Cohort Study on Association of Malignant Neoplasms among the Pulmonary Tuberculosis Patients in Nagoya TB Registry

A total of 3,580 pulmonary TB patients newly registered in Nagoya TB registry in 1979-81 were followed-up until the end of 1983. 489 patients (13.7%) died in the period of observation. O/E ratio of dying from TB was very high, being 15.1 per 100,000 for males and 40.0 for females, and also higher O/E ratio of cancer, heart diseases, pneumonia/bronchitis were shown for both sexes, the risks ranged 1.7 to 3.2. The risk of liver diseases except cancer was high in females only. Lung cancer death showed high O/E ratio of 3.8 for males and 6.4 for females. Observed higher risk of malignant neoplasms related to bone marrow suggest some immune disorders of the patients. Causative factors were discussed.


INTRODUCTION
One of the authors(l) reported excess deaths of lung cancer among pulmonary tuberculosis (TB) patients in the TB Registry of City of Philadelphia, PA, USA by retrospective and prospective studies in 1963-65.The similar results were observed in the studies of Nagoya and Aichi Cancer Registries, Japan, 1965-1967, although the death rate of lung cancer was very low at that time (1,2).Steinitz(3) already reported in 1965 that a significantly high risk of lung cancer deaths among the tuberculosis patients in the Israel TB Registry collating the data of Cancer Registry and that the risk of lung cancer among them was the same magnitude of that of the smokers.The incidence of lung cancer in Nagoya city has continuously been increasing since 1970, and recently the death rate of lung cancer exceeded that of tuberculosis like that of Philadelphia in 1955-65 of the USA.Considering such epidemiological situations, We tried to confirm excess incidence of lung cancer and/or other cancers among pulmonary TB patients registered in Nagoya city (5) and to approach to common causal factors of both cancer and tuberculosis, because it is said that cancer is a specific successor of tuberculosis (2,(6)(7)(8).

SUBJECTS AND METHODS
The TB Registration system in Japan was legally started in 1951 and the registration of TB patients has well established since 1961, supported by governmental subsidy for cost of treatments to the registered patients.Health insurance system for all Japanese in 1961 have promoted complete treatment for the TB patients.The TB patients registered are semianually checked current status of disease up by the local TB specialist committee and public health nurses of district health center visit the patients for care administration.The name of the patients registered is routinely deleted off the registration roll after at least two years' observation from the time at diagnosis of clinically healed.
A total of 4,892 patients aged 30 years and over were newly registered in the Nagoya TB Registry during the three years from January 1,1979 to December 31,1981.Among them 1,309 cases with pleurisy and/or extrapulmonary TB without pulmonary lesions as well as those with tuberculosis lesions healed with or without thoracic surgery were excluded from the cohort; because our main objective in this series was to observe coexistent TB and cancer in the lung.The remaining 3,580 pulmonary TB patients (male; 2,497, females; 1,083) were enrolled in the list of the cohort study, and followed up until the end of 1983.Information of patients were obtained from attending physician's reports to Public Health Center and home visit records prepared by health nurses.When a check-up schedule was interrupted and no further information was obtained either by the failure of the patients to show up the clinic , due to a change of address or to other reasons, inquires on vital status of the patient were sent to local government office of his or her permanent address through Public Health Center.When a patient had died, the cause of death was examined by obtaining death certificate and/or medical records under the permission of the Government and/or the attending physicians.In the case of lung cancer, chest X-rays, cytological and pathological records were reviewed to confirm coexistent TB and cancer, as possible.Secondary lung cancer such as metastatic lesions were excluded from the analyses.
The cases diagnosed as primary lung cancer within one year after the entry of the study were omitted from the present analyses.To examine the effects of severity or activity of TB lesions , excess incidence of cancer was analysed according to the presence of cavitary lesions and/or Koch bacillus in the sputum examination.The history of uses of antituberculous drugs and smoking habits were examined by the records of TB Registry including health nurses' notes.As for lung cancer cases, smoking history and other relevant items were asked to the attending physicians and/or the kins, if possible.
The expected number of deaths from each of the causes of death was calculated using the person-year method based on age-sex specific death rates in the general population of Nagoya City in 1980.The ratio of the observed number (O) to the expected (E) was calculated and statistical significance was tested using the Z value obtained with the following formula; When the expected number was less than ten, the O/E ratio was tested by applying the Poisson's distribution.

RESULTS
The number of the patients in the cohort, the observed person-years and the number of deaths by age and sex were shown in Table 1.The observation period totalled 8,856.50personyears for males and 3,845.75 for females.A total of 489 patients (13.7%) died in this observation period.The crude death rate was 3.96 per hundred person-years for males and 3.59 for females.The death rate increased with age for both sexes except a group of 80 years and over, which rates were all significantly higher than those of general population.The observed and expected number of deaths and O/E ratios by cause of death were shown in Table 2. Significantly higher O/E ratios of death from all causes were computed such as 1.6 for males and 1.9 for females.As was expected, patients with tuberculosis were at a very high risk of dying tuberculosis with O/E ratios of 15.1 for males and 40.0 for females.The O/E ratios for cancer, all sites, heart diseases and pneumonia or bronchitis were significantly high for both sexes, although the risks were not high comparing those of tuberculosis.Females showed higher O/E ratio of dying liver diseases except cancer, being 6.0, while males did not show excess risk.The risk of cerebrovascular diseases was low for both sexes.The risks of death due to heart diseases, pneumonia or bronchitis, and liver diseases except cancer were slightly higher in the patients with cavitary lesions than those without.There was little difference in O/E ratios between the sputum positives and the negatives.Table 3 shows the observed number of deaths and O/E ratios from cancer by site.The O/E ratios for lung cancer were significantly high of 3.8 for males and 6.4 for females.The groups with more active TB lesions showed slightly higher risk than those with less active.
Very high O/E ratio was observed in acute monocytic leukemia, and oral cavity or larynx for males, and malignant lymphoma and multiple myeloma for females, although the numbers of death were small.The O/E ratios of lung cancer in relation to smoking were shown in Table 4 only for females.The smoking rate for females was 14.0%, of which 86% quit smoking after the entry of TB Registry.About half of the cases with lung cancer were smokers and the O/E ratio was 20.0, while that of non-smokers was 4.5 which was still significantly high figure comparing that of female general population in Nagoya.Histological type of five non-smokers in females was as follows; two squamous type, three adenocarcinoma, and that of 4 smokers was two squamous type, one small cell and one adenocarcinoma.As for antituberculous drug, SM used in eight of nine female patients with lung cancer, while it was used in only 50% of non-cancer cases.There was little differences in frequency and duration of prescriptions for other TB drugs including INH between lung cancer patients and non-cancer cases.There was no difference in exposure history to ionizing radiation indicated by medical diagnosis and treatment.
Smoking history for male patients could not completely be obtained, as the smoking rate was 84%.We could not get the information of smoking for 14 cases with coexistent TB and lung cancer.Therefore, the O/E ratios by smoking for males were not shown in this paper.

DISCUSSION
Newly diagnosed TB patients in Japan are registered in the local TB Registry, where the patient resides.A few patients might not be registered by various reasons, but there seemed to be negligible effect on the objective of this study.Governmental subsidy to cover the cost of treatments in part and/or by health insurance system for all Japanese since 1961 have strongly supported TB registration greatly.The patients registered are checked-up semiannually by the local TB specialist committee and have advices on their care by health nurses of Health Center periodically.The above information on the patients were very useful for the follow-up study.The subjects of the study were only pulmonary TB patients who have mostly had chronic relapsed TB lesions in the lung.The observation period of average four years might be rather short.The reasons were that it was difficult to follow-up completely all the cases more than 4 years and that the relative risk of lung cancer could be evaluated in the first four years by the previous studies(1,2), that is, onset of lung cancer and other sites were often observed within 5 years after the entry into the Registry by relapse or worsening TB lesions.All the analyses were done based on causes of death as an endpoint, underlying and contributory causes of death on death certificates were reexamined by other records as possible.Forty five male patients and 14 females with lung cancer in this series were excluded, because of lung cancer occurred within one year, or no active TB lesions in the lungs.The lung cancer cases with active TB alive at the end of 1983 were also excluded in the analyses.
As mentioned before the TB patients have had very high risk of dying from TB even in the days of strong specific treatments, and higherr risk of broncho-pulmonary diseases and pulmonary heart diseases may be an avoidable causes of death considering advanced TB lesions in the lungs.Higher risk of liver diseases might be due to long intake the drugs, but it was shown only in females.Lower risk of cerebrovascular diseases might partly be to related lifestyles of the TB patients including dietary life and working conditions.There were little differences in mortality pattern between active and inactive lesions.
Other several studies in Japan (9)(10)(11)(12) and recently in China(13) showed similar higher risk of lung cancer among TB patients.The risk was higher in females than in males in several studies a significantly high risk of 4.5 in non-smoking female patients was higher than that of passive smoking reported.Those with chronic TB lesions may have some specific factors than smoking.Autopsy cases showed similar association (2.14,15).
Pulmonary fibrosis and sarcoidosis with chronic inflammatory processes (2,16,17) have had a higher risk of lung cancer, but etiological processes are unknown.Fujiki et al. (18) reported a cancer promoting function of trehalose monomycolate extracted from cord factor of Koch bacilli, which may partly explain incidence of cancer in a rather short time from registration.Higher risks of malignant lymphoma, multiple myeloma and monocytic leukemia suggest some immune disorders and/or, bone marrow dysfunctions.Immunodificiency diseases such as dermatomyosites, and progressive systemic sclerosis (19) showed higher incidence of lung cancer, like sarcoidosis.Long standing chronic inflammations from the young may be associated with immune disorders in the adult and the old.There observed no differences in the uses of antituberculous drugs and the duration, ionizing radiation related to medical treatments in any sites of cancer of this series (20,21).
The higher risk of oral/laryngeal cancers in males may be largely due to smoking.And smoking may affect the expected value of male lung cancer as a reference.
Dietary life among TB patients seemed to be different from that of general population, because animal protein/fat rich diet was routinely recommended by the physicians before long.Westernized dietary habits may change cancer pattern in the TB patients than that of general population.
That cancer is a specific successor of tuberculosis in a community and that there may be some susceptibles to dying from cancer in the aged, when they did not die in the younger ages.Considering genetic effect on lung cancer suggested from Coghlan's postulate (6 ,7) .
The proportion of susceptible people(1,2) to both TB and lung cancer in the birth corhort was estimated.
The results were as follows; provided that about 6% of birth cohorts born in 1860 to 1930 in the states of Massachusetts, USA, finally die from TB or lung cancer, the changing pattern of mortality from TB or lung cancer for the last decades could be simulated, however, it is only a result of epidemiological model without biological evidences.