Admissions for acute biliary pancreatitis without necrosis and infection complicated by severe sepsis and septic shock: a national study

Introduction

Acute pancreatitis (AP) is a major healthcare burden in the USA, accounting for over 200,000 admissions and hospital visits each year [1] with a rising trend [2,3]. Obstruction due to gallstone (acute biliary pancreatitis, ABP) is the most common cause of AP, representing 38% of all cases, while alcohol is a close second with 36% [4-6]. Other less common pathologies can involve medications, autoimmune diseases, recent endoscopic retrograde cholangiopancreatography/abdominal surgery, trauma, and infections, among others [7].

Although most patients with ABP experience an event-free recovery, there is a risk of local and systemic complications [8]. One such complication involves severe sepsis with septic shock (SSWSS), which has high mortality and risk of morbidity. In fact, in rare cases, even ABP patients with no necrosis or infection on admission can develop SSWSS. Acute pancreatitis can increase the expression of various proinflammatory factors, such as tumor necrosis factor and interleukin-6, that can cause damage to the intestinal mucosal membranes. The impaired intestinal barrier allows bacterial spread, which plays an important role in the development of pancreatic infection, necrosis, and eventual sepsis [9-12]. Moreover, the leakage of pancreatic enzymes into the pancreatic and neighboring tissues can trigger a direct insult that may complicate with obstruction, inflammation and necrosis, and become a nidus of infection [13]. However, at present, there is a lack of recent data concerning the presence of ABP and complications such as SSWSS. Therefore, we designed a designed a retrospective analysis that would help identify the incidence of SSWSS among ABP cases initially admitted with no necrosis or infections, and the factors that led to SSWSS. We also aimed to estimate the mortality risk associated with SSWSS in ABP patients.

Materials and methods

Results

Baseline patient and hospital characteristics

We evaluated a total of 136,140 adults who were hospitalized for biliary acute pancreatitis without necrosis or infection between 2018 and 2021 (Fig. 1). The median age was 57.00 years, with an interquartile range of 41.00-71.00. Most patients were females (60.6% vs. 39.4% males), who were also younger (females: median 55.00 years, IQR 36.00-69.00, vs. males: median 61.00 years, IQR 47.00-72.00). Most patients (62.0%) were racially White, followed by Hispanics (20.0%), and Blacks (10.4%). Medicare covered 36.4% of cases, while private insurances and Medicaid covered 35.3% and 18.0%, respectively.

Figure 1 Selection of patients from the NIS

NIS, National Inpatient Sample; SSWS, severe sepsis with septic shock

Most patients (27.7%) were from the lower (0-25^th) median household income quartiles. We further found that 28.1% of all hospitalizations took place during the weekend. The majority of patients were treated in hospitals classified as large (45.3%), urban teaching centers (70.0%), and in southern regions of the USA (39.4%).

Most commonly observed comorbidities included hypertension (39.7%), dyslipidemia (33.6%), smoking (31.7%), obesity (30.1%) and diabetes (22.5%), with fewer cases exhibiting chronic kidney disease (9.0%), peripheral vascular disease (2.1%), liver cirrhosis (5.3%), alcohol abuse (5.2%), HIV (0.4%), cachexia (0.2%), drug abuse (3.1%), or COPD (6.7%) (Table 1).

Table 1 Baseline characteristics of patients admitted with acute biliary pancreatitis

Univariate regression analyses

Patient demographics and comorbidities in our univariable analyses that reported a P-value less than 0.05 included sex (odds ratio [OR] 0.377, 95% confidence interval [CI] 0.310-0.459, P<0.001), hypertension (OR 0.577, 95%CI 0.467-0.712; P<0.001), dyslipidemia (OR 1.532, 95%CI 1.267-1.852; P<0.001), smoking (OR 0.485, 95%CI 0.381-0.618; P<0.001), diabetes (OR 2.114, 95%CI 1.741-2.567; P<0.001), liver cirrhosis (OR 10.539, 95%CI 8.660-12.825; P<0.001), obesity (OR 0.786, 95%CI 0.633-0.975; P=0.029), HIV status (OR 12.651, 95%CI 8.011-19.979; P<0.001), COPD (OR 1.813, 95%CI 1.349-2.436; P<0.001) (Table 2).

Table 2 Univariate regression models and the odds ratios of severe sepsis with septic shock among adults with acute biliary pancreatitis

Multivariate regression analyses

Following the previously described multivariate regression model, factors associated with higher odds of SSWSS included chronic kidney disease (adjusted odds ratio [aOR] 4.766, 95%CI 3.685-6.163, P<0.001), liver cirrhosis (aOR 7.942, 95%CI 6.477-9.739, P<0.001) and HIV (aOR 15.308, 95%CI 8.976-26.107, P<0.001). Moreover, patients admitted to medium-sized (aOR 1.877, 95%CI 1.383-2.549, P<0.001) or large hospitals (aOR 2.095, 95%CI 1.568-2.799, P<0.001) were more likely to report SSWSS than those in small hospitals. While no racial differences were seen for Blacks (aOR 0.755, 95%CI 0.526-1.085, P=0.128) and Hispanics (aOR 0.941, 95%CI 0.700-1.264, P=0.684) as compared to Whites, those classified as non-White/Black/Hispanic (aOR 2.128, 95%CI 1.600-2.830, P<0.001) had higher odds of SSWSS. Sex-based differences were also noted, with lower odds among females (aOR 0.456, 95%CI 0.372-0.559, P<0.001). Finally, we also saw socioeconomic disparities, with lower odds among those in the 26^th-50^th percentile (aOR 0.578, 95%CI 0.432-0.775, P<0.001) (vs. 0-25^th), while patients in the 51^st-75^th (aOR 0.761, 95%CI 0.580-1.000, P=0.050) and 76^th-100^th (aOR 0.930, 95%CI 0.708-1.222, P=0.601) percentiles were all comparable (Table 3).

Table 3 Adjusted odds ratio (aOR) of severe sepsis with septic shock among adults with acute biliary pancreatitis

Patients with dyslipidemia (aOR 0.808, 95%CI 0.671-1.120, P=0.048) and smoking (aOR 0.374, 95%CI 0.290-0.483, P<0.001) recorded lower odds of SSWSS. No statistically significant differences were seen for other variables, including hypertension (aOR 0.867, 95%CI 0.671-1.120, P=0.274), diabetes (aOR 1.127, 95%CI 0.910-13.96, P=0.272), obesity (aOR 1.134, 95%CI 0.900-1.428, P=0.286), COPD (aOR 1.202, 95%CI 0.878-1.647, P=0.251) and primary payer insurance type (Table 3). SSWSS events were associated with a longer LOS (median 9.00 days, IQR 5.00-19.00 vs. 4.00, IQR 2.00-5.00; P<0.001) and higher hospital charges (median $178,926 IQR $90,054-318,288 vs. median $47,147, IQR $28,806-73,493; P<0.001). An estimated 65.5% of patients with SSWSS died (vs. 0.4% in the non-SSWSS cohort).

Discussion

To our knowledge, our study is the first extended, national analysis to evaluate the baseline characteristics of patients with biliary acute pancreatitis without necrosis or infection, as well as the factors that influence the occurrence of SSWSS. In total, we evaluated 136,140 adult cases and found that a higher proportion were females, a finding that is comparable with the higher odds of gallstones in women as compared to men, as previously reported [18,19]. We also found that majority of the patients were White, which reflects the US census distribution. In addition, the second most involved racial group were Hispanics, with almost twice the percentage of Blacks. In the core NIS file of all hospitalizations between 2018-2021, Hispanics usually cover 12-13% of sample sizes, while Blacks cover 15-16%. The higher percentage of Hispanic patients with acute biliary pancreatitis compared to Blacks correlates with a study by Figueiredo et al, where Hispanics in the USA were linked with higher odds of reporting gall bladder-related disease; this could be connected to genetic as well as environmental factors, including diet [20].

Our study further found that 27.7% of cases were from people with the lowest household income quartile in the United States. Li et al, in their analysis, found that this group also represented the highest proportion of patients admitted for symptomatic cholelithiasis (32.5%) and cholecystitis (29.1%) [21]. It is vital to understand whether the socioeconomic disparities impacted their access to care and follow up, which in turn led to complications, including acute biliary pancreatitis. Our study also showed similar geographic distributions, with more cases being treated in the southern regions of the USA [21] and in urban hospitals [22], as with other gallbladder and biliary diseases.

A significant portion of patients were also obese (30.1%), which is a major risk factor for gallbladder and biliary complications [23]. We also found a higher prevalence of smoking, which can slow down gallbladder emptying, thus predisposing to gallstone formation [24] and eventual biliary pancreatitis. While the NIS does not include precise coding details for medications, 39.7% of patients had a history of hypertension and 33.6% of them had dyslipidemia. The direct impact of cholesterol and the influence of lipid-lowering medications, such as fibrates, and antihypertensive medications, such as thiazides, on the development of gallstones could predispose such patients to biliary pancreatitis [25-27]. A more thorough comparison, using retrospective records from different centers, could help clarify such issues.

In the second part of our analysis, we found that, among the 136,140 cases who were admitted with no necrosis and infection, there were 435 patients (320 cases per 100,000) who developed SSWSS, potentially attributable to a variety of factors. Higher odds were seen with age, as SSWSS patients were older. Such a finding might be related to the waning immune response seen with aging, as well as the potential presence of multiple comorbidities and age-related changes in pathophysiology [28-30]. In addition, although 60.6% of all admitted patients were female, SSWSS was more likely to involve males than females—possibly because the male patients in our study were older than the female patients.

While events of SSWSS were comparable between Whites, Blacks and Hispanics, other races exhibited higher odds of such events. We also found that those in the 26^th-50^th income quartile had lower odds of SSWSS than those in the lower quartile, with no differences between the top two quartiles. To reduce such disparities, studies at various community levels should be encouraged, to help openly address any issues these patients had in seeking medical care, as well as at hospital level, to identify other factors that contributed to such racial and socioeconomic discrepancies.

Our study further identified that patients in large centers, as compared to small centers, were more prone to experience SSWSS, as were patients in Southern as compared to Northeastern regions. While the center size differences might be linked with the more complex cases being treated, underdiagnosis and underreporting of SSWSS in small centers cannot be ruled out. A center-based retrospective analysis of cases should be encouraged, which could help map the distribution of centers and elucidate any link with a hospital’s region.

In addition, we also noted that some factors, such as chronic kidney disease, liver cirrhosis and HIV, predisposed patients to SSWSS. The liver is a key organ involved in a balanced immune response to trauma and exposure to potential infectious agents; it is therefore vital for physicians to properly assess patients’ liver function and, if necessary, quantify them as high-risk for complications [31]. Similarly, patients with impaired kidney function will have a buildup of toxic materials that can impair healing [32], impact appetite [33] and weaken their immune system [34]. Patients with HIV have a blunted immune system, with a potentially lower T-cell count. The NIS data, however, did not allow us to identify and adjust for T-cell levels or any ongoing antiretroviral therapy. Nevertheless, our finding of a high adjusted odds ratio raises serious concerns about these patients’ outcomes. It is important to understand the key reasons that caused such risks, other than their immune system, and promote adequate changes. We noted that patients who were smokers or had dyslipidemia had lower odds of reporting SSWSS. Prior studies have confirmed that smokers have a downregulated immune response [35-37]. Since part of the pathophysiology of sepsis in acute pancreatitis relies on the immune response and damage caused by cytokines [12], smokers may be experiencing a suppressed immune reaction, which could explain the lower odds seen in our study. Patients with dyslipidemia may be on drugs such as statins. The ASEPSIS trial by Rachoin et al found that statins impaired the progression of sepsis [38]. As the NIS does not include data for the current drugs taken by patients, we were unable to take such potential confounders into account. As the pathophysiology of sepsis in pancreatitis involves a complex interplay of multiple factors, it is imperative to advocate for further research. Such future studies, carried out at multiple centers, should also include a range of medications that might impact the variables of our study.

Finally, we confirmed that SSWSS was a major factor that contributed to death among patients admitted with acute biliary pancreatitis, with a higher hospital charge and a longer stay. To help reduce the burden of healthcare, it is important that physicians are able to identify patients who are at risk for SSWSS. Changes in protocols should be implemented, including a more thorough check of vitals and appropriate identification of the early signs of sepsis.

Our evaluation of data from the biggest inpatient database provided very important findings. However, there were several limitations that need to be addressed. First, the database relies on hospital codes for various procedures and conditions. Mistakes in coding and differences in coding for SSWSS or similar comorbidities might have influenced our results. Moreover, our data did not contain information on the severity at time of admission, the time from symptom to admission, the medications being used, and patients’ laboratory values. With such limitations, we were unable to categorize the ABP patients based on severity, or to adopt various scoring systems, such as the Ranson criteria, APACHE II (Acute Physiology and Chronic Health Evaluation II), BISAP (Bedside Index for Severity in Acute Pancreatitis), or Glasgow scores. A retrospective review of hospital cases in future studies can overcome such limitations and help consolidate our results.

In addition to several retrospective studies, it is important to promote future research to improve our understanding of the various pathways that are involved in the pathogenesis of sepsis in ABP patients. Moreover, additional studies targeting the management of ABP with SSWSS, and factors impacting the patients’ short and long-term outcomes, mortality and healthcare burden, should also be considered.

To conclude, our study confirmed that, while SSWSS is a rare occurrence among patients with acute biliary pancreatitis with no necrosis or infection on admission, it is associated with a range of factors, including sex, race, age, hospital demographics and comorbidities. Moreover, SSWSS is a major predictor of mortality among such patients, and there is an urgent need to conduct more studies so that other factors can be identified and disparities remedied.