Rates of Stress Ulcer Prophylaxis Deprescribing in Children Receiving Usual versus High-Dose Corticosteroids in the Pediatric Intensive Care Unit with Status Asthmaticus

Abstract

Purpose: To compare deprescribing rates of stress ulcer prophylaxis (SUP) between children receiving “usual-dose” (<4 mg/kg/day methylprednisolone equivalents) versus “high-dose” (≥4 mg/kg/day methylprednisolone equivalents) corticosteroids for status asthmaticus in the pediatric intensive care unit (PICU). Methods: This retrospective, cohort study included children <18 years of age receiving corticosteroids for status asthmaticus and SUP from 1/1/2017 to 6/31/2022. The primary objective was to compare the number of children that were deprescribed SUP following transition from the PICU to the floor and at hospital discharge between groups. Secondary objectives included a comparison of SUP-associated adverse events (ADEs) (pneumonia, Clostridium difficile colitis, thrombocytopenia, necrotizing enterocolitis) between groups. Comparisons were performed using exact χ² test or Wilcoxon U-tests as appropriate, with a P value <.05. Results: Ninety-six patients received usual-dose and 57 received high-dose corticosteroids. Eighteen (11.8%) patients were transferred within 24 hours of PICU admission and started on SUP on the floor. Thirteen (8.5%) patients were discharged home from the PICU. The remaining 122 (79.7%) patients were transferred from PICU to the floor and there was no statistical difference for continuation of SUP on the floor between usual-dose versus high-dose group, 58 (76.3%) versus 31 (67.4%) patients, P = .282. Overall, 25 of 153 (16.3%) patients were discharged home on SUP, but there was no difference between groups. SUP-associated ADEs did not differ between groups. Conclusions: SUP continuation during transitions of care in this cohort was common. Assessment of SUP continuation is needed during transitions of care to promote SUP stewardship and limit risk of SUP-associated ADEs.

Introduction

In severe asthma, corticosteroids are often initiated as first-line therapy due to the associated inflammatory process. In the pediatric intensive care unit (PICU), systemic corticosteroids are often given at higher doses to help manage status asthmaticus. ¹ Despite corticosteroids being commonly prescribed for status asthmaticus, there are no guidelines on the starting or maximum corticosteroid doses for critically ill children. The National Asthma Education and Prevention Program (NAEPP) Asthma Management Guidelines currently recommend 1 to 2 mg/kg/day of prednisone equivalents (0.8-1.6 mg/kg/day of methylprednisolone equivalents) divided twice daily to treat non-critically ill children hospitalized with asthma exaberations. ² Giuliano et al ³ surveyed 104 intensivists on their prescribing practices of steroids in status asthmaticus. They concluded that 63% of intensivists use a starting dose of 4 mg/kg/day of intravenous (IV) methylprednisolone, which reflects the standard of care that many children’s hospitals use for children with status asthmaticus in the PICU.

Historically, exposure to >250 mg/day of hydrocortisone or 50 mg/day of methylprednisolone has been associated with an increased risk of stress-related ulceration and gastrointestinal (GI) bleeding in critically ill adults. ⁴ Due to the association of corticosteroids with development of stress ulcers and bleeding, many children with asthma exacerbations and status asthmaticus will subsequently be prescribed gastric acid suppression medications for stress ulcer prophylaxis (SUP).^5,6 Several SUP-associated adverse events (ADEs) have been reported in adults and children including pneumonia, nosocomial infections, Clostridioides difficile colitis, necrotizing enterocolitis, and thrombocytopenia.^6-8

Once initiated on SUP, it is often up to provider discretion to determine the optimal timing of when to discontinue SUP. Several studies in adults and children have explored the continued use of SUP beyond the time-period of greatest risk for GI bleeding or during transitions of care. Studies in critically ill adults have noted a significant number of patients continued on SUP following ICU transfer and subsequent discharge home on these agents without continued risk factors for GI bleeding.^9-13 Two previous studies assessed outcomes related to SUP deprescribing in children admitted to the PICU with status asthmaticus, but they provided limited details regarding the corticosteroid dosing utilized.^5,6 It is possible that patients who were initiated on a higher corticosteroid dose may be more likely to receive a longer course of SUP and may continue these agents at PICU transfer and discharge home due to concerns with rapid tapering or discontinuation. The purpose of this study was to compare deprescribing rates of SUP and the incidence of SUP-associated ADEs between children receiving “usual-dose” (defined as <4 mg/kg/day of methylprednisolone equivalents) versus “high-dose” (defined as ≥4 mg/kg/day of methylprednisolone equivalents) corticosteroids for status asthmaticus in the PICU.

Methods

This was an institutional review board (IRB)-approved retrospective cohort study of patients <18 years old admitted to a 34-bed PICU at a tertiary care academic medical center from January 1, 2017 through June 31, 2022, with a diagnosis of status asthmaticus who received SUP. Patients were identified using ICD-9 or ICD-10 codes for diagnosis of “status asthmaticus” in the Virtual Pediatric Systems, and the hospital electronic medical record (EMR), Meditech® (Medical Information Technology, Inc. Westwood, MA). Patients were excluded if they had a history of GI bleeding, received extracorporeal membrane oxygenation (ECMO), received anticoagulation therapy, received corticosteroids for non-asthma-related indications, did not receive corticosteroids at all during their admission, or received SUP as a home medication prior to hospital admission. Only the first admission during the study timeframe was included for analysis. Patients were divided into two groups: usual-dose corticosteroids group (defined as <4 mg/kg/day of methylprednisolone equivalents) versus high-dose corticosteroids group (defined as ≥4 mg/kg/day of methylprednisolone equivalents).

Data Collection and Study Objectives

Demographics including age, gender, and weight were collected for all patients. Previously established risk factors for GI bleeding were collected including the Pediatric Risk of Mortality (III) scores >10, respiratory failure (defined as mechanical ventilation duration >48 hours), and presence of coagulopathy (defined as a platelet count of <50 000 mm ³ , an International Normalized Ratio of >1.5, or a partial thromboplastin time of >2 times the control value). ⁴ The presence of gastritis or clinically important GI bleeding was collected. Gastritis was identified by utilizing patient documentation within the EMR. Clinically important bleeding was identified based on the definition used by Sochet et al ⁵ defined as an acute drop in hemoglobin (<7 mg/dL), concurrent acute hypotension requiring >20 mL/kg of intravenous volume resuscitation, or hemorrhagic shock requiring packed red blood cells. Data collection for corticosteroids included the dose, route, frequency, and duration. Corticosteroid total daily and cumulative doses were reported as total methylprednisolone dosing equivalents in mg/kg/day and mg/kg respectively using a steroid conversion calculator. ¹⁴ The SUP prophylaxis regimen was collected including the dose, route, frequency, and duration. In addition, several data collection points were collected pertaining to diet including the day of hospitalization that the diet was changed to clear liquids and the day of hospitalization a regular diet was initiated. The presence of SUP-associated ADEs was collected including Clostridioides difficile colitis, thrombocytopenia, Stage 2 or 3 necrotizing enterocolitis, and hospital-acquired pneumonia (defined as a pneumonia diagnosis documented >48 hours after admission to the PICU).

The primary objective was to compare the number of children that were deprescribed SUP following transition from the PICU to the floor and at hospital discharge in the usual-dose versus high-dose groups. Secondary objectives included a comparison of the percent of hospitalization on SUP, the number of patients that continued SUP following initiation of enteral feeds, and the days continued on SUP after initiation of full enteral intake. Additional secondary objectives included the frequency of SUP-related ADEs (pneumonia, Clostridium difficile colitis, thrombocytopenia, necrotizing enterocolitis) and a comparison of the rates of gastritis and clinically important GI bleeding between groups.

Statistical Analysis

Descriptive statistics were computed to summarize patient demographics and clinical characteristics. Unadjusted comparisons were performed using asymptotic or exact χ ² test or Wilcoxon U-tests as appropriate. Analyses were made using SAS v 9.4 (SAS Institute Inc, Cary, NC), with a P value <.05.

Results

A total of 302 patients with status asthmaticus were identified during the study period. There were 149 patients excluded for the following reasons: ECMO (n = 1), concurrent anticoagulation (n = 18), SUP agent as home medication (n = 58), did not receive corticosteroids during admission (n = 26), and patient with >1 admission (n = 26). An additional 20 patients were excluded because they did not receive SUP while on corticosteroids during their admission. One hundred fifty-three patients were analyzed. A total of 96 patients were included in the usual-dose group and 57 patients were included in the high-dose group.

Baseline demographics of the 153 patients are presented in Table 1. The overall median age at PICU admission was 8.5 years (interquartile range [IQR] 5.3-12.3), but there was no difference between groups. There were no statistical differences between groups regarding gender or race/ethnicity. The patients in the usual-dose group had a higher median weight compared to the high-dose group, 34.6 kg versus 27.0 kg, P = .005. There were statistical differences in the median PICU and hospital length of stay (days) between the usual-dose and high-dose groups, 1.1 versus 1.9, P = .003 and 3 versus 4, P = .044, respectfully. Table 1 also provides comparisons of the established risk factors for GI bleeding in critically ill children between groups. The only significant difference noted was the difference between PRISM III scores; there was a significantly lower number of patients in the usual-dose versus high-dose groups with a PRISM III score >10, 2 (1.9%) versus 7 (10.8%), P = 0.014.

Table 1.

Comparison of Demographic, Clinical Characteristics, and Established Risk Factors ⁴ of Gastrointestinal Bleeding Between Groups.

Variables	Usual-dose corticosteroids (n = 96)	High-dose corticosteroids (n = 57)	P-value
	Number (%) or Median (IQR)
Age (years)	9.1 (5.8-12.4)	7.6 (3.9-12.2)	.115
Weight (kg)	34.6 (19.8-59.7)	27 (16.4-42.8)	.005
Female gender	49 (45.4)	30 (46.2)	.849
Race/ethnicity			.419
African American	44 (40.7)	19 (29.2)
White/Caucasian	33 (30.6)	27 (41.5)
American Indian/Alaska Native	6 (5.6)	3 (4.6)
Hispanic	1 (0.9)	0 (0)
Asian American	3 (2.8)	0 (0)
Mixed	4 (3.7)	4 (6.2)
Unknown/unspecified	5 (4.6)	4 (6.2)
Length of stay (d)
PICU	1.1 (0.7-1.9)	1.9 (0.9-3.5)	.003
Hospital	3 (2.1-4.8)	4 (2.8-5.4)	.044
Risk factors for gastrointestinal bleeding 4
Mechanical ventilation >48 h	2 (1.9)	6 (9.2)	.053
Coagulopathy a	0 (0)	2 (3.1)	.137
PRISM III score ≥10	2 (1.9)	7 (10.8)	.014

Note. PICU = Pediatric intensive care unit; PRISM III = Pediatric Risk of Mortality.

^aDefined as a platelet count of <50 000 mm³, an International Normalized Ratio of >1.5, or a partial thromboplastin time of >2 times the control value.

Table 2 provides a comparison of the corticosteroid regimens between groups. The majority (n = 105; 68.6%) of patients in both groups received IV methylprednisolone as the initial corticosteroid with a frequency of every 6 hours. The patients in the usual-dose group also had a significantly lower cumulative mg/kg dose of methylprednisolone equivalents compared to the high-dose group, 6.3 mg/kg versus 11.8 mg/kg, <.001. There was also a significant difference in the median duration in the usual-dose versus high-dose group, 2.6 days versus 3.3 days, P = .030.

Table 2.

Comparison of Corticosteroid Regimens Between Groups (n = 153).

Variables	Usual-dose corticosteroids (n = 96)	High-dose corticosteroids (n = 57)	P-value
	Number (%) or Median (IQR)
Agents
Methylprednisolone	67 (69.8)	38 (66.7)	.687
Prednisone	1 (1)	0 (0)	1.00
Dexamethasone	37 (38.5)	24 (42.1)	.663
Prednisolone	2 (2.1)	3 (5.3)	.362
Hydrocortisone	1 (1)	0 (0)	1.00
Methylprednisolone IV frequency	n = 67	n = 38	.124
One-time dose	16 (16.7)	9 (15.8)
Every 6 h	24 (25)	22 (38.6)
Every 8 h	9 (9.4)	1 (1.8)
Every 12 h	13 (13.5)	5 (8.8)
Every 24 h	5 (5.2)	1 (1.8)
Dosing regimen
Duration (d)	2.6 (1.5-4.7)	3.3 (2.4-5.0)	.030
Cumulative dose in methylprednisolone equivalents (mg/kg)	6.3 (4.3-9.2)	11.8 (8.2-15)	<.001

Transitions of Care with Stress Ulcer Prophylaxis

The majority of patients (n = 135; 88.2%) were initiated on SUP while in the PICU. In the remaining 18 patients, SUP was initiated upon transfer out of the PICU. There was no difference in number of patients in the usual-dose versus high-dose groups who were initiated on SUP on the floor, 12 (12.5%) versus 6 (10.5%), P = .800. Eighty-nine (65.9%) of the 135 patients initiated on SUP in the PICU were transferred to the floor on SUP prophylaxis, but there was no difference between groups. Thirteen (9.6%) of the 135 patients initiated on SUP in the PICU were discharged directly from the PICU and did not transfer to the floor (Table 3). Four (3.0%) of the 135 patients initiated on SUP in the PICU were discharged directly to home on SUP. Overall, 25 (16.3%) patients were discharged home on SUP, but there was no difference between groups on the number of patients discharged home on SUP.

Table 3.

Comparison of Stress Ulcer Prophylaxis in Usual Versus High-Dose Corticosteroid Groups (n = 153).

Variables	Usual-dose corticosteroids (n = 96)	High-dose corticosteroids (n = 57)	P-value
	Number (%) or Median (IQR)
SUP agents			.648
Famotidine	67 (69.8)	41 (71.9)
Ranitidine a	14 (14.6)	10 (17.5)
Pantoprazole	13 (13.5)	6 (10.5)
Omeprazole	2 (2.1)	0 (0)
SUP and transitions of care
SUP continued at transfer from PICU to the floor b	58 (76.3)	31 (67.4)	.282
SUP continued at hospital discharge	17 (17.7)	8 (14.0)	.552
Percentage of hospitalization on SUP (%)	68.5 (42.7-84.5)	63.5 (40.3-80.1)	.609
SUP and diet information
Continued on SUP following initiation of full enteral intake	13 (13.5)	16 (28.1)	.027
Days SUP continued after initiation of full enteral intake	3 (2-4)	3 (2-4.3)	.697

Note. PICU = pediatric intensive care unit; SUP = stress ulcer prophylaxis

^aPatients received ranitidine during the study period before it was taken off of the market.

^bOne hundred twenty-two patients were initiated on SUP in the PICU.

Table 3 provides an overview of SUP regimens; overall, famotidine (n = 108; 70.6%) was the most frequently prescribed. Several patients were transitioned to a different SUP during their stay, but there was no significant difference in the SUP agents that patients received between groups, P = .648. Between the two groups, the median overall percentage of days of SUP exposure during their hospitalization was similar between the usual-dose versus high-dose groups, 68.5% versus 63.5%, P = .609. Data was also collected on the overlap of enteral feeds with SUP. Twenty-nine (19.0%) patients continued to receive SUP despite initiation of a full enteral diet. There was a significantly lower number of patients in the usual versus high-dose group that had their SUP continued despite reaching full enteral feeds, 13 (13.5%) versus 16 (28.1%), P = .027. However, there was no difference in the number of days of SUP after initiation of full enteral intake between groups.

Table 4 provides comparisons of gastritis, clinically important GI bleeding, and SUP-associated ADEs. Only one patient (0.7%) in either group had gastritis and clinically important GI bleeding. Eleven (11.5%) patients experienced a SUP-associated adverse event in the usual-dose group compared to ten (17.5%) patients in the high-dose group. No patients had >1 SUP-associated ADEs. However, there were no significant differences in the number of SUP-associated ADEs between groups.

Table 4.

Comparison of Gastritis, Clinically Important Gastrointestinal Bleeding, and Stress Ulcer Prophylaxis-Associated ADEs Between Groups (n = 153).

Variables	Usual-dose corticosteroids (n = 96)	High-dose corticosteroids (n = 57)	P-value
	Number (%) or Median (IQR)
Gastritis	0 (0)	1 (1.5)	.373
Clinically important gastrointestinal bleeding a	0 (0)	1 (1.5)	.373
Stress-ulcer prophylaxis-associated ADEs:
Hospital-acquired pneumonia b	10 (9.3)	9 (13.8)	.323
Clostridioides difficile colitis	0 (0)	1 (1.5)	.373
Thrombocytopenia c	1 (1.9)	0 (0)	1.00
Necrotizing enterocolitis	0 (0)	0 (0)	—

^aDefined as an acute drop in hemoglobin (<7 mg/dL) or hemorrhagic shock requiring packed red blood cells.

^bDefined as a pneumonia diagnosis documented >48 hours after admission to the PICU.

^cDefined as a platelet count <50 000 cells/mL.

Discussion

To our knowledge, this is one of the first studies in children to explore the deprescribing rates of SUP during transitions from the PICU to the floor and at hospital discharge. We found that most (73%) patients initiated on SUP in the PICU continued this upon transfer to the floor and overall 16.3% were discharged home on SUP. However, despite the receipt of a higher dose of corticosteroids, there was no difference in the deprescribing rates between the usual-dose versus high-dose groups. Another study by Sochet et al ⁵ evaluated GI bleeding outcomes of 217 children with status asthmaticus receiving corticosteroids. They noted 92 (n = 42.4%) of patients within their cohort that received SUP; of these; 28 (30.4%) were discharged home on SUP. The authors did not provide the number of patients who were deprescribed on SUP when transitioned from the PICU to the floor. Previous studies evaluating SUP in critically ill adults noted 15.4% to 63% of patients continued SUP when transferred to the floor, and 17% to 68.8% were discharged home on SUP.^9-13 Our study’s findings seem to fall within these results regardless of the initial dose of corticosteroid selected.

There are several risk factors for GI bleeds in critically ill children including respiratory failure, coagulopathy, and PRISM III scores >10. ⁴ In the present study, we collected data on these risk factors and found a relatively low incidence of these risk factors in both groups. However, all patients received corticosteroids, and higher doses of corticosteroids >50 mg of methylprednisolone per day are considered a potential risk factor in critically ill adults. ⁴ There are a paucity of studies that have explored the rates of GI bleeding in critically ill children receiving corticosteroids. The recommended dose of methylprednisolone per the NAEPP guidelines for asthma exacerbations is 0.8 to 1.6 mg/kg/day, but many clinicians may elect to initiate 4 mg/kg/day of methylprednisolone for children with status asthmaticus admitted to the PICU.^2,3 Children weighing >12.5 kg would easily exceed this established 50 mg/day of methylprednisolone. A few studies have evaluated the impact of GI bleeding in patients with status asthmaticus receiving corticosteroids in the PICU.^5,6 In these studies 30 394 patients were evaluated and approximately one-third of patients received SUP. They found no episodes of GI bleeding and 32 patients (0.1%) had gastritis. One significant limitation of these studies is that they did not describe the corticosteroid dose that patients received. This is important because it may be feasible that patients on higher corticosteroid dosing may have had higher rates of GI bleeding and gastritis. However, we only noted one episode of GI bleeding and gastritis in the high-dose group.

One potential reason why patients with status asthmaticus receiving high-dose corticosteroids may have a low risk of gastritis and GI bleeds is that as these patients clinically improve, they may be initiated on enteral feeds. Some have proposed that enteral feeds may provide a reduced risk of GI bleeding by enhancing GI perfusion, increasing gastric pH, and decreasing mucosal ischemia. ¹⁵ Several studies have evaluated the rates of GI bleeding between critically ill adults receiving either SUP or initiation of early enteral feeds.^15-17 A recent meta-analysis by Huang et al ¹⁸ conducted a meta-analysis of 7 published studies including 889 critically ill adults. They noted no significant difference in GI bleeding rates between groups (relative risk 0.80; 95% Confidence Interval [CI], 0.49-1.31, P = .37). We noted 19% of patients in our study received SUP despite being on full enteral feeds, with a significantly higher number of patients in the high-dose versus usual-dose group continuing on SUP. These results contrast with Sochet et al ⁵ that noted 49% of their cohort continued on SUP despite being on full enteral feeds. It is plausible that providers were aware of some of the protective benefits of enteral feeds and discontinued SUP at a higher rate than the previous study.

In our study, we noted 21 (13.7%) patients with a SUP-associated adverse event with no difference between the usual-dose versus high-dose groups. In two studies evaluating 30 394 critically ill children with critical asthma with and without SUP use, only 24 (0.08%) had documentation of an SUP-associated adverse event with all of these patients developing thrombocytopenia.^5,6 No patients developed nosocomial pneumonia, necrotizing enterocolitis, or Clostridioides difficile colitis. In contrast, we noted 19 (12.4%) patients with nosocomial pneumonia. This is consistent with a previous meta-analysis by Yao et al. ⁸ They conducted a systematic review and meta-analysis of 17 studies including 340 763 critically ill children evaluating outcomes in those with and without SUP. They noted an increased risk of nosocomial pneumonia in patients who received SUP versus without SUP (Mantel-Haenszel risk difference [MHRD] 5.3%; 95% CI: 3.5%-7.0%; I² = 0%).

This study has several limitations. First, this was a single-center study. The corticosteroid dose and SUP initiation were at the discretion of the providers. It is possible that the number of patients receiving the high- versus usual-dose corticosteroids may have been different at other institutions. Second, this is a relatively small sample size. Previous studies and meta-analyses have evaluated GI bleeding and other SUP-associated ADEs in pediatric patients and have included a significantly higher number of patients, with one meta-analyses assessing up to 340 000 patients. ⁸ That being said, the primary objective of this study was focused on deprescribing of SUP at various stages of the patient stay. Previous studies evaluating the deprescribing outcomes have included comparable sample sizes with the present study.^9-13 Third, the primary objective was descriptive focusing on the rates of deprescribing. Thus, a power calculation was not performed.

Conclusions

Initiation of SUP is a common practice in critically ill patients with status asthmaticus receiving corticosteroids. Approximately three-fourths of patients initiated on SUP in the PICU were continued on these agents upon transfer to the floor despite many receiving full enteral nutrition. Additionally, 16% were discharged on SUP without an appropriate indication. Prolonged exposure has been associated with significant morbidities, and in this study 21 (13.7%) had SUP-associated adverse effects.