Background

In early phase of acute pancreatitis (AP), systemic disturbances are the result of the organism’s response to local pancreatic damage. The first 24 hours after clinical manifestation are essential for identifying the patients with risk of complications, multiple organ failure, severe course of disease and mortality [1]. To date, there are no markers ensuring disease course prediction in early phase of AP with a high degree of probability [2, 3]. The AP-specific scoring systems are cumbersome, characterized by limited value, give little additional information and can delay necessary treatment [3, 4]. AP activates the cytokine cascade with clinical and laboratory manifestations of systemic inflammatory response syndrome (SIRS) [5]. SIRS is followed by metabolism restructuring with resting energy expenditure (REE) increase by 20—80% in the first days of disease [6]. Currently, REE analysis is routine in intensive care units. This examination ceases to be some kind of exclusive option requiring special equipment. Indirect calorimetry is widely implemented in modern bedside monitors and ventilators [7]. We found no studies devoted to predictive value of RER in early phase of acute pancreatitis that determined the relevance of our study.

The objective was to assess the predictive value of REE in early phase of AP.

Material and methods

A prospective single-center study was performed at the "Neftyanik" hospital in Tyumen for the period from November 2012 to October 2017. Inclusion criterion was acute pancreatitis with at least one predictor of severe course of disease. Exclusion criteria: age over 80 years, terminal stage of chronic diseases. AP was diagnosed considering typical clinical signs confirmed by laboratory and instrumental data [8]. Serum C-reactive protein (CRP)> 150 mg / L within 48 hours after admission, Acute Physiology And Chronic Health Evaluation (APACHE) II score> 8 and Sepsis-related Organ Failure (SOFA) score > 2 were chosen as predictors of severe acute pancreatitis [2]. REE was measured in 1, 3 and 5 days after admission to the intensive care unit (ICU) using indirect calorimetry (Engstrom Carestation ventilator, metabolic monitoring module manufactured by General Electric, USA). In non-intubated patients (85.7%), indirect calorimetry was performed using a face mask in spontaneous breathing mode (leakage <10%, 30—50 min twice a day). Mortality, form of disease and ventilation over 12 hours were selected as dependent variables to assess prognostic value of RER. In accordance with this approach, the following groups were formed: all patients (n = 56), people with severe AP (n = 23) [8], moderate AP (n = 33), survivors (n = 49), non-survivors (n = 7), ventilation over 12 hours not associated with anesthetic management (n = 24), ventilation less than 12 hours throughout the entire treatment (n = 32). Statistical analysis was carried out using the SPSS software package. Distribution normality was tested using Shapiro-Wilk test. Data are presented as mean with standard deviation (M ± σ) or median with quartiles (Me (Q25; Q75)). Predictive value of REE was analyzed using logistic regression. Thresholds of REE were assessed using ROC analysis. The null hypothesis was rejected at p-value <0.05.

Results

Clinical and laboratory characteristics of patients with AP upon admission to the intensive care unit are summarized in Table 1.

Note: a — serum C-reactive protein on the first day of admission, b — serum C-reactive protein later 48 hours after admission, c — Student's t-test, d — Mann-Whitney U-test.

REE changes in early phase of AP are shown in Table 2. The same values indexed to the body weight are summarized in Table 3.

Note: * — Student's t-test, ** — Mann-Whitney U-test.

Note: * — Student's t-test, ** — Mann-Whitney U-test.

Discussion

According to Table 1, one can conclude that multiple organ failure was significantly more severe in patients with severe AP upon admission to ICU compared to patients with moderate pancreatitis. Non-survivors had significantly higher APACHE II score compared to survivors. Clinical and laboratory parameters were similar in both groups with ventilation over and less than 12 hours (Table 1). It is known that peak values, time of reaching the peak and total duration of metabolic changes following critical state depend on severity and characteristics of the underlying disease [9]. In all groups, maximum REE (including REE indexed to the body weight) was observed on the first day after admission to ICU (Table 2, 3). Our data do not contradict the existing studies. Indeed various authors reported maximum REE within the first days of disease and subsequent gradual decrease over weeks and even months [10]. According to Table 2, significant differences in REE were found only between the groups of severe and moderate forms of disease throughout the follow-up period. These results are similar to the literature data (REE is higher in patients with AP followed by sepsis compared to patients without this complication) [11]. REE indexed to the body weight was similar in all groups (Table 3). Thus, we can conclude that REE changes in early phase of AP do not depend on the patient's weight.

To date, there is no “ideal” predictor capable to predict the course of disease in early phase of AP [2]. Considering significant differences in REE between patients with severe and moderate pancreatitis, we performed a linear regression analysis. The model comprising all REE measurements throughout the entire early phase of AP was significant [OR 1.002, 95% CI 1.001—1.003; p = 0.001] with a correct assessment of results in 67.3% of patients. According to ROC analysis, REE had moderate predictive quality (AUC 0.671, 95% CI 0.586—0.757, p <0.001). The cut-off value was defined as 2030 kcal/day, sensitivity — 72.5%, specificity — 48.5%. To identify the most informative time for REE measurements in early phase of pancreatitis, we performed a linear regression analysis of independent variables (REE in the first, third and fifth days) using backward Wald stepwise method. According to these results, it was the first day of admission to ICU. The model is significant [OR 1.002, 95% CI 1.001—1.003; p = 0.001] with an accuracy of 73.2%. Assessment of a model’s ability to separate positive from negative values required ROC analysis (AUC 0.745, 95% CI 0.608—0.882, p <0.001). The cut-off value of REE was defined as 2495 kcal/day, sensitivity was 73.9%, specificity — 69.7%. Thus, REE within 24 hours after admission may be used as an independent predictor of severe AP along with other predictors and systems (BISAP, HAPS and others) [3, 12].

Conclusion

REE changes in early phase of AP do not depend on the patient's weight. The indicator is an independent predictor with moderate predictive quality for severe pancreatitis throughout the entire initial phase of disease. Moreover, REE has a good predictive quality within the first day of admission to ICU.

The authors declare no conflicts of interest.