Hidalgo et al. BMC Gastroenterology (2023) 23:81 BMC Gastroenterology https://doi.org/10.1186/s12876-023-02730-6 RESEARCH ARTICLE Open Access Impact of comorbidities on hospital mortality in patients with acute pancreatitis: a population‑based study of 110,021 patients Nils Jimmy Hidalgo1,2, Elizabeth Pando1,2* , Rodrigo Mata2, Nair Fernandes2, Sara Villasante2, Marta Barros2, Daniel Herms2, Laia Blanco2, Joaquim Balsells2 and Ramon Charco2 Abstract Background The impact of pre-existing comorbidities on acute pancreatitis (AP) mortality is not clearly defined. Our study aims to determine the trend in AP hospital mortality and the role of comorbidities as a predictor of hospital mortality. Methods We analyzed patients aged ≥ 18 years hospitalized with AP diagnosis between 2016 and 2019. The data have been extracted from the Spanish National Hospital Discharge Database of the Spanish Ministry of Health. We performed a univariate and multivariable analysis of the association of age, sex, and comorbidities with hospital mor- tality in patients with AP. The role of the Charlson and Elixhauser comorbidity indices as predictors of mortality was evaluated. Results A total of 110,021 patients diagnosed with AP were hospitalized during the analyzed period. Hospital mortal- ity was 3.8%, with a progressive decrease observed in the years evaluated. In multivariable analysis, age ≥ 65 years (OR: 4.11, p < 0.001), heart disease (OR: 1.73, p < 0.001), renal disease (OR: 1.99, p < 0.001), moderate-severe liver disease (OR: 2.86, p < 0.001), peripheral vascular disease (OR: 1.43, p < 0.001), and cerebrovascular disease (OR: 1.63, p < 0.001) were independent risk factors for mortality. The Charlson > 1.5 (OR: 2.03, p < 0.001) and Elixhauser > 1.5 (OR: 2.71, p < 0.001) comorbidity indices were also independently associated with mortality, and ROC curve analysis showed that they are useful for predicting hospital mortality. Conclusions Advanced age, heart disease, renal disease, moderate-severe liver disease, peripheral vascular disease, and cerebrovascular disease before admission were independently associated with hospital mortality. The Charlson and Elixhauser comorbidity indices are useful for predicting hospital mortality in AP patients. Keywords Acute pancreatitis, Hospital mortality, Comorbidity, Charlson index, Elixhauser index Background This publication has been made as part of the Doctoral Program inSurgery Acute pancreatitis (AP) is a prevalent acute inflam- and Morphologic Sciences of the Universitat Autònoma deBarcelona, Spain. matory disease that affects the pancreas, with an *Correspondence: increased incidence in recent years [1, 2]. Most cases Elizabeth Pando are mild with a self-limited course [3]. However, elizabeth.pando@vallhebron.cat 1 patients with severe acute pancreatitis have a high Universitat Autonoma de Barcelona, Barcelona, Spain 2 Department of Hepato‑Pancreato‑Biliary and Transplant Surgery, mortality rate (20–50%) [4–6]. For this reason, many Hospital Universitari Vall d’Hebron, Passeig de La Vall d’Hebron, 119‑129. efforts have been made to find predictors of severity 08035 Barcelona, Spain and mortality in patients with AP [7–11] to identify © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 2 of 10 patients who need admission to an intensive care unit extrahepatic bile duct neoplasm, and complications of or specific treatment. the transplanted pancreas. We excluded the population In clinical practice, systems such as the Ranson under 18 years of age because the incidence of AP is score, the Acute Physiology and Chronic Health Evalu- lower in the pediatric population [24], and the etiologies ation II (APACHE II) score, the Computed Tomog- distribution differs from that of adults [25]. raphy Severity Index (CTSI), the Bedside Index for Severity in Acute Pancreatitis (BISAP), and various Variables analyzed biochemical markers are used to predict severe AP The variables included are demographic variables such and mortality [3, 12–16]. However, hospital mortality as age and sex, and AP etiology. We used the ICD-10 in AP could also be related to intrinsic patient char- diagnostic code from each patient to identify the etiol- acteristics, such as individual comorbidities. Most ogy of AP, which includes six categories: biliary, alco- classic scores do not consider comorbidities before hol, idiopathic, drug-related, other, and unspecified. admission, except for APACHE II, but are restricted to Our clinical-administrative database does not have severe chronic diseases. data on diagnostic tests such as ultrasound or magnetic According to some previous studies, patients with cer- resonance imaging that allow the identification of the tain comorbidities, such as obesity [17], hypertriglyceri- etiology in patients with a diagnosis of "unspecified pan- demia [18], chronic renal failure [19], diabetes [20, 21], creatitis. In addition, we did not have data after hospital and systemic lupus erythematosus [22], are associated discharge that could expand the information on the eti- with a higher risk of AP severity and mortality. However, ology. Other variables assessed were clinical variables on few studies currently evaluate the impact of comorbidi- diagnoses and procedures. ties on AP severity and mortality. Our study aimed to determine the relevance of comor- Comorbidity assessment bidities and their indexes (Charlson and Elixhauser) as Comorbidities were identified from the ICD-10 diagno- predictors of hospital mortality in patients with AP. sis codes of each patient. We have used the POA indi- cator (Present on registration) to identify comorbidities Methods and differentiate them from diagnoses produced during Study design hospital admission that could be secondary complica- We carried out a retrospective observational study using tions of AP. The ICD-10 codes used to identify specific the Spanish National Hospital Discharge (Registro de comorbidities are described in the supplementary mate- Actividad de Atención Especializada-Conjunto Mínimo rial (Additional file 1). Básico de Datos, “RAE-CMBD”). The RAE-CMBD col- lects all the administrative data from hospitals (public Calculation of comorbidities indexes and private) in the country [23]. We assessed comorbidity by calculating the Charlson The information collected in this database comes [26] and Elixhauser [27] comorbidity indices. These two from hospital discharge reports made by the physicians indices are used in medical practice to predict mortal- in charge of the patient. This information and record of ity. ICD-10 diagnosis codes described by Quan et al. [28] coded diagnoses are automatically collected by the com- were applied to identify specific comorbidities from the puter software of each center or by technical-adminis- Charlson and Elixhauser indices. trative staff. The Charlson index assigns weights for 17 specific dis- eases, and its value was calculated by adding the weights Study population of each condition as described by Charlson et al. [26]. The study population includes patients diagnosed with The Elixhauser index assigns weights for 30 specific dis- AP and admitted to the Spanish National Health System eases, and its value was calculated using the algorithm hospitals from 2016 to 2019. Since 2016, the RAE-CMDB described by Walraven et al. [29]. has collected 20 diagnoses and 20 procedures from each patient based on the International Classification of Dis- Outcomes eases Version 10 (ICD-10). Outcomes analyzed included pancreatic necrosis The inclusion criteria were: Patients with a primary or length of hospital stay, admission to intensive care unit first registered secondary diagnosis of AP. (ICU), length of ICU stay, and hospital mortality. Since The exclusion criteria were: 1) Patients under 18 years 2018, the definition of pancreatic necrosis and pancre- of age, 2) Patients with a diagnosis primary or first reg- atic necrosis infection has been included in the ICD-10 istered secondary of pancreatic neoplasm, chronic diagnosis code for AP: AP without necrosis or infec- pancreatitis, pancreatic cyst, pancreatic pseudocyst, tion (K85. × 0), AP with uninfected necrosis (K85. × 1) Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 3 of 10 and AP with infected necrosis (K85. × 2). In addition, during the period are shown in Table 1. The mean age we did not have information on the percentage exten- was 64.32 ± 17.94 years, with a slight progressive decrease sion of pancreatic necrosis. Therefore, the data on throughout the years. The 53.3% of patients were 65 years pancreatic necrosis were only used for the descriptive or older. Male sex prevalence was 53.1%, which signifi- analysis of the evolution of AP in the period analyzed. cantly increased during the study period (p = 0.043). The most frequent etiologies of AP were biliary (41.2%) and Statistical analysis alcohol (7.9%). The Charlson and Elixhauser comorbidity We used the Kruskal–Wallis test for continuous vari- indices values progressively increased in the last years of ables and the Linear-by-Linear association test for cat- the study period (Table 1). egorical variables to analyze the characteristics and results of patients with AP during the years evaluated (2016–2019). General outcomes The analysis of risk factors for hospital mortality was The proportion of patients who required ICU admis- performed by applying the chi-square test for categorical sion was 4.7%, with no differences in its prevalence variables and the Student’s t-test or the Mann–Whitney by year studied. The mean length of hospital stay was U test for continuous variables. 9.38 ± 12.22 days, showing a significant decrease over Univariate and multivariable analysis of the factors the period. Pancreatic necrosis was reported in 8.5% in associated with hospital mortality was performed using 2018 and 9.2% in 2019 (Table 1). logistic regression. We performed three multivariable analyses of significant variables in the univariate analysis. Mortality The first analysis included: age ≥ 65 years, sex, and spe- Mortality was 3.8% in all the population and significantly cific comorbidities. In the second analysis, the Charlson decreased over time, from 4.1% in 2016 to 3.5% in 2019 comorbidity index replaced the specific comorbidities. (p < 0.001) (Table 1). The third analysis replaced specific comorbidities with the Elixhauser comorbidity index. Impact of Age, sex, and etiology Receiver-operating characteristic (ROC) curves were Age and male sex were higher in the non-survivors drawn to analyze the in-hospital mortality prediction compared to the group of survivors (78.02 ± 13.24 vs. capacity of the Charlson and Elixhauser comorbid- 63.78 ± 17.89, p =  < 0.001 and 51.1% vs. 53.2%, p = 0.007, ity indices, and the area under the curve (AUC) was respectively). The prevalence of pancreatitis of biliary, described. A Delong test [30] was performed to com- alcoholic, and drug-related etiology was lower in the pare the AUC. We used the Youden’s index to identify group of non-survivors (p < 0.001) (Table 2). the best cut-off point for the Charlson and Elixhauser comorbidity indices. Impact of comorbidities and indexes Statistical analyses were performed using IBM SPSS Non-survivor patients presented a higher percentage in 20.0 (IBM Corp. in Armonk, NY) and Stata version 16 all comorbidities except for obesity (Table 2). (Stata, College Station, Texas, USA). Statistical signifi- Median values of Charlson and Elixhauser indexes cance was set at p < 0.05. were significantly higher in the group of non-survivors compared with survivors (1.76 ± 2.1 vs. 0.92 ± 1.43, Ethical aspects p < 0.001 and 7.15 ± 7.4 vs. 3.29 ± 5.8, p < 0.001 respec- Our study follows the principles of the Declaration of tively) (Table 2). The analysis of the comorbidities Helsinki for research on human beings. The data was included in the Charlson and Elixhauser indices accord- extracted from the Spanish Ministry of Health register, ing to hospital mortality is described in the supplemen- which is anonymous following Spanish legislation. Iden- tary material (Additional file 1). tifying patients at the individual or reporting unit level with the data obtained is impossible. Multivariable analysis Results Logistic regression analysis was performed using the General population characteristics best cut-off point obtained by Youden’s index (J = 1.5 Between January 1, 2016, and December 31, 2019, a total points for both Charlson and Elixhauser indices). After of 125,622 cases with the diagnosis of AP were identified. multivariable logistic regression analysis, we found that After applying inclusion and exclusion criteria, 110,021 factors independently associated with mortality were patients were included (Fig. 1). The demographic and age 65 years or older (OR 4.11, 95% CI 3.75–4.5), heart clinical characteristics of the population and its variations disease (OR 1.73, 95% CI 1.62–1.86), renal disease (OR Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 4 of 10 Fig. 1 Case Selection Flow Chart. AP: acute pancreatitis IDC-10: 10th revision of the International Statistical Classification of Diseases 1.99, 95% CI 1.74–2.07), moderate-severe liver disease disease, and age ≥ 65 years were independently associated (OR 2.86, 95% CI 2.35–3.47), peripheral vascular dis- with mortality in AP. Charlson and Elixhauser comorbid- ease (OR 1.43, 95% CI 1.25–1.64), and cerebrovascular ity indices were independently associated with mortality. disease (OR 1.63, 95% CI 1.4–1.9). Arterial hyperten- Advanced age has been extensively studied as a marker sion has been found to be a protective factor in the of severity and mortality in AP. Most studies report population. The Charlson Index > 1.5 points (OR 2.03, longer hospitalization [31, 32] and higher overall mortal- 95% CI 1.9–2.16) and Elixhauser Index > 1.5 points (OR ity from AP in elderly patients [33–36]. However, other 2.71, 95% CI 2.53–2.9) were independently associated studies have observed that older patients may have a with mortality (Table 3). more severe course of AP but do not present increased mortality [37]. Likely explanations explaining advanced AUC analysis age as a risk factor for mortality include a proinflamma- The Elixhauser comorbidity index exhibited a slightly tory state in older people [38] and increased production higher AUC value in predicting hospital mortality (AUC: of cytokines in elderly patients with sepsis [39]. Other 0.666, 95% CI 0.657 – 0.674) than the Charlson comor- reasons would be delayed diagnosis and treatment due to bidity index (AUC: 0.633, 95% CI 0.623 – 0.641). When less clinical and analytical expression [40, 41]. performing the Delong test to compare these AUC, it was The increase in life expectancy and the aging of the observed that this difference is significant (p < 0.001). The population have been associated with the increase in ROC curves and AUC for the Charlson and Elixhauser patients with comorbidities [42, 43], so determining its comorbidity indices to predict hospital mortality are impact on AP becomes more necessary. The importance shown in Fig. 2. of comorbidities in predicting outcomes in other dis- eases that require acute hospital admission is well known Discussion [44–46]. However, few studies analyze the impact of Our study found that pre-admission comorbidities such comorbidities on severity and mortality in AP patients as heart disease, kidney disease, moderate-severe liver [47, 48]. Additionally, few studies have incorporated disease, peripheral vascular disease, cerebrovascular comorbidities in their clinical models when evaluating Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 5 of 10 Table 1 Characteristics of patients hospitalized with a diagnosis of acute pancreatitis Total (N: 110,021) 2016 (N: 26,952) 2017 (N: 22,170) 2018 (N: 29,785) 2019 (N: 31,114) p value Age, years Mean ± SD 64.32 ± 17.94 64.54 ± 17.99 64.34 ± 17.82 64.33 ± 17.97 64.11 ± 17.96 0.017 Age ≥ 65 years, N (%) 58,612 (53.3) 14,679 (54.5) 11,806 (53.3) 15,832 (53.2) 16,295 (52.4) < 0.001 Sex, N (%) Male 58,457 (53.1) 14,178 (52.6) 11,765 (53.1) 15,899 (53.4) 16,615 (53.4) 0.043 Female 51,564 (46.9) 12,774 (47.4) 10,405 (46.9) 13,886 (46.6) 14,499 (46.6) 0.043 Charlson Index Mean ± SD 0.95 ± 1.47 0.88 ± 1.41 0.91 ± 1.43 0.97 ± 1.5 1.02 ± 1.53 < 0.001 Elixhauser Index Mean ± SD 3.43 ± 5.87 3.24 ± 5.71 3.2 ± 5.66 3.49 ± 5.93 3.72 ± 6.08 < 0.001 Pancreatitis etiology, N (%) Biliary 45,281 (41.2) 9,909 (36.8) 8,263 (37.3) 13,354 (44.9) 13,755 (44.2) < 0.001 Alcohol 8,658 (7.9) 1,708 (6.3) 1,477 (6.7) 2,625 (8.8) 2,848 (9.2) < 0.001 Medications 1,086 (1) 197 (0.7) 177 (0.8) 332 (1.1) 380 (1.2) < 0.001 Idiopathic 2,456 (2.2) 380 (1.4) 366 (1.7) 821 (2.8) 889 (2.9) < 0.001 Other 4,029 (3.7) 947 (3.5) 909 (4.1) 1,038 (3.5) 1,135 (3.6) 0.761 Not specified 48,511 (44.1) 13,811 (51.2) 10,978 (49.5) 11,615 (39) 12,107 (38.9) < 0.001 Pancreatic necrosis, N (%) - NR NR 2,523 (8.5) 2,873 (9.2) Infected necrosis, N (%) - NR NR 918 (3.1) 1,069 (3.4) ICU admission, N (%) 5,155 (4.7) 1,187 (4.4) 1,093 (4.9) 1,445 (4.9) 1,430 (4.6) 0.383 ICU stay (days) Mean ± SD 13.05 ± 22.27 13.15 ± 22.05 12.65 ± 22.06 13.35 ± 21.57 13.97 ± 23.31 0.181 Hospital stay (days) Mean ± SD 9.38 ± 12.22 9.73 ± 12.11 9.43 ± 12.02 9.35 ± 12.35 9.08 ± 12.29 < 0.001 Mortality, N (%) 4,153 (3.8) 1,095 (4.1) 884 (4) 1,097 (3.7) 1,077 (3.5) < 0.001 SD Standard deviation, NR Not reported, ICU Intensive care unit determinants of AP severity. Our study is one of the first are independently associated with hospital mortality in studies in the literature that put in relevance the role of AP after adjusting for age and sex. Future studies that comorbidities in AP. expand knowledge of the effects of comorbidities on In the last period of our study, we observed an increase complications and mortality in patients with AP will in comorbidities and the values of the Charlson and improve the identification of patients at risk and their Elixhauser comorbidity indices. These trends could be quality of care. explained by the increase in life expectancy and the prev- Regarding other comorbidities, our results align with alence of chronic diseases in the European population in previous studies stating that pre-existing heart and renal the last decades [49, 50]. However, despite the increase diseases predict mortality in patients with AP [19, 47, 55]. in comorbidities, hospital mortality has decreased in the One hypothesis is that intravascular depletion and aggres- period studied. The decrease in mortality is likely due to sive fluid resuscitation cause decompensation of previous the advances in critical care medicine, step-up approach heart and renal disease [55]. to treat infected necrosis, and surgical and endoscopic Similarly, to Frey et al. [47], we found an association new approaches [3, 51]. between liver disease, peripheral vascular disease, and Few studies had previously assessed comorbidity cerebrovascular disease with mortality. However, Murata indexes’ role in predicting mortality in patients with et al. found no association between these diseases with AP. Previous studies have observed that more comor- mortality [55]. The worse results of AP in patients with bidities are associated with organ failure and mortality liver diseases such as cirrhosis could be explained in patients with AP [47, 52]. In our study, we analyzed because they present an inflammatory syndrome with the Charlson and Elixhauser indices which are good arterial vasodilation and release of proinflammatory predictors of mortality in other diseases [53, 54]. Our cytokines that increase the severity of AP [56]. In addi- results showed that values > 1.5 points for both indices tion, acute pancreatitis produces significant stress that Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 6 of 10 Table 2 Demographic characteristics of acute pancreatitis according to survivor and non-survivors Total Survivors Non-survivors p value Age, mean ± SD 64.32 ± 17.94 63.78 ± 17.89 78.02 ± 13.24 < 0.001 Sex, N (%) Male 58,457 (53.1) 56,335 (53.2) 2,122 (51.1) 0.007 Female 51,564 (46.9) 49,533 (46.8) 2,031 (48.9) 0.007 Comorbidities, N (%) Arterial hypertension 51,532 (46.8) 49,011 (46.3) 2,521 (60.7) < 0.001 Heart disease 18,696 (17) 17,202 (16.2) 1,494 (35.9) < 0.001 Chronic pulmonary disease 9,234 (8.4) 8,739 (8.3) 495 (11.9) < 0.001 Renal disease 8,193 (7.4) 7,399 (7) 794 (19.1) < 0.001 Moderate or severe liver disease 1,498 (1.4) 1,375 (1.3) 123 (3) < 0.001 Diabetes mellitus 20,259 (18.4) 19,315 (18.2) 944 (22.7) < 0.001 Obesity 9,681 (8.8) 9,327 (8.8) 354 (8.5) 0.523 Peripheral vascular disease 3,361 (3.1) 3,091 (2.9) 270 (6.5) < 0.001 Cerebrovascular disease 2,220 (2) 2,023 (1.9) 197 (4.7) < 0.001 Rheumatic disease 1,335 (1.2) 1,261 (1.2) 74 (1.8) 0.001 Charlson Index, mean ± SD 0.95 ± 1.47 0.92 ± 1.43 1.76 ± 2.1 < 0.001 Elixhauser Index, mean ± SD 3.43 ± 5.87 3.29 ± 5.75 7.15 ± 7.38 < 0.001 Pancreatitis etiology, N (%) Biliary 45,281 (41.2) 44,041 (41.6) 1,240 (29.9) < 0.001 Alcohol 8,658 (7.9) 8,533 (8.1) 125 (3) < 0.001 Medications 1,086 (1) 1,070 (1) 16 (0.4) < 0.001 Idiopathic 2,456 (2.2) 2,333 (2.2) 123 (3) 0.001 Other 4,029 (3.7) 3,768 (3.6) 261 (6.3) < 0.001 Not specified 48,511 (44.1) 46,123 (43.6) 2,388 (57.5) < 0.001 ICU admission, N (%) 5,155 (4.7) 3,684 (3.5) 1,471 (35.4) < 0.001 SD Standard deviation, ICU Intensive care unit could decompensate underlying chronic comorbidities Other risk factors classically related to mortality in AP and increase the risk of death. patients failed to represent a risk factor in our population. Other major comorbidities before admission, such as This was relevant to the role of AP etiology, in which the chronic lung disease, were not independently associated literature reports controversial results, identifying a more with mortality, coinciding with the results of Murata et al. severe course and higher mortality in alcoholic pancrea- [55]. Similar controversial results were found regarding titis [58, 59]. In contrast, others observed greater severity obesity, in which previous studies have observed that obe- in biliary pancreatitis [60, 61] or no relation with mortal- sity is a risk factor for developing local and systemic com- ity between both aetiologies [37, 62–64]. Our study did plications and mortality in patients with AP [17, 57–59]. not observe that acute pancreatitis’s biliary or alcoholic We did not find an association between hospital mortality etiology was associated with higher hospital mortality. and obesity, but this result has to be taken carefully due However, in our study, 44.3% of the patients were classi- to the potential information bias because our results are fied as "unspecified acute pancreatitis,” limiting the preci- based on the history of obesity recorded in the medical- sion of our results and constituting a bias regarding the administrative database and not on the BMI at hospital real impact of etiology in AP mortality. In addition, our admission. In the same line, we did not find a relation database does not include other etiologies of acute pan- between diabetes mellitus and mortality after the mul- creatitis such as those caused by hypercalcemia, after tivariable analysis. The literature remains controversial, trauma, viral infections, anatomical variants, iatrogenic with reports describing diabetes mellitus as a risk factor after endoscopic retrograde cholangiopancreatography for mortality [20, 21]. However, Frey et al. found that dia- or endoscopic ultrasound-guided interventions [65–67]. betes increased the risk of multiple organ failure but was Our study is subject to some limitations. The data anal- not associated with mortality [47]. ysis from a clinical-administrative base has low level of Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 7 of 10 Table 3 Multivariable analysis showing association of proposed granularity and does not include some clinical results of risk factors with hospital mortality in acute pancreatitis interest, such as severity or the evolution of the patient Multivariable analysis in the medium or long term after their hospital stay. Our study could not identify pancreatic necrosis in the first two OR (95% CI) p value years because it began to be considered in the ICD-10 in Variables of analysis 1 2018. Another limitation is the potential underreporting of Age ≥ 65 years 4.11 (3.75–4.5) < 0.001 information because the hospital discharge report may be Sex Female 0.96 (0.9–1.03) 0.236 incomplete or poorly registered by the technical-adminis- Arterial hypertension 0.89 (0.84–0.96) 0.003 trative staff. In our study, we could not identify the etiology Heart disease 1.73 (1.62–1.86) < 0.001 of AP in many patients due to a lack of precise coding. Chronic pulmonary disease 1.09 (0.99–1.21) 0.085 No studies have been published to validate the use of Renal disease 1.99 (1.74–2.07) < 0.001 ICD-10 codes for identifying patients with AP using the Moderate or severe liver disease 2.86 (2.35–3.47) < 0.001 RAE-CMBD database. However, there are recent stud- Diabetes mellitus 0.94 (0.87–1.01) 0.09 ies from Danish [68] and US [69] databases with PPV of Peripheral vascular disease 1.43 (1.25–1.64) < 0.001 97.3% and 87%, respectively. A recent meta-analysis rec- Cerebrovascular disease 1.63 (1.4–1.9) < 0.001 ommends using ICD codes only in incident cases of AP Rheumatic disease 1.13 (0.89–1.44) 0.317 in adults, where it reaches a PPV of 78% [70]. However, Variables of analysis 2 this may be because the studies analyzed used ICD-8, Age ≥ 65 years 4.41 (4.04–4.8) < 0.001 ICD-9, and ICD-10 codes, and the PPV is higher when Sex Female 0.96 (0.9–1.03) 0.962 using the ICD-10 because it is more specific and includes Charlson Index ≥ 1.5 2.03 (1.9–2.16) < 0.001 the etiology of pancreatitis [70, 71]. In addition, the stud- Variables of analysis 3 ies were carried out in different hospitals in several coun- Age ≥ 65 years 4.23 (3.88–4.61) < 0.001 tries, contributing to the heterogeneity. Sex Female 0.98 (0.92–1.05) 0.581 We used the primary and the first registered sec- Elixhauser Index ≥ 1.5 2.71 (2.53–2.9) < 0.001 ondary diagnoses to reduce the bias of not includ- ing patients with an initial diagnosis of cholelithiasis/ OR Odds ratio, CI Confidence interval Fig. 2 ROC curve and AUC (Area Under the Curve) of the Charlson Comorbidity Index, and the Elixhauser Comorbidity Index in predicting the hospital mortality rate in patients with acute pancreatitis Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 8 of 10 choledocholithiasis and AP. A recent study validating Funding Not applicable. ICD codes using primary and secondary diagnoses observed a PPV of 97.3% for AP [68]. Availability of data and materials The main strength of our study is its large sample The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. size which provides strong statistical power. The RAE- CMBD database is a mandatory registry for the Spanish National Health System, which covers almost 100% of Declarations admissions in Spain, reinforcing the external validity of Ethics approval and consent to participate our results. In addition, the database has several inter- Our study follows the principles of the Declaration of Helsinki for research on human beings. The data was extracted from the Spanish Ministry of Health nal audit mechanisms and has proven its usefulness for record, which is anonymous. Identifying the patients at the individual or health research [53, 72, 73]. reporting unit level is impossible as this database is subject to the Civil Service Statistics Law 12/1989. Obtaining informed consent or approval from an ethics committee was not necessary. Conclusions Consent for publication Comorbidities such as heart disease, kidney disease, Not applicable; this manuscript does not contain individual personal data. moderate-severe liver disease, peripheral vascular dis- Competing interests ease, cerebrovascular disease, and advanced age were The authors declare that they have no competing interests. independently associated with mortality in AP. The Charlson and Elixhauser comorbidity indices are useful Received: 20 September 2022 Accepted: 16 March 2023 for predicting hospital mortality in these patients. Abbreviations AP Acute pancreatitis References APACHE II Acute physiology and chronic health evaluation II 1. Peery AF, Crockett SD, Murphy CC, Lund JL, Dellon ES, Williams JL, et al. CTSI Computed tomography severity index Burden and cost of gastrointestinal, liver, and pancreatic diseases in BISAP Bedside index for severity in acute pancreatitis the United States: Update 2018. Gastroenterology. 2019;156:254-272. RAE-CMBD Registro de Actividad de Atención Especializada-Conjunto Mínimo e11. Básico de Datos 2. Forsmark CE, Vege SS, Wilcox CM. Acute Pancreatitis. N Engl J Med. ICD-10 International classification of diseases version 10 2016;375:1972–81. POA Present on registration 3. Banks PA, Bollen TL, Dervenis C, Gooszen HG, Johnson CD, Sarr MG, et al. ICU Intensive care unit Classification of acute pancreatitis–2012: revision of the Atlanta classifica- ROC Receiver-operating characteristic tion and definitions by international consensus. Gut. 2013;62:102–11. AUC​ Area under the curve 4. Yadav D, Lowenfels AB. Trends in the epidemiology of the first attack of OR Odds ratio acute pancreatitis: a systematic review. Pancreas. 2006;33:323–30. CI Confidence interval 5. Shen H-N, Lu C-L, Li C-Y. Epidemiology of first-attack acute pancreatitis in Taiwan from 2000 through 2009: a nationwide population-based study. Pancreas. 2012;41:696–702. Supplementary Information 6. van Dijk SM, Hallensleben NDL, van Santvoort HC, Fockens P, van Goor H, The online version contains supplementary material available at https://​doi.​ Bruno MJ, et al. Acute pancreatitis: recent advances through randomised org/​10.​1186/​s12876-​023-​02730-6. trials. Gut. 2017;66:2024–32. 7. Petrov MS, Shanbhag S, Chakraborty M, Phillips ARJ, Windsor JA. Organ failure and infection of pancreatic necrosis as determinants of mortality Additional file 1: Table S1. The ICD-10 codes used to identify spe- in patients with acute pancreatitis. Gastroenterology. 2010;139:813–20. cific comorbidities. Table S2. Association of comorbidities of Charlson 8. Papachristou GI, Muddana V, Yadav D, O’Connell M, Sanders MK, Slivka A, Comorbidity Index and hospital mortality in acute pancreatitis. Table S3. et al. Comparison of BISAP, Ranson’s, APACHE-II, and CTSI scores in pre- Association of comorbidities of Elixhauser Comorbidity Index and hospital dicting organ failure, complications, and mortality in acute pancreatitis. mortality in acute pancreatitis. Am J Gastroenterol. 2010;105:435–41 (quiz 442). 9. Gravante G, Garcea G, Ong SL, Metcalfe MS, Berry DP, Lloyd DM, et al. Acknowledgements Prediction of mortality in acute pancreatitis: a systematic review of the This publication has been made as part of the Doctoral Program in Surgery and published evidence. Pancreatology. 2009;9:601–14. Morphologic Sciences of the Universitat Autònoma de Barcelona, Spain. Depart- 10. Pando E, Alberti P, Hidalgo J, Vidal L, Dopazo C, Caralt M, et al. The role of ments of Surgery and Morphological Science at the Universitat Autònoma. extra-pancreatic infections in the prediction of severity and local compli- cations in acute pancreatitis. Pancreatology. 2018;18:486–93. Authors’ contributions 11. Alberti P, Pando E, Mata R, Vidal L, Roson N, Mast R, et al. Evaluation of the NJH: project development, data collection and analysis, manuscript writing modified computed tomography severity index (MCTSI) and computed and editing. EP: project development, data collection and analysis, manuscript tomography severity index (CTSI) in predicting severity and clinical writing and editing. RM: data collection and analysis, manuscript writing and outcomes in acute pancreatitis. J Dig Dis. 2021;22:41–8. editing. NF: data collection, manuscript editing. SV: data collection, manuscript 12. Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH. Acute pancreatitis: editing. MB: data collection, manuscript editing. DH: data collection, manu- value of CT in establishing prognosis. Radiology. 1990;174:331–6. script editing. LB: data collection, manuscript editing. JB: project development, 13. Larvin M, McMahon MJ. APACHE-II score for assessment and monitoring manuscript editing. RC: project development, manuscript writing and editing. of acute pancreatitis. Lancet. 1989;2:201–5. All authors read and approved the final manuscript. Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 9 of 10 14. Ranson JH, Rifkind KM, Roses DF, Fink SD, Eng K, Localio SA. Objective 36. Malik AM. Biliary pancreatitis. Deadly threat to the elderly. Is it a real early identification of severe acute pancreatitis. Am J Gastroenterol. threat? Int J Health Sci. 2015;9:35–9. 1974;61:443–51. 37. Gullo L, Migliori M, Oláh A, Farkas G, Levy P, Arvanitakis C, et al. Acute 15. Wu BU, Johannes RS, Sun X, Tabak Y, Conwell DL, Banks PA. The early pancreatitis in five European countries: etiology and mortality. Pancreas. prediction of mortality in acute pancreatitis: a large population-based 2002;24:223–7. study. Gut. 2008;57:1698–703. 38. Starr ME, Ueda J, Yamamoto S, Evers BM, Saito H. The effects of aging on 16. Pando E, Alberti P, Mata R, Gomez MJ, Vidal L, Cirera A, et al. Early pulmonary oxidative damage, protein nitration, and extracellular super- changes in Blood Urea Nitrogen (BUN) can predict mortality in acute oxide dismutase down-regulation during systemic inflammation. Free pancreatitis: comparative study between BISAP score, APACHE-II, and Radic Biol Med. 2011;50:371–80. other laboratory markers-a prospective observational study. Can J 39. Turnbull IR, Clark AT, Stromberg PE, Dixon DJ, Woolsey CA, Davis CG, et al. Gastroenterol Hepatol. 2021;2021:6643595. Effects of aging on the immunopathologic response to sepsis. Crit Care 17. Martínez J, Johnson CD, Sánchez-Payá J, de Madaria E, Robles-Díaz G, Med. 2009;37:1018–23. Pérez-Mateo M. Obesity is a definitive risk factor of severity and mortal- 40. Lyon C, Clark DC. Diagnosis of acute abdominal pain in older patients. ity in acute pancreatitis: an updated meta-analysis. Pancreatology. Am Fam Physician. 2006;74:1537–44. 2006;6:206–9. 41. Potts FE, Vukov LF. Utility of fever and leukocytosis in acute surgical 18. Lloret Linares C, Pelletier AL, Czernichow S, Vergnaud AC, Bonnefont- abdomens in octogenarians and beyond. J Gerontol A Biol Sci Med Sci. Rousselot D, Levy P, et al. Acute pancreatitis in a cohort of 129 patients 1999;54:M55–8. referred for severe hypertriglyceridemia. Pancreas. 2008;37:13–22. 42. Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complica- 19. Lankisch PG, Weber-Dany B, Maisonneuve P, Lowenfels AB. Frequency tions of multiple chronic conditions in the elderly. Arch Intern Med. and severity of acute pancreatitis in chronic dialysis patients. Nephrol 2002;162:2269–76. Dial Transplant. 2008;23:1401–5. 43. Starfield B, Lemke KW, Bernhardt T, Foldes SS, Forrest CB, Weiner JP. 20. Huh JH, Jeon H, Park SM, Choi E, Lee GS, Kim JW, et al. Diabetes mellitus Comorbidity: implications for the importance of primary care in “case” is associated with mortality in acute pancreatitis. J Clin Gastroenterol. management. Ann Fam Med. 2003;1:8–14. 2018;52:178–83. 44. Mnatzaganian G, Ryan P, Norman PE, Hiller JE. Accuracy of hospital 21. Mikó A, Farkas N, Garami A, Szabó I, Vincze Á, Veres G, et al. Preexist- morbidity data and the performance of comorbidity scores as predictors ing diabetes elevates risk of local and systemic complications in of mortality. J Clin Epidemiol. 2012;65:107–15. acute pancreatitis: systematic review and meta-analysis. Pancreas. 45. Ho T-W, Tsai Y-J, Ruan S-Y, Huang C-T, Lai F, Yu C-J, et al. In-hospital and 2018;47:917–23. one-year mortality and their predictors in patients hospitalized for first- 22. Pascual-Ramos V, Duarte-Rojo A, Villa AR, Hernández-Cruz B, Alarcón- ever chronic obstructive pulmonary disease exacerbations: a nationwide Segovia D, Alcocer-Varela J, et al. Systemic lupus erythematosus as population-based study. PLoS ONE. 2014;9:e114866. a cause and prognostic factor of acute pancreatitis. J Rheumatol. 46. Pocock SJ, Wang D, Pfeffer MA, Yusuf S, McMurray JJV, Swedberg KB, 2004;31:707–12. et al. Predictors of mortality and morbidity in patients with chronic heart 23. Ministerio de Sanidad C y BS. Registro de Actividad de Atención failure. Eur Heart J. 2006;27:65–75. Especializada del Conjunto Mínimo Básico de Datos (RAE-CMBD). 2021. 47. Frey C, Zhou H, Harvey D, White RH. Co-morbidity is a strong predictor of https://​www.​mscbs.​gob.​es/​estad​Estud​ios/​estad​istic​as/​estad​istic​as/​ early death and multi-organ system failure among patients with acute estMi​niste​rio/​Solic​itudC​MBD.​htm. pancreatitis. J Gastrointest Surg. 2007;11:733–42. 24. Majbar AA, Cusick E, Johnson P, Lynn RM, Hunt LP, Shield JPH. 48. Akshintala VS, Hutfless SM, Yadav D, Khashab MA, Lennon AM, Makary Incidence and Clinical Associations of Childhood Acute Pancreatitis. MA, et al. A population-based study of severity in patients with acute on Pediatrics. 2016;138:1198. chronic pancreatitis. Pancreas. 2013;42:1245–50. 25. Sakorafas GH, Tsiotou AG. Etiology and pathogenesis of acute pancrea- 49. Roser M, Ortiz-Ospina E, Ritchie H. Life Expectancy. OurWorldInData.org. titis: current concepts. J Clin Gastroenterol. 2000;30:343–56. Retrieved from: “https://​ourwo​rldin​data.​org/​life-​expec​tancy.” 2013. 26. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of clas- 50. Eurostat. Mortality and life expectancy statistics. https://​ec.​europa.​eu/​ sifying prognostic comorbidity in longitudinal studies: development euros​tat/​stati​stics-​expla​ined/​index.​php?​title=​Morta​lity_​and_​life_​expec​ and validation. J Chronic Dis. 1987;40:373–83. tancy_​stati​stics. 2022. 27. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for 51. Pitchumoni CS, Patel NM, Shah P. Factors influencing mortality in acute use with administrative data. Med Care. 1998;36:8–27. pancreatitis: can we alter them? J Clin Gastroenterol. 2005;39:798–814. 28. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi J-C, et al. 52. Weitz G, Woitalla J, Wellhöner P, Schmidt KJ, Büning J, Fellermann K. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 Comorbidity in acute pancreatitis relates to organ failure but not to local administrative data. Med Care. 2005;43:1130–9. complications. Z Gastroenterol. 2016;54:226–30. 29. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modifica- 53. Pedraza-Serrano F, Jiménez-García R, López-de-Andrés A, Hernández-Bar- tion of the Elixhauser comorbidity measures into a point system for rera V, Esteban-Hernández J, Sánchez-Muñoz G, et al. Comorbidities and hospital death using administrative data. Med Care. 2009;47:626–33. risk of mortality among hospitalized patients with idiopathic pulmonary 30. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas fibrosis in Spain from 2002 to 2014. Respir Med. 2018;138:137–43. under two or more correlated receiver operating characteristic curves: 54. Menendez ME, Neuhaus V, van Dijk CN, Ring D. The Elixhauser a nonparametric approach. Biometrics. 1988;44:837–45. comorbidity method outperforms the Charlson index in predict- 31. Murata A, Matsuda S, Mayumi T, Yokoe M, Kuwabara K, Ichimiya Y, et al. ing inpatient death after orthopaedic surgery. Clin Orthop Relat Res. Effect of hospital volume on clinical outcome in patients with acute 2014;472:2878–86. pancreatitis, based on a national administrative database. Pancreas. 55. Murata A, Ohtani M, Muramatsu K, Matsuda S. Influence of comorbid- 2011;40:1018–23. ity on outcomes of older patients with acute pancreatitis based on 32. McNabb-Baltar J, Ravi P, Isabwe GA, Suleiman SL, Yaghoobi M, Trinh a national administrative database. Hepatobiliary Pancreat Dis Int. Q-D, et al. A population-based assessment of the burden of acute 2015;14:422–8. pancreatitis in the United States. Pancreas. 2014;43:687–91. 56. Bernardi M, Moreau R, Angeli P, Schnabl B, Arroyo V. Mechanisms 33. Tonsi AF, Bacchion M, Crippa S, Malleo G, Bassi C. Acute pancreatitis at of decompensation and organ failure in cirrhosis: From peripheral the beginning of the 21st century: the state of the art. World J Gastro- arterial vasodilation to systemic inflammation hypothesis. J Hepatol. enterol. 2009;15:2945–59. 2015;63:1272–84. 34. Gardner TB, Vege SS, Chari ST, Pearson RK, Clain JE, Topazian MD, et al. 57. Shin KY, Lee WS, Chung DW, Heo J, Jung MK, Tak WY, et al. Influence of The effect of age on hospital outcomes in severe acute pancreatitis. obesity on the severity and clinical outcome of acute pancreatitis. Gut Pancreatology. 2008;8:265–70. Liver. 2011;5:335–9. 35. Xin M-J, Chen H, Luo B, Sun J-B. Severe acute pancreatitis in the 58. Moran RA, García-Rayado G, de la Iglesia-García D, Martínez-Moneo E, elderly: etiology and clinical characteristics. World J Gastroenterol. Fort-Martorell E, Lauret-Braña E, et al. Influence of age, body mass index 2008;14:2517–21. and comorbidity on major outcomes in acute pancreatitis, a prospective Hidalgo et al. BMC Gastroenterology (2023) 23:81 Page 10 of 10 nation-wide multicentre study. United European Gastroenterol J. 2018;6:1508–18. 59. Hong S, Qiwen B, Ying J, Wei A, Chaoyang T. Body mass index and the risk and prognosis of acute pancreatitis: a meta-analysis. Eur J Gastroenterol Hepatol. 2011;23:1136–43. 60. Ranson JH, Rifkind KM, Roses DF, Fink SD, Eng K, Spencer FC. Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet. 1974;139:69–81. 61. Frey CF. Gallstone pancreatitis. Surg Clin North Am. 1981;61:923–38. 62. Samanta J, Dhaka N, Gupta P, Singh AK, Yadav TD, Gupta V, et al. Compara- tive study of the outcome between alcohol and gallstone pancreatitis in a high-volume tertiary care center. JGH Open. 2019;3:338–43. 63. Andersen AM, Novovic S, Ersbøll AK, Hansen MB. Mortality in alcohol and biliary acute pancreatitis. Pancreas. 2008;36:432–4. 64. Bálint ER, Fűr G, Kiss L, Németh DI, Soós A, Hegyi P, et al. Assessment of the course of acute pancreatitis in the light of aetiology: a systematic review and meta-analysis. Sci Rep. 2020;10:17936. 65 Gardner TB. Acute Pancreatitis. Ann Intern Med. 2021;174:ITC17-32. 66. Kochar B, Akshintala VS, Afghani E, Elmunzer BJ, Kim KJ, Lennon AM, et al. Incidence, severity, and mortality of post-ERCP pancreatitis: a system- atic review by using randomized, controlled trials. Gastrointest Endosc. 2015;81:143-149.e9. 67. Facciorusso A, di Maso M, Serviddio G, Larghi A, Costamagna G, Mus- catiello N. Echoendoscopic ethanol ablation of tumor combined with celiac plexus neurolysis in patients with pancreatic adenocarcinoma. J Gastroenterol Hepatol. 2017;32:439–45. 68. Kirkegård J, Mortensen MR, Johannsen IR, Mortensen FV, Cronin-Fenton D. Positive predictive value of acute and chronic pancreatitis diagnoses in the danish national patient registry: A validation study. Scand J Public Health. 2020;48:14–9. 69. Floyd JS, Bann MA, Felcher AH, Sapp D, Nguyen MD, Ajao A, et al. Valida- tion of acute pancreatitis among adults in an integrated healthcare system. Epidemiology. 2023;34:33–7. 70. Xiao AY, Tan ML, Plana MN, Yadav D, Zamora J, Petrov MS. The use of international classification of diseases codes to identify patients with pancreatitis: a systematic review and meta-analysis of diagnostic accu- racy studies. Clin Transl Gastroenterol. 2018;9:191. 71. Razavi D, Ljung R, Lu Y, Andrén-Sandberg A, Lindblad M. Reliability of acute pancreatitis diagnosis coding in a National Patient Register: a validation study in Sweden. Pancreatology. 2011;11:525–32. 72. Ribera A, Marsal JR, Ferreira-González I, Cascant P, Pons JM, Mitjavila F, et al. Predicting in-hospital mortality with coronary bypass surgery using hospital discharge data: comparison with a prospective observational study. Rev Esp Cardiol. 2008;61:843–52. 73. Guillaumes S, Hoyuela C, Hidalgo NJ, Juvany M, Bachero I, Ardid J, et al. Inguinal hernia repair in Spain. A population-based study of 263,283 patients: factors associated with the choice of laparoscopic approach. Hernia. 2021;25:1345–54. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions