90 homosexual men: relation to generalized lymphadenopathy and the acquired immunodeficiency syndrome. N Engl J Med
5. Obrams GI, Grufferman S. Epidemiology of HIV-associated non-Hodgkin's lymphoma. Cancer Surv
6. Harnly ME, Swan SH, Holly EA, Kelter A, Padian N. Temporal trends in the incidence of non-Hodgkin's lymphoma and selected malignancies in a population with a high incidence of acquired immunodeficiency syndrome (AIDS). Am J Epidemiol
7. Horn PL, DeLorenze GN, Brown SR, Holly EA, West DW. Re: "Temporal trends in the incidence of non-Hodgkin's lymphoma and selected malignancies in a population with a high incidence of acquired immunodeficiency syndrome (AIDS)" [letter]. Am J Epidemiol
8. Rabkin CS, Biggar RJ, Horm JW. Increasing incidence of cancers associated with the human immunodeficiency virus epidemic. Int J Cancer
9. Biggar RJ, Rabkin CS. The epidemiology of acquired immunodeficiency syndrome-related lymphomas. Curr Opin Oncol
10. Rabkin CS, Goedert JJ. Risk of non-Hodgkin's lymphoma and Kaposi's sarcoma in homosexual men. Lancet
11. Hessol NA, Katz MH, Liu JY, Buchbinder SP, Rubino CJ, Holmberg SD. Increased incidence of Hodgkin disease in homosexual men with HIV infection. Ann Intern Med
12. Holly EA, Lele C, Bracci P. Non-Hodgkin's lymphoma in homosexual men in the San Francisco Bay Area: occupational, chemical and environmental exposures. J Acquire Immune Defic Syndr Hum Retrovirol
13. Waksberg J. Sampling methods for random digit dialing. J Am Stat Assoc
14. Hartge P, Brinton LA, Rosenthal JF, Cahill JI, Hoover RN, Waksberg J. Random digit dialing in selecting a population-based control group. Am J Epidemiol
15. Olson SH, Kelsey JL, Pearson TA, Levin B. Evaluation of random digit dialing as a method of control selection in case-control studies. Am J Epidemiol
16. Lele C, Holly EA. Roseman DS, Thomas DB. Comparison of control subjects recruited by random digit dialing and area survey. Am J Epidemiol
17. Breslow NE, Day NE. Statistical methods in cancer research, vol 1: the analysis of case-control studies
. IARC Scientific Publication 32. Lyon: International Agency for Research on Cancer, 1980.
18. Walter SD. The estimation and interpretation of attributable risk in health research. Biometrics
19. Beral V, Peterman T, Berkelman R. Jaffe H. AIDS-associated non-Hodgkin lymphoma. Lancet
20. Levine AM. Non-Hodgkin's lymphomas and other malignancies in the acquired immune deficiency syndrome. Semin Oncol
21. Scherr PA, Hutchison GB, Neiman RS. Non-Hodgkin's lymphoma and occupational exposure. Cancer Res
22. James K, Harvey J, Bradbury AW, Hargreave TB, Cullen RT. The effect of seminal plasma on macrophage runction—a possible contributory factor in sexually transmitted disease. AIDS Res
23. Turner MJ, White JO, Soutter WP. Human seminal plasma inhibits the lymphocyte response to infection with Epstein-Barr virus. Gynecol Oncol
24. Kelly RW, Quayle AJ, Wallace EM, Wu FCW, Hargreave TB. James K. Immunosuppression by seminal plasma from fertile and infertile men: inhibition of natural killer cell function correlates with seminal PG concentration. Prostaglandins Leukot Essent Fatty Acids
25. Skibinski G, Kelly RW, Harrison CM, McMillan LA, James K. Relative immunosuppressive activity of human seminal prostaglandins. J Reprod Immunol
25a. Detels R, Schwartz K, Visscher BR. Fahey JL, Greene RS, Gottlieb MS. Relation between sexual practices and T-cell subsets in homosexually active men. Lancet
25b. Witkin SS, Sonnabend J. Immune responses to spermatozoa in homosexual men. Fertil Steril
25c. Witkin SS, Bongiovanni AM, Yu IR, Goldstein M, Wallace J. Sonnabend J. Humoral immune responses in healthy heterosexual, homosexual and vasectomized men and in homosexual men with the acquired immune deficiency syndrome. AIDS Res
26. Richards JM, Bedford JM, Witkin SS. Rectal insemination modifies immune responses in rabbits. Science
27. Ng VL, Hurt MH, Fein CL, et al. IgMs produced by two acquired immune deficiency syndrome lymphoma cell lines: Ig binding specificity and VH-gene putative somatic mutation analysis. Blood
28. Ng VL, Hurt MH, Herndier BG, Fry KE, McGrath MS. VH gene use by HIV type 1-associated lymphoproliferations. AIDS Res Hum Retroviruses
29. Paffenbarger Jr RS, Wing AL, Hyde RT. Characteristics in youth predictive of adult-onset malignant lymphomas, melanomas, and leukemias: brief communication. J Natl Cancer Inst
30. Cartwright RA, McKinney PA, O'Brien C, et al. Non-Hodgkin's lymphoma: case control epidemiological study in Yorkshire. Leuk Res
31. Martinez FD. Role of viral infections in the inception of asthma and allergies during childhood: could they be protective? Thorax
32. Bernstein L, Ross RK. Prior medication use and health history as risk factors for non-Hodgkin's lymphoma: preliminary results from a case-control study in Los Angeles County. Cancer Res
33. La Vecchia C, Negri E, Franceschi S. Medical history and the risk of non-Hodgkin's lymphomas. Cancer Epidemiol Biomarkers Prev
34. Tielsch JM, Linet MS, Szklo M. Acquired disorders affecting the immune system and non-Hodgkin's lymphoma. Prev Med
35. Doody MM, Linet MS, Glass AG, et al. Leukemia, lymphoma, and multiple myeloma following selected medical conditions. Cancer Causes Control
36. Mosbech H, Dahl R, Malling HJ, Pedersen S, Svendsen UG. Allergy to insect stings. Ugeskr Laeger
37. Moore RD, Kessler H, Richman DD, Flexner C, Chaisson RE. Non-Hodgkin's lymphoma in patients with advanced HIV infection treated with zidovudine. JAMA
38. Franceschi S, Serraino D, Bidoli E, et al. The epidemiology of non-Hodgkin's lymphoma in the northeast of Italy: a hospital-based case-control study. Leuk Res
39. Stanciu L. Immunomodulation by histamine. Ann Biol Clin
40. Avella J. Binder HJ, Madsen JE, Askenase PW. Effect of histamine H2
-receptor antagonists on delayed hypersensitivity
41. White WB, Ballow M. Modulation of suppressor-cell activity by cimetidine in patients with common variable hypogammaglobulinemia. V Engl J Med
42. Chuang LF, Killam KF, Chuang RY. Opioid dependency and Thelper cell functions in rhesus monkey. In Vivo
43. Flores LR, Wahl SM, Bayer BM. Mechanisms of morphineinduced immunosuppression: effect of acute morphine administration on lymphocyte trafficking. J Pharmacol Exp Ther
44. Di Francesco P, Pica F, Croce C, Favalli C, Tubaro E, Garaci E. Effect of acute or daily cocaine administration on cellular immune response and virus infection in mice. Nat Immun Cell Growth Regul