Article Text
Abstract
Background Several studies have shown that smoking increases the risk of chronic pancreatitis. However, the impact of smoking on the development of acute pancreatitis has not been fully studied.
Objective To clarify the association between cigarette smoking, smoking cessation and the risk of acute pancreatitis.
Design A follow-up study was conducted of 84 667 Swedish women and men, aged 46–84, during 12 years to study the association between smoking status, smoking intensity and duration, duration of smoking cessation and the risk of acute pancreatitis. Only those with the first event of the disease and no previous history of acute pancreatitis were included. Cox proportional hazards models were used to estimate rate ratios (RRs) with 95% CI for different smoking-related variables, adjusted for age, gender, body mass index, diabetes, educational level and alcohol consumption.
Results In total, 307 cases with non-gallstone-related and 234 cases with gallstone-related acute pancreatitis were identified. The risk of non-gallstone-related acute pancreatitis was more than double (RR=2.29; 95% CI 1.63 to 3.22, p<0.01) among current smokers with ≥20 pack-years of smoking as compared with never-smokers. The corresponding risk among individuals with ≥400 g monthly consumption of alcohol was increased more than fourfold (RR=4.12; 95% CI 1.98 to 8.60, p<0.01). The duration of smoking rather than smoking intensity increased the risk of non-gallstone-related acute pancreatitis. After two decades of smoking cessation the risk of non-gallstone-related acute pancreatitis was reduced to a level comparable to that of non-smokers. There was no association between smoking and gallstone-related acute pancreatitis.
Conclusion Smoking is an important risk factor for non-gallstone-related acute pancreatitis. Early smoking cessation should be recommended as a part of the clinical management of patients with acute pancreatitis.
- Acute pancreatitis
- smoking
- smoking cessation
- epidemiology
- abdominal surgery
Statistics from Altmetric.com
Significance of this study
What is already known about this subject?
Smoking is a risk factor for chronic pancreatitis.
The effect of smoking on the risk of acute pancreatitis has rarely been studied.
The relative effect of smoking intensity and smoking duration on development of the disease is unknown.
Possible benefits of smoking cessation on the development of acute pancreatitis have not been fully investigated.
What are the new findings?
Smoking is an important risk factor for the development of non-gallstone-related acute pancreatitis, but there is no association between smoking and gallstone-related disease.
The duration of smoking rather than the smoking intensity increases the risk of non-gallstone-related acute pancreatitis.
After two decades of smoking cessation the risk of non-gallstone-related acute pancreatitis is reduced to a level comparable to that of never-smokers.
How might it impact on clinical practice in the foreseeable future?
Smoking cessation should be recommended and initiated in all patients with non-gallstone-related acute pancreatitis.
Introduction
Acute pancreatitis is a common gastrointestinal diagnosis, which in 20% of patients progresses to a life-threatening condition with multi-organ failure.1 2 The incidence of the disease has grown steadily during the past decades in Europe and North America.3–8 During recent years, smoking has emerged as an important risk factor in the pathophysiology of chronic pancreatitis.9–12 In experimental animal studies, smoking has been shown to induce pathological and functional changes in the pancreas.13–15 Whereas smoking has been observed to accelerate the development of alcohol-related chronic pancreatitis,16 17 smoking cessation has been reported to reduce the disease progression.18
Two studies have investigated the role of smoking in the aetiology of acute pancreatitis,9 19 both showing smoking to be an independent risk factor of acute pancreatitis. These studies, however, had some limitations. Pack-years of smoking—that is, the average number of packs of cigarettes smoked per day multiplied by the number of years a person has smoked, were used to evaluate dose–response association between smoking and the risk of acute pancreatitis.9 19 Therefore, the relative effects of smoking intensity and duration cannot be discriminated by pack-years. The effect of smoking cessation on the development of acute pancreatitis was investigated in one study.9 However, there was no information on the duration of smoking cessation making it difficult to assess the possible benefits of smoking cessation on the development of acute pancreatitis. Moreover, in one of the two studies there was a lack of information about the amount of alcohol consumed.19
We used a large population-based cohort of Swedish men (COSM) and women enrolled in a prospectively collected database to study the association between smoking status, lifetime exposure of cigarette smoking expressed as smoking intensity and duration, the duration of smoking cessation and the risk of acute pancreatitis.
Methods
Study population
The study population consisted of participants in the Swedish Mammography Cohort (SMC)20 and the COSM21 described in detail elsewhere. Briefly, the SMC was established between 1987 and 1990, when all women (90 303) born between 1914 and 1948 and residing in central Sweden (Västmanland and Uppsala counties) received a mailed questionnaire on diet, body size and education. A total of 66 651 women (74%) returned the questionnaire. In the fall of 1997, all surviving participants (56 030) received a new expanded questionnaire that included ∼350 items concerning anthropometry, diet, alcoholic beverages and other lifestyle factors, including smoking status and medical history of diabetes mellitus (additional information about the history was obtained by computerised linkage of the study population with the Swedish Patient Register22); 38 988 women returned a completed questionnaire. The COSM was initiated in the fall of 1997, when all men (100 303) born between 1918 and 1952 and residing in central Sweden (Västmanland and Örebro counties) received a mailed questionnaire that was identical (except for some sex-specific questions) to the SMC 1997 questionnaire; 48 645 men (48.5%) answered the questionnaire. The cohorts represent well the Swedish population for age distribution, relative body weight, educational level23 24 as well as incidence rates of acute pancreatitis. For example, in 1998 the age-adjusted incidence rates (per 100 000 people) of acute pancreatitis in Sweden (excluding individuals between 0 and 49 years)25 and in this cohort were 66 versus 64 for men and 45 versus 46 for women, respectively. The study was approved by the regional ethical board at Karolinska Institute.
Ascertainment of cases with acute pancreatitis
Established in 1964, the National Board of Health and Welfare has collected data on individual hospital discharges in the Swedish Inpatient Register. The national coverage of this register was 60% in 1969, 85% in 1983 and 100% since 1987.22 In addition to the national registration number (NRN) (uniquely identifying each resident in Sweden), each record contains as many as eight discharge diagnoses coded according to the current International Classification of Diseases (ICD) and up to 12 surgical codes assigned according to the Swedish Classification of Operations and Major Procedures. The register has been evaluated for validity and completeness, and the codes for the main diagnoses were correct at three-digit level for 92–94% of the records for surgical patients. For surgical procedures (excluding endoscopies or biopsies), the codes were incorrect in 2% of the records and were missing in 5.3%.26
Through the linkage of the cohort to the Inpatient Register, all individual discharges from hospital due to acute pancreatitis (ICD 10: K850–9, ICD-9: 577A, ICD-8: 577.00–08) and chronic pancreatitis (ICD-10: K86–9, ICD-9: 577B, ICD-8: 577.09–19) were identified. Only the first-time event of acute pancreatitis after 1 January 1998 with no previous history of chronic pancreatitis was included in the analyses. Gallstone-related pancreatitis was classified as that requiring surgical procedures owing to gallstone-related disease such as cholecystectomy (JKA20, JAK21), endoscopic-retrograde cholangiography (JKE00, JKE02, JKE12, JKE18) or percutaneous cholecystostomy (JKB30) within 3 months after the diagnosis of acute pancreatitis using the Swedish Classification of Operations and Major Procedures. All other patients were classified as acute non-gallstone-related pancreatitis.
Follow-up of the cohorts
The SMC and COSM cohorts were followed-up between 1 January 1998 and 31 December 2009. Information for the cohorts on the dates of death was obtained from the Swedish Death Register at Statistics Sweden, and information about cancer diagnoses from the Swedish Cancer Register.
Exclusions
Excluded from the databases were individuals with previous diagnosis of cancer before 1 January 1998 (except for non-melanoma skin cancer), participants who returned an incomplete questionnaire or had an erroneous or missing NRN, and no follow-up was carried out on these individuals. In addition, exclusions included those diagnosed with pancreatic cancer after entry to the cohorts (n=227). After these exclusions, 84 667 participants (38 886 women and 45 781 men), aged 46–84 years in 1998, remained for analyses in this study cohort (figure 1).
Assessment of smoking exposure
The 1997 questionnaire included detailed information about lifetime history of smoking, age when starting smoking and age at smoking cessation. Pack-years were estimated from the smoking history by multiplying the number of years of smoking by the average number of packs of cigarettes smoked. Years since smoking cessation were divided into <10, 10–14.9, 15–19.9, 20–24.9, 25–29.9, ≥30 years.
Statistical analysis
Person-years were calculated from 1 January 1998 until the date of death, admission to hospital due to a discharge diagnosis code of acute pancreatitis or 31 December 2009, whichever occurred first. A Cox proportional hazard regression model was used to estimate the incidence rate ratios (RRs) and 95% CI of the association between smoking status (current, former, never-smoker), pack-years of smoking (current <20 pack-years, current ≥20 pack-years; former <20 pack-years, former ≥20 pack-years, never-smokers), years since smoking cessation (<10, 10–14.9, 15–19.9, 20–24.9, 25–29.9, ≥30 years) and the risk of acute pancreatitis. The multivariable model was adjusted for age (continuous), gender, educational level (primary school, high school, university, missing data), body mass index (kg/m2) (<25, 25–29.9, ≥30, missing data), diabetes and total consumption of alcohol per month (g/month) (≤100, >100–199, 200–299, 300–399, ≥400, missing data). The choice of covariables included in the model was based on clinically relevant factors and before conducting the analysis.
To study the relative importance of smoking intensity and smoking duration, the effect of smoking intensity assessed as the mean number of cigarettes smoked daily during lifetime (<10 or ≥10 cigarettes per day) and years of smoking duration (<20 or ≥20 years) were separated in a model adjusting for confounders mentioned above. All analyses were performed separately for gallstone-related pancreatitis and non-gallstone-related acute pancreatitis. The proportional hazard assumption was evaluated by regressing scaled Schoenfeld residuals against survival time. There was no evidence of departure from the assumption.
Stratified analyses were performed to investigate the role of monthly alcohol consumption (<400 g, ≥400 g) as a potential effect modifier for the relation between smoking-related variables and acute pancreatitis. The choice of this cut-off point was based on the recommendations of the American Heart Association and World Cancer Research Fund/American Institute of Cancer Research's recommendations on alcohol consumption. The ‘allowed’ range of alcohol consumption is 1–2 standard drinks per day, which translates to approximately 400–750 g per month.27 28 We chose the lower limit as the limit for stratification. Since smoking status at baseline might have been secondary to preclinical or chronic illness, a sensitivity analysis was performed by excluding the first 2 years of follow-up. Furthermore, to investigate if baseline smoking habits changed during the follow-up, stratified analyses were performed by separating the analyses over those who developed acute pancreatitis at <6 or ≥6 years of follow-up. All statistical analyses were performed using Stata, V.11.0 (StataCorp).
Results
During a median of 12 years of follow-up (981 187 person-years), 541 patients with first-time acute pancreatitis were identified. Among them 307 (56.7%) had non-gallstone-related pancreatitis of whom 198 (64.5%) were men and 109 (35.5%) were women. Of the remaining 234 cases with gallstone-related pancreatitis, 102 (43.6%) were men and 132 (56.4%) women. The distribution of baseline characteristics according to smoking status is shown in table 1. Current and former smokers were less likely to be women, to have university education and were more likely to consume ≥400 g of alcohol per month.
The multivariable-adjusted RRs and the corresponding 95% CI for the risk of acute pancreatitis with respect to smoking status, pack-years of smoking and the duration of smoking cessation are shown in table 2. When adjusted for potential confounders, current and former smokers were more likely to develop acute non-gallstone-related pancreatitis as compared with never-smokers. This risk was the highest (RR=2.29; 95% CI 1.63 to 3.22, p<0.01) among current smokers with ≥20 pack-years of smoking as compared with never-smokers. The risk of acute non-gallstone-related pancreatitis was reduced to a level comparable to that of never-smokers after 20 years of smoking cessation. There was no statistically significant association between smoking status, pack-years of smoking or smoking cessation and the risk of gallstone-related pancreatitis.
The risk of acute pancreatitis with respect to the average number of cigarettes smoked per day and the duration of smoking is shown in table 3. There was no statistically significant association between duration of smoking <20 years, irrespective of smoking intensity, and the risk of non-gallstone-related pancreatitis. However, this risk was significantly increased among those who had smoked ≥20 years compared with never-smokers. The intensity and duration of smoking seem to have an additive effect on this association, increasing the risk of non-gallstone-related pancreatitis to more than twofold (RR=2.20; 95% CI 1.61 to 3.00, p<0.01). There was no significant association between smoking duration or intensity and the risk of gallstone-related acute pancreatitis.
Alcohol consumption modified the effect between smoking and the risk of non-gallstone-related pancreatitis (table 4). Current smokers consuming <400 g alcohol per month had a 63% (RR=1.63; 95% CI 1.09 to 2.45, p=0.02) increased risk of non-gallstone-related pancreatitis compared with never-smokers. The corresponding risk among those with ≥400 g of monthly alcohol consumption was more than twofold (RR=2.10; 95% CI 1.38 to 3.19, p<0.01). Among those with <400 g of monthly alcohol consumption, the risk of non-gallstone-related pancreatitis reduced to a level comparable to that of never-smokers after 10 years of smoking cessation. Those consuming ≥400 g of monthly alcohol achieved this risk reduction after at least double the duration of smoking cessation.
To investigate if the smoking characteristics at baseline were secondary to a preclinical or chronic disease, the first 2 years of follow-up were excluded. This had little influence on the observed associations. For instance, the multivariable RR for acute non-gallstone-related pancreatitis increased by more than twofold (RR=2.16; 95% CI 1.49 to 3.12, p<0.01) in current smokers with ≥20 pack-years of smoking compared with never-smokers. To evaluate if the smoking status at baseline had changed during the study period, separate analyses were performed for those developing acute pancreatitis <6 or ≥6 years of follow-up. Again, the results did not differ substantially. For example, this risk among those with ≥20 pack-years of smoking who developed acute pancreatitis at <6 years of follow-up was more than twofold (RR=2.21; 95% CI 1.34 to 3.67, p<0.01) compared with never-smokers. The corresponding risk in those who developed the disease at ≥6 years of follow-up was practically identical (RR=2.34; 95% CI 1.46 to 3.75, p<0.01). Finally, even though alcohol information was frequently missing, especially in never-smokers, there was no statistically significant association between individuals with missing alcohol information (RR=1.35; 95% CI 0.94 to 1.93, p=0.11) and the risk of acute non-gallstone-related pancreatitis after adjusting for confounders.
Discussion
In this large population-based cohort study of men and women, cigarette smoking was found to be an independent risk factor for the first event of acute non-gallstone-related pancreatitis. Duration of smoking increased the risk of the disease, whereas smoking intensity potentiated this association. Two decades of smoking cessation were required to reduce the risk of non-gallstone-related pancreatitis to the level comparable to never-smokers. The observed risks were substantially higher in individuals consuming ≥400 g of alcohol per month, corresponding to approximately one or more standard drink of alcohol/day. There was no association between smoking and the risk of gallstone-related pancreatitis.
Some methodological aspects deserve attention. The major strengths of this study are the prospective population-based design, its large size, the relatively large number of cases with acute pancreatitis and the completeness of case ascertainment, which together minimise potential recall and selection biases, and improve the generalisability of the findings. Nevertheless, this study has some potential limitations. The estimates of lifetime cigarette smoking were based on retrospective recall. As data on smoking status were collected only at baseline, random misclassification of current and past smoking might have occurred because participants underestimated their smoking habits. Furthermore, the smoking habits might have changed during the follow-up period of 12 years. However, there was no such indication when stratifying the analyses over a median follow-up time of <6 or ≥6 years.
Although the Swedish Patient Register has been proved to be highly accurate,26 the validity of the diagnosis acute pancreatitis has not been evaluated specifically in this register. There might therefore be some misclassification of the outcome. However, since such potential misclassification would be irrespective of the smoking status, the estimates would be biased towards the null. Although all patients had a first event of acute pancreatitis, we cannot rule out the possibility that some of these individuals might have had chronic pancreatitis which was clinically presented as acute pancreatitis. However, by excluding cases with recurrent pancreatitis we reduced this possibility.
While controlling for known risk factors of acute pancreatitis, we cannot rule out the possibility that our findings were affected by residual confounding. For instance hypercalcaemia, extreme levels of serum triglycerides and cystic fibrosis are described as risk factors for acute pancreatitis.29 Yet, the first two comorbidities explain <1% of all cases with acute pancreatitis and cystic fibrosis should not influence the observed associations in this study population. One proxy for high levels of triglycerides may be obesity and the results did not change markedly after adjustment for this factor.
Finally, even though individuals with missing alcohol information did not have increased risk of acute non-gallstone-related pancreatitis, there might be some residual confounding by alcohol in our analyses. However, since missing data on alcohol use was most frequently present in never-smokers, any residual confounding by alcohol use would result in an underestimate of the observed associations.
Two population-based cohort studies of men and women found smokers to have an increased risk of acute pancreatitis.9 19 An increase in the number of pack-years of smoking was positively associated with the risk of the disease.9 Nevertheless, a new finding in this study was that the duration of smoking, rather than smoking intensity, increased the risk of acute pancreatitis. Smoking intensity had, however, an additive effect on this association. Importantly, after two decades of smoking cessation the risk of acute non-gallstone-related pancreatitis was reversed to a level comparable to that of never-smokers. This risk reduction was achieved even more quickly if the monthly alcohol consumption was <400 g However, the reduced risk of acute pancreatitis in those abstaining from smoking does not necessarily rule out a subclinical chronic disease. In agreement with previous results, gallstone-related pancreatitis was not associated with smoking, indicating different pathophysiological mechanisms.9 19
Several experimental studies on rat models have investigated the effect of smoking on the pancreas showing increased inflammatory activity. Exposure to smoking in rats results in infiltration of the pancreas by lymphocytes and plasma cells.30 In addition, as a sign of oxidative stress, the glutathione peroxidase activity is suppressed and interleukin 6 levels are increased.30 Focal inflammation, decreased number of acinar structures and upregulation of genes expressing digestive enzymes were observed in rats after heavy exposure to cigarette smoke.13 15 30 There was also an imbalance between production of the digestive enzyme trypsinogen and the inhibitory antiprotease pancreatic inhibitory trypsin inhibitor, which in turn might be responsible for the morphological damage to the pancreas.13 Exposure to cigarette smoke results also in reduced blood flow and ischaemia in pancreas of rats.31 Alcohol enhances the inflammatory effect of smoking in the pancreas of rats, resulting in more pronounced ischaemia and leucocyte sequestration.31
In conclusion, smoking was shown to be an independent risk factor of non-gallstone-related acute pancreatitis. Duration of smoking increases this risk, whereas smoking intensity potentiates this association. Most importantly, smoking cessation reverses the risk of non-gallstone-related acute pancreatitis to a level comparable to that of never-smokers. This reduced risk was achieved much more quickly in those with <400 g of monthly alcohol consumption. Thus, smoking cessation should be a part of the clinical management of patients with the disease and initiated as soon as possible in patients who have had acute pancreatitis.
References
Footnotes
Funding This work was supported by research grants from the Swedish Cancer Foundation and the Swedish Research Council Committee for Infrastructure. The first author was supported as a postdoctoral researcher by a grant from Olle Engkvist Byggmästare Foundation and the Swedish Society of Medicine. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript.
Competing interests None.
Ethics approval Institutional ethical committee at Karolinska Institutet.
Provenance and peer review Not commissioned; externally peer reviewed.