Trends and seasonality in hospitalizations for acute alcohol-related and biliary pancreatitis in the USA

Abstract

Background

The incidence of acute pancreatitis (AP) is characterized by circannual and geographical variation. The aim of this study was to describe seasonal variation and trends in hospitalizations for AP in the USA with respect to AP etiology.

Methods

The Nationwide Inpatient Sample data (2000–2016) from the Healthcare Cost and Utilization Project were used. The study population included all primary hospitalizations for AP. Biliary AP (BAP) and alcohol-induced AP (AAP) were distinguished by diagnostic and procedural ICD codes. Seasonal trend decomposition was performed.

Results

There was a linear increase in annual incidence (per 100 000 population) of AAP in the USA (from 17.0 in 2000 to 22.9 in 2016), while incidence of BAP, equaled 19.9 in 2000, peaked at 22.1 in 2006 and decreased to 17.4 in 2016. AP incidence demonstrated 18% annual incidence amplitude with summer peak and winter trough, more prominent in AAP. In 2016, within AAP, the highest incidence (per 100 000 population) was noted among African-Americans (up to 50.4), followed by males aged 56–70 years (26.5) and Asians of low income (25.5); within BAP, above the average incidence was observed in Hispanic (up to 25.8) and Asian (up to 25.0) population. The most consistent and rapid increase in AP incidence was noted in males aged 56–70 years with an alcoholic etiology (average 6% annual incidence growth).

Conclusions

The incidence and annual trends of AP vary significantly among demographic and socioeconomic groups and this knowledge may be useful for the planning of healthcare resources and identification of at-risk populations.

Introduction

Acute pancreatitis (AP) ranks seventeenth among all causes of hospitalizations in the USA and first among all gastrointestinal diseases [1]. Two major causes of AP, biliary stones and alcohol use, comprise 28%−42% and 20%−40% of all AP admissions, respectively [2]. The incidence, demographics, and etiology of AP vary substantially in different populations as have been described in several recent epidemiological studies [2], [3], [4].

Over the past few decades, the incidence of AP has continued to increase worldwide [4], [5], [6], [7], [8], [9]. In the USA, the incidence increased from 55 to 90 per 100 000 population from 1993 to 2014 [1]. The obesity epidemic in the USA, which predisposes patients to gallstones formation and higher triglyceride levels, and a wider use of serum pancreatic enzymes assays, are among possible explanations of the observed rise in AP incidence [4,10]. The epidemiology and trends of AP in the USA in the last decade with respect to age groups, sex, race and socioeconomic status (SES) are not well documented.

AP incidence is not only changing over the course of years, but is also subject to prominent circannual, or seasonal, fluctuations. Knowledge of these fluctuations can help with healthcare resources planning and perhaps can improve the understanding of potential etiological factors of AP. Seasonal variations in incidence of various pathologies, including acute surgical conditions, are well-established [[11], [12]–13]. The seasonal variation of AP has also been described in different countries [6,8,[14], [15], [16]–17]. Despite certain population-specific discrepancies, most groups reported a predisposition to AP during warm seasons, particularly in alcohol-related AP (AAP), compared to biliary AP (BAP). Data on the association of peaks in AP incidence with festival seasons appear to be conflicting [8,18]. Other studies have investigated the correlation between metrics and patterns of alcohol consumption and AP incidence with discordant results [[18], [19], [20], [21], [22]–23]. Although there are several studies on the seasonality of AP from Europe [2–4,6], data from the USA are relatively lacking [7]. Another critical point is the methodology used for analysis of incidence: most studies used raw AP incidence and conventional statistical approaches to compare interseason or intermonth variability. However, it is important to understand that the variability in incidence is better assessed with a time-series analysis, such as seasonal decomposition procedures [24]. This approach is widely used in econometrics and natural sciences for forecasting and allows one to expose nonobvious seasonality.

The aims of this study were to detect populations that pose a prominent impact into overall AP incidence and with most rapid increase in AP rates, as well as description of seasonal variation of AP admissions, using modern methods of time-series analysis.