Upper gastrointestinal bleeding is common in critically ill children, with an incidence of approximately 10% that increases to approximately 51% with mechanical ventilation. The risk factors include coagulopathy, organ failure, respiratory failure, high-pressure ventilator settings, and a high Paediatric Risk of Mortality (PRISM) score of 10 or greater.1,2 Some of these episodes are clinically significant, causing shock or requiring blood transfusion.

Although not fully understood, the pathophysiology of stress ulcers is thought to be different from that of peptic ulcers, involving factors such as mucosal ischemia and reperfusion injury. Acid-suppressive drugs have the potential to prevent stress ulcers, and an intragastric pH of more than 4 was been shown to be protective.3 Consequently, stress ulcer prophylaxis (SUP) has gained popularity in adult and paediatric intensive care units, with mechanical ventilation being the most frequent indication. However, the routine use of SUP in all critically ill patients is also very common.4,5

Several agents have been used for SUP, including proton-pump inhibitors (PPIs), histamine-2 receptor antagonists (H2RAs) and sucralfate, with the former exhibiting the greatest efficacy in adult studies.6 Omeprazole is the PPI most frequently used for SUP, although this is not one of the licensed indications for this drug.

Stress ulcer prophylaxis is not without risk: the increase of gastric pH achieved by PPIs or H2RAs may favour gastric colonization by pathogens. Subsequent regurgitation or retrograde colonization to the pharynx and trachea may increase the risk of ventilator-associated pneumonia (VAP).7,8 In addition, it has been proposed that the loss of the bacteriostatic effect of the acidic gastric juice, combined with the frequent use of antibiotics in critically ill patients could increase the risk of Clostridium difficile-associated diarrhoea.9

The previous literature on the usefulness of SUP in critically ill children is limited, and data from randomized controlled trials (RCTs) are scarce. Therefore, the aim of our trial was to evaluate the safety and efficacy of SUP with omeprazole, the agent used most commonly for SUP in critically ill children.

We conducted a randomized controlled open-label trial in the paediatric intensive care unit (PICU) of Menoufia University Hospital between January 2019 and September 2020. The study protocol was approved by the Medical Research Ethics Committee of the Menoufia School of Medicine, and we obtained written informed consent from the parents of participants.

Critically ill children admitted to the PICU were eligible for the study. The exclusion criteria included age less than 1 month or greater than 16 years; presence of upper gastrointestinal bleeding at the time of PICU admission; underlying haemorrhagic disorders (eg, haemophilia or purpura) and administration of SUP before PICU admission. We also excluded patients with paediatric Sequential Organ Failure Assessment (pSOFA) scores of 16 or greater, as there is evidence suggesting that they are at higher risk of gastrointestinal bleeding and therefore all of them received SUP.

At admission, patients were randomly assigned to the prophylaxis group or the control group. We used block randomization to achieve a homogeneous allocation of participants in the 2 study groups using the website https://www.graphpad.com/quickcalcs/randMenu/.

We calculated that a sample size of 72 patients in each group would offer a power of 84% with a type I error probability (alpha level) of 0.05 and an expected proportion of gastrointestinal bleeding of 1/2.5.

Patients in the prophylaxis group received omeprazole within 4 h of PICU admission at a dose of 1 mg/kg/day delivered intravenously. Patients in the control group were not given anything for SUP. Participants were not blinded to treatment allocation.

If a patient in the control group developed gastrointestinal bleeding during the PICU stay, omeprazole was given to treat it. If a patient in the prophylaxis group developed gastrointestinal bleeding while taking omeprazole, the dose was increased up to 3.5 mg/kg/day.

Treatment with omeprazole was discontinued when patients improved, as evinced by no longer needing vasoactive medications, weaning to low levels of ventilatory support and progression of enteral feeding to two thirds of the required energy intake.

Patients underwent a thorough evaluation, and the diagnosis of sepsis was based on the criteria of the International Paediatric Sepsis Consensus Conference.10 We assessed the severity of illness severity at admission by means of the pSOFA score (possible value range, 0–24 points)11 at the end of the first 24 h. The follow-up for each patient ended at the discharge from the PICU, and the primary outcome was the development of upper gastrointestinal bleeding during SUP or of infections such as hospital-acquired pneumonia (HAP), VAP, central line-associated bloodstream infection (CLABSI), and C. difficile-associated diarrhoea. The secondary outcomes were PICU mortality, length of PICU stay, and duration of mechanical ventilation.

Overt upper gastrointestinal bleeding was defined as the presence of haematemesis, melena or flecks of blood in nasogastric aspirates (coffee ground vomitus). Clinically significant gastrointestinal bleeding was defined as bleeding causing development or worsening of shock or requiring non-elective transfusion of packed red blood cells.

We defined VAP according to the criteria established by the Centers for Disease Control and Prevention (CDC)12 as pneumonia occurring more than 2 calendar days after initiation of invasive mechanical ventilation and the ventilator being in place on the date of the event or the day before. We considered pneumonia that developed 2 days after hospital admission in the absence of endotracheal intubation healthcare-associated pneumonia (HAP).

We defined CLABSI as presence of a primary bloodstream infection in patients that had a central line within the 48-h period before the development of the infection and was not bloodstream-related to an infection at another site.13 We diagnosed C. difficile-associated diarrhoea by means of stool toxin testing.

Stress ulcer prophylaxis is a common intervention in critically ill children, and its use varies considerably between PICUs, which reflects the uncertainty about its risks and benefits.14

Ours is one of the few paediatric RCTs that addresses this issue, and the only one to our knowledge assessing the usefulness of omeprazole.15

In our study, we found that SUP with omeprazole failed to reduce the incidence of gastrointestinal bleeding, while a systematic review that pooled data from 2 small paediatric RCTs found that SUP was significantly more effective in preventing gastrointestinal bleeding compared with “no treatment”. However, when these 2 studies were pooled with an additional small RCT comparing “treatment” vs “placebo”, the authors did not find a significant difference.16

We ought to highlight that we only included children with mild to moderate organ dysfunction (pSOFA score < 16), so it is possible that more severely ill children may in fact benefit from SUP. Furthermore, the sample under study was heterogeneous, and it is likely that the usefulness of SUP varies between patient subsets, as demonstrated by a recent observational study that found no instances of clinically important bleeding in children admitted for status asthmaticus that did and did not receive SUP.17

Ventilator-associated pneumonia has been a major concern discouraging SUP based on the findings of observational studies.18 However, in our study, SUP was not associated with a significant increase in the incidence of VAP or HAP, which was consistent with a retrospective paediatric study that compared ranitidine and sucralfate with no prophylaxis.7 In another paediatric trial, none of the patients that underwent SUP developed pneumonia caused by an organism previously isolated from the stomach.19 Similarly, another paediatric RCT failed to find significant differences in the incidence of VAP in groups that received ranitidine, omeprazole, sucralfate or no treatment.15

Likewise, a meta-analysis of adult RCTs did not find an increased risk of pneumonia in critically ill patients associated with the use of PPIs,6 but a larger meta-analysis concluded that PPIs may be associated with higher risk of pneumonia (OR, 1.39; 95% CI, 0.98–2.10).20

The findings of the latter studies suggest a non-existent or, at most, weak association of SUP with VAP. Nevertheless, strong concerns linger, and the issue can only be settled with the performance of larger paediatric RCTs. It important to keep in mind that SUP is not the sole risk factor implicated in VAP, with more important factors including prior antibiotic therapy, steroid therapy, bloodstream infections, genetic syndromes and reintubation,21 and therefore large trials would need to be conducted to establish the small contribution of SUP, assuming it exists.

Another concern related to PPIs is the risk of C. difficile-associated diarrhoea suggested by previous observational studies.22 In our study, however, none of the patients developed C. difficile-associated diarrhoea, in agreement with a recent paediatric multicentre observational study that found a very low incidence of C. difficile-associated diarrhoea (1%) in mechanically ventilated children that received SUP.14 Furthermore, a meta-analysis of RCTs in critically ill adult patients found no association of PPI or H2RA with C. difficile infection.20

While our study did not find an association of omeprazole with VAP or C. difficile-associated diarrhoea, it did find a significantly higher incidence of CLABSI in children that underwent SUP. This is of utmost importance, as it suggests that the use of SUP in children with non-severe critical illness is morally questionable, as it puts them at risk of a serious form of nosocomial infection associated with an increased morbidity and mortality, in addition to increasing health care costs.

The association of bloodstream infections with SUP could result from bacterial overgrowth in the stomach and duodenum made possible by gastric acid suppression, followed by bacterial translocation through the damaged intestinal epithelial barrier under conditions of mucosal ischemia.23 However, this cannot explain the development of CLABSI in our study. One explanation worth considering is that PPIs also have anti-inflammatory effects, including inhibition of production of pro-inflammatory cytokines. Furthermore, PPIs inhibit neutrophil proton pumps and therefore interfere with neutrophil function. The clinical consequences of these anti-inflammatory effects are not clear. Theoretically, they may promote healing of acid peptic disorders, but they might also predispose to the development of infections, especially in patients with liver disease.24 A multicentre point-prevalence study in the general ICU population found that SUP was associated with ICU-acquired infections, such as bloodstream and urinary tract infections.25 On the other hand, a recent multicentre retrospective study did not find a similar association.26

In addition, PPIs have an antimicrobial effect on the gut flora through the blocking of H+-ATPase in some bacteria and fungi, which may be beneficial if they destroy pathogens but detrimental if they destroy useful organisms.24

The net impact of these effects on the risk of bloodstream infections needs to be addressed in larger paediatric trials.

We ought to clarify that the term “CLABSI” used in this study differs from the term “catheter-related bloodstream infection”, as the latter denotes that the catheter is the source of bloodstream infection; and we cannot exclude the possibility that some of the bloodstream infections in our sample originated from sources other than catheters, including the gastrointestinal tract on account of acid suppression. In any case, our data revealed an association of SUP with bloodstream infections and showed that VAP and C. difficile-associated diarrhoea are not the only types of infections that may occur with SUP.

It is fair to assume that the safety and efficacy of SUP will be reflected on the general PICU outcome measures. In agreement with the findings of a meta-analysis of adult RCTs, we found did not find significant differences in mortality, length of PICU stay, length of hospital stay and duration of mechanical ventilation between the prophylaxis and control groups.20 Similarly, a paediatric RCT did not find an association between acid-suppressive drugs or sucralfate with mortality.15

Until the controversy surrounding the safety of SUP is resolved, it would be more prudent to consider nonpharmacological strategies to prevent bleeding secondary to stress ulcers, such as prompt resuscitation of patients with shock and early initiation of enteral feeding, both of which contribute to maintaining splanchnic blood flow.27,28

Furthermore, there is evidence that while upper gastrointestinal bleeding is common in critically ill children, clinically significant bleeding is rare. Therefore, it has been suggested that SUP be restricted to patients with at least 2 risk factors for significant bleeding.1 In this regard, in our study we found that gastrointestinal bleeding was more likely to develop in patients with higher pSOFA scores, that required mechanical ventilation or with a diagnosis of sepsis. The pSOFA score is a measure of organ failure, a factor found to be associated with an increased risk of gastrointestinal bleeding in previous studies,2,29 although these studies did not quantify the level of organ dysfunction using the pSOFA or similar scores.

However, in the multivariate analysis we found that mechanical ventilation was the only independent predictor of bleeding, which corroborated the findings of a previous retrospective study.30 We ought to highlight that the proportion of patients that required mechanical ventilation did not differ significantly in the prophylaxis and control groups, which suggests that this variable did not affect our findings regarding SUP safety and efficacy.

Other paediatric observational studies have reported a higher incidence of gastrointestinal bleeding associated with high PRISM scores, coagulopathy, respiratory failure, pneumonia and polytrauma.1,31

The notion that SUP is not justified except in children with certain risk factors is supported by the findings of a meta-analysis of RCTs in critically ill adults, which concluded that SUP with PPIs and H2RAs likely achieves important reductions in gastrointestinal bleeding in patients at increased risk of bleeding, but not in patients at low risk of bleeding.20 As a consequence, some authors have proposed the development of guidelines and educational programs aimed at PICU providers as a valuable tool to achieve a more rational use of acid-suppressive medications.32

The main limitation of our study is that the treatment was not masked, which could be a source of ascertainment bias. However, the same limitation applies to the vast majority of paediatric SUP trials in the previous literature.16 In addition, we only included patients with mild to moderate degrees of organ dysfunction at admission, so our findings cannot be generalized. However, this does not make them worthless, as non-severely ill patients, to whom the findings do apply, constitute the largest group of patients admitted to many PICUs, from which we can glean a clear message: routine SUP should be avoided in non-severely ill children.

Another limitation is that we did not compare omeprazole with other drugs used for SUP. In addition, the sample was heterogeneous and not large enough to reliably assess the useful of omeprazole in different subsets of patients. Nevertheless, this heterogeneity is what characterizes most PICU admissions in the real world. Lastly, we did not evaluate the influence of some factors, like enteral nutrition, on the study results.

In children with mild-to-moderate organ dysfunction at the time of PICU admission, SUP with omeprazole was not effective in preventing upper gastrointestinal bleeding. Furthermore, it was associated with bloodstream infection. Our findings speak strongly against the common practice of routinely administering PPIs to all critically ill children. It seems prudent to narrow down the indications for SUP exclusively to patients requiring mechanical ventilation. At the same time, it is clear that larger RCTs are needed to assess the usefulness of SUP in critically ill children more rigorously.

The authors have no conflicts of interest to declare.