Life-style and environmental factors in the development of nasal NK/T-cell lymphoma: A case–control study in East Asia



Cases of nasal NK/T-cell lymphoma (NKTCL) occur occasionally in Asian and Latin American countries but rarely in Western countries. The etiological role of life-style and environmental factors in nasal NKTCL was investigated. Five university hospitals in Japan and one each in Korea and China participated in this study; a total of 88 cases and 305 hospital controls were accrued during 2000–2005. The odds ratio (OR) of NKTCL obtained after adjustments of age, sex and country was 4.15 (95% confidence interval (CI), 1.74–9.87) for farmers, 2.81 (CI, 1.49–5.29) for producers of crops, 4.01 (CI, 1.99–8.09) for pesticide users, 11.65 (CI, 1.17–115.82) for residents near garbage burning plants, 2.95 (CI, 1.25–6.95) for former drinkers, and 0.49 (CI, 0.23–1.04) for current smokers. The ORs for crop producers, who minimized their exposure to pesticides by using gloves and glasses, and sprinkling downwind at the time of pesticide use, were 3.30 (95% CI, 1.28–8.54), 1.18 (95% CI, 0.11–12.13) and 2.20 (95% CI, 0.88–5.53), respectively, which were lower than those for producers who did not take these precautions. Exposure to pesticides and chemical solvents could be causative of NKTCL. Taken together, life-style and environmental factors might be risk factors for NKTCL. © 2006 Wiley-Liss, Inc.

Sinonasal NK/T-cell lymphoma (NKTCL) is a lethal midline granuloma that produces necrotic and granulomatous lesions in the upper respiratory tract, especially in the nasal cavity. Histologically, the disease usually shows a polymorphous pattern of proliferation involving large atypical cells, small lymphocytes, macrophages, and plasma cells. Thus, the term polymorphic reticulosis was initially given to this disease.1 The disease usually shows resistance to chemotherapy and pursues a highly aggressive course. However, local control of the upper respiratory lesions with radiotherapy combined with systemic chemotherapy gave a relatively favorable outcome.2

Epidemiological studies have revealed that the disease occurs much more frequently in Asian countries such as Japan, China and Korea3, 4 and Latin American countries5 than in Western countries and is closely associated with Epstein-Barr virus (EBV) infection.6 Meanwhile, different frequencies of p53 and c-kit gene mutations among patients with sinonasal NKTCL in Japan, China and Korea were reported.7, 8 These findings might suggest a causative role for some genetical, environmental and lifestyle factors in sinonasal lymphomagenesis.

Recently, we reported the first case of familial sinonasal NKTCL, affecting a father and 1 of his 6 children.9 They used large amounts of pesticides in a green house. An increase in the risk of developing non-Hodgkin's lymphomas (NHL) among individuals exposed to pesticides was reported.10, 11, 12 In addition, a correlation of exposure to certain pesticides and organochlorines with increased titers of antibodies to EBV was reported.13 Therefore, an epidemiological study to elucidate whether socioenvironmental ambient factors contribute to the development of sinonasal NKTCL was conducted.

Patients and methods

Criteria for sinonasal NKTCL

Patients were admitted with necrotic and granulomatous lesions in the upper respiratory tract. Histologically there was necrotic change of varying degrees and a polymorphous pattern of proliferation involving large atypical cells with an occasional multilobated nucleus and various numbers of lymphocytes, plasma cells and macrophages. An angiocentric pattern of proliferation was occasionally observed. The large atypical cells were negative for CD20 and positive for CD3ε, CD45RO, CD56 and/or TIA-1. For detection of EBV, in situ hybridization using an EBER-1 probe was performed on paraffin-embedded histologic sections from 80 cases, giving positive results in 76 cases (95.0%).


Eligible cases were defined as new patients who were diagnosed with NKTCL at age 20 years or older and were admitted to the hospitals of Osaka University, University of Ryukyus, Asahikawa Medical College, and St. Marianna University in Japan, Yonsei University in Seoul, Korea, and China Medical University in Shenyang, China. One hundred and twenty-six patients with NKTCL were admitted to these hospitals during the period from March 2000 to March 2005, but 2 of the patients were excluded from further analysis because adequate data could not be obtained. In China, 34 cases were originally registered for the study, but 26 cases were excluded because of lack of immunohistochemical confirmation of the NK/T-cell nature of tumor cells, an in situ hybridization study for EBV, or data in the questionnaire. As a result, 88 eligible cases were accumulated: 65 from Japan, 15 from Korea and 8 from China. The age range and sex ratio (M/F) among Japanese, Korean and Chinese patients were 25–89 (median 60) years and 1.41, 25–77 (median 43) years and 0.88 and 26–77 (median 43.5) years and 3.00, respectively (Table I).

Table I. Brief Summary of Cases and Controls from the Different Countries
NumberAge (y)M/F ratioNumberAge (y)M/F ratioType of disease
Japan656025–891.411735520–780.71Inflammation of ear (62), nose (49), larynx (7), and tonsil (20), Hearing problem (14), Vertigo (8), Others (13)
Korea154325–770.881124520–731.95Inflammation of tonsil (32) and nose (32), Nasal polyp (32), Septal deviation (16)
China843.526–773.00204740–741.86Inflammation of tonsil (7), nose (6), and ear (3), Others (4)


Controls were obtained by a consecutive series, i.e., from subjects aged 20 years or older, with inflammatory diseases, hearing problems, benign cystic diseases, and medical check up in the otolaryngological regions, who were admitted to the same hospitals and during the same period as the NKTCL patients (Table I). They were selected at random for the present study. They were included such that the number of controls should be 3–4 times the number of NKTCL cases. A total of 305 controls including normal individuals were accrued during the same period as the admitted cases. The age and sex ratio (M/F) of controls were 21–78 (median 51) years and 1.09. The kinds of occupations in the cases, controls and general population by the district studied are shown in Table II. In Japan, the proportion of farmers among the controls was 1.5% (2/137) and 5.6% (2/36) in urban areas (Osaka University and St. Marianna University) and rural areas (Asahikawa Medical University and University of Ryukyu), respectively. The proportion in the general population is 0.5–1.0% and 5.5–6.1% in urban and rural areas, respectively.14 In Korea, the proportion of farmers among the controls was 5.4% (6/112), which was higher than that in the general population (0.1% in Seoul area)15; whereas in China, the proportion of farmers among the controls was 10% (2/20) lower than that in the general population (35.7% in Shenyang, China).16

Table II. Ratio of Farmers Among the Cases and Controls
HospitalDistrictCasesControlsProportion of farmers in general population
Total No.No. of farmersProportion of farmersTotal No.No. of farmersProportion of farmers
Osaka UniversityOsaka (Japan)38410.5%12121.7%0.5%1
St. Marianna UniversityKanagawa (Japan)300.0%160.0%.1%1
Asahikawa Medical SchoolHokkaido (Japan)5120.0%16212.5%6.1%1
Ryukyus UniversityOkinawa (Japan)19526.3%2000%5.5%1
Yonsei UniversitySeoul (Korea)15213.3%11265.4%0.1%2
China Medical UniversityShenyang (China)8225.0%20210.0%35.7%3
Total 881415.9%305123.9% 

Data collection and analysis

A structured self-administered questionnaire was filled out by the patients with NKTCL and the controls at the time of admission. The questionnaire contained information on the following items: past medical history, familial medical history, history of alcohol drinking and cigarette smoking, occupation, cultivation of crops, pesticide use (herbicide, insecticide and fungicide) and environment of residence. Current drinkers or smokers were defined as those who have continued drinking or smoking and former drinkers or smokers, as those who discontinued drinking or smoking before visiting the hospital.

The software package STATA (Stata, College Station, TX) was used for statistical analysis. Statistical significance and the 95% confidence interval (CI) of the odds ratio (OR) were calculated using an unconditional logistic regression model adjusted for age, sex and country and judged by using a 2-sided test. Age-adjustment was conducted by using categorical dummy variables of less than 40, 40–59 and 60 and over.


A comparison of smoking and drinking habits between cases and controls is made in Table III. For calculation of the OR, individuals who were not exposed to each factor or individuals for whom information on exposure to each factor was unknown were categorized as unknown. There was no association between tobacco smoking and NKTCL development. The risk of NKTCL was increased in former drinkers (OR, 2.95; 95% CI, 1.25–6.95), but not in current drinkers, when compared with the risk among nondrinkers. Among current or former drinkers, the OR was 1.23 for hard liquor drinkers and 1.48 for frequent drinkers (more than 3 days per week), but results were not statistically significant. As for the risk estimation of pesticide use for the development of NKTCL, 9 items of occupation were set in the questionnaire, i.e., farmer, forest, fisheries, chemical factory service, staff except chemical factory service, self-employed individuals, special workers and other occupations including individuals who hold no work. Special workers include doctors, lawyers, certified public accountants, estate surveyors, licensed tax accountants and judicial scriveners. Based on the chance of exposure to chemicals, occupations were categorized into 4 groups (Table IV). Farming was strongly associated with the risk of NKTCL (OR, 4.15; 95% CI, 1.74–9.87), whereas self-employment and special work were inversely associated with the risk of NKTCL (OR, 0.32; 95% CI, 0.12–0.85 and OR, 0.36; 95% CI, 0.14–0.88) (Table IV).

Table III. Risk of Nasal NK/T-Cell Lymphoma in Relation to Smoking and Drinking Habits
 Number of cases N = 88Number of controls N = 305Odds ratio195% CI
  • CI, confidence interval.

  • 1

    Adjusted for age (>40, 40–59, <60), sex (male/female), drinking/smoking habits and country (Japan, Korea and China).

  • 2

    2 More than 3 days per week.

Tobacco smoking
 Former smoker25680.880.441.74
 Current smoker15830.490.231.04
Alcohol drinking
 Former drinker14212.951.256.95
 Current drinker351461.080.592.00
 Type of drinker
  Hard liquor drinker11441.230.552.73
  Frequent drinker229691.480.822.68
Table IV. Risk of Nasal NK/T-Cell Lymphoma in Relation to Occupation
 Number of cases N = 88Number of controls N = 305Odds ratio195% CI
  • CI, confidence interval.

  • 1

    Adjusted for age (<40, 40–59, >60), sex (male/female) and country (Japan, Korea and China).

Chemical factory service471.530.425.57
Self-employed individuals5470.320.120.85
Special workers6490.360.140.88

To further evaluate the factors correlated with the development of NKTCL, the state of cultivation of crops and pesticide use were analyzed (Table V). Cultivation of crops was associated with the disease (OR, 2.81; 95% CI, 1.49–5.29). The risk was higher (OR: 5.08; 95% CI 2.47–10.43) among individuals who had cultivated crops for more than 5 years than among individuals without experience in cultivation or with less than 5 years of experience. The OR for garden farming was 2.66 (95% CI, 1.34–5.27). Pesticide users were at greater risk (OR, 4.01; 95% CI, 1.99–8.09) than nonusers. All kinds of pesticides increased the risk for NKTCL; the risk was highest for fungicides (OR, 6.05; 95% CI, 1.98–18.46). As for the method of sprinkling, dusting was a risk factor (OR, 6.94; 95% CI, 1.71–28.20). The ORs for the pesticide users who avoided exposure to pesticides by using gloves and glasses and sprinkling downwind were 3.30 (95% CI, 1.28–8.54), 1.18 (95% CI, 0.11–12.13) and 2.20 (95% CI, 0.88–5.53), respectively, which were lower than those for users who did not, i.e., 4.76 (95% CI, 1.93–11.72), 4.52 (95% CI, 2.17–9.42) and 8.45 (95% CI, 3.01–23.70), respectively. As for masks, the OR in the pesticide user who wore mask was higher than those who did not. The OR for residents near (within 500 m of) a garbage incineration plant, golf course and chemical factory was 11.65 (95% CI, 1.17–115.82), 2.21 (95% CI, 0.58–8.42) and 0.32 (95% CI, 0.04–2.68) with a p-value of 0.036, 0.243 and 0.296, respectively. These findings, however, are of limited value because the number of cases and controls was small, i.e., 4 and 1, respectively.

Table V. Risk of Nasal NK/T-Cell Lymphoma in Relation to Cultivation of Crops and Pesticide Use
 Odds ratio (OR)1Number of cases N = 88Number of controls N = 30595% CI
  • CI, confidence interval.

  • 1

    Adjusted for age (<40, 40–59, >60), sex (male/female) and country (Japan, Korea and China).

Cultivation of crops
 At present2.8127361.495.29
 More than 5 years5.0824192.4710.43
Cultivation method
 Garden farming2.6622281.345.27
 Green house1.84350.408.36
 Type of pesticide
 Method of sprinkling
  Granular powder2.73430.5712.93
  Gloves used3.3010111.288.54
  Gloves not used4.7613121.9311.72
  Mask used5.4414102.2013.47
  Mask not used2.829131.087.37
  Glasses used1.18130.1112.13
  Glasses not used4.5222202.179.42
  Sprinkling downward attended2.209160.885.53
  Sprinkling downward not attended8.451473.0123.70


The present study suggested that life-style and environmental factors are risk factors for NKTCL, which supports our supposition based on gene mutation studies in East Asian countries.7, 8

The incidence of NHL has increased in the past 3 decades worldwide, especially in rural areas, perhaps associated in part with agricultural practices.17 Environmental and life-style factors including infectious organisms, drugs, solvents and other chemicals together with occupation and genetical factors may play a role in the etiology of NHL.18 More recent data indicate that the increasing incidence of NHL has declined in Western countries because of regulations regarding exposure to chemicals.19 Sinonasal NKTCL is closely associated with EBV infection.6 A majority of previous epidemiological studies on NHL described risk factors for NHL as a whole but not for specific disease types. Patients with sinonasal NKTCL are reported to be clustered in Asian and Latin American countries,3, 4, 5 but are quite rare in Western countries. Thus, this is the first report of an epidemiological study on the relationship between life-style and environmental factors and the development of sinonasal NKTCL.

Previous studies showed that smoking tobacco could be a risk factor for NHL, especially follicular NHL, in women.20, 21 The present study revealed that smoking was not associated with the risk of developing NKTCL. Former but not current drinkers showed an increased risk for NKTCL. However, an inverse association of wine consumption with the development of NHL was reported.22 In the current study, the questionnaires were filled out by the patients at the time of admission for the NKTCL. The patients might have stopped smoking and drinking when the prodromal symptoms such as nasal obstruction and discharge appeared. This might lower the frequency of smoking and drinking among the patients with NKTCL, thus in part explaining why tobacco smoking and current drinking were not risk factors, but the odds ratio was high in the former drinkers.

Increased risk for chronic lymphocytic leukemia and NHL in farmers was reported.23, 24 In the present study, occupations were categorized into 4 groups based on the chance of exposure to chemicals (Table IV). Risk for NKTCL was elevated in farmers compared with nonfarmers. This might suggest that agricultural work is associated with increased risk of NKTCL. Indeed, cultivation of crops was associated with the disease, which was more prominent in individuals who had been cultivating crops for more than 5 years. From these findings, exposure to pesticides and chemical solvents could be causative factors for NKTCL. This was further supported by the following findings: pesticide users were at greater risk and this risk was reduced with the use of protective equipment and careful sprinkling of pesticides. Previous studies showed an increased risk of NHL in individuals using pesticides, especially phenoxyacetic acid-type herbicides.17, 19

Increased risk for cancer development, especially malignant lymphomas, was reported in a population exposed to dioxins, including accidental exposure, in Soveso, Italy.25, 26, 27 Baccarelli et al. reported that exposure to dioxins was associated with increased numbers of circulating t(14;18)—positive lymphocytes.28 t(14;18) is commonly found in neoplastic cells of follicular lymphomas. Chiu et al.29 reported that pesticides were associated with the risk of developing t(14;18)-positive NHL, but not t(14;18)-negative NHL, suggesting an etiological role for pesticides in the development of specific subsets of NHL. A recent study in France suggested that environmental dioxins increase the risk of non-Hodgkin's lymphoma among populations living in the vicinity of municipal solid waste incinerators.30 In the present study, the residents near a garbage incineration plant showed a marked increase in risk for NKTCL, suggesting that the product of the incineration process such as dioxins might be causative of NKTCL. But the present results are of limited value because of the small number of cases and controls examined and lack of data on specific chemicals such as tissue concentrations of dioxin and polychlorinated biphenyl (PCB).

Taken together, environmental factors might be involved in the occurrence of NKTCL, an aggressive lymphoma. Because the number of patients diagnosed annually is quite small even in East Asian countries, the number of cases and controls accrued in the current study was not large enough to confirm the relationship between pesticide and chemical solvent use and the development of NKCTL, or to make a comparison of the proportion of farmers among cases and controls. Therefore, further accumulation of cases and controls is necessary to obtain a definite conclusion.