ORIGINAL PAPER
Increased risk of acute pancreatitis following pneumococcal pneumonia: a nationwide cohort study S.-W. Lai,1,2 C.-L. Lin,3,4 K.-F. Liao,5,6 C.-L. Ma6
1
SUMMARY
What’s known
Background: The purpose of this study was to evaluate the risk of acute pancreatitis following pneumococcal pneumonia in Taiwan. Methods: We undertook a retrospective cohort study using the hospitalisation claims data of the Taiwan National Health Insurance Program. We identified 16709 subjects aged 20–84 with the first-attack of pneumococcal pneumonia between 1998 and 2010 as the pneumonia group and we randomly selected 66836 subjects without a history of pneumonia as the non-pneumonia group. Both groups were matched for gender, age and index year. We examined the incidence of acute pancreatitis by the end of 2010 and we used a multivariable Cox proportional hazards regression model to calculate the hazard ratio (HR) and 95% confidence interval (95% CI) of acute pancreatitis associated with pneumococcal pneumonia and other comorbidities. Results: Subjects with pneumococcal pneumonia had higher incidence of acute pancreatitis, when compared with non-pneumonia subjects (2.41 vs. 1.47 per 1000 person-years, crude HR 1.65, 95% CI = 1.38, 1.97). The highest risk of developing acute pancreatitis occurred during the first 3 months after diagnosing pneumococcal pneumonia (crude HR 4.11, 95% CI 1.98, 8.52). After adjusted for potential confounders, the adjusted HR of acute pancreatitis was 1.51 (95% CI 1.25, 1.82) for the pneumonia group, as compared with the non-pneumonia group. Conclusions: Overall, this study reveals a 51% increased hazard of acute pancreatitis following infection with pneumococcal pneumonia. Patients with pneumococcal pneumonia should receive close surveillance for risk of developing acute pancreatitis during the first 3 months after diagnosing pneumococcal pneumonia.
Introduction Pneumococcal pneumonia is a common disease with a major concern because of its significant morbidity and mortality. In acute stage of pneumococcal pneumonia, septic shock, acute respiratory failure, bacteraemia, pleural effusion and empyema are frequently found (1,2). The mortality rate is about 1.3–36%, depending on host factors, pneumococcal serotypes and disease severity (1–3). Recent research has revealed subsequent severe outcomes after acute stage. Pneumococcal pneumonia correlates with increased risk of stroke, acute coronary syndrome and lung cancer (4–6), but acute pancreatitis is not included till now. Acute pancreatitis is another public health concern because of its significant morbidity and mortality. Extensive evidence has shown that some factors, including alcoholism, biliary stone, diabetes mellitus,
ª 2015 John Wiley & Sons Ltd Int J Clin Pract, May 2015, 69, 5, 611–617. doi: 10.1111/ijcp.12590
• •
Pneumococcal pneumonia is a common disease with a major concern because of its significant morbidity and mortality. Pneumococcal pneumonia correlates with increased risk of stroke, acute coronary syndrome and lung cancer, but acute pancreatitis is not included till now.
What’s new
• •
Pneumococcal pneumonia is associated with 1.51-fold increased hazard of developing acute pancreatitis. The highest risk of developing acute pancreatitis occurred during the first 3 months after diagnosing pneumococcal pneumonia.
hepatitis C infection and hypertriglyceridaemia, significantly correlate with increased risk of acute pancreatitis (7–9), but pneumococcal pneumonia is not mentioned till now. Because pneumococcal pneumonia and acute pancreatitis are both inflammatory diseases, we make a novel hypothesis that there could be a relation between pneumococcal pneumonia and acute pancreatitis based on inflammatory theory that pneumolysin and/or other virulence factors released from the pneumococcus presumably may be involved in the pathogenesis of pancreatic inflammation (10–12). If pneumococcal pneumonia significantly correlates with increased risk of developing acute pancreatitis, close surveillance for risk of developing acute pancreatitis in patients with pneumococcal pneumonia can be suggested. Therefore, with the aim of better understanding this issue, we conducted a nationwide cohort study.
School of Medicine, China Medical University, Taichung, Taiwan 2 Department of Family Medicine, China Medical University Hospital, Taichung, Taiwan 3 Department of Public Health, China Medical University, Taichung, Taiwan 4 Management Office for Health Data, China Medical University Hospital, Taichung, Taiwan 5 Graduate Institute of Integrated Medicine, China Medical University, Taichung, Taiwan 6 Department of Internal Medicine, Taichung Tzu Chi General Hospital, Taichung, Taiwan Correspondence to: Kuan-Fu Liao, Department of Internal Medicine, Taichung Tzu Chi General Hospital, No.66, Sec. 1, Fongsing Road, Tanzi District, Taichung City, 427, Taiwan Tel.: + 886 4 2205 2121 Fax: + 886 4 2203 3986 Email: [email protected]. tw
Disclosure The authors disclose no conflicts of interest.
611
612
Pneumococcal pneumonia and acute pancreatitis
Materials and methods Data sources We undertook a retrospective nationwide cohort study using the database of the Taiwan National Health Insurance Program. This program started in March 1995 and almost 99% of 23 million citizens joined this program (13). This program provided outpatient, hospitalisation and emergency services. The diagnostic codes are based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9 codes). The details of this program have been published previously (14–17). This study was approved by the Institutional Review Board (IRB) of China Medical University and Hospital (CMU-REC-101-012).
Participants Figure 1 illustrates flowchart of participant selection in this study. We identified subjects aged 20–84 who were newly and mainly diagnosed pneumococcal pneumonia from the hospitalisation dataset between 1998 and 2010 as the pneumonia group (ICD-9 codes 481). The date of diagnosing pneumococcal pneumonia was defined as the index date. To
increase the statistical power, fourfolds of comparison subjects without a history of pneumonia were randomly selected from all National Health Insurance beneficiaries as the non-pneumonia group (ICD-9 codes 480–487). The non-pneumonia group was matched for the pneumonia group by gender, age (every 5-year span) and the index year of diagnosing pneumococcal pneumonia. In both groups, subjects with history of acute pancreatitis, chronic pancreatitis or pancreatic cancer at baseline, or missing information about gender or age, were excluded from this study.
Main outcome measurement All study subjects were monitored until the date on which they were newly and mainly diagnosed with acute pancreatitis based on the hospitalisation dataset (ICD-9 codes 577.0), or being censored because of loss of follow-up, death or withdrawal from insurance, or to the end of 2010.
Comorbidity assessment Baseline comorbidities were analysed as follows: alcoholism (ICD-9 codes 291, 303, 305.00, 305.01, 305.02, 305.03, 790.3 and V11.3), biliary stone (ICD-9
A cohort from people enrolled in the Taiwan National Health Insurance Program N = 23,740,000 Subjects with newly diagnosed pneumococcal pneumonia in 1998–2010 (n = 44,621) Subjects with acute pancreatitis, chronic pancreatitis or pancreatic cancer before index date (n = 414), under 20 years or more than 84 years of age (n = 27,406), or missing information about sex or age (n = 92) were excluded.
Finally, 16,709 pneumococcal pneumonia subjects without acute pancreatitis at baseline
Subjects without pneumonia
Using the same exclusion criteria as the pneumococcal pneumonia group and 1 : 4 randomly matched with sex-, age-, and index year
Finally, 66,836 comparison subjects without acute pancreatitis at baseline
Total, 83,545 pneumococcal pneumonia subjects and comparison subjects in the study cohort. Follow–up until diagnosis of acute pancreatitis, or the time of subject being censored, or to the end of 2010 Figure 1 Flowchart of participant selection in this study ª 2015 John Wiley & Sons Ltd Int J Clin Pract, May 2015, 69, 5, 611–617
Pneumococcal pneumonia and acute pancreatitis
codes 574), diabetes mellitus (ICD-9 codes 250), hepatitis B (ICD-9 codes V02.61, 070.20, 070.22, 070.30 and 070.32), hepatitis C (ICD-9 codes V02.62, 070.41, 070.44, 070.51 and 070.54) and hypertriglyceridaemia (ICD-9 codes 272.1), which were associated with acute pancreatitis found in previous studies (7–9, 18).
Statistical analysis The v2 test was used to compare categorical variables for demographic characteristics and baseline comorbidities between the pneumonia group and the nonpneumonia group. The overall and follow-up periodspecific incidence rates of acute pancreatitis were examined for both groups. The incidence rate of acute pancreatitis was calculated as the number of acute pancreatitis subjects identified during the follow-up, divided by the total follow-up person-years for each group. The incidence rates of acute pancreatitis between the pneumonia group and the non-pneumonia group were compared by an univariable Cox proportional hazards regression model in all study subjects, and in gender-, age- and follow-up period-specific subgroups. To calculate the hazard ratio (HR) and 95% confidence interval (95% CI) of acute pancreatitis associated with pneumococcal pneumonia and other comorbidities, initially, all variables were included in an univariable Cox proportional hazards regression model. Variables detected significantly in the univari-
able analysis were further included in a multivariable Cox proportional hazards regression model after simultaneously adjusting for demographic characteristics and comorbidities. The time-to-event analysis (Kaplan–Meier method) and the log-rank test were used to examine the cumulative incidences for the pneumonia group and the non-pneumonia group. All statistical analyses were performed by using the SAS software version 9.1 (SAS Institute Inc., Cary, NC). The tests of significance were two-sided at p < 0.05.
Results Baseline characteristics of the study population There were 16,709 subjects in the pneumonia group and 66,836 subjects in the non-pneumonia group with equal distributions in gender and age (Table 1). The mean ages (SD) were 63.4 (16.8) years in the pneumonia group and 62.5 (16.7) years in the nonpneumonia group. There were higher prevalence rates of alcoholism, biliary stone, diabetes mellitus, hepatitis B, hepatitis C and hypertriglyceridaemia in the pneumonia group than those in the non-pneumonia group (p < 0.0001 for all).
Overall and follow-up period-specific incidences of acute pancreatitis Figure 2 illustrates that the cumulative incidence of acute pancreatitis was higher in the pneumonia group than that in the non-pneumonia for 0.48% at
Table 1 Baseline characteristics between pneumococcal pneumonia group and non-pneumonia group
Gender Male Female Age group (years) 20–39 40–64 65–84 Baseline comorbidities Alcoholism Biliary stone Diabetes mellitus Hepatitis B Hepatitis C Hypertriglyceridaemia
Pneumococcal pneumonia, N = 16,709
Non-pneumonia, N = 66,836
n
%
n
%
11,037 5672
66.05 33.95
44,148 22,688
66.05 33.95
0.99
2196 5034 9479
13.14 30.13 56.73
8791 20,129 37,916
13.15 30.12 56.73
0.99
296 979 4396 395 419 222
1.77 5.86 26.31 2.36 2.51 1.33
120 1685 4646 326 344 255
0.18 2.52 6.95 0.49 0.51 0.38
p value
< < < < <
Increased risk of acute pancreatitis following pneumococcal pneumonia: a nationwide cohort study S.-W. Lai,1,2 C.-L. Lin,3,4 K.-F. Liao,5,6 C.-L. Ma6
1
SUMMARY
What’s known
Background: The purpose of this study was to evaluate the risk of acute pancreatitis following pneumococcal pneumonia in Taiwan. Methods: We undertook a retrospective cohort study using the hospitalisation claims data of the Taiwan National Health Insurance Program. We identified 16709 subjects aged 20–84 with the first-attack of pneumococcal pneumonia between 1998 and 2010 as the pneumonia group and we randomly selected 66836 subjects without a history of pneumonia as the non-pneumonia group. Both groups were matched for gender, age and index year. We examined the incidence of acute pancreatitis by the end of 2010 and we used a multivariable Cox proportional hazards regression model to calculate the hazard ratio (HR) and 95% confidence interval (95% CI) of acute pancreatitis associated with pneumococcal pneumonia and other comorbidities. Results: Subjects with pneumococcal pneumonia had higher incidence of acute pancreatitis, when compared with non-pneumonia subjects (2.41 vs. 1.47 per 1000 person-years, crude HR 1.65, 95% CI = 1.38, 1.97). The highest risk of developing acute pancreatitis occurred during the first 3 months after diagnosing pneumococcal pneumonia (crude HR 4.11, 95% CI 1.98, 8.52). After adjusted for potential confounders, the adjusted HR of acute pancreatitis was 1.51 (95% CI 1.25, 1.82) for the pneumonia group, as compared with the non-pneumonia group. Conclusions: Overall, this study reveals a 51% increased hazard of acute pancreatitis following infection with pneumococcal pneumonia. Patients with pneumococcal pneumonia should receive close surveillance for risk of developing acute pancreatitis during the first 3 months after diagnosing pneumococcal pneumonia.
Introduction Pneumococcal pneumonia is a common disease with a major concern because of its significant morbidity and mortality. In acute stage of pneumococcal pneumonia, septic shock, acute respiratory failure, bacteraemia, pleural effusion and empyema are frequently found (1,2). The mortality rate is about 1.3–36%, depending on host factors, pneumococcal serotypes and disease severity (1–3). Recent research has revealed subsequent severe outcomes after acute stage. Pneumococcal pneumonia correlates with increased risk of stroke, acute coronary syndrome and lung cancer (4–6), but acute pancreatitis is not included till now. Acute pancreatitis is another public health concern because of its significant morbidity and mortality. Extensive evidence has shown that some factors, including alcoholism, biliary stone, diabetes mellitus,
ª 2015 John Wiley & Sons Ltd Int J Clin Pract, May 2015, 69, 5, 611–617. doi: 10.1111/ijcp.12590
• •
Pneumococcal pneumonia is a common disease with a major concern because of its significant morbidity and mortality. Pneumococcal pneumonia correlates with increased risk of stroke, acute coronary syndrome and lung cancer, but acute pancreatitis is not included till now.
What’s new
• •
Pneumococcal pneumonia is associated with 1.51-fold increased hazard of developing acute pancreatitis. The highest risk of developing acute pancreatitis occurred during the first 3 months after diagnosing pneumococcal pneumonia.
hepatitis C infection and hypertriglyceridaemia, significantly correlate with increased risk of acute pancreatitis (7–9), but pneumococcal pneumonia is not mentioned till now. Because pneumococcal pneumonia and acute pancreatitis are both inflammatory diseases, we make a novel hypothesis that there could be a relation between pneumococcal pneumonia and acute pancreatitis based on inflammatory theory that pneumolysin and/or other virulence factors released from the pneumococcus presumably may be involved in the pathogenesis of pancreatic inflammation (10–12). If pneumococcal pneumonia significantly correlates with increased risk of developing acute pancreatitis, close surveillance for risk of developing acute pancreatitis in patients with pneumococcal pneumonia can be suggested. Therefore, with the aim of better understanding this issue, we conducted a nationwide cohort study.
School of Medicine, China Medical University, Taichung, Taiwan 2 Department of Family Medicine, China Medical University Hospital, Taichung, Taiwan 3 Department of Public Health, China Medical University, Taichung, Taiwan 4 Management Office for Health Data, China Medical University Hospital, Taichung, Taiwan 5 Graduate Institute of Integrated Medicine, China Medical University, Taichung, Taiwan 6 Department of Internal Medicine, Taichung Tzu Chi General Hospital, Taichung, Taiwan Correspondence to: Kuan-Fu Liao, Department of Internal Medicine, Taichung Tzu Chi General Hospital, No.66, Sec. 1, Fongsing Road, Tanzi District, Taichung City, 427, Taiwan Tel.: + 886 4 2205 2121 Fax: + 886 4 2203 3986 Email: [email protected]. tw
Disclosure The authors disclose no conflicts of interest.
611
612
Pneumococcal pneumonia and acute pancreatitis
Materials and methods Data sources We undertook a retrospective nationwide cohort study using the database of the Taiwan National Health Insurance Program. This program started in March 1995 and almost 99% of 23 million citizens joined this program (13). This program provided outpatient, hospitalisation and emergency services. The diagnostic codes are based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9 codes). The details of this program have been published previously (14–17). This study was approved by the Institutional Review Board (IRB) of China Medical University and Hospital (CMU-REC-101-012).
Participants Figure 1 illustrates flowchart of participant selection in this study. We identified subjects aged 20–84 who were newly and mainly diagnosed pneumococcal pneumonia from the hospitalisation dataset between 1998 and 2010 as the pneumonia group (ICD-9 codes 481). The date of diagnosing pneumococcal pneumonia was defined as the index date. To
increase the statistical power, fourfolds of comparison subjects without a history of pneumonia were randomly selected from all National Health Insurance beneficiaries as the non-pneumonia group (ICD-9 codes 480–487). The non-pneumonia group was matched for the pneumonia group by gender, age (every 5-year span) and the index year of diagnosing pneumococcal pneumonia. In both groups, subjects with history of acute pancreatitis, chronic pancreatitis or pancreatic cancer at baseline, or missing information about gender or age, were excluded from this study.
Main outcome measurement All study subjects were monitored until the date on which they were newly and mainly diagnosed with acute pancreatitis based on the hospitalisation dataset (ICD-9 codes 577.0), or being censored because of loss of follow-up, death or withdrawal from insurance, or to the end of 2010.
Comorbidity assessment Baseline comorbidities were analysed as follows: alcoholism (ICD-9 codes 291, 303, 305.00, 305.01, 305.02, 305.03, 790.3 and V11.3), biliary stone (ICD-9
A cohort from people enrolled in the Taiwan National Health Insurance Program N = 23,740,000 Subjects with newly diagnosed pneumococcal pneumonia in 1998–2010 (n = 44,621) Subjects with acute pancreatitis, chronic pancreatitis or pancreatic cancer before index date (n = 414), under 20 years or more than 84 years of age (n = 27,406), or missing information about sex or age (n = 92) were excluded.
Finally, 16,709 pneumococcal pneumonia subjects without acute pancreatitis at baseline
Subjects without pneumonia
Using the same exclusion criteria as the pneumococcal pneumonia group and 1 : 4 randomly matched with sex-, age-, and index year
Finally, 66,836 comparison subjects without acute pancreatitis at baseline
Total, 83,545 pneumococcal pneumonia subjects and comparison subjects in the study cohort. Follow–up until diagnosis of acute pancreatitis, or the time of subject being censored, or to the end of 2010 Figure 1 Flowchart of participant selection in this study ª 2015 John Wiley & Sons Ltd Int J Clin Pract, May 2015, 69, 5, 611–617
Pneumococcal pneumonia and acute pancreatitis
codes 574), diabetes mellitus (ICD-9 codes 250), hepatitis B (ICD-9 codes V02.61, 070.20, 070.22, 070.30 and 070.32), hepatitis C (ICD-9 codes V02.62, 070.41, 070.44, 070.51 and 070.54) and hypertriglyceridaemia (ICD-9 codes 272.1), which were associated with acute pancreatitis found in previous studies (7–9, 18).
Statistical analysis The v2 test was used to compare categorical variables for demographic characteristics and baseline comorbidities between the pneumonia group and the nonpneumonia group. The overall and follow-up periodspecific incidence rates of acute pancreatitis were examined for both groups. The incidence rate of acute pancreatitis was calculated as the number of acute pancreatitis subjects identified during the follow-up, divided by the total follow-up person-years for each group. The incidence rates of acute pancreatitis between the pneumonia group and the non-pneumonia group were compared by an univariable Cox proportional hazards regression model in all study subjects, and in gender-, age- and follow-up period-specific subgroups. To calculate the hazard ratio (HR) and 95% confidence interval (95% CI) of acute pancreatitis associated with pneumococcal pneumonia and other comorbidities, initially, all variables were included in an univariable Cox proportional hazards regression model. Variables detected significantly in the univari-
able analysis were further included in a multivariable Cox proportional hazards regression model after simultaneously adjusting for demographic characteristics and comorbidities. The time-to-event analysis (Kaplan–Meier method) and the log-rank test were used to examine the cumulative incidences for the pneumonia group and the non-pneumonia group. All statistical analyses were performed by using the SAS software version 9.1 (SAS Institute Inc., Cary, NC). The tests of significance were two-sided at p < 0.05.
Results Baseline characteristics of the study population There were 16,709 subjects in the pneumonia group and 66,836 subjects in the non-pneumonia group with equal distributions in gender and age (Table 1). The mean ages (SD) were 63.4 (16.8) years in the pneumonia group and 62.5 (16.7) years in the nonpneumonia group. There were higher prevalence rates of alcoholism, biliary stone, diabetes mellitus, hepatitis B, hepatitis C and hypertriglyceridaemia in the pneumonia group than those in the non-pneumonia group (p < 0.0001 for all).
Overall and follow-up period-specific incidences of acute pancreatitis Figure 2 illustrates that the cumulative incidence of acute pancreatitis was higher in the pneumonia group than that in the non-pneumonia for 0.48% at
Table 1 Baseline characteristics between pneumococcal pneumonia group and non-pneumonia group
Gender Male Female Age group (years) 20–39 40–64 65–84 Baseline comorbidities Alcoholism Biliary stone Diabetes mellitus Hepatitis B Hepatitis C Hypertriglyceridaemia
Pneumococcal pneumonia, N = 16,709
Non-pneumonia, N = 66,836
n
%
n
%
11,037 5672
66.05 33.95
44,148 22,688
66.05 33.95
0.99
2196 5034 9479
13.14 30.13 56.73
8791 20,129 37,916
13.15 30.12 56.73
0.99
296 979 4396 395 419 222
1.77 5.86 26.31 2.36 2.51 1.33
120 1685 4646 326 344 255
0.18 2.52 6.95 0.49 0.51 0.38
p value
< < < < <
