New Diagnosis of Chronic Pancreatitis
This is a retrospective cohort study. Data for this study were obtained from a national cohort of Veterans Health Administration inpatient and outpatient records using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses maintained by the Veterans Health Administration national medical care data sets from fiscal year 1998. ICD-9 codes were recorded if a patient had any diagnosis or symptom during each visit/encounter. The initial cohort identified for the study was 534,956 patients.
All Veterans Affairs (VA) patients from the study cohort above the age of 25 years were eligible for the study. Patients with only one visit (n=810) and patients with < 3 years of follow-up in the VA database were excluded from the study (60,086). Patients with pancreatic cysts (n=881), patients under 25 years of age (n=1,143) and, among CP patients, patients diagnosed with PaCa either before CP or within 7 days of CP (n=43) were excluded from the analysis as well. Patients had to have at least two visits to the VA from fiscal year 1998 to 2007 and be present in the VA medical system for more than 3 years to be included in the study.
VA patients with CP diagnosis were further grouped into preexisting CP and new diagnosis of CP. Pre-existing CP: any patient with CP diagnosis within first 3 years of entry into the database was grouped as preexisting CP. For these patients, a 3-year washout period was applied from the date of CP diagnosis noted within the database by us. They were included for follow-up in our study after the 3-year washout period. Patients in this group diagnosed with PaCa within the first 3 years (i.e., washout period) were deleted (n=18). Thus, all patients classified as preexisting CP had had the diagnosis of CP at least for 3 years. New diagnosis of CP: patients in whom CP diagnosis was noted for the first time more than 3 years after entry into the database were regarded to have new diagnosis of CP. The remaining VA patients in the database were those without a diagnosis of CP during the washout and the follow-up periods.
A total of 471,992 subjects were included in the final cohort (Figure 1). This study was approved by the Institutional Review Board of Veterans Administration St Louis Health Care System.
(Enlarge Image)
Figure 1.
Study cohort. CP, chronic pancreatitis; FY, fiscal year; PaCa, pancreatic cancer; Pts, patients; VA, Veterans Affairs.
The primary outcome of interest was PaCa, defined as presence of 2 or more ICD codes 157.0, 157.1, 157.2, 157.3, and 157.9 less than 1 year apart. For patients with new CP diagnosis, duration of the follow-up was calculated from the time of diagnosis of CP to the PaCa diagnosis. For the patients with preexisting CP and those without CP, duration of the follow-up was calculated from the day after the 3-year washout period to the diagnosis of PaCa.
The primary predictor of interest was new diagnosis of CP (ICD code 577.1). History of significant alcohol intake (Alcohol abuse/dependence, ICD codes 303, 3030, 3039, and 3050), history of smoking (nicotine dependence, ICD codes 305.1 or V15.82), age at the time of entry into the study, race, and sex were the secondary predictors of interest.
Demographic characteristics (e.g., race, sex) and secondary predictors (e.g., history of alcohol and smoking) were assessed using frequencies and proportions. Age was calculated using mean and s.d. The proportion of patients with PaCa who were initially misdiagnosed as having CP and proportion of patients with new diagnosis of CP subsequently diagnosed to have PaCa were estimated. Incidence rates for PaCa (per 1,000 person-years (py)) among patients with new diagnosis of CP, preexisting CP, and remaining patients in the database each year and for the entire study period were estimated as well. Rate ratios with 95% confidence intervals (CIs) were evaluated using Poisson regression after adjusting for race, sex, age, and history of smoking and alcohol. First-year incidence of PaCa by age was calculated in patients with new diagnosis of CP, preexisting CP, and remaining veterans in the database. Stratified analyses were done to check for the role of alcohol abuse and heavy use of tobacco on new diagnosis of CP and PaCa and their association was evaluated using Cochran–Mantel–Hansel χ test. Time interval between the new diagnosis of CP and subsequent development of PaCa was also calculated. Follow-up time for the study participants ended on the date of diagnosis of pancreatic cancer, last visit to the VA, death, or 30 September 2007 (whichever occurred first). All analyses were conducted using SAS version 9.2 (SAS, Cary, NC).
Methods
Design and Participants
This is a retrospective cohort study. Data for this study were obtained from a national cohort of Veterans Health Administration inpatient and outpatient records using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses maintained by the Veterans Health Administration national medical care data sets from fiscal year 1998. ICD-9 codes were recorded if a patient had any diagnosis or symptom during each visit/encounter. The initial cohort identified for the study was 534,956 patients.
All Veterans Affairs (VA) patients from the study cohort above the age of 25 years were eligible for the study. Patients with only one visit (n=810) and patients with < 3 years of follow-up in the VA database were excluded from the study (60,086). Patients with pancreatic cysts (n=881), patients under 25 years of age (n=1,143) and, among CP patients, patients diagnosed with PaCa either before CP or within 7 days of CP (n=43) were excluded from the analysis as well. Patients had to have at least two visits to the VA from fiscal year 1998 to 2007 and be present in the VA medical system for more than 3 years to be included in the study.
VA patients with CP diagnosis were further grouped into preexisting CP and new diagnosis of CP. Pre-existing CP: any patient with CP diagnosis within first 3 years of entry into the database was grouped as preexisting CP. For these patients, a 3-year washout period was applied from the date of CP diagnosis noted within the database by us. They were included for follow-up in our study after the 3-year washout period. Patients in this group diagnosed with PaCa within the first 3 years (i.e., washout period) were deleted (n=18). Thus, all patients classified as preexisting CP had had the diagnosis of CP at least for 3 years. New diagnosis of CP: patients in whom CP diagnosis was noted for the first time more than 3 years after entry into the database were regarded to have new diagnosis of CP. The remaining VA patients in the database were those without a diagnosis of CP during the washout and the follow-up periods.
A total of 471,992 subjects were included in the final cohort (Figure 1). This study was approved by the Institutional Review Board of Veterans Administration St Louis Health Care System.
(Enlarge Image)
Figure 1.
Study cohort. CP, chronic pancreatitis; FY, fiscal year; PaCa, pancreatic cancer; Pts, patients; VA, Veterans Affairs.
Measurements
The primary outcome of interest was PaCa, defined as presence of 2 or more ICD codes 157.0, 157.1, 157.2, 157.3, and 157.9 less than 1 year apart. For patients with new CP diagnosis, duration of the follow-up was calculated from the time of diagnosis of CP to the PaCa diagnosis. For the patients with preexisting CP and those without CP, duration of the follow-up was calculated from the day after the 3-year washout period to the diagnosis of PaCa.
The primary predictor of interest was new diagnosis of CP (ICD code 577.1). History of significant alcohol intake (Alcohol abuse/dependence, ICD codes 303, 3030, 3039, and 3050), history of smoking (nicotine dependence, ICD codes 305.1 or V15.82), age at the time of entry into the study, race, and sex were the secondary predictors of interest.
Statistical Analyses
Demographic characteristics (e.g., race, sex) and secondary predictors (e.g., history of alcohol and smoking) were assessed using frequencies and proportions. Age was calculated using mean and s.d. The proportion of patients with PaCa who were initially misdiagnosed as having CP and proportion of patients with new diagnosis of CP subsequently diagnosed to have PaCa were estimated. Incidence rates for PaCa (per 1,000 person-years (py)) among patients with new diagnosis of CP, preexisting CP, and remaining patients in the database each year and for the entire study period were estimated as well. Rate ratios with 95% confidence intervals (CIs) were evaluated using Poisson regression after adjusting for race, sex, age, and history of smoking and alcohol. First-year incidence of PaCa by age was calculated in patients with new diagnosis of CP, preexisting CP, and remaining veterans in the database. Stratified analyses were done to check for the role of alcohol abuse and heavy use of tobacco on new diagnosis of CP and PaCa and their association was evaluated using Cochran–Mantel–Hansel χ test. Time interval between the new diagnosis of CP and subsequent development of PaCa was also calculated. Follow-up time for the study participants ended on the date of diagnosis of pancreatic cancer, last visit to the VA, death, or 30 September 2007 (whichever occurred first). All analyses were conducted using SAS version 9.2 (SAS, Cary, NC).
Source...