Risk and Protective Factors Related to Mortality from Pneumonia among Middle-aged and Elderly Community Residents: The JACC Study

BACKGROUND There have been few systematic investigations into risk and protective factors for pneumonia related mortality for community residents. This study investigated these factors utilizing a large cohort study on Japanese community residents. METHODS Subjects, 110,792 individuals (aged 40-79 years) enrolled in 1988-1990, were followed until death, or when they moved away from the surveyed communities, or the end of 2003. Pneumonia death was defined following 480-486 (International Classification of Diseases, 9th Revision) or J12-J18 (10th Revision). Age-adjusted and multivariate hazard ratios were calculated along with 95% confidence intervals using the Cox proportional hazards model. RESULTS With 1,112,747 person-years of the study, a total of 1,246 persons died of pneumonia. We found history of blood transfusion (multivariate hazard ratio=2.0 [95% confidence interval: 1.7-2.4]) was a potent novel risk factor. Walking 0.5-1 hour/day (0.8 [0.6-1.0]), 1+ hour/day (0.7 [0.6-0.8]), and/or a history of pregnancy (0.6 [0.4-0.9]) were found to reduce pneumonia mortality. A large body mass index (BMI) (25+kg/m2) was a protective factor (0.7 [0.5-0.8]), while low BMI (<18) was confirmed as a risk one (2.1 [1.7-2.6]). Smoking was an important preventable risk factor (1.6 [1.3-1.9], population attributable risk proportion=14%), and its cessation reduced risk (0.7 [0.5-1.0]) to levels comparable to never-smokers (0.7 [0.5-1.0]). CONCLUSIONS The risk and protective factors ascertained here for pneumonia mortality among community residents, history of blood transfusion, large BMI, and walking habits, warrant further study. Smoking cessation may effectively reduce pneumonia mortality.


Demographic Characteristics
Demographic characteristics of the cohort are provided in Table  1. During the observation period of 1,112,747 person-years, 16.0 % (10,367 men and 7,330 women) of the participants died, and 4.3 % (1,773 men and 2,992 women) were lost to follow-up because they moved out of the study areas. A total of 1,246 (791 men and 455 women) died of pneumonia. AMR for pneumonia increased with age, and AMRs in men were more than twice those in women for all age classes (Table 2). Generally, AMRs of the cohort were stably lower than those of the national average, suggesting a possibility that the present study failed to detect some of the pneumonia cases in the cohort. This is especially true for patients as well as case-control studies have been conducted, [1][2][3][4][5] there have been few systematic investigations into community residents' risk factors: smoking, 6-8 medical history such as heart disease, stroke, and diabetes mellitus, 7,8 expiratory volume, 8 and lower body mass index (BMI) 7,9 are some. These studies were relatively small population studies, and lifestyle influences were largely uncharted. Some lifestyle habits have been reported to be protective against physical decline (including immune response). [10][11][12] The primary aim of the present study, therefore, was to identify factors determining susceptibility or resistance to pneumonia death among community residents. To achieve this, our study has taken advantage of the data available from a large cohort study.

Study Cohort
The Japan Collaborative Cohort Study for Evaluation of Cancer Risk (JACC Study), sponsored by the Ministry of Education, Science, Sports and Culture of Japan, was established from 1988 through 1990 in 45 areas in Japan. 13,14 A total of 110,792 individuals (46,465 men and 64,327 women, aged 40-79 years) participated in municipal health screening examinations and completed a self-administered questionnaire.
Smoking and alcohol-consumption statuses were divided into three categories (current, past, never). Playing sports was categorized as <1, 1-2, 3-4, or >4 hours a week. Similarly, walking habit was categorized as <0.5, 0.5, 0.5-1, or >1 hour a day. Medical histories were inquired about, using a yes/no question as to whether the participant had a particular medical history. Those with nonmarked or missing data in the questionnaire were not used in the analyses.

Outcomes
Participants were followed up until death, or till they moved away from the surveyed community, or to the end of 2003. For mortality surveillance in each community, investigators systematically reviewed the death certificates of any participants, all of which were forwarded to the public health center in the area of residency. Mortality data were sent centrally to the Ministry of Health and Welfare and the underlying causes of death were coded for the National Vital Statistics according to the International

Risk Reduction by Smoking Cessation
In the final model, we examined the association of smoking cessation with pneumonia mortality risk. We used current smokers as the reference category with adjustment of significant variables (baseline age, low BMI, little walking, and history of stroke, diabetes mellitus, tuberculosis, cancer, and blood transfusion).
Our results showed that ex-smokers significantly reduced the risk of pneumonia mortality (multivariate hazard ratio = 0.7 [95% confidence interval: 0.5-1.0]) to levels comparable to those in never-smokers (0.7 [0.5-1.0]). Even when early deaths within 5 years of follow-up were excluded, the significant associations between cessation of smoking and reduction of pneumonia mortality risk (0.7 [0.5-1.0]) were confirmed.
For ex-smokers, smoking cessation for longer than 5 years before the baseline entry significantly reduced the risk of pneumonia mortality (0.7 [0.5-1.0]) to levels comparable to those in never-smokers. However, we could not accurately test the association between smoking cessation for 0-1 year, or 2-5 years with pneumonia mortality risk, because of the small sample sizes (0-1 years, n=12; 2-5 years, n=33). Therefore, we could not determine an effective minimum cessation period that significantly reduced pneumonia mortality risk.
those who moved from their original residence to long-term hospitalization.

Detection of Potential Risk and Protective Factors by Age-adjusted Analyses
Potential risk and protective factors for pneumonia death from the age-adjusted analyses are shown in Table 3. BMI was associated with pneumonia mortality: Low BMI elevated the risk while the reverse was true for large BMI. Current smoking and ex-drinking habits showed significant risk associations. With regards to medical history; stroke, myocardial infarction, diabetes mellitus, tuberculosis, cancer, and blood transfusion were all found to be associated with mortality risk, while a history of pregnancy in women significantly reduced mortality risk. Exercising (playing sports and walking) showed protective trends, although walking less than 0.5 hour/day habits increased mortality risk.

Multivariate Analyses of Possible Risk and Protective Factors
In our multivariate analyses of medical histories and lifestyle factors (playing sports and walking), baseline age, sex, history of diabetes mellitus were adjusted, and participants with a history of stroke were excluded from the analyses. However, even with these adjustments most risk/protective factors identified by Table  3 remained unchanged (Table 4). Associations were also confirmed when early deaths within 5 years of follow-up were excluded (low BMI, multivariate hazard ratio=2.0 [95% confidence interval: 1.6-2.6]; smoking habit, 1.     27 Our data on Japanese did not support such a trend, although the range of our data on BMI was limited to between 10 and 33 in the cohort of community residents surveyed here. We also found novel factors related to lifestyle: Exercise habits might be understood as an intermediate variable of being healthy, contrary to previous findings. Exercise habits were reported to reduce a wide range of mortalities, 28 and some reports suggest that physical activity is associated with natural killer cell activity. 29 These factors might also lead to functional independence in daily life. Adults aged 60 years and older who were functionally independent before admission were reported to be more likely to present with less severe pneumonia symptoms than patients who were functionally dependent before admission. 30 Rather more systematic and detailed surveys will be needed to better elucidate the effects of these aspects of lifestyle. The effects of alcohol consumption are also equivocal: one study reports chronic alcoholism increases pneumonia mortality risk, 31 while a significant relationship between drinking habits and pneumonia infection was excluded in another. 27 Our results from multivariate analyses showed ex-drinkers were associated with increased pneumonia mortality risk, while current-drinkers and even heavy-drinkers were not significantly associated with mortality risk. This might mean that a behavioral problem of a predisposition to alcohol abuse or other unspecified conditions that required quitting a drinking habit might be an important factor, as suggested in another study. 32 Although we could not find an effective minimum smoking cessation period, we did demonstrate that smoking cessation might be effective in reducing, to a certain degree, the risk of pneumonia mortality. While the preventable portion of pneumonia infection by smoking cessation in the general population aged 18 to 64 years is reported to be 51%, 33 our results for mortality showed that smoking cessation was still a significant measure (14%) against fatal conditions. We confirmed that various medical histories are associated with pneumonia mortality risk, 6,7,17 Additionally, our results newly revealed that some medical histories were associated with pneumonia mortality risk. Further etiological research will be needed on these newly identified medical histories.
The present study has several limitations. First, we did not detect any incident diseases or smoking status following baseline entry. Therefore any smoking cessation during follow-up was not included. Second, we could not take into account more detailed classifications of pneumonia death, potential confoundings from vaccinations, or other unmeasured items that might affect smoking cessation or pulmonary disease histories. 7,8 For example, compared with current smokers, ex-smokers may be more health conscious, while some of them quit smoking because of illness; the net effect of these factors is uncertain. 34 Furthermore, we could not confirm the reason for the blood transfusion, their frequency, the number of transfused blood units, or if there were unknown infectious agents. These factors might contain some unrecognized In this large cohort study, we found risk and protective factors that were significantly associated with pneumonia: Blood transfusion history was found to be a newly recognized risk factor while history of pregnancy, large BMI, and daily walking habits were protective factors. We also confirmed that low BMI was a risk factor while cessation of smoking was a protective one which might significantly reduce pneumonia mortality risk. These findings were observed using data from a cohort study of community residents and thus provide useful measures for intervention in respect to fatal pneumonia in the elderly.
History of blood transfusion might be related with immunomodulation. Before 1990, blood transfusions in Japan were conducted without white blood cell filters, irradiation, or screening for any bacterial or viral infections. 18 Evidence from a variety of sources suggests that microchimerism by allogeneic blood transfusion increases the incidence of immunomodulation, [19][20][21][22] and allogeneic donor leucocytes have been considered one of the causes of the immunomodulation effect of transfusion. 23 Such modulation, combined with aging, might result in an increased susceptibility to fatal conditions. It should be considered, however, that the experience of pregnancy, which is known to induce microchimerisms from fetuses, 24 was a protective factor. This conflicting observation, also reported elsewhere, 25 suggested that similar antigen exposure induces either a protective or detrimental effect to the host depending on the milieu of the immune system. Although the Pneumonia Severity Index (PSI) does not account for immunosuppression, pneumonia patients with immunosuppression such as hematological malignancies or bone marrow transplantation were reported to have significantly greater mortality. 26 On the other hand, blood transfusion also could be considered as a surrogate marker for a variety of underlying factors that added to the pneumonia mortality risk. These underlying factors might be cofounders of the association between transfusion and pneumonia mortality risk, being associated with both the need for transfusion and the pneumonia mortality risk. Although we excluded the effect of medical histories surveyed in the questionnaire, there might be unidentified medical histories that engendered a spurious association between transfusion and pneumonia mortality risk. The persistent immunological influences of allogeneic blood transfusions before 1990 and its association with pneumonia mortality warrant further investigation.
BMI and some lifestyle factors were also found to be significantly associated with pneumonia mortality risk. Low BMI has been associated with increased risk of infection, possibly due to malnutrition or underlying illness, 7, 9 although obesity has also been suggested to elevate the risk associated with impaired Tand/or B-cell function. 27 A cohort study in the United States reported a U-shaped relationship between BMI and immune dysfunction, showing that not only a low BMI but also a BMI with 27.0+ raise the pneumonia infection risk for men aged 40 to 75 confounding effects related to the associations we found. Finally, there is a possibility that some of the pneumonia cases had been lost, specifically those who left their original communities to undergo long-term hospitalization.
In conclusion, the present study demonstrated possible risk and protective factors for pneumonia death based on a prospective mega-cohort of community residents. Our results showed that a history of blood transfusion significantly increased pneumonia mortality risk, and that low BMI and smoking habit were confirmed as risk factors. Contrarily, smoking cessation, walking habit, pregnancy in women, and a large BMI were found to reduce pneumonia mortality risk. Confirmation of the associations between the risk/protective factors suggested here and pneumonia mortality in other populations, and investigations related to their operative mechanisms await further study.