Non-Hodgkin lymphoma

IN THE UNITED STATES, the increase in non-Hodgkin's lymphoma (NHL) mortality between 1950 and 1980 occurred disproportionately in the agricultural midwest.[1] In several epidemiologic studies, researchers examined NHL risk from pesticide exposures in this geographic area,[2-7] but excess risk remains substantially unexplained.[8]

The increase in fertilizer use[9] during the past 40 y has contributed to nitrate contamination of drinking water in rural areas. Some individuals have hypothesized that nitrate may be a risk factor for cancer through its role as a precursor compound in the formation of N-nitroso compounds, many of which are animal carcinogens.[10] The results of a recent population based case-control study in Nebraska suggested that long-term consumption of nitrate in drinking water has contributed to NHL risk.[11]

We further explored the association between nitrate in drinking water and NHL by evaluating the risk of NHL from average long-term consumption of nitrate in drinking water in Minnesota--a state with heavy nitrogen fertilizer use.[12]

Method

Study framework. The data collected in the present study supplement the data collected in the 1980 study in which in-person interviews of 329 incident cases of NHL and 642 population-based controls were conducted. The design and data collection in the original study, which also included cases of leukemia, are described in detail elsewhere.[2]

Briefly, in the original study, the investigators ascertained all cases of NHL among white men aged 30 y or older who were diagnosed between 1980 and 1982 from Minnesota hospital and pathology laboratory records. Eligibility was limited to cases who resided in areas of Minnesota other than the four largest cities. Four experienced pathologists reviewed all NHL diagnoses and classified them according to the working formulation for NHL.[13] We selected controls randomly from white men who lived in Minnesota and frequency-matched them to NHL and leukemia cases by age in 5-y groups and by vital status at the time of interview.

We conducted in-person interviews with a structured questionnaire. The questionnaire included information about sociodemographic, medical, behavioral, and occupational characteristics, as well as questions about consumption of beverages and dairy products. The questionnaire also contained references to lifetime residential history and historical sources of drinking water. We then asked the participants to provide the location and period of each residence at which they had lived for at least 1 y, as well as what served as their primary source of drinking water at each residence (i.e., community supply, private well, bottled water, and other source).

Study population. Given that historic nitrate levels in private wells are unknown, we restricted our study to cases and controls who had reported using Minnesota community water supplies or bottled water for at least 90% of the years between 1947 and 1980. A total of 73 cases and 147 controls who used water from 157 communities in Minnesota met this criterion. To check our analysis, we also assessed risk with alternative inclusion criteria for the study population, based on varying percentages of years (i.e., 60%-100%) that community water supplies were used in Minnesota.

Classification of exposure to waterborne nitrates. We obtained records on laboratory nitrate measurements (on the basis of nitrogen in nitrate) in Minnesota community water supplies from the Minnesota Department of Health (MDH) for the period between 1947 and 1984. We used nitrate measurements from both public wells and distribution points (e.g., schools) to calculate annual nitrate levels in the drinking water of each community between 1947 and 1980. For any nondetectable nitrate value, we assigned a value of one-half the detection threshold to limit misclassification.

To fill in years for which data were missing and to use all information in the estimation of exposure in years with data, we calculated a time-weighted nitrate average for each year in each community. Measurements within 2 y of the year being estimated were weighted as 1; [is greater than or equal to] 2-5 y, as 0.5; and [is greater than] 5-10 y, as 0.2. We weighted measurements that exceeded 10 y as 0.01 to ensure an estimate for early years (i.e., when some towns had no records). We accorded additional weight to distribution-point measurements, which reflected actual tap-water delivered to community residents.

We linked participants' reported residential histories with estimated nitrate level by year in each community of residence. For any year during which participants reported the use of bottled water, we assumed the nitrate level was 0. We then calculated an estimated arithmetic average nitrate exposure for cases and controls for the period 1947 through 1975 to allow a 5-y lag period from 1980--the earliest date of diagnosis. We also calculated exposure on the basis of an unlagged period.

We set the exposure referent at [is less than or equal to] 0.5 mg/l--half the detection limit prior to about 1975. The remaining average values were grouped into tertiles, with the highest category set at [is greater than] 1.5 mg/l.

To provide a further check on our analysis, we used other algorithms, with different weights applied to nitrate measurements, to estimate average exposure. We also examined the association with maximum nitrate exposure for the years 1947-1975. In setting the maximum exposure reference value at [is less than or equal to] 0.5 mg/l, we grouped the remaining values by tertiles, with the highest category set at [is greater than] 1.9 mg/l.

Statistical methods. We used logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs).[14] To examine potential confounders, we used questionnaire information about age (as a categorical variable), marital status, education, cancer history, occupational exposure, smoking and drinking histories, farming history, use of hair dye, pesticide use, and proximity to industrial sites. Inasmuch as adjusting for potential confounders did not affect the crude OR appreciably, only age-adjusted ORs and 95% CIs are presented herein.

Results

Characteristics identified in the questionnaire were compared with respect to cases and controls (Table 1). Among other things, cases were more likely than controls to be younger, to have a family history of cancer, and to have an occupation associated previously with NHL.

Table 1.--Characteristics of Cases and Controls in the Community Water Analysis(*)

(*) Study participants = individuals who spent at least 90% of their lives between the years 1947 and 1980 consuming community water supplies in Minnesota or bottled water anywhere.

([dagger]) Occupations identified in published literature as associated with non-Hodgkin's lymphoma, including machinists, photoengravers, cosmetologists, embalmers, meat workers, rubber workers, foundry workers, pulp and paper mill workers, leather workers, aircraft and parts workers, and fabricated metal industry workers.

We found no association between NHL and long-term average nitrate exposure from community water supplies within the exposure range of this study (i.e., 0.1-7.2 mg/l [Table 2]). The age-adjusted association (i.e., OR) was 0.3 in the highest category (i.e., [is greater than] 1.5 mg/l), with a median nitrate average of 2.4 mg/l.

Table 2.--Age-Adjusted Odds Ratios (ORs) and 95% Confidence Intervals (CIs) for Non-Hodgkin's Lymphoma and Long-Term Average Nitrate Exposure from Community Water Supplies(*)

(*) Study participants = individuals who spent at least 90% of their lives between the years 1947 and 1980 consuming community water supplies in Minnesota or bottled water anywhere.

([dagger]) Average, based on period between the years 1947 and 1975.

([double dagger]) Categories originally set at 0.5 to [is less than or equal to] .07 and 0.7 to [is less than or equal to] 1.5 were merged because the ORs were similar.

Results were similar for the alternative-exposure algorithms tested. No association with drinking water nitrate was found when we unlagged the exposure variable. When we modified the inclusion criteria (based on percentage of years community water supplies were used), we found no association between NHL and average waterborne nitrate. When direct and surrogate responders were analyzed separately, we again failed to observe an association. Finally, NHL was not associated with maximum nitrate exposure.

Discussion

Our study had several advantages over many previous studies in which cancer risk versus nitrate in drinking water has been studied. In our study, exposure classification was better than in previous studies, which have been ecologic or have been based on short-term measurements of nitrate[15-22] (i.e., designs more vulnerable to misclassification).

We obtained nitrate levels in community water supplies for more than 30 y and linked them with each participant's individual residential history. The method of inputing missing data was the same for cases and controls; therefore, the likelihood of differentially biased estimates was reduced. In only one other study[11] have investigators considered individual exposure histories, and they found a significant association between NHL and long-term average exposures [is greater than or equal to] 4 mg/l among community water supply users; such a level, however, exceeded the estimated exposure of most of the participants in our study and, therefore, does not conflict with our findings.

An additional advantage of our study was that the questionnaire provided information about many risk factors potentially applicable to our population, although none of the factors proved to be confounders with respect to waterborne nitrate. Also, whereas we did not include other waterborne contaminants in our model, other contaminants have not been linked strongly to NHL.[23] Moreover, given our negative finding, had we controlled for other factors that might have been associated with both NHL and nitrate, we likely would not have found a positive finding in our study.

Our findings are also supported by several alternative analyses of our data, which yielded similar results to our initial analysis. There was no association with NHL, regardless of (a) the algorithm selected for the estimation of exposure, (b) whether direct responders or surrogate responders were analyzed, or (c) whether other inclusion criteria were used. Risk of NHL was also not associated with long-term maximum nitrate exposure.

Nonetheless, limitations in our ability to estimate nitrate exposure did exist. Sources of potential exposure misclassification included the absence of information about the amount of water consumed, the contribution of different community wells to the drinking water distributed, and the nitrate intake from drinking water at the worksite. Another potential source of misclassification was the sparseness of community nitrate measurements in much earlier years. Given that the resulting misclassification presumably did not differentiate between cases and controls, it is unlikely that our findings masked a "true" positive association.

The results of government surveys indicate that of the 219 million Americans who use community-water supplies, approximately 92 million ingest drinking water with nitrate levels below 0.3 mg/l.[24] The results of our study suggest that concentrations of nitrate in water of 2.4 mg/l do not pose a risk of NHL and, therefore, provide some assurance of safety to many Americans.

Copyright is not claimed; study is in public domain.

The authors wish to express their appreciation to the participants of the study for their contributions; the Minnesota Department of Health, particularly Jerry Smith and Jean Kahilianen, for their general guidance and assistance with data; Bob Banks of Information Management Services (IMS) for his programming support; Lucy Bill of West-at, Inc., and Casey Boudreau and Diane Gezelter of SRA, Inc., for their contributions to the collection and preparation of data; and Steven Mark and Sholom Wacholder of the National Cancer Institute for their statistical advice.

Submitted for publication March 23, 1999; revised; accepted for publication September 29, 1999.

Requests for reprints should be sent to Dr. D. M. Freedman, Radiation Epidemiology Branch, National Cancer Institute, EPS 7087, 6120 Executive Blvd., Bethesda, MD 20892.

References

[1.] Pickle LW, Mason TJ, Howard N, et al. Atlas of U.S. Cancer Mortality among Whites: 1950-1980. Washington, DC: U.S. Department of Health and Human Services (DHHS), 1987; DHHS pub no 87-2900.

[2.] Cantor KP, Blair A, Everett G, et al. Pesticides and other agricultural risk factors for non-Hodgkin's lymphoma among men in Iowa and Minnesota. Cancer Res 1992; 52:2447-55.

[3.] Hoar SK, Blair A, Holmes FF, et al. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 1986;256: 1141-47.

[4.] Weisenburger DD. Environmental epidemiology of non-Hodgkin's lymphoma in Eastern Nebraska. Am J Ind Med 1990; 18:303-05.

[5.] Zahm SH, Weisenburger DD, Babbitt PA, et al. A case-control study of non-Hodgkin's lymphoma and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) in Eastern Nebraska. Epidemiology 1990; 1:349-56.

[6.] Zahm SH, Blair A. Pesticides and non-Hodgkin's lymphoma. Cancer Res 1992; 52(suppl):5485s-88s.

[7.] Zahm SH, Weisenburger DD, Cantor KP, et al. Non-Hodgkin's lymphoma and the use of atrazine: results from three case-control studies. In: McDuffie HH, Dosman JA, Semchuk KM, Olenchock SA, Senthilselvan A (Eds). Agricultural Health and Safety: Workplace, Environment, Sustainability. Boca Raton, FL: Lewis Pub., 1995; pp 151-56.

[8.] Hartge P, Devesa SS. Quantification of the impact of known risk factors on time trends in non-Hodgkin's lymphoma incidence. Cancer Res 1992; 52(suppl):5566s-69s.

[9.] Hallberg GR. Nitrate in ground water in the United States. In: Follett RF (Ed). Nitrogen Management and Ground Water Protection. Amsterdam, The Netherlands: Elsevier Pub., 1989; pp 35-74.

[10.] National Academy of Sciences. The Health Effects of Nitrate, Nitrite, and N-Nitroso Compounds. Washington, DC: National Academy Press, 1981; pub no 5-1-5-78.

[11.] Ward MH, Mark SD, Cantor KP, et al. Drinking water and nitrate and the risk of non-Hodgkin's lymphoma. Epidemiology 1996; 7:465-71.

[12.] Berry JT. Commercial Fertilizers 1994. Muscle Shoals, Alabama: Tennessee Valley Authority, Environmental Research Center; 1994.

[13.] Rosenburg SA, Berard CW, Brown BW, et al. National Cancer Institute sponsored study of classification of non-Hodgkin's lymphomas: summary and description of a working formulation for clinical usage. Cancer 1982; 49:2112-35.

[14.] SAS Institute, Inc. SAS User's Guide. Cary, NC: SAS Institute, Inc., 1990.

[15.] Suarez-Varela MM, Gonzalez AL, Perez MLT, et al. Concentration of nitrates in drinking water and its relationship with bladder cancer. J Environ Pathol Toxicol Oncol 1993; 12:229-36.

[16.] Heckman CW, dos Campos JLE, Hardoim EL. Nitrite concentration in well water from Pocone, Mato Grosso, and its relationship to public health in rural Brazil. Bull Environ Contam Toxicol 1997; 58:8-15.

[17.] Rademacher JJ, Young TB, Kanarek MS. Gastric cancer mortality and nitrate levels in Wisconsin drinking water. Arch Environ Health 1992; 27:292-94.

[18.] Cuello C, Correa P, Haenszel W, et al. Gastric cancer in Colombia: suspect environmental agents. J Natl Cancer Inst 1976; 57:1015-20.

[19.] Jensen OM. Nitrate in drinking water and cancer in Northern Jutland, Denmark, with special reference to stomach cancer. Ecotoxicol Environ Safety 1982; 6:258-67.

[20.] Weisenburger DD. Potential health consequences of ground-water contamination by nitrates in Nebraska. In: Bogarki I, Kuzelka RD (Eds). Nitrate Contamination Exposure, Consequence, and Control. Berlin, Germany: Springer-Verlag, 1991; pp 309-15.

[21.] Wu Y, Chen J, Oshima H, et al. Geographic association between urinary excretion of N-nitroso compounds and oesophageal cancer mortality in China. Int J Cancer 1993; 54:713-19.

[22.] Law G, Parslow R, McKinney P, et al. Non-Hodgkin's lymphoma and nitrate in drinking water: a study in Yorkshire, United Kingdom. J Epidemiol Commun Health 1999; 53:383-84.

[23.] Cantor K. Drinking water and cancer. Cancer Causes Control 1997; 8:292-308.

[24.] National Research Council. Nitrate and Nitrite in Drinking Water. Washington, DC: National Academy Press, 1995; pub no 1-63.

D. MICHAL FREEDMAN Radiation Epidemiology Branch

KENNETH P. CANTOR MARY H. WARD Occupational Epidemiology Branch National Cancer Institute Bethesda, Maryland

KATHY J. HELZLSOUER The Johns Hopkins University School of Hygiene and Public Health Department of Epidemiology Baltimore, Maryland

COPYRIGHT 2000 Heldref Publications
COPYRIGHT 2000 Gale Group