Int. J. Cancer: 71, 517–520 (1997) r 1997 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer CANCER RISK AFTER RENAL TRANSPLANTATION IN JAPAN Yoshihiko HOSHIDA1, Hideaki TSUKUMA2, Yutaka YASUNAGA1, Ning XU1, Masaki Q. FUJITA1, Takaomi SATOH3, Yasuji ICHIKAWA4, Kenji KURIHARA5, Masaaki IMANISHI6, Tsuyoshi MATSUNO7 and Katsuyuki AOZASA1,* 1Department of Pathology, Osaka University Medical School, Suita, Osaka, Japan 2Department of Field Research, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka, Japan 3Department of Surgery, Sendai Shakai Hoken Hospital, Sendai, Miyagi, Japan 4Department of Urology, Hyogo Prefectural Nishinomiya Hospital, Nishinomiya, Hyogo, Japan 5Department of Pathology, Uwajima City Hospital, Uwajima, Ehime, Japan 6Department of Urology, Kinki University, Sayama, Osaka, Japan 7Department of Surgery, Okayama University, Okayama, Japan Excess of cancer in patients receiving renal transplantation is well-known in Western countries, but information in Japan remains limited. Our study examined whether excess risk is found in patients receiving renal transplantation in Japan. Between 1970 and 1995, 1155 males and 589 females underwent renal transplantation in 6 hospitals, and a total of 12,982 person-years of observation was accumulated. Malignancies developed in 2.6% of patients; O/E ratio was 2.78. Median interval from renal transplantation to tumor development was 58 months. The interval in the patients receiving medication with cyclosporine-A (CyA) (median, 42.5 months) was significantly shorter than that with non-CyA (median, 95.5 months). Median age at the diagnosis of malignancy was 40 years, which is much younger than that in the general population. Relative risk was highest in renal cancer, followed by thyroid cancer, malignant lymphoma and uterine cancer. A distribution of malignancies was different from that reported from Western countries. These findings showed the excess risk of malignancies in Japan with renal transplants, especially in male patients, similar to that observed in Western countries, though the types of malignancy were different. Int. J. Cancer 71:517–520, 1997. r 1997 Wiley-Liss, Inc. Excess of cancer in patients receiving renal transplantation is well known in Western countries: the frequency of cancer development was reported to be 6% in the United States (Penn, 1994) and 8.3% in the Nordic countries (Birkeland et al., 1995). These studies showed a 4.5- to 6.3-fold increase in risk for cancer compared with the general population. Depression of immune surveillance owing to the use of immunosuppressive agents was considered to cause cancer development (Kinlen et al., 1979). In the countries mentioned, cancers of the skin and lips (non-melanotic) and malignant lymphoma were the commonest malignancies. In Japan, though information on malignancies after renal transplants remains limited, some studies have suggested a pattern of cancer development different from that in Western countries, gastric cancer being the commonest malignancy (Ochiai et al., 1987; Gunji et al., 1990). In these studies, however, histologic diagnoses of malignancies were not confirmed, and relative risk for cancer development was not shown. In order to qualify and quantify the risk for cancer associated with renal transplants, a nation-wide study in Japan was carried out in 1995 and 1996. We reviewed all the histologic specimens of developed cancers, and examined the excess risk of malignancies. PATIENTS AND METHODS Study subjects Renal-transplant patients in 6 hospitals in Japan (Osaka University Hospital, Sendai Shakai Hoken Hospital, Hyogo Prefectural Nishinomiya Hospital, Uwajima City Hospital, Kinki University Hospital and Okayama University Hospital) were studied regarding the O/E ratio of developing malignancies. These hospitals were representative institutes for renal transplantation in Japan, the number of transplant patients in each hospital being 417, 381, 285, 316, 144, and 201, respectively. Patients were admitted to the hospitals during the period 1970 to 1995; all patients were followed up to the end of 1995 for vital status and development of malignancies. No special methods for detection of cancer were used if the patients were without symptoms. Diagnosis of malignancy was made on the basis of histologic examination of biopsy materials in each hospital. All the histologic specimens were collected, and reviewed by one of us (Y.H.). Consequently, all the original histologic diagnoses were confirmed. Statistical analysis The observed number (O) of malignancies developed following renal transplants was compared with the expected number (E) in the general population, according to the selected site of malignancy and number of years elapsed since transplantation. The relative risk (RR) was shown by the O/E ratio. Each individual was considered to be at risk from the date of the first renal transplantation until death. The expected number of cases with malignancies was calculated by applying age- and calendar-year-specific incidence rates from Osaka Cancer Registry, which was founded in 1962 for the purpose of registering all malignant tumors and benign intracranial tumors arising in Osaka Prefecture, Japan (1990 Census population, 8.7 millions). A computer program developed by Monson (1984) was used for these calculations. The significance of the O/E ratio was tested by Poisson-distribution analysis (Bailar et al., 1954), while 95% confidence intervals (CI) were calculated using the CIA statistical package (Gardner et al., 1989). The interval between renal transplantation and tumor development in patients medicated with CyA and with non-CyA was evaluated by the Chi-squared test. Histological and immunohistochemical analysis Histologic specimens obtained by biopsy or autopsy were fixed in 10% formalin and routinely paraffin-embedded. Histologic sections, 6-µm thick, were stained with hematoxylin and eosin. Immunohistochemical procedures were performed for cases of malignant lymphoma with monoclonal antibodies Mx-panB (CD20) (Kyowa Medex, Tokyo, Japan, diluted at 1:50), MB-1 (Bioscience, Emmenbrucke, Switzerland, diluted at 1:50), MT-1 (CD43) (Bioscience, diluted at 1:50), UCHL-1 (CD45RO) (Dakopatts, Glostrup, Denmark, diluted at 1:100), and OPD4 (Dakopatts, diluted at 1:100). Mx-panB and MB-1 react with human B lymphocytes and MT-1, UCHL-1, and OPD4 with T lymphocytes. RESULTS In all, 1155 males and 589 females underwent renal transplantation during the period 1970 to 1995 in 6 hospitals. Up to December Contract grant sponsor: Ministry of Education, Science and Culture; Contract grant numbers: 08457061, 08670202, and 08770126. *Correspondence to: Department of Pathology, Osaka University Medical School, 2-2 Yamadaoka, Suita, Osaka 565, Japan. Fax: 81-6-879-3719. Received 12 September 1996; revised 13 December, 1996 HOSHIDA ET AL. 518 31, 1995, 229 patients had died, 1441 patients were alive, and 74 patients were lost to follow-up. When these 74 patients were included among the living patients, a total of 12,982 person-years of observation were accumulated. Ages of patients at transplantation ranged from 3 to 65 (median 36) years. As immunosuppressive agents, cyclosporine A (CyA) was used in 1166 patients, and other agents (non-CyA) in 578. O/E ratios of tumor development At 6 hospitals, 46 of 1744 patients (2.6%) were diagnosed as having malignancies (Table I). The O/E ratios according to the elapsed years from renal transplantation are shown in Table II; risk of tumor development in transplant patients was 2.78-fold higher than in the general population ( p , 0.01). Except for uterine cancer, only male patients showed increased risk. Risk is highest in the first year after transplantation and decreases through the time course. The O/E ratios by site of malignancy are shown in Table III. RR was highest in renal cancer (79.96, p , 0.01), followed by thyroid cancer (12.43, p , 0.01), malignant lymphoma (11.10, p , 0.01) and uterine cancer (4.12, p , 0.05). Age of patients at transplantation and at diagnosis of malignancy ranged from 15 to 51 (median 36) years and from 19 to 63 (median 40) years respectively. The O/E ratio was significantly high between 25 and 50 years of age at the time of diagnosis of malignancy, especially between 25 and 30 years of age (Table IV). Intervals between renal transplantation and tumor development ranged from 35 days to 162 months (median 58 months). The risk was the highest in the first year after renal transplantation: 5 malignancies, including 2 cases of renal cancer and one case each of malignant lymphoma, thyroid cancer, and seminoma. The intervals in patients medicated with CyA (32 cases) and non-CyA (14 cases) ranged from 5 to 132 months (median 42.5 months) and 35 days to 162 months (median 95.5 months) respectively. This difference in the interval was statistically significant by Chi-squared analysis ( p , 0.05). As shown in Table V, O/E ratio was higher in patients with CyA medication than in those with non-CyA medication, especially in renal and skin cancers. Histology of malignancies Histologic types of malignancy are summarized in Table VI. All renal cancers were renal-cell carcinomas. Gastric cancers were classified as tubular adenocarcinoma in 5 cases and poorly differentiated carcinoma in one case. None of the gastric cancers in the present study showed a lympho-epitheliomatous pattern of proliferation. The primary site of malignant lymphoma was extranodal in all 5 cases, the ileum in 2 cases, and lung, skin and spleen in 1 case each. Histologically, all were non-Hodgkin’s lymphoma (NHL), and 4 were diffuse lymphoma of the large-cell type, while TABLE I – MALIGNANCIES DEVELOPED IN RENAL-TRANSPLANT PATIENTS Type of tumor Renal cancer Gastric cancer Non-Hodgkin’s lymphoma Uterine cancer Thyroid cancer Breast cancer Bladder cancer Sarcoma2 Other3 Total Interval for tumor development (months) Number of cases (%) Median age1 (years) Hemodialysis Transplantation 15 (32.6) 6 (13.0) 5 (10.9) 41.0 41.0 40.0 114 140 98 35 97 57 4 (8.7) 3 (6.5) 2 (4.3) 2 (4.3) 3 (6.5) 6 (13.0) 46 31.0 38.0 43.0 43.5 40.5 40.0 40.0 99 61 117 129 75 98 97 94 33 36 63 38 58 58 1Median age at diagnosis of malignancy.–2Hepatic leiomyosarcoma (1), hepatic osteosarcoma (1), and retroperitoneal liposarcoma (1).– 3Liver cancer (1), tongue cancer (1), salivary gland cancer (1), colon cancer (1), seminoma (1) and skin cancer (1). one was of the pleomorphic-cell type. Immunohistochemistry revealed that 3 cases were B-cell type, and 2 T-cell type. All the thyroid cancers were papillary carcinoma, and liver cancer was hepatocellular carcinoma. There was one case of liposarcoma in the retroperitoneal cavity and one case of leiomyosarcoma and one of osteosarcoma primarily arising in the liver. This is extremely rare among the general population. Among cases with uterine cancers, one was adenocarcinoma of the endometrial-cell type arising from the corpus, and 3 were epidermoid carcinoma of invasive type arising from the cervix. One of the breast cancers was Paget’s disease. Original renal diseases in transplants Causes of renal failure in current patients included chronic glomerulonephritis in 27 cases (IgA glomerulonephritis in one, others in 26), acute glomerulonephritis in 3, polycystic kidney in 2, congenital hypoplastic kidney, chronic pyelonephritis, and iatrogenic bilateral hydronephrosis in 1 case each, and unknown causes in 11. One of the 27 cases with chronic glomerulonephritis showed a nephrotic syndrome. DISCUSSION The frequency of malignancies in current patients with renal transplantation was 2.6%, lower than that reported from Western countries (6–8%; Penn, 1994; Birkeland et al., 1995). The RR for tumor development in renal transplants was 2.78-fold higher than that in the general population in Osaka, Japan, which was also lower than the value reported from England (6.3-fold higher) (Gaya et al., 1995) and the Nordic countries (4.6 in males and 4.5-fold in females) (Birkeland et al., 1995). When comparing RR in Japan with that in Western countries, however, the following points should be taken into account: (i) we used the expected number of malignancies from the Osaka Cancer Registry, which is higher than that in other districts of Japan (Parkin et al., 1992); (ii) 74 patients lost to follow-up were conventionally included in the living patients, resulting in an increase of person-years of observation. These factors lowered the O/E ratio of malignancies in the current study. The current study revealed that renal cancer (32.6%) was commonest, followed by gastric cancer (13.0%), malignant lymphoma (10.9%) and uterine cancer (8.7%). As for detection bias, special procedures in the detection of malignancy including renal cancer were not employed in the current series. Earlier studies in Japan have shown that the commonest cancer in renal-transplant patients were hepatocellular carcinoma (18–19%) and gastric cancer (16–18%) (Ochiai et al., 1987; Gunji et al., 1990). Although this difference could not be explained, our current data were based on a large number of cases with careful histologic diagnosis. As a background, the frequency of hepatitis B- and C-virus-associated hepatocellular carcinoma (HCC) among all HCC cases in the US and in Japan were 38 and 27% (Yu et al., 1990) and 21 and 51% (Tanaka et al., 1991) respectively, showing a relatively similar frequency of combined hepatitis-B- and C-virus-associated HCC. Gastric cancer is a common malignancy in Japan, not only in the general population but also in renal-transplant patients. Epidemiological studies have suggested that Japanese-style dietary habits represent a major risk factor in gastric cancer (Inoue et al., 1994). RR for development of malignancies in the current cases was quite different regarding primary sites from that in Western countries (Penn, 1994; Birkeland et al., 1995; Gaya et al., 1995): lower frequency of skin cancer, absence of Kaposi’s sarcoma, and higher frequency of renal and thyroid cancer in Japan. Incidence rates of non-melanotic skin cancers in males and females in the general population are reported to be 1.2 and 0.9 in Osaka, Japan (Parkin et al., 1992) and 144.9 and 73.3 among ‘‘Anglo-Saxons’’ (Waterhouse et al., 1976), showing much higher rates in ‘‘AngloSaxons.’’ This may underlie the lower frequency of skin cancer in the current series. In registration of non-melanotic skin cancers, CANCER RISK AFTER RENAL TRANSPLANTATION 519 TABLE II – OBSERVED (O) AND EXPECTED (E) NUMBERS OF MALIGNANCIES IN RELATION TO CALENDAR YEAR FROM THE TIME OF RENAL TRANSPLANTATION Year Total Males E O/E (95% CI) O 5 1.46 1–4 18 5.52 5–9 16 5.61 10–14 6 3.04 3.421 (1.07–7.68) 3.082 (1.79–4.78) 3.032 (1.43–4.07) 1.97 (0.59–3.51) 0 1p O Females E O/E (95% CI) O E O/E (95% CI) 4 1.01 1 0.45 14 3.84 4 1.68 11 3.89 5 1.72 4 2.30 3.941 (1.03–9.66) 3.392 (1.83–5.62) 3.082 (1.20–4.31) 1.74 (0.37–3.48) 2 0.75 2.23 (0.06–12.10) 2.38 (0.62–5.79) 2.91 (0.89–6.41) 2.68 (0.31–9.26) , 0.05.–2p , 0.01. TABLE III – OBSERVED (O) AND EXPECTED (E) NUMBERS OF MALIGNANCIES BY SITE IN RENAL-TRANSPLANT PATIENTS Total Site Males Females O O/E (95% CI) O O/E (95% CI) O O/E (95% CI) All sites 46 34 6 Liver 3 Breast 2 Uterus 4 Urinary bladder 2 2.911 (1.73–3.49) 1.50 (0.42–3.02) 0.96 (0.10–2.91) 0.00 (0–369.00) 0.00 (0–0) 7.14 (0.73–21.90) 91.631 (44.20–130.00) 12.63 (0.28–61.90) 11.151 (2.95–27.70) 12 Stomach 2.781 (1.80–3.28) 1.40 (0.45–2.68) 1.36 (0.24–3.36) 1.53 (0.18–5.27) 4.122 (1.07–10.00) 6.61 (0.69–20.60) 79.961 (39.98–114.95) 12.431 (2.38–33.70) 11.101 (3.53–25.40) 2.461 (1.21–4.09) 1.05 (0.03–5.57) 8.14 (0.20–42.90) 1.53 (0.18–5.27) 4.122 (1.07–10.00) 0.00 (0–184.00) 0.00 (0–123.00) 12.332 (1.42–42.50) 10.89 (0.25–55.70) Kidney 15 Thyroid gland 3 Malignant lymphoma 5 1p 5 2 0 0 2 15 1 4 1 1 2 4 0 0 2 1 , 0.01.–2p , 0.05. TABLE IV – OBSERVED (O) AND EXPECTED (E) NUMBERS OF MALIGNANCIES IN RELATION TO AGE AT DIAGNOSIS OF MALIGNANCY Total Male Female Age (years) O E O/E (95% CI) O E O/E (95% CI) O E O/E (95% CI) 25–30 4 0.34 2 0.20 0.14 7 0.92 5 0.50 1 0.43 35–40 7 2.01 5 1.09 2 0.92 40–45 9 2.87 7 1.68 2 1.19 45–50 8 2.63 6 1.72 9.992 (1.15–34.40) 12.101 (3.06–22.00) 4.582 (1.40–10.10) 4.171 (1.55–7.92) 3.482 (1.15–6.80) 2 30–35 11.701 (3.11–29.30) 7.591 (2.93–15.00) 3.491 (1.33–6.80) 3.131 (1.36–5.64) 3.032 (1.20–5.49) 2 0.91 14.122 (1.61–48.20) 2.35 (0.58–12.70) 2.18 (0.25–7.53) 1.69 (0.20–5.92) 2.20 (0.26–7.60) 1p , 0.01.–2p , 0.05. basal-cell carcinoma may not be included in some registries in Japan. This may also lower the recorded frequency of skin cancer in Japan. Among normal Japanese and Western populations, the incidence rate of renal cancer was higher in the West than in Japan, while that of gastric cancer was higher in Japan than in the West, and that of thyroid cancer was similar (Parkin et al., 1992); that of cancer of the uterus was higher in Japan (Parkin et al., 1992). Sarcomas developed in 3 patients, 2 in the liver and 1 in the retroperitoneal cavity. Penn (1994) reported an increased incidence of sarcomas in organ-transplant recipients. The median interval between renal transplantation and tumor development was 58 months. The risk was highest in the first year after renal transplantation. No special detection procedures were carried out and no attention was paid to development of malignancy during this period. The interval in patients receiving medication with CyA (median 42.5 months) was significantly shorter than that with non-CyA (median 95.5 months) ( p , 0.05), suggesting that medication with CyA accelerates tumor development in renal-transplant patients. Tumors developed in the first year in 3 of 32 patients (9%) and in 2 of 14 patients (14%) receiving CyA and non-CyA medication respectively. O/E ratio in patients with CyA medication was markedly higher for those with renal and skin cancers than for those without. Median age at diagnosis of malignancy in the current series was 40 years, which is much younger than that in general in Japan (64 years): fewer than 5% of cancer patients are diagnosed below 40 years of age as having the same types of malignancy as observed in HOSHIDA ET AL. 520 TABLE V – OBSERVED (O) AND EXPECTED (E) NUMBERS OF MALIGNANCIES IN RELATION TO CYCLOSPORINE-A (CyA) MEDICATION Site All sites CyA medication O 30 Skin 1 Kidney 12 Thyroid gland 1 Malignant lymphoma 4 1p O/E (95% CI) Non-CyA medication O O/E (95% CI) 3.451 16 2.062 (2.13–4.51) (1.18–3.34) 79.132 0 0.00 (2.53–557.00) (0–369.00) 1 122.13 3 33.951 (44.30–150.00) (5.62–79.70) 7.67 2 18.322 (0.18–39.80) (2.06–66.15) 17.401 1 4.59 (4.36–41.00) (0.11–24.20) , 0.01.–2p , 0.05. TABLE VI – HISTOLOGIC TYPES OF MALIGNANCY Site Tongue Salivary gland Stomach Colon Liver Retroperitoneum Skin Breast Uterus Cervix Corpus Testis Urinary bladder Kidney Thyroid gland Malignant lymphoma Total Histologic type Squamous-cell carcinoma Duct carcinoma Adenocarcinoma Adenocarcinoma Hepatocellular carcinoma Leiomyosarcoma Osteosarcoma Liposarcoma Squamous-cell carcinoma Solid tubular carcinoma Paget’s carcinoma Squamous-cell carcinoma Adenocarcinoma Seminoma Transitional-cell carcinoma Renal-cell carcinoma Papillary carcinoma Non-Hodgkin’s lymphoma Malignant lymphomas that developed in current cases shared common features with those reported from the Western countries, except for the rarity of lymphoma of the central nervous system: large-cell morphology, preponderance of B-cell immunophenotype, and extranodal origin. Penn (1994) reported that treat with CyA decreased the incidence of skin cancer, but increased that of NHL and Kaposi’s sarcoma. In contrast, Gaya et al. (1995) reported that the frequency of NHL was higher in patients receiving azathioprine and prednisolone than those receiving CyA. In the current cases, the number of patients developing NHL and renalcell carcinoma was higher in patients receiving CyA than in others. The nephrotoxicity of CyA might be a partial reason (Klintmalm et al., 1981). In conclusion, our study shows increased incidence of malignancies in Japanese renal-transplant patients as observed Western countries, although RR is lower in Japan. Types of malignancy differ between Japanese and Western patients. Number of cases 1 1 6 1 1 1 1 1 1 1 1 ACKNOWLEDGEMENTS 3 1 1 2 15 3 5 46 the current series (Research Group for Population-Based Cancer Registration in Japan, 1987). In patients with renal transplants, the risk of malignancy was higher than in the general population between 25 and 50 years of age. We thank the following pathologists and physicians for allowing us to study their cases: Drs. H. Satoh (Sapporo Municipal Hospital), T. Entani (Hiraga General Hospital), Y. Taguma and S. Onodera (Sendai Shakai Hoken Hospital), Y. Kondo and D. Ozaki (Chiba University), K. Hamaguchi (Sakura National Hospital), T. Wakabayashi (Tokyo University), M. Kobayashi (Tokyo Women’s Medical College), S. Kawamura (Toho University), T. Kameya (Kitasato University), Y. Osamura and M. Yasuda (Tokai University), N. Itoh (Shinshu University), M. Kuroda (Fujita Health University), S. Murakami (Social Insurance Chukyo Hospital), S. Miwa (Toyama Prefectural Hospital), T. Yasumura and Y. Tsuchihashi (Kyoto Prefectural University), A. Okuyama (Osaka University), C. Yutani (National Cardiovascular Center), T. Uesugi (Kinki University), H. Yasoshima (Hyogo College of Medicine), T. Kohro (Hyogo Prefectural Nishinomiya Hospital), Y. Sadahira (Kawasaki Medical School), K. Matsui (San-in Rosai Hospital), E. Tahara and H. Yokozaki (Hiroshima University), M. Tsuneyoshi (Kyushu University), and K. Irie (Saga Prefectural Hospital Koseikan). This work was supported in part by grants from the Ministry of Education, Science and Culture (08457061, 08670202, 08770126), Japan. REFERENCES BAILAR, J.C. and EDERER, F., Significance factors for the ratio of a Poisson variable to its expectation. Biometrics, 20, 639–643 (1954). BIRKELAND, S.A. and 18 OTHERS, Cancer risk after renal transplantation in the Nordic countries, 1964–1986. Int. J. Cancer, 60, 183–189 (1995). GAYA, S.B.M., REES, A.J., LECHLER, R.I., WILLIAMS, G. and MASON, P.D., Malignant disease in patients with long-term renal transplants. Transplantation, 59, 1705–1709 (1995). GARDNER, M.J., GARDNER, S.B. and WINTER, P.D., Confidence-Interval analysis (CIA). Brit. med J., 46–56 (1989). GUNJI, Y., OCHIAI, T., ASANO, T., SUZUKI, T., ENOMOTO, K., HORI, S., SHIMADA, H., YASUMOTO, A. and ISONO, K., Development of malignant tumor in renal-transplant recipients and their NK and LAK activity. Jap. J. clin. Immun., 13, 236–245 (1990). (In Japanese). INOUE, M., TAJIMA, K., HIROSE, K., KUROISHI, T., GAO, C.M. and KITOH, T., Life-style and sub-site of gastric cancer-joint effect of smoking and drinking habits. Int. J. Cancer, 56, 494–499 (1994). KINLEN, L.J., SHEIL, A.G.R., PETO, J. and DOLL, R., Collaborative United Kingdom-Australasian study of cancer in patients treated with immunosuppressive drugs. Brit. med. J., 2, 1461–1466 (1979). KLINTMALM, G.B.M., IWATSUKI, S. and STARZL, T.E., Nephrotoxicity of cyclosporin A in liver- and kidney-transplant patients. Lancet, I, 470–471 (1981). MONSON, R.R., Analysis of relative survival and proportional mortality. Comput. Biomed. Res., 7, 325–332 (1984). OCHIAI, T., ASANO, T. and ISONO, K., Development of malignancies in Japanese renal-transplant recipients. Transplantation Proc., 14, 2967–2970 (1987). PARKIN, D.M., MUIR, C.S., WHELAN, S.L., GAO, Y.T., FERLAY, J. and POWELL, J., Cancer Incidence in Five Continents, Vol. VI, IARC Scientific Publication 120, IARC, Lyon (1992). PENN, I., Depressed immunity and the development of cancer. Cancer Detect. Prevent., 18, 241–252 (1994). RESEARCH GROUP FOR POPULATION-BASED CANCER REGISTRATION IN JAPAN, Cancer incidence and incidence rates in Japan in 1987: estimates based on data from nine population-based cancer registries. Jap. J. clin. Oncol., 22, 437–442 (1992). TANAKA, K. and 14 OTHERS, Hepatitis C and hepatitis B in the etiology of hepatocellular carcinoma in the Japanese population. Cancer Res., 51, 2842–2847 (1991). WATERHOUSE, J., MUIR, C.S., CORREA, P. and POWELL, J., Cancer incidence in five continents. Vol. III, IARC Scientific Publication 15, IARC, Lyon (1976). YU, M.C., TONG, M.J., COURSAGET, P., ROSS, R.K., GOVINDARAJAN, S. and HENDERSON, B.E., Prevalence of hepatitis-B and -C viral markers in black and white patients with hepatocellular carcinoma in the United States. J. nat. Cancer Inst., 82, 1038–1041 (1990).