Acute pancreatitis (AP) is the most common cause of pancreatic disease in children. Its incidence has increased worldwide in the last two decades and is estimated at 3.6–13.2 cases per 100,000 children1,2. There is evidence that the number of cases has been rising in Mexico in recent years3. The etiology of AP varies at the different hospital centers, but the most common causes are those of biliary, anatomic, and genetic origin; up to 20% of cases are idiopathic3–5. AP is responsible for $200 million USD in hospital expenses, with a median hospitalization cost of $20,000 USD6. Studies conducted on children show different factors associated with days of hospital length of stay (LOS), that include: chronic illnesses (oncologic, gastrointestinal, neurologic, hematologic), parenteral nutrition (PN), Hispanic ethnicity, pseudocyst or pancreatic necrosis, malnutrition, male sex, disease severity, systemic inflammatory response syndrome (SIRS), comorbidities, and fasting time6–8.

The European Society for Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) and the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) recently published guidelines for the treatment of AP in children, in addition to nutritional recommendations that support starting early enteral nutrition (EN), reserving the use of PN for severe cases, and administering antibiotics only in cases of infected necrosis9,10.

The aim of the present study was to identify factors associated with LOS in pediatric patients hospitalized for AP at a tertiary care hospital in Mexico City, Mexico.

Fifty-one events of AP were registered, corresponding to 32 patients that met the selection criteria. Half of the events occurred in females (51%, n = 26/51). The median age at diagnosis was 11.5 years (IQR 7.8–14.3 years). A total of 68.2% of the cases had recurrence. Seventy-eight percent of the events involved abdominal pain, accompanied by vomiting in 49% (n = 25/51) and isolated abdominal pain without vomiting in 29.4% (n = 15/51). A total of 62.7% (n = 32/51) of the events were considered mild, 29.4% (n = 15/51) moderately severe, and 7.8% (n = 4/51) severe. Anatomic/obstructive etiology (pancreas divisum, annular pancreas, accessory pancreatic duct, cholelithiasis, biliary sludge, choledochal cyst) was the most frequent (58.8%, n = 30/51) and the remaining causes were: idiopathic (17.6%, n = 9/51), traumatic (9.8%, n = 5/51), toxic (9.8%, n = 5/51), and metabolic/toxic (3.9%, n = 2/51). Comorbidities were not registered in 27.4% (14/51) of the patients; 14 (27.4%) patients had recurrent pancreatitis; the combination of annular pancreas/recurrence and pancreas divisum/recurrence presented in 6 and 3 patients, respectively. Likewise, 6 (11.7%) patients presented with obesity (4/51 with recurrence and 2/51 as an isolated event). A total of 62.7% (n = 32/51) of the patients presented with no complications (local or systemic), pancreatic cyst or pseudocyst was documented in 19.6% (n = 10/51), 2 (3.9%) patients presented with pancreatic necrosis, one (2%) patient had an isolated peripancreatic collection, and the rest of the events (n = 6, 11.7%) involved combinations of the abovementioned, plus organ failure (kidney, heart, lung) or SIRS. The median hospital LOS was 8 days (IQR 4–14 days), and the mean hospital LOS was 11.76 ± 13.45 days.

Initially, only 21.5% (n = 11/51) of the patients received fluid resuscitation with Ringer’s lactate solution (RL). For pain management, 49% (n = 25/51) of the cases received opioids and non-opioids, together, whereas 29.4% (n = 15/51) received non-opioids and 13.7% (n = 7/51) received opioids. Antibiotics were indicated for and administered to 49% (n = 25/51) of the cases.

Fasting was indicated for all the patients upon hospital admission. Thirty-two (62.7%) episodes were managed with EF (23 patients with mild AP, 8 with moderately severe AP, and only one patient classified as severe AP). Subdividing the 32 admissions that received EF, 21 received it within the first 48 h of admission (17 with mild AP and 4 with moderately severe AP) and 11 between >48 h and <72 h from admission (6 with mild AP, 4 with moderately severe AP, and 1 with severe AP). Clear liquid diet was the most frequently indicated (76.4%, n = 39/51), followed by normal diet (15.6%, n = 8/51), and administration was oral in 92.1% (47/51). The registers of nutrient proportions or the caloric supply administered were not available. PN was indicated in 10 patients (19.6%), corresponding to one mild AP event, 6 moderately severe events, and 3 severe events. The indication for PN was attributed to prolonged fasting, pancreatic-cutaneous fistula, and severe AP, and was combined with oral or enteral feeding.

By subgrouping the study population according to hospital LOS, Table 1 summarizes the demographic and clinical characteristics, in which there were statistically significant differences in relation to age, recurrence (two or more events), grade, etiology, comorbidities, and complications. Likewise, there were significant differences between groups regarding biochemical values at the time of admission, with respect to levels of creatinine, BUN, TB, AST, ALT, albumin, Hb, Hct, platelets, and calcium (Table 2). Table 3 describes the treatment per LOS, in which hospital treatment was heterogeneous for fluid resuscitation, antibiotic use, and time to oral/enteral refeeding.

Table 4 shows the univariate analysis of the variables and their association with PLOS. The variables associated with PLOS underwent a multivariate analysis, in which antibiotic use was significantly associated with a greater risk for PLOS (OR 31.71; 95% CI 2.71–370.65; p = 0.006); likewise, EF (within the first 72 h of admission) was associated with a decrease in the days of hospital LOS (OR 0.05; 95% CI 0.004–0.63; p = 0.02) (Table 5).

The etiology of AP is varied and involves regional changes that depend on the type of population that is treated at each hospital. In our study, there was a high number of events of anatomic/obstructive etiology, at over 50% (n = 30/51), and when itemized, those with an obstructive cause were the most frequent, contrasting with other national and international studies. In a study conducted at a pediatric referral hospital in Mexico, only 20% of the AP etiologies were obstructive and the most frequent cause was idiopathic, at 34%3. Our results also differ importantly from a previous study carried out at our institution, in which more than 50% of the cases of AP were attributable to the use of chemotherapeutic agents4. Nevertheless, it is important to remember that the present study did not consider patients that developed AP during their hospitalization due to other diagnoses, which we believe contributed to our finding a different etiology as most frequent. In addition, there was recurrence in 68.2% of our cases.

A multicenter review that evaluated the factors associated with days of hospital LOS, in 7693 discharges of 5507 unique patients (12 years of age, IQR 8–15 years; female sex 55%, n = 4230), concluded that the presence of chronic illnesses (oncologic, gastrointestinal, neurologic, hematologic), PN, and Hispanic ethnicity were factors independently associated with LOS in AP6. Likewise, reports described in the literature, focused on the development of predictive models of severity, have enumerated factors associated with the increase in days of hospital LOS7,8,13–15. One study found that the presence of pseudocyst or pancreatic necrosis, malnutrition, male sex, and disease severity had an influence on prolonging LOS7. Grover et al.15 identified the presence of two or more SIRS criteria (OR 7.99; p = 0.045) and the requirement of intensive care at admission (OR 12.06; p = 0.027) as independent predictors for longer LOS. In our work, we identified TB (p = 0.01), ALT (p = 0.05), albumin (p = 0.003), platelets (p = 0.05), and LDH (p = 0.03) as factors for PLOS. However, when those variables were evaluated in the multivariate analysis, they were not statistically significant, similar to the results of the abovementioned studies, in which no associated biomarker could be identified. Another factor associated with PLOS was the presence of severe AP (p = 0.05). That association was previously established by the Cincinnati group13, utilizing the same methodology as ours to classify AP severity, and documented a higher median LOS (6.1 days; IQR 2.9–14.5 days) in patients with severe AP, compared with the mild AP subgroup (3.9, IQR 2.6–6.4 days; p = 0.0497).

Taking the above into account, one report described a median LOS of 4 days (IQR 3–7 days)6 and another documented a median of 5 days (IQR 3–10 days)16. Both reports contrasted with our findings that were conditioned by the inappropriate trend in the use of antibiotics in practically half of the patients, in which only 60% were regimens authorized by the infectious diseases service. Even though there is no protocol for starting antibiotics, generally they are frequently utilized (31%–82%), up to 90% are started within the first three days, and up to 3 different regimens can be prescribed17,18. Neither the NASPGHAN nor the American College of Gastroenterology recommends the prophylactic use of antibiotics, even in severe AP or the presence of necrosis (given that the majority of cases are not infected); they are only recommended in cases of infected necrosis9,17, and there were no such cases in our study.

Importantly, only 11 of our patients (21.5%), all corresponding to the PLOS group, received aggressive fluid resuscitation upon hospital admission. Aggressive fluid resuscitation is the cornerstone in the treatment of AP, to prevent potential complications, such as necrosis and organ failure, due to an alteration in the pancreatic microcirculation secondary to events that include hypovolemia, increased capillary permeability, and the formation of microthrombi. Current recommendations state that crystalloids should be started, according to hydration status and hemodynamic status, and with a fluid bolus of 10–20 ml/kg, when there are signs of hemodynamic compromise9. Both saline solution (SS) and RL have been evaluated. One study on adults showed that RL, compared with SS, significantly reduced the incidence of SIRS and the development of post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis, but another study found no differences in relation to mortality and LOS duration19. Likewise, early EN in combination with aggressive fluid resuscitation (fluid >1.5–2 times requirements in the first 24 h) has been shown to reduce the days of hospital LOS and the presence of severe disease, and no association with pulmonary complications or incidence of re-admission20. In that respect, in adults, EN has been shown to be an integral part of treatment and has been associated with a lower incidence of infection, mortality, multiple organ failure, fewer days of hospital LOS, and a decrease in associated complications. Consequently, the intestine plays an important role as an immunologic barrier and EN facilitates that barrier function, preserving the microbiota, preventing bacterial translocation, and lowering the systemic inflammatory response20,21. With respect to nutritional treatment, in cases of mild AP, except in the presence of direct contraindications (ileus, fistula, or abdominal compartment syndrome), the ESPGHAN and NASPGHAN recommend that EN be started, combined with a normal oral diet, as soon as possible; ideally within the first 48–72 h of admission. Specialized formulas (elemental, semi-elemental, polymeric) are not necessary, nor is the use of immunoenhanced nutrients9,10. Starting EN within the first 48 h in pediatric patients has been shown to be a safe and effective strategy that reduces the days of hospital LOS, the incidence of severe AP, and the need for admission to the intensive care unit. Likewise, in patients with severe AP, it should be attempted within the first 72 h, once hemodynamic stability has been achieved9,10. However, in our study, nutrition was begun within that time span in only one patient categorized as having severe AP. In the study by Szabo et al.22 on 201 children, EN was started within the first 48 h of admission in 75% of the patients, resulting in a statistically significant difference in the days of hospital LOS in that subgroup, compared with the children that remained fasting (2.9 vs 4.9 days, p < 0.0001).

Nutritional treatment indication in our institution is different between the hospitalization services, resulting in heterogeneous established managements. Even though fat content has been reported to not increase days of hospital LOS, pain severity, or lipase levels10,23, the tradition of indicating fasting or a low-fat diet persists, under the concept that “resting the gut” prevents stimulating the pancreas, and thus permits faster healing. Evidence justifies the use of EN over PN, for its safety profile, low cost, and lower risk of infection. Likewise, NP is reserved for severe events, for cases in which oral nutrition or EN (nasogastric or nasojejunal) is not tolerated, when it is not possible to begin oral feeding or EN within a 5–7-day period, or administer it in combination with EN, when caloric requirements are not met9,10,24.

There was one death secondary to septic shock at day 8 of hospitalization, corresponding to a female patient, 12 years and 2 months of age, that was transferred after having spent two weeks in a secondary care hospital. Death due to AP is rare, with a mortality rate that varies from 0% to 21%, as a result of multiple organ failure25,26.

Our data reaffirm findings from a recent mini meta-analysis, in which the early start of EN significantly reduced days of hospital LOS, compared with nil per os (SD = 0.806, p = 0.034), and was the only treatment modality showing a clear benefit in improving results in pediatric patients with AP20. Despite the current international guidelines that attempt to provide a consensus on the recommendations for treating AP in children27, there is still a discrepancy among established treatments, especially in nutritional management. With that in mind, it is worth mentioning that our subgroup with PLOS inversely presented with less recurrence and fewer comorbidities and complications, possibly resulting from the heterogeneity in the decision-making for managing AP. Multidisciplinary coordination in the hospitalization treatment of AP would very likely benefit those patients.

The limitations of our case series include its small sample size and retrospective study design.

In conclusion, our study results showed that early nutrition was associated with shorter hospital LOS and that the use of antibiotics was associated with longer hospital stay.

No financial support was received in relation to this study/article.

The authors declare that there is no conflict of interest.