Acute Kidney Injury Complicating Severe Acute Pancreatitis: Clinical Profile and Factors Predicting Mortality

Introduction

Acute pancreatitis (AP) is an inflammatory condition involving the pancreas with severity ranging from mild disease to severe acute pancreatitis (SAP). SAP is an acute clinical syndrome with multiple organ failure and high mortality in the range of 7% to 47%.[1] Acute kidney injury (AKI) is a frequent complication of SAP carrying a dismal prognosis, when kidney replacement therapy is required, with mortality ranging from 25% to 75%.[2] A ten-fold increase in mortality has been shown to be associated with AKI in pancreatitis.[3] Retrospective data by Kumar et al.[1] and Manokaran et al.[4] reported SAP-AKI in 19.4% and 32% with a mortality of 57.1% and 12.5%, respectively. However, SAP-AKI is greatly understudied, as the clinical characteristics and underlying processes of AKI coexisting with SAP stay elusive. Hence, we conducted a prospective study to determine the complications, factors predicting mortality, and renal outcomes among SAP-AKI.

Materials and Methods

We prospectively analyzed (n = 68) patients with SAP-AKI admitted at Madras Medical College between September 2015 and September 2019 in our tertiary care center. The inclusion criteria were as follows: patients with AKI as per Kidney Disease: Improving Global Outcomes (KDIGO)[5] staging presenting with clinical, biochemical, and imaging criteria for acute pancreatitis (ICD-9 [International Classification of Diseases, Ninth Revision]; Code 577). The exclusion criteria were as follows: patients with chronic pancreatitis and known preexisting kidney disease. Demographic, clinical, laboratory, and imaging data were obtained. AP was diagnosed in patients who met two of the following three criteria: (1) abdominal pain (acute onset of a persistent, severe, epigastric pain often radiating to the back), (2) serum lipase/amylase levels at least three times greater than the upper limit of normal, and (3) imaging evidence of AP. SAP was defined as per Revised Atlanta classification criteria-2012 of persistent organ failure >48 hours and local complications (peripancreatic fluid collection, pseudocyst, necrotic collection, and walled-off necrosis).[6] Bedside Index of Severity for Acute Pancreatitis score (BISAPs)[7] defined by blood urea nitrogen >25 mg/dL, impaired mental status, systemic inflammatory response syndrome (SIRS) ≥2/4 present, age >60 years, and pleural effusion – a scoring tool to assess the probability of in-hospital mortality within 24 hours of presentation was applied. Kidney replacement therapy (KRT) was done when indicated. Renal biopsy was done in cases with AKI persisting more than 2 weeks. The cohorts were grouped as survivors and deceased [Figure 1]. The survivors were followed every 3 months, and blood pressure, urine analysis, urine spot protein–creatinine ratio (PCR), and serum creatinine tests were done during the visits. Glomerular filtration rate was estimated using Chronic Kidney Disease – Epidemiologic Collaboration Creatinine (CKD-EPI) equation (2009). The study protocol was approved by the Institutional Ethics Committee, Madras Medical College, registration number- ECR/270/inst./TN/2013/RR-16.

Close

Figure 1

Flow of the Study

Export to PPT

Results

A total of 68 patients with diagnosis of SAP-AKI were included during the study period. This constituted 2.14% (68 of 3,169) of all-cause AKIs. Basic demographic details are shown in Table 1. Oliguria (urine output <500 mL/day) was noted in n = 40 (58.8%). Volume overload in 21 (30.8%; P = 0.031), shock in 10 (14.7%; P = 0.014), and respiratory failure requiring mechanical ventilation in 6 (8.8%; P = 0.001) were associated with significant mortality. Imaging noted in SAP were edematous interstitial pancreatitis in n = 53 (77.9%), necrotizing pancreatitis in 14 (20.6%), emphysematous pancreatitis in 1 (1.4%), and pseudocyst in 7 (10.3%). Necrotizing pancreatitis (39% vs. 11%, P = 0.01) was associated with significant mortality. BISAPs >3 (86.3% vs. 60%, P = 0.038) predicted increased mortality. An ROC (receiver operating characteristics) on BISAPs was performed, with sensitivity of 97.8% and specificity of 52.2% (area under the curve [AUC] = 0.728 [0.587–0.869]; P = 0.002) for predicting mortality in SAP-AKI in our cohort [Figure 2]. The mean time to AKI occurrence from symptom onset is 2.58 ± 3.4 days. KDIGO Stage I in n = 1 (1.5%), KDIGO Stage II in 14 (20.6%), and KDIGO Stage III in 53 (77.9%), with higher KDIGO stage (95.6% vs. 68.8%, odds ratio [OR] = 9.935; 95% confidence interval [CI; 1.215–81.219]; P = 0.03) predicting mortality. Seven patients had clinical evidence of thrombotic microangiopathy (TMA) in the form of hemolytic anemia, raised lactate dehydrogenase, peripheral smear suggesting schistiocytes (>5%), and new onset or worsening hypertension, of which three had biopsy-proven TMA. KRT was done in 40 (59%), among which hemodialysis in 36 (90%) and 4 (10%) required acute intermittent PD, an increased need for KRT (78.2% vs. 48.8%; OR = 3.764; 95% CI [1.191–11.891]; P = 0.024) was seen among the deceased population. The local and systemic complications are given in Table 2. The occurrence of hyperamylasemia (> 350 IU/mL; P = 0.02), serum calcium <8 mg/dL (P = 0.03), hypoalbuminemia (P = 0.05), international normalized ratio (INR) > 1.1 (P = 0.014) at presentation predicted risk of mortality. By logistic regression, the occurrence of necrotizing pancreatitis, circulatory failure, peak creatinine >3 mg/dL, BISAPs >3, need of KRT, and KDIGO stage III independently predicted the risk for mortality [Table 3]. We reported 23 (33.8%) deaths during hospital stay. The deceased patients had a shorter median intensive care unit length of stay (LOS; 5 vs. 10; P < 0.001).

Close

Figure 2

BISAPs (receiver operating characteristics) for predicting mortality. BISAPs = Bedside Index of Severity for Acute Pancreatitis score

Export to PPT

Among the total survivors (n = 45) at discharge, 16 followed up as per protocol with median duration of 3.5 months. The mean time to nadir creatinine was 22.7 ± 9.6 days. Of these, three progressed to chronic kidney disease (CKD), one in CKD Stage 2 and the other two in CKD Stage 4.

Renal biopsy was done in seven patients [Table 4], three had TMA [Pictures 1 and 2], of which one developed thrombotic thrombocytopenic purpura (TTP) with bilateral cortical blindness due to bilateral lateral geniculate bodies (LGB) infarction, requiring plasma exchange with complete recovery. ADAMTS 13 activity was normal. Two had nonrecovery progressing to CKD Stage 2 and Stage 4 and the other in CKD Stage 4 with inconclusive pathology.

Close

Picture 1

Fibrin thrombi seen occluding the capillary lumen (MT 400x, RENOPATH, Chennai)

Export to PPT

Close

Picture 2

Fibrin thrombi seen occluding the arterioles, with fibrous intimal proliferation (PAS 100x. RENOPATH, Chennai)

Export to PPT

Discussion

AKI is a common complication in patients with SAP with incidence ranging from 14% to 42%.[38] In our study, age and gender did not correlate with mortality. Li et al.,[8] Zhou et al.,[9] and Devani et al.[10] reported a positive correlation with age and female gender with mortality. The occurrence of comorbidities in our study was not associated with mortality, which was similar to study done by Zhou et al.[9]; however, Kumar et al.[1] and Manokaran et al.[4] reported alcohol addiction and diabetes mellitus as significant risk factors for death. We reported chronic and binge alcohol intake as the commonest etiology, which was similar to Kumar et al.[1]; however, Company et al.[11] and Li et al.[8] reported biliary disease, followed by alcohol. Gougol et al.[12] reported post-ERCP (endoscopic retrograde cholangiopancreatography) status, followed by alcohol addiction and biliary disease.

AKI is a major complication that occurs in patients with AP due to many factors such as hypovolemia due to vascular permeability and systemic inflammation, septic shock, vasoconstriction, abdominal compartment syndrome, and reduced tissue perfusion, which ultimately causes organ damage and multiorgan failure.[13] Several studies have asserted that renal failure is always one of the main drivers of multiorgan failure in SAP.[9] In our study, higher deaths occurred in AKI with volume overload, circulatory and respiratory failure. Gougol et al.[12] reported organ failure in 22% (111 of 500) presenting with AP, among which 15% (17 of 111) had isolated renal failure. Kumar et al.[1] and Zhou et al.[9] in their study reported shock (9 vs. 0; P < 0.001;OR = 202) and acute respiratory distress syndrome (P = 0.009) with mortality, respectively.

The BISAPs has been developed to identify patients at risk for mortality or severe disease early during the course of AP. In our study, a BISAPs >3 was significantly associated with mortality with a sensitivity of 97% and specificity of 52%, predicting the probability of death. In a meta-analysis by Gao et al.[7] stated that BISAPs >3 had an overall sensitivity of 56%, with a specificity 91% for mortality. Our results established higher amylase >350 IU/mL, severe hypocalcemia (<7 g/dL), hypoalbuminemia (serum albumin <3.5 mg/dL), peak creatinine >3 mg/dL, deranged coagulation at admission with mortality. Manokaran et al.,[4] in their study of 32 patients with SAP-AKI, reported that serum amylase >200 IU/mL and serum creatinine >2.4 mg/dL had four times increased risk of mortality compared to without AKI.

In our study, KDIGO Stage III had the highest mortality similar to Zhou et al.[9] Our results showed KRT in 59%, higher in deceased 78.2%, similar to Zhou et al.[9] with 414 patients wherein KRT was done in 59% (245), with 69% (89) among the deceased. Manokaran et al.[4] reported that 6.25% (2 of 32) required KRT. We reported an overall mortality of 23 (33.8%), similar to Zhou et al.,[9] Company et al.,[11] Kong et al.[14] and Kumar et al.[1] reported mortality of 44.9%, 30.5%, 31%, and 57.1%, respectively, in SAP-AKI. The major cause of death may be due to sepsis and multiorgan failure. TMA was the common finding in our cohort, with seven patients having clinical TMA, of which three underwent kidney biopsy. All of them developed TMA after SAP, pointing to secondary HUS rather than aHUS with pancreatic involvement. In a systematic review, Chaudhry and Sawier[15] showed 27 cases with TMA preceded by AP described in the literature, the mechanism of which is less clearly understood.

Thachil[16] explained that inflammatory release of cytokines like interleukin 6 (IL-6), IL-1, IL-8, and tumor necrosis factor alpha (TNF-α) stimulates the release of ultra-large vWF (von Willebrand Factor) from endothelium leading to deficiency of ADAMTS-13 protease, which gets consumed quickly. Similarly, Swisher et al.[17] hypothesized the presence of circulating pancreatic protease-mediated direct endothelial injury modifying the circulating vWF to ultra-large vWF resistant to proteolytic cleavage or a relative deficiency of ADAMTS13 with subsequent platelet aggregation. Serum ADAMTS13 activity estimation was done in one of our patients with severe renal and neurological involvement who improved with plasma exchange despite having normal ADAMTS13 levels. Chaudhry and Sawier[15] even highlighted the ability of TTP to occur despite normal ADAMTS13 activity and suggested that a more complex model may underpin the pathologic process. Furthermore, the role of complement activation in SAP has been demonstrated in necrotizing pancreatitis models, wherein direct cleavage of C3 and C5 by pancreatic enzyme trypsin with significantly low C3 levels causes microvascular damage.[16] Similarly Ruiz-Torres et al.[18] established the role of complements and neutrophils in microvascular damage from sera from patients with TMA shown to cause C3 and membrane attack complex deposition and surface expression of P selectin on the human endothelial cell line. There is also evidence from animal models that pancreatic endothelial nitric oxide synthase is decreased in AP predisposing to the development of TMA.[19] Three patients progressed to CKD in the follow-up cohort.

There are certain limitations to our study. First, due to lack of exact and complete urine volumes, we relied on serum creatinine obtained during the relevant period and possibly could not include patients fulfilling AKI criteria with urine output. Second, risk factors for AKI in SAP could not be studied as patients having SAP without AKI were not included. Third, a relatively small sample size involving only single-center data cannot be extrapolated to the general population. Fourth, our long-term follow-up data remain insufficient as fewer patients reviewed stating logistic difficulties.

Conclusions

SAP with AKI constituted 2.14% of all-cause AKI. High mortality of 33.8% was seen. We found necrotizing pancreatitis, circulatory failure, peak creatinine >3 mg/dL, BISAPs >3, need for KRT, and KDIGO Stage III as significant risk factors associated with high mortality among SAP presenting with AKI. Three patients progressed to CKD.