Abstract
Background
Early prediction of persistent organ failure (POF) is important for triage and timely treatment of patients with acute pancreatitis (AP).
Methods
All AP patients were consecutively admitted within 48 h of symptom onset. A nomogram was developed to predict POF on admission using data from a retrospective training cohort, validated by two prospective cohorts. The clinical utility of the nomogram was defined by concordance index (C-index), decision curve analysis (DCA), and clinical impact curve (CIC), while the performance by post-test probability.
Results
There were 816, 398, and 880 patients in the training, internal and external validation cohorts, respectively. Six independent predictors determined by logistic regression analysis were age, respiratory rate, albumin, lactate dehydrogenase, oxygen support, and pleural effusion and were included in the nomogram (web-based calculator: https://shina.shinyapps.io/DynNomapp/). This nomogram had reasonable predictive ability (C-indexes 0.88/0.91/0.81 for each cohort) and promising clinical utility (DCA and CIC). The nomogram had a positive likelihood ratio and post-test probability of developing POF in the training, internal and external validation cohorts of 4.26/31.7%, 7.89/39.1%, and 2.75/41%, respectively, superior or equal to other prognostic scores.
Conclusions
This nomogram can predict POF of AP patients and should be considered for clinical practice and trial allocation.
Introduction
Acute pancreatitis (AP) has a spectrum of severity ranging from mild to critical.
1
The early prediction of the severity of AP informs clinical decisions about triage, transfer, and intervention.2
Early prediction is also important in the research setting, where the accurate allocation of patients into trial arms based on predicted severity is important for testing treatments for AP. The key determinants of AP severity are organ failure and infected pancreatic necrosis.3
Persistent organ failure (POF) is the most important determinant for mortality,4
, 5
, 6
the leading cause of death from AP7
and is the basis for the severe grade of AP in the revised Atlanta classification (RAC)8
).Given the importance of predicting severity and since Ranson defined his score in 1974,
9
there have been more than 20 prognostic scores have been studied for AP severity prediction.10
However, their clinical utility is limited by an accuracy of predicting POF of around 75% and many are cumbersome to use.11
Recent systematic reviews and meta-analyses conclude that current early predictors of POF,12
infected pancreatic necrosis,12
and mortality10
are not sufficiently accurate for decision making in individual patients. The ideal predictor of POF would be applied on patient admission and within 24 h of the onset of symptoms. It would be cost-effective, easy to use and have an accuracy between 95 and 100%. Considerable progress has been made in identifying serum biomarkers to stratify early risk and severity in patients with AP.13
However, due to their rapid time-course changes in serum concentration, non-specificity, cost, complexity, and suboptimal accuracy, none of these biochemical biomarkers have been adopted into routine clinical practice. In recent years, different nomograms have been applied for predicting severity and mortality,14
, 15
, 16
, 17
splanchnic vein thrombosis,18
abdominal infection,19
oral refeeding intolerance,20
success of catheter drainage in necrotizing AP,21
,22
and risk of new-onset diabetes after AP,23
as well as computed tomography index for assessing AP outcomes24
(Table 1). Of 4 studies14
, 15
, 16
, 17
that used a nomogram to predict severity and mortality, 3 studies used online Critical Care Database (Medical Information Mart for Intensive Care III database [MIMIC-III],14
,16
eICU Collaborative Research Database [eICU-CRD]15
) and the fourth used retrospective data.17
None of the 4 studies reported the duration of symptoms prior to hospital admission. No study to date has used training and validation cohorts to develop a nomogram to predict POF in AP patients.Table 1Summarize of the present developed nomograms in AP
| Nomograms for predicting SAP or mortality | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Predicted outcome | Study design | Symptom onset time | Data collection time | Indicators of the Nomogram | Training cohort | Validation cohort | Main findings | ||||
| N | C-index | N | C-index | ||||||||
| Jiang et al. 2019 [ 14 ] | Mortality | 30-day | MIMIC-III | NA | Within 24 hWithin 24 h | age, ALT, RDW, BUN | 228 | 0.751 | 114 | 0.875 | Nomogram > BISAP > SOFA > SIRS for long-term mortality both in cohorts. |
| 180-day/ 1-year | age, WBC, RDW, SCR | 0.758 | 0.856 | ||||||||
| Li et al. 2020 [ 15 ] | In-hospital mortality | eICU-CRD | NA | Within 24 h | age, BUN and lactate | 378 | 0.896 | 123 | 0.892 | The nomogram and APACHE IV demonstrated comparable power in predicting in-hospital mortality. | |
| Xu et al. 2020 [ 16 ] | SAP | MIMIC-III | NA | On admissionOn admission | SOFA, hemoglobin, albumin, TBIL, BUN | 708 | 0.855 | 477 | 0.879 | Nomogram > SOFA / OASIS for SAP and mortality in both cohorts. | |
| In-hospital mortality | Age, SOFA, WBC, TBIL, albumin | 0.821 | 0.822 | ||||||||
| Cao et al. 2021 [ 17 ] | SAP | Retrospective | NA | Within 24 h | Sex, Ca2+, SCR, NE%, LYMPH%, EO% | 571 | 0.69 | 150 | 0.71 | No comparison with others. | |
| Other nomograms in AP | |||||||||||
| Predicted outcome | Statistical methods | Indicators of the Nomogram | Training cohort | Validation cohort | |||||||
| N | C-index | N | C-index | ||||||||
| Hollemans et al. 2016 [ 21 ] | Success of catheter drainage in infected necrotizing pancreatitis | Univariate and multivariate logistic regression | Sex, percentage of pancreatic necrosis, density collection, multiple organ failure | 130 | 0.76 | NA | NA | ||||
| Zhou et al. 2016 [ 18 ] | Symptomatic splanchnic vein thrombosis in necrotizing acute pancreatitis | Univariate and multivariate logistic regression | Balthazar's CT score, intra-abdominal pressure and superior mesenteric vein thrombosis | 104 | 0.842 | NA | NA | ||||
| Bevan et al. 2017 [ 20 ] | Oral feeding intolerance | Univariate and multivariate logistic regression | Day 2 GCSI nausea/vomiting subs-core, etiology | 217 | NA | NA | NA | ||||
| Ma et al. 2019 [ 23 ] | New-onset diabetes mellitus after the first attack of acute pancreatitis | Univariate and multivariate logistic regression | Age, BMI, glucose, triglyceride, and LDL-C | 616 | 0.686 | NA | NA | ||||
| Bellam et al. 2019 [ 22 ] | Success of percutaneous catheter drainage in patients with acute pancreatitis having acute fluid collection | Univariate and multivariate logistic regression | volume of collection after PCD and organ failure resolution after PCD | 51 | 0.915 | NA | NA | ||||
| Gupta et al. 2021 [ 24 ] | Mortality | Binomial logistic regression | Pancreatic necrosis, ascites, pleural effusion | 103 | 0.79 | 20 | 0.74 | ||||
| ICU stay | Number of collection, pleural effusion | 0.66 | 0.70 | ||||||||
| hospital stay≥4 weeks | Pancreatic necrosis, number of collection, amount PE | 0.75 | 0.77 | ||||||||
| Readmission | Number of collection, coeliac artery | 0.70 | 0.52 | ||||||||
| ICU stay≥2 weeks | Pancreatic necrosis, largest dimension of collection, pleural effusion | 0.83 | 0.45 | ||||||||
| SAP | Number of collection, liver steatosis | 0.64 | 0.69 | ||||||||
| Zhu et al. 2021 [ 19 ] | Intra-abdominal infection | LASSO regression | Intra-abdominal pressure, APACHE II score, CTSI, ICU admission, and severity grade | 417 | 0.99 | 294 | 0.98 | ||||
SAP, severe acute pancreatitis; MIMIC-III, Medical Information Mart for Intensive Care III; NA, not available; ALT, alanine aminotransferase; RDW, red cell distribution width; BUN, blood urea nitrogen; WBC, white blood cell count; SCR, serum creatinine; BISAP, Bedside Index of Severity in Acute Pancreatitis; SOFA, Sequential Organ Failure Assessment; SIRS, Systemic Inflammatory Response Syndrome; ICU, intensive care unit; eICU-CRD, eICU Collaborative Research Database; APACHE, Acute Physiology, Age, and Chronic Health Evaluation; TBIL, total bilirubin; OASIS, Outcome and Assessment Information Set; NE%, neutrophil ratio; LYMPH%, lymphocyte Ratio; eosinophil ratio, EO%; GCSI, Gastroparesis Cardinal Symptom Index; LDL-C, low density lipoprotein-cholesterol; PCD, percutaneous catheter drainage; CTSI, computed tomography severity index.
The aim of this study was (1) to develop and validate a nomogram for predicting POF on admission in patients with AP, and (2) to compare the performance of the nomogram with conventional prognostic scores for predicting POF.
Methods
Study design and ethics
This study followed the STROBE guidelines
25
for observational studies. The study protocol was approved by respective Institutional Review Board in two tertiary hospitals. Data were obtained from two AP cohorts in West China Hospital of Sichuan University (WCH/SCU; Chengdu). The training cohort was from a retrospective dataset from 1st July 2009 to 30th June 2013.6
The internal validation cohort was from a prospective dataset from 1st January 2016 and 31st August 2017.26
The external validation cohort was from a prospective dataset from The First Affiliated Hospital of Nanchang University (Nanchang) from 1st January 2005 to 31st December 2012.27
Inclusion and exclusion criteria
The inclusion criteria were (1) diagnosis of AP diagnosed by the RAC criteria,
8
(2) age >18 and ≤80 years old, and (3) duration of abdominal pain from onset to admission ≤48 h, (4) data to calculate conventional prognostic scores within 6 h of admission.The exclusion criteria were (1) patients admitted to another hospital first and then transferred, (2) patients re-admitted during the same episode of AP, (3) patients with an etiology related to chronic pancreatitis, pancreatic neoplasia, pancreatic trauma, or pregnancy, (4) advanced pulmonary, cardiac, renal diseases (chronic kidney disease stage 4–5), liver cirrhosis (modified Child–Pugh grade 2–3) or malignancy, and (5) if the nomogram could not be completed because of missing data.
Definitions, variables, and outcome measures
Data collection: Experienced doctoral students and resident doctors specializing in management of AP were stringently trained for data collection according to pre-defined proforma and standard operating procedures in both centers. Each proforma was checked and signed off by attending or more senior doctors. The data collected on admission included age, gender, data for Charlson co-morbidity index, and duration of abdominal pain prior to admission. Vital signs, laboratory parameters (indices of routine blood, liver and renal function, myocardial enzymes, blood lipids and glucose, electrolyte levels and etc.), details of treatment (especially use of oxygen support, vasopressors, dialysis, antibiotic, and steroids), surgical interventions, and clinical outcomes were also recorded.
Presence of pleural effusion that were mostly obtained from nonenhanced computed tomography scan on admission. Patients without overt symptoms of pleural effusion (dyspnea, cough, and occasionally sharp, non-radiating, pleuritic chest pain) were not referred to having X-ray or CT scan in up to 10% of overall patients and these patients were considered as pleural effusion free.
Etiology
Hypertriglyceridemia as the etiology was defined as admission serum triglyceride (TG) level >11.3 mmol/l (1000 mg/dl) or TG > 5.65 mmol/l (500 mg/dl) with lipemic serum when other causes have been ruled out.
28
, 29
, 30
Biliary etiology was considered if gallstones or biliary sludge was present on radiological imaging or had an admission alanine aminotransferase level of over three times the upper limit of normal.31
Alcohol excess was diagnosed if with drinking history >35 standard drinks per week for >5 years32
or was deduced according to our center's own practice according to the drinking history after ruling out other etiologies.6
Organ failure, complications, and mortality
POF was defined as at least one organ systems (respiratory, circulatory, and renal) with a SOFA score of ≥2 and lasting ≥48 h.
1
Acute necrotic collection and acute peripancreatic fluid collection were defined as per RAC criteria.8
Major infection included infected pancreatic necrosis, bacteremia, or pneumonia alone or in combination.33
Mortality was recorded during the index admission or at 3 months of follow up for those who were automatically discharged with persistent multiple organ failure and had high possibility of death. Length of hospital stay were recorded for the index hospitalization.Prognostic scores
The prognostic scores that were compared with the nomogram in this study were National Early Warning Score (NEWS),
34
Systemic Inflammatory Response Syndrome (SIRS),11
Bedside Index for Severity in Acute Pancreatitis (BISAP),11
modified Glasgow criteria,11
Acute Physiology and Chronic Health Examination (APACHE) II,11
and Sequential Organ Failure Assessment (SOFA).11
These were all calculated on admission (or within 6 h of admission in rare cases).Statistical analysis
Continuous variables were expressed as median with 25th-75th percentile and were compared by Mann–Whitney U test (2 groups) or Kruskal–Wallis H test (3 groups). Categorical data were reported as number with percentage and were compared by means of χ2 or Fisher's exact test.
Independent risk factors for POF were identified and assessed significance of each by univariate logistic regression analysis in the training cohort. All variables significantly associated with POF were candidates for stepwise multivariate analysis, the results of which were used to formulate a nomogram. The predictive performance of the nomogram was measured by concordance index (C-index) and calibration with 1000 bootstrap samples to decrease the overfit bias.
35
The clinical utility of the nomogram was assessed by decision curve analysis (DCA) and clinical impact curve (CIC).36
DCA is a simple, novel method of evaluating predictive models with decision analyses of diagnostic and prognostic tests by using a risk-benefit ratio.37
CIC shows the estimated number considered at high risk for each risk threshold and depicts the proportion of those that are true positives.36
The area under curve (AUC) of the receiver operating characteristic (ROC) curves with 95% confidence intervals (CI) was calculated for potential predictors. An AUC of 0.5–0.7 was taken to represent no significant discrimination, 0.7 to 0.8 acceptable, 0.8 to 0.9 excellent, and >0.9 outstanding.
38
The AUC of two predictors were compared by method of DeLong.39
The optimum cut-off for each predictor was determined by the ROC curve. The sensitivity, specificity, positive likelihood ratio (PLR), and negative likelihood ratio (NLR) were calculated by clinical decision analysis. A positive likelihood ratio is the probability that a positive test would be expected in a patient divided by the probability that a positive test would be expected in a patient without the endpoint of interest (e.g. POF). A negative likelihood ratio is the converse and if < 0.1 effectively rules out the endpoint. The post-test probabilities for each predictor were derived from the PLR (and NLR), using the pre-test probability of the particular endpoint from the relevant cohort and the standard likelihood ratio nomogram.40
The analyses were performed using IBM SPSS Statistics V26.0 software (SPSS, IBM Corp; Armonk, New York, USA) and R software version 4.0.4 (http://www.Rproject.org). Two-tailed P value with statistical significance set at < 0.05 was used for all tests.
Results
Patient characteristics
There were 816, 398, and 880 eligible patients in the training, internal validation and external validation cohorts, respectively. The demographic and clinical outcome profiles are outlined in Table 2 and Online Supplementary Tables S1a–c. These results are similar to previous published studies.
6
,26
,27
The median age was 45 years old and about 70% patients were male in both training and internal validation cohorts, while median age was 50 years old and 61% patients were male in the external validation cohorts. Hypertriglyceridemia (32.7% and 42%) was the most common etiology in the training and internal validation cohorts. Biliary (57.5%) was the most common etiology in the external validation cohort.Table 2Baseline characteristics and clinical outcomes in training and validation cohorts
| Variables | Training cohort | Validation cohorts | |
|---|---|---|---|
| n = 816 | (Internal) n = 398 | (External) n = 880 | |
| Demographics | |||
| Age, years∗ | 45 (37–55) | 45 (38–51) | 50 (41–62) |
| Male | 568 (69.6) | 273 (68.6) | 533 (60.6) |
| Charlson comorbidity index∗ | 1 (0–1) | 1 (0–2) | 0 (0–0) |
| Etiology | |||
| Biliary | 230 (28.2) | 78 (19.6) | 506 (57.5) |
| Hypertriglyceridemia | 267 (32.7) | 167 (42.0) | 184 (20.9) |
| Alcohol | 30 (3.7) | 31 (7.8) | 61 (6.9) |
| Unknown or others | 289 (35.4) | 122 (30.7) | 129 (14.7) |
| Time to admission, hours∗ | 12 (7–24) | 15 (9–24) | NA |
| Clinical outcomes | |||
| Persistent organ failure | 80 (9.8) | 30 (7.5) | 178 (20.2) |
| Respiratory | 77 (9.4) | 30 (7.5) | 169 (19.2) |
| Circulatory | 14 (1.7) | 2 (0.5) | 12 (1.4) |
| Renal | 10 (1.2) | 3 (0.8) | 28 (3.2) |
| Need for HDU/ICU | 59 (7.2) | 29 (7.3) | 168 (19.1) |
| Local complication | |||
| Acute peripancreatic fluid collection | 163 (20.0) | 119 (29.9) | 332 (37.7) |
| Acute necrotic collection | 59 (7.2) | 55 (13.8) | 204 (23.2) |
| Major infection | 45 (5.5) | 20 (5.0) | 53 (6.0) |
| Infected pancreatic necrosis | 12 (1.5) | 9 (2.3) | 36 (4.1) |
| Bacteremia | 15 (1.8) | 6 (1.5) | 13 (1.5) |
| Lung | 35 (4.3) | 9 (2.3) | 18 (2.0) |
| Necrosectomy | 14 (1.7) | 13 (3.3) | 17 (1.9) |
| Mortality | 10 (1.2) | 4 (1.0) | 20 (2.3) |
| RAC | |||
| Mild | 508 (62.3) | 190 (47.7) | 328 (37.3) |
| Moderately severe | 228 (27.9) | 178 (44.7) | 374 (42.5) |
| Severe | 80 (9.8) | 30 (7.5) | 178 (20.2) |
| Length of hospital stay (days)∗ | 9 (6–13) | 8 (6–11) | 8 (6–13) |
NA, not available; HDU, high dependency unit; ICU, intensive care unit; RAC, revised Atlanta classification.
Values in parentheses are percentages unless indicated otherwise ∗values are median with 25th-75th percentile.
Developing the prediction nomogram
Candidate variables for the nomogram were drawn from demographic, vital signs, blood and biochemical tests, on admission. The independent prognostic factors for POF after univariate and multivariate logistic regression analysis were age (OR 1.03 [95% CI 1.01–0.05]; P = 0.01), respiratory rate (1.25 [1.10–1.42]; P = 0.001), albumin (0.92 [0.87–0.98]; P = 0.013), lactate dehydrogenase (LDH; 1.002 [1.000–1.003]; P = 0.036), oxygen support (5.17 [2.91–9.20]; P < 0.001), and pleural effusion (3.61 [1.97–6.61]; P < 0.001) (Table 3). These variables were then used to construct the predictive nomogram (Fig. 1a). The total score was calculated as: ([age × 0·3224] – 4.8358]) + ([respiratory rate × 2.5] – 25]) + (43.518 – [0·7912 × albumin]) + (LDH × 0.0196) + 15.521 (if oxygen support) + 12.551 (if exist of pleural effusion). To facilitate the clinical application of our findings, we developed a mobile terminal-based calculator (https://shina.shinyapps.io/DynNomapp/) of the predictive nomogram (Fig. 1b and c).
Table 3Univariate and multivariate logistic regression analysis with stepwise variable selection
| Variables | Univariate analysis | Multivariate analysis | ||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Demographics | ||||
| Age, years | 1.036 (1.019–1.053) | <0.001 | 1.030 (1.007–1.053) | 0.01 |
| Gender | 1.263 (0.777–2.053) | 0.346 | ||
| Charlson comorbidity index | 0.96 (0.684–1.346) | 0.811 | ||
| Etiology | 0.861 (0.711–1.043) | 0.127 | ||
| Time to admission, hours | 1.005 (0.984–1.026) | 0.658 | ||
| Admission vital signs | ||||
| Body temperature | 1.377 (0.902–2.102) | 0.138 | ||
| Respiratory rate | 1.352 (1.207–1.514) | <0.001 | 1.249 (1.099–1.420) | 0.001 |
| Heart rate | 1.029 (1.016–1.042) | <0.001 | 1.006 (0.989–1.023) | 0.499 |
| Admission laboratory parameters | ||||
| Amylase, IU/l | 1.000 (1.000–1.001) | 0.001 | 1.000 (1.000–1.000) | 0.448 |
| Lipase, IU/l | 1.000 (1.000–1.000) | 0.096 | ||
| White blood cell, 109/l | 1.112 (1.062–1.165) | <0.001 | 1.085 (0.854–1.379) | 0.505 |
| Neutrophils, 109/l | 1.115 (1.063–1.169) | <0.001 | 0.974 (0.755–1.257) | 0.839 |
| Lymphocytes, 109/l | 0.958 (0.709–1.293) | 0.777 | ||
| NLR | 1.007 (0.996–1.018) | 0.186 | ||
| Platelet, 109/l | 1.000 (0.997–1.003) | 0.868 | ||
| Hematocrits, I/l | 2.973 (0.023–377.4) | 0.659 | ||
| Hemoglobin, g/l | 1.011 (1.000–1.022) | 0.057 | ||
| ALT, IU/l | 1.001 (0.999–1.002) | 0.247 | ||
| AST, IU/l | 1.001 (1.000–1.002) | 0.125 | ||
| Albumin, g/l | 0.884 (0.844–0.925) | <0.001 | 0.924 (0.868–0.984) | 0.013 |
| Glucose, mmol/l | 1.063 (1.021–1.106) | 0.003 | 1.003 (0.945–1.066) | 0.915 |
| Urea, mmol/l | 1.219 (1.124–1.321) | <0.001 | 1.105 (0.973–1.256) | 0.123 |
| Creatinine, μmol/l | 1.012 (1.005–1.018) | <0.001 | 1.001 (0.992–1.009) | 0.904 |
| Triglycerides, mmol/l | 1.017 (0.993–1.041) | 0.162 | ||
| Cholesterol, mmol/l | 1.041 (0.977–1.109) | 0.211 | ||
| Lactate dehydrogenase, IU/l | 1.003 (1.002–1.004) | <0.001 | 1.002 (1.000–1.003) | 0.036 |
| Calcium, mmol/l | 0.093 (0.033–0.263) | <0.001 | 0.719 (0.174–2.970) | 0.648 |
| Oxygen supporting | 5.722 (3.486–9.393) | <0.001 | 5.170 (2.905–9.202) | <0.001 |
| Pleural effusion | 3.802 (2.318–6.237) | <0.001 | 3.611 (1.973–6.610) | <0.001 |
NLR, neutrophil to lymphocyte ratio; LDH, lactate dehydrogenase; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
Figure 1Nomogram for predicting POF in patients with AP and dynamic web-based calculator. Patient prognostic values are located on the axis of each variable; a line is then drawn upwards at a 90° angle to determine the number of points for that particular variable. The sum of these numbers is located on the total points axis, and a line is drawn at a 90° angle downward to the “Predicted SAP” axis to determine the likelihood of POF (a). Alternatively, POF rate can be ascertained from the online calculator with 95% CI in both graphical (b) and numerical (c) summary
Concordance and clinical utility of the nomogram
In the training cohort there was good concordance (C-index) between predicted POF and actual POF at 0.88 (Fig. 2a) and shown by the calibration curve (Fig. 2b). The C-index was 0.91 for the internal validation cohort (Figure 2c and d) and 0.81 for the external validation cohort (Figure 2e and f) compared with the training cohort.
Figure 2Assessment of the nomogram in all three cohorts. The accuracy of the model was determined using AUC analysis and the distribution of the predicted probabilities presented by calibration curve for cohorts of the training (a, b), internal validation (c, d), and external validation (e, f)
The nomogram of training cohort (Fig. 3a and b), internal validation cohort (Fig. 3c and d), and external validation cohort (Fig. 3e and f) exhibited better or equal clinical utility (DCA) with the other prognostic scores and had good clinical net benefits (CIC) for the identification of SAP patients. Of particular note, the nomogram outperformed all other indices in the internal validation cohort suggesting that it could help clinicians to obtain maximum benefit when making clinical decisions as it showed more benefit than the extreme situation of diagnosing POF in all patients or none.
Figure 3Clinical utility of the nomogram. The DCA and CIC of the nomogram for predicting POF in cohorts of training (a, b), internal validation (c, d), and external validation (e, f)
Prediction of POF, major infection, and mortality by nomogram
Prediction of POF
The pre-test probability of POF in the training, internal validation and external validation cohorts was 9.8% (80/816), 7.5% (30/398) and 20.2% (178/880) patients, respectively (Table 2).
The nomogram had a positive likelihood ratio and post-test probability of developing POF in the training, internal validation and external validation cohorts of 4.26 and 31.7% (259/816), PLR 7.89 and 39.1% (156/398) and PLR 2.75 and 41% (361/880) patients, respectively (Table 4).
Table 4Predictive value of the nomogram and clinical prognostic scores for persistent organ failure
| AUC | P value | Cut-off | Sensitivity (%) | Specificity (%) | PLR | NLR | Post_Prob_Pos (%) | Post_Prob_Neg (%) | |
|---|---|---|---|---|---|---|---|---|---|
| Training cohort (9.8% pre-test probability) | |||||||||
| Nomogram | 0.88 (0.86–0.91) | <0.001 | 91.2 | 78.8 | 81.5 | 4.26 | 0.26 | 31.7 | 2.8 |
| NEWS | 0.85 (0.82–0.87) | <0.001 | 4 | 66.3 | 88 | 5.54 | 0.38 | 37.6 | 4 |
| BISAP | 0.89 (0.87–0.91) | <0.001 | 2 | 78.8 | 89.1 | 7.24 | 0.24 | 44.0 | 2.5 |
| APACHE II | 0.79 (0.76–0.82) | <0.001 | 6 | 61.3 | 83 | 3.61 | 0.47 | 28.2 | 4.9 |
| SIRS | 0.84 (0.81–0.86) | <0.001 | 2 | 91.3 | 74.6 | 3.59 | 0.12 | 28.1 | 1.3 |
| Glasgow | 0.75 (0.72–0.78) | <0.001 | 2 | 61.3 | 78.8 | 2.89 | 0.49 | 23.9 | 5.1 |
| SOFA | 0.87 (0.84–0.89) | <0.001 | 1 | 85 | 81.2 | 4.53 | 0.18 | 33.0 | 1.9 |
| Validation cohort | |||||||||
| Internal validation (7.5% pre-test probability) | |||||||||
| Nomogram | 0.91 (0.88–0.94) | <0.001 | 79.2 | 90 | 88.6 | 7.89 | 0.11 | 39.1 | 0.9 |
| NEWS | 0.79 (0.75–0.83) | <0.001 | 4 | 70.0 | 75.0 | 2.80 | 0.40 | 18.5 | 3.1 |
| BISAP | 0.75 (0.71–0.79) | <0.001 | 2 | 46.7 | 91.9 | 5.72 | 0.58 | 31.7 | 4.5 |
| APACHE II | 0.66 (0.61–0.71) | 0.005 | 6 | 36.7 | 89.4 | 3.46 | 0.71 | 21.9 | 5.4 |
| SIRS | 0.68 (0.63–0.72) | <0.001 | 2 | 70.0 | 61.4 | 1.81 | 0.49 | 12.8 | 3.8 |
| Glasgow | 0.69 (0.64–0.73) | <0.001 | 3 | 46.7 | 80.7 | 2.42 | 0.66 | 16.4 | 5.1 |
| SOFA | 0.73 (0.68–0.77) | <0.001 | 2 | 63.3 | 75.8 | 2.62 | 0.48 | 17.5 | 3.7 |
| External validation (20.2% pre-test probability) | |||||||||
| Nomogram | 0.81 (0.79–0.84) | <0.001 | 69.4 | 79.8 | 70.9 | 2.75 | 0.29 | 41 | 6.7 |
| NEWS | 0.76 (0.73–0.79) | <0.001 | 5 | 62.4 | 78.6 | 2.92 | 0.48 | 42.5 | 10.8 |
| BISAP | 0.75 (0.72–0.78) | <0.001 | 2 | 53.4 | 84.3 | 3.41 | 0.55 | 46.3 | 12.3 |
| APACHE II | 0.75 (0.72–0.78) | <0.001 | 9 | 56.2 | 80.8 | 2.92 | 0.54 | 42.6 | 12.1 |
| SIRS | 0.73 (0.70–0.76) | <0.001 | 2 | 68.0 | 69.2 | 2.21 | 0.46 | 35.9 | 10.5 |
| Glasgow | 0.78 (0.75–0.81) | <0.001 | 3 | 59.6 | 82.5 | 3.4 | 0.49 | 46.3 | 11.1 |
| SOFA | 0.80 (0.77–0.83) | <0.001 | 2 | 63.5 | 85.0 | 4.24 | 0.43 | 51.8 | 9.8 |
AUC, area under curve; PLR, positive likelihood ratio; NLR, negative likelihood ratio; Post_Prob_Pos, post-test probability of a positive test; Post_Prob_Neg, post-test probability of a negative test; NEWS, National Early Warning Score; BISAP, Bedside Index for Severity in Acute Pancreatitis; APACHE II, Acute Physiology and Chronic Health Examination II; SIRS, Systemic Inflammatory Response Syndrome; SOFA, Sequential Organ Failure Assessment.
The nomogram had a negative likelihood ratio and post-test probability of not developing POF in the training, internal validation and external validation cohorts of 0.26 and 2.8% (23/816), PLR 0.11 and 0.9% (4/398) and PLR 0.29 and 6.7% (59/880) patients, respectively (Table 4).
Comparison with other prognostic systems for POF
Training Cohort: Comparing with the other prognostic scores, the nomogram had the second highest AUC (AUC 0.88 [0.86–0.91]; the AUC for BISAP was 0.89 [0.87–0.91]. BISAP had the highest predictive value (AUC 0.89 [0.87–0.91], PLR 7.24) for POF compared with the other indices (Table 4; upper panel) and followed by the nomogram (AUC 0.88 [0.86–0.91], PLR 4.26), both higher than APACHE II (0.79 [0.76–0.82], PLR 3.61) and Glasgow (0.75 [0.72–0.78], PLR 2.89) (P < 0.05, Online Supplementary Table S2) in training cohort.
Internal validation cohort: The nomogram showed the best predictive value (AUC 0.91 [0.88–0.94], PLR 7.89, Table 4; middle panel) in internal validation cohort and higher than all other clinical scoring systems (P < 0.05; Online Supplementary Table S2), followed by NEWS (0.79 [0.75–0.83], PLR 2.80) and BISAP (0.75 [0.71–0.79], PLR 5.72). In addition, the NLR (0.11) of the nomogram ranked the best among all the comparative indices, with the lowest post-test probability of not being POF (0.9%) indicating a high negative prediction value.
External validation cohort: The nomogram also ranked the top AUC (0.81 [0.79–0.84]), lowest NLR (0.29) and post-test probability of not being POF (6.7%) among all the indices in the external validation cohort (Table 4; lower panel), with the AUC higher than NEWS, BISAP, APACHE II and SIRS (P < 0.05; Online Supplementary Table S2).
Subgroup analyses compared predictive value of the nomogram on POF between different causes of AP patients. The nomogram showed no significant difference in predictive values for predicting POF in both hypertriglyceridemia and non-hypertriglyceridemia (biliary, alcohol and others) patients in all three cohorts (Table 5).
Table 5Accuracy of nomogram in predicting POF between hypertriglyceridemia and non-hypertriglyceridemia associated-acute pancreatitis patients
| Etiology | Pre-test probability | AUC | P value | Cut-off | Sensitivity (%) | Specificity (%) | PLR | NLR | Post_Prob_Pos (%) | Post_Prob_Neg (%) | P value |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Training cohort | |||||||||||
| HTG | 12.4% | 0.89 (0.84–0.92) | <0.001 | 71.7 | 100 | 62.0 | 2.63 | 0.0 | 27.0 | 0.0 | 0.998 |
| Non-HTG | 8.56% | 0.89 (0.86–0.91) | <0.001 | 91.2 | 83.0 | 78.7 | 3.89 | 0.22 | 26.7 | 2.0 | |
| Validation cohort | |||||||||||
| Internal validation cohort | |||||||||||
| HTG | 8.98% | 0.91 (0.86–0.95) | <0.001 | 82.5 | 93.3 | 90.8 | 10.13 | 0.07 | 50.0 | 0.7 | 0.970 |
| Non-HTG | 6.49% | 0.92 (0.87–0.95) | <0.001 | 79.2 | 86.7 | 90.3 | 8.91 | 0.15 | 38.2 | 1.0 | |
| External validation cohort | |||||||||||
| HTG | 25% | 0.86 (0.80–0.91) | <0.001 | 76.3 | 76.1 | 82.6 | 4.37 | 0.29 | 59.3 | 8.8 | 0.115 |
| Non-HTG | 19% | 0.80 (0.77–0.83) | <0.001 | 69.6 | 77.3 | 70.9 | 2.66 | 0.32 | 38.3 | 7.0 | |
POF, persistent organ failure; AUC, area under curve; PLR, positive likelihood ratio; NLR, negative likelihood ratio; Post_Prob_Pos, post-test probability of a positive test; Post_Prob_Neg, post-test probability of a negative test; HTG, hypertriglyceridemia.
Prediction of major infection
There was 5.5% (45/816), 5.0% (20/398) and 6.0% (53/880) had infection in the training, internal and external validation cohorts, respectively. In the training cohort, the accuracy of these predictors was generally low with BISAP (0.75 [0.72–0.78], PLR 4.29) having the highest predictive value compared with the other indices in training cohort (Online Supplementary Table S3; upper panel), but there was no statistical difference between any of them (Online Supplementary Table S2). In both the internal and external validation cohorts, the nomogram showed the best predictive value (highest AUC 0.78 [0.73–0.82]/0.80 [0.77–0.83]; lowest NLR 0.47/0.22; lowest post-test probability of not being major infection 2.4%/1.4%; Online Supplementary Table S3; middle and lower panels).
Prediction of mortality
There was 1.2% (10/816), 1.0% (4/398) and 2.3% (20/880) mortality in the training, internal validation and external validation cohorts, respectively. All the deaths were from patients with POF in the three cohorts (12.5%, 13.3% and 11.2%; Online Supplementary Tables S1a–c). In both the training and external cohorts, the nomogram (AUC 0.88 [0.85–0.90], 0.89 [0.87-0.91]) showed the equivalent predictive values for mortality with most of the clinical scoring systems (Online Supplementary Table S4). In the internal validation cohort, the nomogram had the highest predictive values (0.99 [0.98–1.00], PLR 32.8, NLR 0.26) for mortality, followed by BISAP (0.95 [0.92–0.97], PLR 9.85), and NEWS (0.90 [0.86–0.93], PLR 29.6, NLR 0.26), all higher than the remaining clinical scores (Online Supplementary Table S4).
Data integrity
Since there was substantial amount of missing data, we assessed whether the patients included in all cohorts were different from those who were excluded due to missing data. There were no significant differences in age, hypertriglycedemic etiology, persistent organ failure, major infection, and mortality (Online Supplementary Table S5).
Discussion
In this study, we developed and validated a mobile terminal-based nomogram for predicting POF, major infection, and mortality using six independent prognostic factors (age, respiratory rate, albumin, LDH, oxygenation, and pleural effusion) that readily available on admission. This nomogram was found to be superior with the both internal and external validation cohort compared with other prognostic scores recommended for clinical use for predicting POF and major infection. Subgroup analysis comparing hypertriglyceridemia and non-hypertriglyceridemia etiology did not significantly change the predictive values of the nomogram for predicting POF, indicating its suitability for heterogenous etiologies. As POF is the diagnostic criteria for SAP, this nomogram is recommended for routine clinical use to predict SAP on admission or to rule it out (validation NLR 0.11/0.29). Therefore, these findings encourage the use of the simple-to-use web-based nomogram before new markers are developed and introduced in our settings. The consecutive nature of patient recruitment and short time from onset of pain to admission, stringently applied in two large different Chinese centers, adds strength to these conclusions.
Age is recognized as an individual risk factor for increased severity of AP and has been used by several prognostic scores
10
,11
and practice guidelines.41
To investigate the role of age and comorbidity in the severity of AP, Frey et al.42
carried out a retrospective study in 84,713 patients with a first-attack AP. They found that the 65 to 75 age group, and age >75 are strong predictors of early death with an odds ratio (OR) of 2.6 and 5.2, respectively. Similar findings also applied to patients with two chronic comorbidities (OR: 3.5) or ≥ 3 comorbidities (OR: 7.4). Moreover, the mortality rate was only 0.1% (14/14,280) for younger patients (age <55) without chronic comorbidities compared to 5.9% (701/24,852) for elderly patients (age >64) with ≥3 comorbidities in the first 14 days. In addition, they showed that recent cancer, heart failure, and renal and liver diseases are strongly correlated with outcomes. Further, in acute interstitial AP, which is known to have low mortality, the Charlson comorbidity index was strongly associated with adverse clinical outcomes.43
A recent review has concluded that age is associated with frequent cases of severe AP, leading to increased multiorgan failure and high mortality.44
Because of the significant impact of the degree and number of comorbidities on clinical outcomes, we therefore excluded patients with advanced (end-stage) comorbidities with an emphasizing on assessing the intrinsic prognostic factors for AP severity.Hypoalbuminemia occur in critically ill patients due to several factors including dilution from resuscitation, increased interstitial loss, altered liver function, and catabolic nutritional state.
45
It is strongly associated with poor clinical outcomes in acutely ill patients and it has also been shown to independently associated with POF and mortality in AP patients.46
Whitcomb et al.47
has recently found that albumin dropped rapidly in AP patients with multiple organ failure resulting in unregulated capillary leak with continued loss of larger plasma proteins. In contrast, the plasma albumin levels only dropped slightly in patients without multiple organ failure who tended to recover quickly unless they develop complications (infected pancreatic necrosis or sepsis). The therapeutic effect of albumin in inflammatory states is not only by affecting plasma volume dilation, but also by regulating inflammation and oxidative stress.48
Hypalbuminemia has recently been shown to be associated with severe AP and increased mortality in a cohort of 1149 patients.49
This is consistent with the inclusion of serum albumin level in a number of prognostic indices, including the Glasgow criteria.16
Raised lactate levels has been observed in many critical acute illness situations including sepsis
50
and AP.51
Elevated lactate may serve as a protective mechanism and has been shown to reduce Toll-like receptor and inflammasome-mediated pancreatic and liver injury via its receptor GPR81. LDH can reversibly catalyze the oxidation of lactate to pyruvate and has been employed by Ranson, Glasgow, Japanese Severity Score,11
and a nomogram15
for early AP severity prediction and reported as a simple and useful parameter for predicting POF52
and pancreatic necrosis.53
Respiratory rate and oxygen support reflect respiratory status and with respiratory failure as one most common organ dysfunction in AP,
6
the early recognition of respiratory dysfunction is considered important. Our results showed a positive association between the development of POF and an increased respiratory rate or requirement for oxygen support on admission. Therefore, both respiratory rate and oxygen support have been adopted in NEWS34
for their convenience and prognostic value. The significance of pleural effusion in AP patients has been long reported, and is part of BISAP.54
Our results showed that pleural effusion (odds ratio: 3.61 95%CI 1.97–6.6, P < 0.001) was an independent predictor for SAP, consistent with a previous study.55
In our univariate analysis before establishing the predictive nomogram, we also found white blood cell count, glucose, urea (or blood urea nitrogen), creatinine, and ionized calcium were independent individual prognostic factors for POF, consistent with previously published literature.
11
However, when these parameters were fitted into our multiple logistic regression model, they only had negligible impact on the final nomogram. Unlike most previously published studies included high proportion of SAP patients, we used a training consecutive cohort constituted only up to 10% of SAP patients which may partially explain different weights of individual prognostic factors in varied epidemiology situations. For example, three of the four existing predictive nomograms for AP severity were conducted in the ICU settings which cannot be generalizable for emergence departments or general wards where patients are primarily admitted.Some of the six independent prognostic factors of our nomogram are included as part of NEWS (respiratory rate and oxygen support) and BISAP (age, respiratory rate, and pleural effusion), the two that we found had justifiable good predictive values for POF in both our training and internal validation cohorts. However, the nomogram was simpler than BISAP and more AP-specific than NEWS to quantitatively predict clinical outcomes in a personalized way. In addition, we validated the nomogram in another Chinese tertiary hospital with the etiological composition different from ours. The results showed that the nomogram with stable clinical applicability in both AP cohorts with hypertriglyceridemia (internal validation) or biliary (external validation) as the main etiology, adding strength to its applicability.
Our study also has some limitations. Firstly, excluding missing data may contribute to varied accuracy of the nomogram. However, we found majority of the key parameters of baseline and outcomes, albeit not all (i.e. gender and length of hospital stay), were similar between those included and those with missing data in all cohorts. Secondly, the nomogram model was developed mainly based on the variables that were easy to get in our retrospective sets, but did not include other factors that may influence the precision of the model. For example, the oxygenation index was not included in our analysis because only paucity data were available. In a most recent study,
56
the authors found that oxygenation index had low prognostic power (AUC 55.3%) for acute respiratory distress syndrome. Thirdly, the nomogram did not have specific markers for circulatory and renal failure. The reasons for this may be attributed to low incidence of circulatory and renal failure of the study population and at the early disease stage respiratory failure commonly precedes other organ failures.4
,5
In the last, the lack of international validation may limit the extrapolation and generalizability of the nomogram.Conclusions
Our nomogram based on six readily available factors accurately predicted POF on admission in patients with AP. This nomogram can be routinely used for early AP severity prediction in our clinical practice if further validated in multiple center studies.
Authors’ contributions
QX, WH, and WHe obtained funding, and concepted, designed and supervised the study. NS collected, analyzed all retrospective and prospective data of Chengdu center, and drafted the manuscript. XZ collected and analyzed all retrospective data. WHe, LX, YZ, and NL provided and analyzed external validation datasets. LL, WC, LY, XY (Xinmin Yang) and RZ supported collecting data. PZ assisted data analysis. LD, TJ, ZL and KJ audited data quality. GJ and XY (XiaonanYang) supervised the patients’ treatment. WH and QX designed, TW and PS assisted, RS reviewed the proforma and e-database. VKS, RS, and JAW had important intelligence input. WH and JAW critically revised the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
This study was approved by the Institutional Review Board of West China Hospital of Sichuan University and The First Affiliated Hospital of Nanchang University.
Consent for publication
All authors have provided consent for publication of the manuscript.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Acknowledgements
The authors thank all the participants and attending physicians for their contributions. This study was supported by NZ-China Strategic Research Alliance 2016 Award (China: 2016YFE0101800, QX, TJ, WH and LD; New Zealand: JAW); Key Research and Development Program of Science and Technology Department of Sichuan Province (2020YFS0235, NS; 2019YFS0259, XZ); National Natural Science Foundation of China (81973632, WH; 81774120, QX; 81860122, WHe); and National Institute for Health Research ( NIHR ) Senior Investigator Award (RS).
Conflicts of interest
None to declare.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
- Multimedia component 1
- Multimedia component 2
References
- Determinant-based classification of acute pancreatitis severity: an international multidisciplinary consultation.Ann Surg. 2012; 256: 875-880
- Acute pancreatitis: a review.JAMA. 2021; 325: 382-390
- Organ failure and infection of pancreatic necrosis as determinants of mortality in patients with acute pancreatitis.Gastroenterology. 2010; 139: 813-820
- The role of organ failure and infection in necrotizing pancreatitis: a prospective study.Ann Surg. 2014; 259: 1201-1207
- Impact of characteristics of organ failure and infected necrosis on mortality in necrotising pancreatitis.Gut. 2019; 68: 1044-1051
- Duration of organ failure impacts mortality in acute pancreatitis.Gut. 2020; 69: 604-605
- Organ Failure Due to Systemic Injury in Acute Pancreatitis Gastroenterology. 2019; 156: 2008-2023
- Classification of acute pancreatitis--2012: revision of the Atlanta classification and definitions by international consensus.Gut. 2013; 62: 102-111
- Prognostic signs and the role of operative management in acute pancreatitis.Surg Gynecol Obstet. 1974; 139: 69-81
- Prediction models of mortality in acute pancreatitis in adults: a systematic review.Ann Intern Med. 2016; 165: 482-490
- Comparison of existing clinical scoring systems to predict persistent organ failure in patients with acute pancreatitis.Gastroenterology. 2012; 142 (quiz e1415-1476): 1476-1482
- Predictors of severe and critical acute pancreatitis: a systematic review.Dig Liver Dis. 2014; 46: 446-451
- Early laboratory biomarkers for severity in acute pancreatitis; A systematic review and meta-analysis.Pancreatology. 2020; 20: 1302-1311
- Prognostic nomogram for acute pancreatitis patients: an analysis of publicly electronic healthcare records in intensive care unit.J Crit Care. 2019; 50: 213-220
- Early prediction of in-hospital mortality in acute pancreatitis: a retrospective observational cohort study based on a large multicentre critical care database.BMJ Open. 2020; 10e041893
- Deceased serum bilirubin and albumin levels in the assessment of severity and mortality in patients with acute pancreatitis.Int J Med Sci. 2020; 17: 2685-2695
- Establishment and verification of a nomogram for predicting severe acute pancreatitis.Eur Rev Med Pharmacol Sci. 2021; 25: 1455-1461
- Predicting the clinical manifestations in necrotizing acute pancreatitis patients with splanchnic vein thrombosis.Pancreatology. 2016; 16: 973-978
- Intra-abdominal infection in acute pancreatitis in eastern China: microbiological features and a prediction model.Ann Transl Med. 2021; 9: 477
- Nomogram for predicting oral feeding intolerance in patients with acute pancreatitis.Nutrition. 2017; 36: 41-45
- Predicting success of catheter drainage in infected necrotizing pancreatitis.Ann Surg. 2016; 263: 787-792
- Predictors of outcome of percutaneous catheter drainage in patients with acute pancreatitis having acute fluid collection and development of a predictive model.Pancreatology. 2019; 19: 658-664
- Nomogram for predicting diabetes mellitus after the first attack of acute pancreatitis.Eur J Gastroenterol Hepatol. 2019; 31: 323-328
- Development and validation of a computed tomography index for assessing outcomes in patients with acute pancreatitis: "SMART-CT" index.Abdominal radiology (New York). 2021; 46: 1618-1628
- Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.BMJ. 2007; 335: 806-808
- Early rapid fluid therapy is associated with increased rate of noninvasive positive-pressure ventilation in hemoconcentrated patients with severe acute pancreatitis.Dig Dis Sci. 2020; 65: 2700-2711
- A study on the etiology, severity, and mortality of 3260 patients with acute pancreatitis according to the revised Atlanta classification in jiangxi, China over an 8-year period.Pancreas. 2017; 46: 504-509
- Elevated serum triglycerides are independently associated with persistent organ failure in acute pancreatitis.Am J Gastroenterol. 2015; 110: 1497-1503
- Systematic review of hypertriglyceridemia-induced acute pancreatitis: a more virulent etiology?.Pancreatology. 2016; 16: 469-476
- Clinical profile and natural course in a large cohort of patients with hypertriglyceridemia and pancreatitis.J Clin Gastroenterol. 2017; 51: 77-85
- Etiology and diagnosis of acute biliary pancreatitis.Nat Rev Gastroenterol Hepatol. 2010; 7: 495-502
- Patients with sentinel acute pancreatitis of alcoholic etiology are at risk for organ failure and pancreatic necrosis: a dual-center experience.Pancreas. 2016; 45: 997-1002
- Early versus on-demand nasoenteric tube feeding in acute pancreatitis.N Engl J Med. 2014; 371: 1983-1993
- Standardising the assessment of acute- illness severity in the NHS. Report of a working party.Royal College of Physicians, London2012
- Towards better clinical prediction models: seven steps for development and an ABCD for validation.Eur Heart J. 2014; 35: 1925-1931
- Assessing the clinical impact of risk prediction models with decision curves: guidance for correct interpretation and appropriate use.J Clin Oncol. 2016; 34: 2534-2540
- Decision curve analysis: a novel method for evaluating prediction models.Med Decis Making. 2006; 26: 565-574
- Receiver operating characteristic curve in diagnostic test assessment.J Thorac Oncol. 2010; 5: 1315-1316
- Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach.Biometrics. 1988; 44: 837-845
- Refining clinical diagnosis with likelihood ratios.Lancet. 2005; 365: 1500-1505
- WSES guidelines for the management of severe acute pancreatitis.World J Emerg Surg. 2019; 14: 27
- Co-morbidity is a strong predictor of early death and multi-organ system failure among patients with acute pancreatitis.J Gastrointest Surg : official journal of the Society for Surgery of the Alimentary Tract. 2007; 11: 733-742
- An assessment of the severity of interstitial pancreatitis.Clin Gastroenterol Hepatol. 2011; 9: 1098-1103
- Acute pancreatitis in elderly patients.Gastroenterology. 2021; 161: 1736-1740
- Albumin administration in the acutely ill: what is new and where next?.Crit Care. 2014; 18: 231
- Serum albumin is independently associated with persistent organ failure in acute pancreatitis.Chin J Gastroenterol Hepatol. 2017; 20175297143
- Severe acute pancreatitis: capillary permeability model linking systemic inflammation to multiorgan failure.Am J Physiol Gastrointest Liver Physiol. 2020; 319: G573-G583
- Albumin in decompensated cirrhosis: new concepts and perspectives.Gut. 2020; 69: 1127-1138
- Hypoalbuminemia affects one third of acute pancreatitis patients and is independently associated with severity and mortality.Sci Rep. 2021; 1124158
- The third international consensus definitions for sepsis and septic shock (Sepsis-3).JAMA. 2016; 315: 801-810
- Elevated arterial lactate level as an independent risk factor for pancreatic infection in moderately severe acute pancreatitis.Pancreatology. 2019; 19: 653-657
- Serum hydroxybutyrate dehydrogenase as an early predictive marker of the severity of acute pancreatitis: a retrospective study.BMC Gastroenterol. 2020; 20: 393
- Serum C-reactive protein, procalcitonin, and lactate dehydrogenase for the diagnosis of pancreatic necrosis.Cochrane Database Syst Rev. 2017; 4: CD012645
- The early prediction of mortality in acute pancreatitis: a large population-based study.Gut. 2008; 57: 1698-1703
- Factors predicting the severity of acute pancreatitis in elderly patients.Aging Clin Exp Res. 2021; 33: 183-192
- A simple-to-use web-based calculator for survival prediction in acute respiratory distress syndrome.Front Med. 2021; 8604694
Article Info
Publication History
Published online: June 05, 2022
Accepted:
May 31,
2022
Received:
August 31,
2021
Identification
Copyright
© 2022 The Authors. Published by Elsevier Ltd on behalf of International Hepato-Pancreato-Biliary Association Inc.
User License
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
01494-0&id=gr1.jpg){kind=link}
01494-0&id=gr2.jpg){kind=link}
01494-0&id=gr3.jpg){kind=link}