Study design biases in pancreatic inflammatory diseases
- Giulia Martina Cavestro1,
- Alberto Mariani1,
- Satish K Singh2,
- Paolo Giorgio Arcidiacono1,
- Pier Alberto Testoni1
- 1Vita-Salute San Raffaele University Division of Gastroenterology and Gastrointestinal Endoscopy, Scientific Institute San Raffaele, Milano, Italy
- 2Section of Gastroenterology, Boston University School of Medicine and Boston Medical Center, Boston, Massachusetts, USA
- Correspondence to Dr Giulia Martina Cavestro, Vita-Salute San Raffaele University Division of Gastroenterology and Gastrointestinal Endoscopy, Scientific Institute San Raffaele, Via Olgettina 60 Milano, Italy;
Contributors All authors contributed as commentators.
- Accepted 21 March 2012
- Published Online First 1 June 2012
The recent report by Sadr-Azodi et al entitled ‘Cigarette smoking, smoking cessation and acute pancreatitis: a prospective population-based study,’1 suggests that smoking is an important risk factor for non-gallstone-related acute pancreatitis.
In this study, the examined populations consisted of participants from the Swedish Mammography Cohort (SMC) established between 1987 and 1990. A questionnaire pertaining to diet, body size and education was mailed to 90 303 women born between 1914 and 1948; 66 651 women (74%) returned the questionnaire. In 1997, all surviving participants (56 030) received a new expanded questionnaire regarding anthropometric and lifestyle factors, including smoking status: 38 988 women returned a completed questionnaire. In addition in 1997, a questionnaire identical to the one sent to the SMC was mailed to a cohort of 100 303 Swedish men (COSM) born between 1918 and 1952 to which 50 000 responded. A diagnosis of acute pancreatitis was based on individual hospital discharge diagnoses from the Swedish inpatient registry. Only subjects with a first-time diagnosis of acute pancreatitis after January 1998 without a previous history of chronic pancreatitis were included in the analysis.
We have several concerns with this study. As stated in the Swedish National Data Service website, the general aim of the SMC is to assess relationships between a number of modifiable factors and the occurrence of several major chronic diseases. COSM is a multidisciplinary longitudinal project on the association of lifestyle and genetic factors with morbidity and mortality in men. Both of these cohorts studied middle-aged and older populations by excluding, by design, individuals <50 years of age, thus introducing age-related bias. Beyond this, the two cohorts were primarily chosen to study the development of chronic diseases like chronic pancreatitis—not to study acute diseases like acute pancreatitis. The authors neglected to examine the association of smoking habits with chronic pancreatitis where the pathophysiology is equally, if not more mechanistically, linked to smoking than acute pancreatitis. Furthermore, it is well known that acute pancreatitis may simply represent an early clinical manifestation of chronic pancreatitis. It is unclear if the patients reported with a first attack of acute pancreatitis received initial and/or follow-up cross-sectional imaging to define whether their disease progressed or was part of the disease spectrum of chronic pancreatitis. Indeed, in a recent prospective 30-year follow-up study of the Danish registries,2 24.1% of patients progressed to chronic pancreatitis within 3.5 years of their first admission for acute pancreatitis. This is in agreement with Ammann et al 3 who found a mean interval of 5.5 years to progression to chronic pancreatitis, although more recently Lankisch et al have suggested that progression can occur over an even longer time-frame of a decade.4 Finally, in a recent retrospective report of the Swedish National Patient Registry,5 coding was found to be highly accurate for the diagnosis of acute pancreatitis, but there were a significant number of patients with chronic pancreatitis who were coded as having acute pancreatitis.
Overall, we believe that an accurate classification of the disease is crucial to better define which subjects, carrying certain morphological conditions in association with exogenous factors, are at risk of developing acute versus chronic pancreatitis. Optimal classification of pancreatic inflammatory disease requires both clinical and radiological assessment criteria. More specifically, patient follow-up by means of cross-sectional imaging is required in order to correctly classify acute versus chronic pancreatitis. The diagnosis of acute pancreatitis based only on individual hospital discharge codes can, we believe, grossly underestimate the prevalence of chronic pancreatitis in a subject population.
Competing interests None.
Provenance and peer review Not commissioned; internally peer reviewed.