Acute pancreatitis (AP) is an acute necroinflammation of the pancreas. AP is a leading cause of hospitalization for gastrointestinal disorders (1). Although AP is often mild, the severe form (SAP) develops in 15–20% of patients and has up to 40% mortality (2). AP is diagnosed if at least two of the following three criteria are fulfilled: (1) clinical (abdominal pain consistent with acute pancreatitis), (2) laboratory (serum lipase or amylase >3-fold higher than the upper limit of normal), (3) imaging (characteristic findings of AP by contrast-enhanced computed tomography (CECT), magnetic resonance imaging (MRI) or transabdominal ultrasonography) (based on the International Association of Pancreatology and the American Pancreatic Association IAP/APA guidelines) (3). The incidence of AP ranges from 4.6 to 100 per 100,000 persons in Europe (4), and from 13 to 45 per 100,000 persons in the USA (5). The overall mortality ranges from 2 to 35% depending on whether organ failure or necrotizing pancreatitis occurs (6). The Hungarian Pancreatic Study Group (HPSG) was established in 2011. It developed a prospective electronic registry (https://tm-centre.org/en/registries/acute-pancreatitis-registry/) and collected data of 1,224 patients which served as a basis for this study.

The risk of developing AP is multifactorial. Among the most common causes are age (5) and lifestyle factors such as obesity and alcohol consumption (7). In a previous study, performed on the same cohort, we demonstrated that comorbidities predict mortality and severity of AP (8). Although early severity assessment is crucial for prognosis of outcome, there is no “gold standard” prognostic score for AP (9). Several prognostic systems are based solely on the evaluation of pancreatic necrosis and organ failure [e.g., Atlanta classification, Determinant based classification (DBC), Amylase and BMI (CAB) or the contrast-enhanced Computer Tomography base Severity Index (CTSI)]. Some score systems include other variables (PANC3: Htk, BMI, chest X-ray), CRP: inflammation (C-reactive protein). On the other hand, PANC4 is already extended with the inclusion of blood urea nitrogen (BUN). Some scoring systems include serum creatinine, eGFR or BUN [Revised Atlanta classification (RAC), Ranson's criteria, bedside index of severity in AP (BISAP), Acute Physiology and Chronic Health Examination II (APACHEII), Glasgow score, Harmless AP (10), Charlson Comorbidity Index (CCI)] (9). Thus, several presently used formulas do not include renal function and those which incorporate the most common comorbidities, and thus include renal function to predict AP prognosis, include it only as a simple score without analyzing the severity of renal functional impairment.

AP can be associated with circulatory shock-induced organ failure, including renal damage, potentially leading to acute kidney injury (AKI). The exact mechanism of AP-induced AKI (or renal failure) is yet to be fully understood, but besides hypotension the common causes can be systemic inflammation with cytokine production and free radical species, leading to renal parenchymal damage and hypoxemia (11, 12). Tissue damage and hypoxia stimulate further inflammation and oxidative stress leading to a vicious cycle. Although, organ damage or failure is well-known to severely impact the outcome of AP, the effects of early renal impairment on AP outcome have not been studied in depth before.

The aim of this study was to investigate possible associations between AP outcome and the clinical gold standard biomarker of renal function, the chronic kidney disease epidemiology (CKD-EPI) formula-based estimated glomerular filtration rate (eGFR) measured within the first 24 h after hospitalization.Article types.

Patients with AP were enrolled to the multicenter-based electronic database of the Hungarian Pancreatic Study Group (HPSG) – acute pancreatitis Registry. The data were collected from 2012 to 2017. AP diagnosis was established according to the IAP/APA guidelines (3). The following symptoms were considered in each patient: abdominal pain (clinical symptom), pancreatic enzyme elevation at least three times above the upper limit (laboratory data) and morphological changes (imaging). Patients were included in the registry if at least 2 out of the 3 criteria were fulfilled. In order to gain a more detailed insight into the possible associations and predictive role of estimated glomerular filtration rate (eGFR), the present study is based on data of all patients included in the registry, irrespective of the pancreatic enzyme elevation at the time of eGFR test.

Eighty-six parameters were registered altogether. Seventy-seven percentage of the requested data were provided by the investigators. The remaining missing data were either not measured or not investigated. For further details, see Parniczky et al. (13).

In this study 1,224 patients were included. Serum creatinine based eGFR was measured within 24 h following hospitalization. Patients were divided into 3 groups based on their eGFR values derived from their respective serum creatinine values and calculated according to the chronic kidney disease epidemiology (CKD-EPI) formula, by gender:

The correction factor for black race was not necessary in this registry. Creatinine was measured by the modified Jaffe reaction on a Hitachi 911 automatic analyzer (Roche Diagnostics, Mannheim, Germany).

The 3 groups were the following: patients with eGFR <30 mL/min (n = 44), eGFR: 30–90 mL/min (n = 630) and eGFR >90 mL/min (n = 550). A matched control group was selected from patients with eGFR >90 mL/min (n = 39) to match by age (± 5 years) and sex with the eGFR <30 mL/min group. Most evaluations were performed including the matched control group eGFR >90 mL/min (n = 39) and the non-matched patients with eGFR >90 mL/min (n = 511) were not included in the study.

The severity of AP was evaluated based on the revised Atlanta classification (14). Mild AP was characterized by the lack of organ failure, lack of local or systemic complications and if resolved within the first week. Moderate AP was characterized by transient organ failure, local complications or exacerbation of any co-morbid disease. Severe AP was characterized by persistent (>48 h) failure of one or more organs.

Analyzed parameters were defined as follows: Organ failure was defined according to the modified Marshal scoring system (15). New onset diabetes was defined as no diabetes (lack of documented or reported diabetes history, lack of antidiabetic medication and HbA1C in reference range) at the time of hospitalization due to AP, but diabetes developed as a consequence of the AP episode and the patient was released with appropriate antidiabetic medication. Comorbidities, such as peptic ulcer, myocardial infarction and vascular disease were recorded if the patient reported it or the patient's medical documentation included it. Symptoms and complaints (weight loss, loss of appetite, nausea, vomiting) as well as life-style attributes (smoking, alcohol consumption) were recorded based on the patients' reports.

This is the first paper describing the associations between AP and biomarkers of renal function, assessed by creatinine-based eGFR within 24 h of hospitalization, in patients enrolled in our prospective AP registry, a large cohort with high quality data. Renal dysfunction is regarded as a late complication in AP. However, according to our analysis, even early and mild to moderate renal functional impairment has a prognostic value for both severity and outcome.

Better initial renal function was associated with better outcome, such as longer survival and shorter length of hospitalization. In line with this observation, eGFR strongly correlated with severity of AP and blood levels of exocrine pancreatic enzyme levels. Furthermore, eGFR was significantly associated with subfebrility/fever, systemic inflammatory parameters (CRP, WBC count) hyperlipidemia and vascular complications (atherosclerosis, myocardial infarction, heart- and renal-failure). Observational studies, such as the current registry analysis do not enable to establish causative relationships (16). Thus, we can only conclude that survival was better if initial eGFR was >60 mL/min, length of hospitalization increased if eGFR was <90 mL/min, severe pancreatitis was associated with eGFR<30 mL/min and blood glucose was elevated if eGFR was <30 mL/min. Blood amylase and lipase activities were significantly elevated if initial renal function was slightly impaired (eGFR <90 mL/min). Several investigated parameters are either comorbidities or complications, such as peripheral vascular disease and consequent myocardial infarctions, which were more common in low eGFR patients. Although, causative relationships are not proven, preexisting atherosclerosis is a known risk factor for inferior outcome of AP and is associated with low eGFR. However, the observed association raises the possibility that complications like myocardial infarction are more likely to occur if renal function is impaired. The lack of association between eGFR reductions and increases in pancreatic damage markers (amylase, lipase) is best explained by the rapid decline of pancreatic enzymes in the blood after pancreatitis, often resulting in reference range values already at the time of diagnosis.

In our registry, AP severity was defined based on the Atlanta scoring system, including renal failure (eGFR <15 ml/min) and this could influence the association between severity and eGFR. However, renal failure was 1 of 4 complications (Heart-, Respiratory- Renal- and Systemic organ- failure) and further 6 pancreas related complications (Local pancreatic-, Fluid collection, Pseudocyst, Necrosis, Diabetes, and other local complications) were included in the severity score. Furthermore, only 13 patients had an eGFR <15. Thus, the detected associations between the initial eGFR and severity is not likely to be due to the inclusion of renal failure in the definition of severity. Association of renal function and pancreatitis outcome has been reported before: eGFR <90 mL/min was a predictor of pancreatic necrosis (17) and a BUN increase >1.8 mmol/L had a high predictive value as a single parameter for severe AP (18). It has been reported recently that within the first week of the disease, acute kidney injury (AKI) is among the most common causes of mortality in AP, termed as AP-renal syndrome (19). A recent study suggested high serum creatinine to be a marker of pancreatic necrosis (20). Although most severity scores include eGFR, a detailed renal functional evaluation is still not regarded as an important prognostic parameter for AP severity and outcome (9).

In such a large, multicenter cohort demographic distribution may have a significant effect. However, as most centers included in our registry were central-eastern European with only 2 patients from Japan (Supplementary Figure 1) the cohort was homogenously Caucasian. Average age was 65 spanning from 15 to 95 and a bit over the half of the patients were males (Supplementary Figure 3). The leading cause of AP was biliary complications (54.29%). Alcohol induced AP and idiopathic forms (together: 36.49%) were less common than biliary complications. These observations are in line with literature data. In several studies, gallstones were the most frequent etiology of AP followed by alcohol-induced AP (21–23). Liver function (as assessed by serum bilirubin) did not correlate with eGFR, suggesting that impaired liver function did not contribute to the more severe outcome of the patients with eGFR <30 ml/min. As – similarly to the kidney – the liver is affected in AP by hypoxia due to centralization of the circulating blood and by toxicity of circulating exocrine enzymes, we have observed a non-significant association of ALAT and γ-GT with eGFR suggesting that functional impairment of the kidney and liver damage were parallel.

We detected no association between the cause of AP and initial eGFR in this registry. One report suggests that biliary pancreatitis could lead to renal insufficiency (24). Our analysis did not support this observation, probably due to a relatively low number of patients with renal impairment in our cohort.

In our study, severe initial renal dysfunction (eGFR < 30 ml/min) was associated with significant elevation of systemic inflammatory markers, such as CRP and white blood cell count. Similar observations have been reported before describing the role of pancreatic enzyme release due to autodigestion of acinar cells of the pancreas showing that AP can trigger AKI by various stimuli such as inflammation, vasoconstriction and hypoxia (11, 12). Moreover, as amylase is partly excreted by the kidneys serum amylase concentration can be even more increased in patients with initial renal dysfunction (25). However, the importance of amylase plasma concentration is limited for establishing the diagnosis of pancreatitis (26). Erythrocyte sedimentation rate (ESR), another common inflammatory marker was not assessed in the present study, as only a minority of patients had ESR data in our cohort.

Although one would expect an association between the anemia status (Hb, Hct) and reduced eGFR, especially as eGFR was strongly associated with the severity of acute pancreatitis, our data did not support any association between eGFR and Hct. A slight correlation between eGFR and Hb supports that Hb is a more accurate parameter to describe anemia status than Hct. A possible conclusion from the observation that eGFR was not strongly associated with anemia is that AP influenced eGFR and not the other way around, as acute kidney injury is expected to increase plasma erythropoietin concentration. As the patients were hospitalized after the diagnosis of AP, we have no data about their eGFR before the onset of the disease.

Limitation of the study is that we have no creatinine values before the onset of the AP episode. Although renal impairment was probably a consequence of AP in most cases, as peripheral vascular disease was seldom present in the cohort, we were not able to differentiate between preexisting renal disease or acute decline of renal function due to AP. Thus, the causative role of renal impairment could not be investigated. However, registry analysis is not suitable for establishing causative linkages, anyway. Thus, despite this limitation, our presented data are conclusive about the predictive value of initial eGFR irrespective whether the cause of low eGFR is a preexisting renal impairment or is caused by AP.

eGFR is a good predictor of the prognosis of acute pancreatitis and should be regarded as an important prognostic index independent of other complications of acute pancreatitis. As even mild to moderate eGFR reduction predicted survival, severity, comorbidities and outcome, a detailed evaluation of renal function is suggested for improving the prediction of acute pancreatitis prognosis.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

NFar, GS, and PHa: conceptualization. PT, NFar, DN, GS, and PHa: data curation. NFar, DN, GS, ZS, PHe, and PHa: formal analysis. PHe: funding acquisition. ÁV, RH, LC, DI, FI, VD-V, MP, JH, MV, KG, TN, ZM, NFal, IT, AI, SG, PJH, AS, AP, and ZS: investigation. TP, NFar, DN, GS, ZS, PHe, and PHa: methodology. PJH, AS, AP, ZS, and PHe: project administration. PHe: resources. PHa and PHe: supervision. NFar, AS, PHa, and PHe: validation. PT, GS, and PHa: writing–original draft. PHa: writing–review & editing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.