Association between smoking status and inpatient outcomes of acute cholangitis in the United States: a propensity matched analysis

Introduction

Acute cholangitis (AC) is an infection that is superimposed on biliary tract stasis [1]. It is a relatively uncommon disease, with fewer than 200,000 cases per year in the United States (US) [2]. AC has been associated with high morbidity and mortality, with mortality rates widely ranging from 2-65% [3].

Smoking is known to have an impact on the inpatient outcomes of various medical conditions, including infections [4,5]. It affects innate as well as humoral immune systems, and has been shown to disrupt immunological homeostasis through various mechanisms, including its effects on nuclear factor-κB (NF-κB) and MAPK signaling pathways, as well as histone modification. This can result in either an exaggerated immunological response, leading to autoimmunity, or a weakened immune response that increases susceptibility to infections [6]. Smoking also has been identified as a risk factor for hepatobiliary diseases, including gallstone disease and possibly AC, yet it is unclear what effects smoking has on inpatient outcomes of AC [7-9].

Smoking remains a significant public health issue in the US, despite the fact that smoking rates have been declining in the US in recent years. According to the Centers for Disease Control, 12.5% of adults were current smokers in the US in 2020 [10]. Therefore, it is important to investigate the effects of smoking on inpatient outcomes of AC, as smoking is a major cause of preventable deaths in the US. We analyzed the National Inpatient Sample (NIS), the largest inpatient database in the US, to compare AC-related inpatient outcomes between smokers and non-smokers.

Materials and methods

Statistical analysis

Statistical analysis was conducted using STATA version 17.0 software (StataCorp LLC, College Station, TX, USA). We performed propensity-matched analysis to minimize the effects of imbalances in patient and hospital characteristics and comorbidities between the 2 cohorts [12,13]. Each hospitalization was assigned a propensity score using a multivariate logistic regression model that incorporated patient and hospital characteristics and comorbidities shown in Tables 1 and 2. Propensity scores were matched between the 2 cohorts using 1:1 matching within 0.1 standard deviation of the calculated propensity scores. The balance of covariates between the cohorts was analyzed and visualized using 2-way plots, as shown in Fig. 1 and 2. We used the chi-square test for categorical variables and the Wilcoxon rank-sum (Mann-Whitney) test for continuous variables, given the nonparametric nature of the NIS database. Categorical variables were reported as frequency (N) and percentage (%), while continuous variables were reported as median with interquartile range (IQR). A P-value ≤0.05 was considered statistically significant, and all P-values were two-sided. The study was exempt from institutional review board approval and patient consent as the NIS databases are publicly available and contain de-identified patient information.

Table 1 Biodemographic and hospital characteristics of acute cholangitis-related hospitalizations in the United States in 2017-2020 stratified by smoking status

Table 2 Elixhauser comorbidities among the patients hospitalized with acute cholangitis in 2017-2020 in the United States stratified by smoking status

Figure 1 Covariate balance before propensity matching in the present study

Figure 2 Covariate balance after propensity matching in the present study

Results

Between 2017 and 2020, we identified 43,725 patients who met our inclusion criteria. Prior to matching, 4345 (11.03%) patients were active smokers, while the rest were non-smokers. Using 1:1 matching, we paired 3960 hospitalizations from the smoker cohort (cases) with an equal number from the non-smoker cohort (controls). Propensity scores were generated using hospital-level variables (Table 1) and comorbidities (Table 2), and covariate balance was assessed with two-way plots (Fig. 1, before matching, and Fig. 2, after matching).

No significant differences in major comorbidities were observed between the matched cohorts (P>0.05). Most patients were white, hospitalized on weekdays and treated in large teaching hospitals (Table 1). The most prevalent risk factors for AC were bile duct calculi, bile duct stricture or stenosis, cholelithiasis, other bile duct disorders and bile duct malignancy (Table 3). Smokers showed a stronger association with bile duct calculi (74.37% vs. 69.29%, P<0.001) and other bile duct disorders, including clots, parasites and extrinsic compression (6.82% vs. 5.05%, P=0.011).

Table 3 Risk factors for acute cholangitis among acute cholangitis-related hospitalizations in the United States in 2017-2020 stratified by smoking status

After propensity matching, a higher proportion of smokers died, but the difference was not statistically significant-although it nearly reaching significance (1.01% vs. 0.63%, P=0.063). No significant difference in LOS or MIC was found between the 2 cohorts (P>0.05). Smoking was associated with a higher risk of complications, including sepsis without shock (0.88% vs. 0.51%, P=0.042), sepsis with shock (1.26% vs. 0.51%, P<0.001), biliary pancreatitis (6.57% vs. 4.42%, P<0.001), and myocardial infarction (6.19% vs. 3.54%, P<0.001). Additionally, smokers had a higher need for inpatient ERCP (72.85% vs. 63.76%, P<0.001) and early ERCP (50.76% vs. 42.32%, P<0.001) compared to non-smokers (Table 4).

Table 4 Primary and secondary outcomes of acute cholangitis related hospitalizations in the United States in 2017-2020 stratified by smoking status

Discussion

Smoking affects both innate and humoral immunity. It mediates its impact on immunity through several mechanisms, including the regulation of NF-κB and MAPK pathways, as well as histone modification. This can result in either an unregulated and abnormally exacerbated immune response, potentially leading to autoimmunity, or a weakened immune response, which can cause infections [6]. Previous studies have demonstrated that smoking is associated with longer hospital stays, greater morbidity and higher inpatient mortality compared to non-smokers [5,14,15]. Moreover, research has shown a connection between smoking status and autoimmune diseases [16,17]. Smoking may elevate the risk of developing gallbladder disease, through an unknown mechanism, and it also increases the risk of diabetes and reduces plasma high-density lipoproteins, both of which are potential risk factors for gallbladder disease development. A higher risk of gallbladder disease may heighten the risk of developing AC in smokers. Although smoking has been linked to gallbladder disease, and may increase the risk of AC, its effects on inpatient outcomes related to AC remain largely unknown. Therefore, we conducted a retrospective study on hospitalizations related to acute cholangitis in the United States between 2017 and 2020 to better understand the relationship between smoking status and inpatient outcomes for acute cholangitis.

Prior to matching, the non-smoker AC cohort had a significantly larger number of patients compared to the smoker AC cohort. The non-smoker AC patients were generally older (67±0.08 vs. 56±0.27 years), resulting in a higher proportion of patients with an Elixhauser Comorbidity Index of 3 or greater. The smoker AC cohort contained more males than females, which could be attributed to sex-related differences in smoking prevalence within the US [18,19].

In the matched cohorts, the most prevalent risk factor for AC was bile duct calculi, bile duct stricture/stenosis, cholelithiasis, and other bile duct disorders (including clots, parasites, and exogenous compression). Consistently with previous studies, bile duct calculi were identified as the most common cause of AC, followed by biliary stricture and malignancy [3]. Bile duct calculi and other bile duct disorders were more frequently observed in AC patients who were smokers compared to non-smokers.

The AC patients who were smokers had a higher proportion of in-hospital deaths; however, the difference was not statistically significant, though it nearly reached significance (1.01% vs. 0.63%, P=0.063). No statistically significant differences were observed in the LOS or morbidity between the two cohorts (P>0.05). Nevertheless, smoking in AC patients was associated with a greater need for inpatient ERCP (72.85% vs. 63.76%, P<0.001) and early ERCP (50.76% vs. 42.32%, P<0.001), and higher odds of developing complications, including sepsis without shock (0.88% vs. 0.51%, P=0.042), sepsis with shock (1.26% vs. 0.51%, P<0.001), biliary pancreatitis (6.57% vs. 4.42%, P < 0.001) and myocardial infarction (6.19% vs. 3.54%, P<0.001) compared to non-smokers. In the past, smoking has been identified as a predictor of the need for urgent endoscopic intervention [20]. The greater need for ERCP and early ERCP, as well as the higher odds of complications, probably indicate more severe disease in smokers.

This study, based on the NIS database, has several strengths and limitations. One significant strength is the large sample size, which enhances the generalizability of the findings to a broader patient population. Furthermore, the NIS database provides a wealth of information on patient demographics, diagnoses and procedures, enabling researchers to analyze a wide range of variables.

However, there are some limitations to consider. The retrospective nature of the study may have introduced selection bias and limited the ability to establish causal relationships. The NIS database relies on administrative codes for diagnoses and procedures, which can lead to potential inaccuracies and misclassification [21]. As the NIS is based on discharge records, there may be a lack of granularity in the data, such as detailed clinical information, laboratory results or longitudinal follow up [22]. Lastly, unmeasured confounders may impact the validity of the study’s conclusions [23]. Despite these limitations, the study offers valuable insights into the subject matter, which can be further explored and validated through additional research.

In conclusion, our study demonstrated that smoking did not result in greater mortality, LOS or MIC among patients hospitalized with acute cholangitis. However, it was associated with a higher risk of complications, such as sepsis, biliary pancreatitis and myocardial infarction, as well as the need for inpatient ERCP and early ERCP. These findings emphasize the importance of further research to better understand the relationship between smoking and inpatient outcomes of acute cholangitis.