Prediction of Severity of Acute Pancreatitis Using Total Serum Calcium and Albumin-Corrected Calcium: A Prospective Study in Tertiary Center Hospital in Nepal

Introduction Total calcium (TC) and albumin-corrected calcium (ACC) are easily accessible AP severity tests in the Primary Health Care Center of Nepal. The aim of the study was to evaluate TC and ACC as prognostic severity markers in acute pancreatitis (AP). Methods All patients admitted in Tribhuvan University Teaching Hospital with the diagnosis of AP were studied prospectively over a period of one year from January 2015 to January 2016. TC and ACC were measured in the first 24 hours of admission in each patient. The modified Marshall score was determined at admission and at 48 hours and at any point of time during admission as per the need of the patient. Severity of acute pancreatitis was defined as per the Revised Atlanta Classification 2012. Results 80 patients of AP were included in the study. Among them, 14% were categorized as having severe AP. The mean total calcium was 8.22, 7.51, and 6.98 for mild, moderate, and severe AP, respectively, which was significant at 0.001. Conclusion TC and ACC, measured within the first 24 hours, are useful severity predictors in acute pancreatitis.


Introduction
Acute pancreatitis (AP) is a major cause of mortality and morbidity worldwide. In approximately one-third of the patients, acute severe pancreatitis may develop, producing progressive organ dysfunction usually caused by a rapidly progressive in ammatory response, which is associated with a prolonged hospital stay and signi cant morbidity and mortality. Patients with mild AP have mortality rates of less than 1%, but it is rapidly increased up to 10-30% in cases with severe AP [1]. So, early detection of severe pancreatitis is essential for proper care and management and to limit its complications. e 1992 Atlanta guidelines was the rst clinically based and universally applicable classi cation which de ned the severity and the complications of AP [2]. e Atlanta classi cation was revised in the year 2012, which categorizes AP into 3 states: mild, moderate, and severe based on the presence of organ failure or local complications [3]. e organ failure is determined by the modi ed Marshall scoring system [4].
Early detection of severe AP is important so as to deliver proper care to the patient and to avoid its complications. Total calcium and albumin-corrected calcium are simpli ed markers that can be readily measured and can be easily calculated and interpreted by any health persons. It has been evaluated as a mortality prognostic factor and has also been evaluated as a predictor of severe AP with infection. Albumin-corrected calcium has also been associated with severity [5].

Materials and Methods
Patients with a diagnosis of AP admitted in the surgical and medical wards of Tribhuvan University Teaching Hospital over a period of one year (January 2015-January 2016) were studied. Demographic variables including age, sex, and etiological factors were recorded. Diagnosis of acute pancreatitis, its severity, and local and systemic complications were de ned as per the Revised Atlanta Classi cation 2012. e diagnosis of acute pancreatitis was made if it ful lled two of the following three features: (1) upper abdominal pain of acute onset often radiating to the back, (2) serum amylase or lipase activity greater than 3 times normal, and (3) ndings on cross-sectional abdominal imaging consistent with acute pancreatitis [3]. e vitals (blood pressure, pulse, respiratory rate, and temperature) and urine output were recorded four hourly and as necessary from the time of admission until 72 hours.
Routine hematological and biochemical parameters were measured using an autoanalyzer machine available in the hospital laboratory. Serum calcium was calculated using Biotech 3000 Analyzer using cresolphthalein method, and serum albumin was calculated using bromocresol green method.
In order to evaluate TC and ACC as prognostic factors of severity, the lowest TC values were collected within the rst 24 h of hospital admittance. ese values were then corrected based on the serum albumin level, thus obtaining the ACC using the following formula: (ACC � TC + (0.8 * [4 -serum albumin concentration])) [6]. e modi ed Marshall score was determined at admission and at 48 hours and at any point of time during admission as per the need of the patient [3]. Severity of AP as de ned by the Revised Atlanta Classi cation was taken into account [3]. Ethical clearance was obtained from the Institutional Review Board. Informed consent was obtained from each patient.

Collection of Data.
Data were collected in pro forma from all the admitted patients with the diagnosis of acute pancreatitis. Biliary etiology was con rmed as the cause of AP in patients with gall bladder stone or bile duct stone on imaging. A combination of age, sex, and laboratory markers was used in case of di culty to predict a biliary etiology. Alcoholic etiology was de ned on the basis of a history of chronic alcohol intake or recent alcohol intake in the week prior to admission while AP of other etiologies was excluded. When a diagnosis could not be made through a history, physical examination, laboratory studies, and imaging modalities, those cases were designated as idiopathic pancreatitis.

Statistical Analysis.
One-way ANOVA and the chisquare test were employed to establish the statistical signi cance of the di erences between groups, based on the characteristics of the analyzed variables. Statistical signicance was determined with p < 0.05. ROC curves were used in order to establish the possible cuto values for TC and ACC. e maximum cuto value was utilized to calculate the sensitivity (S), speci city (Sp), positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), and negative likelihood ratio (NLR) of each criterion by means of contingency tables. All the statistical analyses were done using Statistical Package for the Social Sciences Programme v.21 (IBM SPSS).

Results
Within the study period, a total of eighty patients were included in the study. e age of the patients ranged from 19 to 87 years. AP was common in 41-50 years of age group with a mean age of 47.82 ± 15.91 years. ere was no signi cant di erence in the age of patients in each severity grade (p value: 0.242). Forty-six patients (58%) who presented with acute pancreatitis were female, whereas thirtyfour (42%) were male. 55 (69%) of the patients belonged to the mild group, 14 (17.0%) in moderately severe group, and 11 (14.0%) in severe group. 47 (59%) patients were admitted with biliary etiology and 33 (41%) with nonbiliary etiology. Among nonbiliary patients, alcohol-induced pancreatitis was the most common (n � 20). In 10% of the cases, etiology could not be identi ed (n � 8). ere was no signi cant association between etiology and severity of the disease (p value > 0.05).
e mean duration of stay in severe pancreatitis was 14.45 ± 4.39 days. e mean total calcium was 8.22, 7.51, and 6.98 for mild, moderate, and severe AP, respectively, which was signi cant at 0.001. As compared to TC, mean values of ACC were 8.15, 7.41, and 7.01 for mild, moderate, and severe AP, respectively, which were also signi cant at 0.002 (Table 1).
A receiver-operating characteristic (ROC) analysis of the total calcium was analyzed for severe acute pancreatitis, AUC of which came out to be 0.787, which was signi cant (p value: 0.001) ( Figure 1, Table 2). Similarly, ROC curve of albumin-corrected calcium analyzed for severe acute pancreatitis showed AUC of 0.781, which was also signi cant (p value: 0.002) (Figure 1, Table 2). e ROC curves of TC and ACC were comparable.
When the coordinates on the curve were analyzed ( Table 3), TC of 8.20 mg/dl was computed as cuto for severe AP with sensitivity of 96%, speci city of 54.5%, PPV of 49%, and NPV of 96.8% (Table 3). Also, when the coordinates on the curve for ACC were analyzed (Table 3), ACC of 7.72 mg/dl was computed as cuto for severe AP with sensitivity of 88%, speci city of 69.1%, PPV of 56.4%, and NPV of 92.7% (Table 3).

Discussion
Acute pancreatitis is a common surgical emergency with mortality of severe attacks, reaching up to 30%-50% [7]. is subgroup of patients needs to be identi ed early in the course of the disease and needs to be aggressively treated to prevent mortality. Proper identi cation of the mild disease is also necessary to avoid unnecessary over treatment, thereby reducing the nancial implications.
Severity assessment in acute pancreatitis was rst started in 1974 by Ranson et al. [8].
e Ranson, Glasgow, and APACHE II score are few of the commonly used scoring systems [9]. Limitations of these scoring systems include delay in complete scoring where it takes 48 hours to complete Ranson and Glasgow scoring systems to complete the assessment, while APACHE II score is very cumbersome to calculate [9]. Hypocalcemia is one of the components of Ranson's scoring system done to assess the severity of pancreatitis. Ammori et al. reported that hypocalcemia was more frequent during severe attack as compared to mild attack of pancreatitis (86% versus 39%, p < 0.001) [10]. Prevalence of hypocalcemia ranges between 15% and 88% in 2 Surgery Research and Practice critically ill patients depending on the setting and cuto s used [11,12]. Proposed mechanisms for hypocalcemia in early phase are autodigestion of mesenteric fat by pancreatic enzymes and release of free fatty acids, which form calcium salts, transient hypoparathyroidism, and hypomagnesemia [13][14][15]. Later stages of pancreatitis are frequently complicated by sepsis. Whitted et al. proposed that increased circulating catecholamines in sepsis cause a shift of circulating calcium into the intracellular compartment, leading to relative hypocalcemia. is causes increased PTH secretion by negative feedback loop, leading to further increase in intracellular calcium overload, oxidative stress, and cell death [16]. Hypomagnesaemia-induced impaired PTH secretion and action, relative PTH de ciency, and vitamin D de ciency are some of the other plausible causes. e largest multicenter study conducted in four hospitals of Australia on a cohort of 7024 patients showed that iCa < 0.8 mmol/L was an independent predictor of mortality in intensive care unit (ICU) patients [17]. In a study by Chhabra et al. [18], patients with hypocalcemia in acute pancreatitis were found to have a signi cantly higher frequency of persistent organ failure and need for intervention as well as mortality compared with patients with normal serum calcium levels. e low corrected serum calcium levels had a sensitivity of 81.3% and a speci city of 87.6%, whereas lower ionized serum calcium levels had a sensitivity of 81.3% and speci city of 77.5% for prediction of mortality [18].
In our study, severity of AP was not related to the age of the patient. ere was no di erence between sex and the severity of AP in our study. Similarly, there was no association between etiology and severity of AP. In a review article by Meher et al. [19] the total calcium and albumincorrected calcium were considered as emerging potential biomarkers for prediction of severity in AP.
Our study showed decreasing TC and ACC for increasing severity of the disease. Although this study failed to di erentiate between moderate and mild AP, cuto of 8.20 mg/dL predicted the occurrence of severe AP with sensitivity of 96%, speci city of 54.50%, positive predictive value of 49%, and negative predictive value of 96.8%. When 7.5 mg/dl is taken as a cuto value for the total serum calcium in our study, the sensitivity and negative predictive values decreased up to 68% and 84%, respectively, and while comparing the sensitivity, speci city, and predictive values of TC with that of the study done by Gutierrez-Jimenez et al. [5], our sensitivity was much higher (96% versus 67%) and a positive predictive value (49% versus 27%) was obtained when the cuto value was taken as 8.20 mg/dl. e higher sensitivity and positive predictive value could be due to the higher cuto value in our study. e higher cuto value for TC in our study could be due to a higher normal range of calcium in our lab which is in the range of 8.4-10.4 mg/dl. But the exact reason could not be found.
Similarly, when ACC of 7.72 mg/dl was computed as cuto for severe AP, we found sensitivity of 88%, speci city     [5], our sensitivity was much higher (88% versus 67%) and a positive predictive value (56.4% versus 40%) was obtained when the cuto value was taken as 7.72 mg/dl. As explained above, the higher sensitivity and PPV could be due to the higher cuto value in our study. e higher cuto value for ACC in our study could be due to a higher normal range of calcium in our lab which is in the range of 8.4-10.4 mg/dl. When we compare the sensitivity, speci city, and predictive values of TC and ACC, TC seems to better predict the severity of acute pancreatitis as the value of ACC varied with various other parameters including the nutritional status and chronic liver disease, and it also takes a little time for albumin to get depleted in diseases.
A large number of health care centers have access to TC and ACC use, but not as many have access to the resources required for using the APACHE-II scale and other AP severity markers, such as C-reactive protein, interleukin 6, or procalcitonin. Serum calcium and albumin for calculating ACC are simple biochemical markers that are routinely determined in the majority of hospital centers. eir use as prognostic factors of severity in AP would be valuable for identifying those persons who require intensive care, even at the primary and secondary care center levels.

Conclusion
Serum calcium and albumin-corrected calcium obtained within the rst 24 hours of hospital admission are useful predictors of severity in acute pancreatitis.
is will not replace the currently accepted scoring systems, but these are simpli ed markers that can be readily measured and can be easily calculated and interpreted by any health sta . With an adequate interpretation of their cuto points, they can be used routinely in every case of acute pancreatitis to assess its severity, predict complications, and identify the patients who require intensive care support even in primary and secondary care centers. Disclosure e manuscript was previously presented in the XI National Conference of the Society of Surgeons of Nepal.

Conflicts of Interest
e authors declare that they have no con icts of interest.