Efficacy of proton pump inhibitors in acute pancreatitis: a narrative review

Benjamin Nicholas Hughes

BPharmSci, MPharm | Acting Medicines Strategy and Outpatients Team Leader, Surgical Treatment and Rehabilitation Service

Francis Michael Wheelahan

BSc, MD | Hospital Medical Officer, Alfred Hospital Melbourne

Aim: To evaluate the benefit of proton pump inhibitors (PPIs) in acute pancreatitis.

Data sources: A literature search of PubMed and Cochrane library was conducted in May 2022 to search for relevant articles that present evidence for PPIs in acute pancreatitis.

Study selection: Articles were considered relevant if they examined PPI treatment during the management of acute pancreatitis in human participants. Four studies were found, two open label randomised controlled trials, one observational study, and one cohort study.

Results: No study showed any benefit for PPI use in any primary outcomes for patients with acute pancreatitis.

Acute Pancreatitis (AP) is the inflammation of the pancreas and often other nearby organs. Patients will generally present with central epigastric intense pain which can radiate to the back. This can also be accompanied by fever, nausea, vomiting, and tenderness when touching the abdomen. Elevated lipase, c-reactive protein (CRP) and white blood cell (WBC) counts are often found on testing. Acute pancreatitis is often caused by excessive alcohol consumption, gallstones, certain medicines, or idiopathic causes.¹ Acute pancreatitis may develop into more severe syndromes if left untreated. These include, pancreatic pseudocysts, necrotising pancreatitis, acute lung injury, acute renal failure, and multi organ failure. Although authors often quote a mortality rate for acute pancreatitis to be between 10–15%,² these numbers arrived from studies published in the 1990s. More recent studies show a mortality rate of 4–7%. The reduction in mortality for acute pancreatitis can be attributed to more accurate severity assessment tools (leading to faster intensive care unit admissions) and more advanced surgical, endoscopic, and angiographic tools.² General treatments for acute pancreatitis include fluid administration and analgesia. In practice, proton pump inhibitors (PPIs) are often prescribed as part of acute pancreatitis treatment.³ PPIs are medicines that reduce the secretion of gastric acid to aid in the treatment of GORD, peptic ulcers, and oesophageal damage.⁴ Most PPIs have similar efficacy and adverse effect profiles. A theoretical benefit of the use of PPIs in acute pancreatitis has been postulated and tested in animal studies, through mechanisms that include pancreatic anti-secretory effects and a pronounced inhibitory reactivity towards hydroxyl radicals, however this hasn’t been replicated in human trials.⁵

There is a well-documented overuse of PPIs within Australia. From 1995–2006, PPI use in Australia has increased by 1318%.⁶ There were over 19 million prescriptions filled in 2013–2014 for PPIs in Australia. PPIs are commonly one of the most prescribed items on the Pharmaceutical Benefit Scheme (PBS).⁷ Despite increasing the prescribing requirements via the PBS in May 2019, overall PPI prescriptions and usage remained unchanged.⁸

Although generally well tolerated, the use of PPIs in patients with acute pancreatitis is not without risks. PPIs may increase the risk for gastrointestinal infections (specifically Clostridium difficile)⁹ and can lead to micronutrient deficiencies (more so with long term use).¹⁰ Other risks include increased polypharmacy, increased cost to healthcare systems, and increased workload for healthcare staff when using intravenous formulations. Other adverse drug effects of PPIs include but are not limited to: iron deficiency; hypomagnesaemia; reduced calcium absorption; vitamin and nutrient deficiencies; community acquired pneumonia; acute kidney injury; chronic kidney disease; kidney failure; cardiovascular disease; gastric cancer; and all-cause mortality increase with long term use. Studies show that, although PPIs have relatively quick half-lives of 1–2 hours,¹¹ the effects of PPIs can be seen for up to 2 weeks after caseation due to the strong binding affinity.¹²

The introduction of PPIs in the treatment of acute pancreatitis has been a clinical issue of some contention. Although commencement of an intravenous PPI at the initial presentation of acute pancreatitis is seen regularly in clinical practice, current clinical guidelines do not endorse this. This review will discuss the evidence for PPI use in acute pancreatitis which will help inform clinician decision making.

Data sources

A literature search of PubMed and Cochrane library was conducted in May 2022 to search for relevant articles that present evidence for PPIs in acute pancreatitis. The reference lists of these relevant articles were explored to identify further evidence. Key search words included relevant terms for PPI and acute pancreatitis such as: proton pump inhibitor, acid suppression therapy, and pancreatitis.

Study selection

Articles were considered relevant if they examined PPI treatment during the management of acute pancreatitis in human participants. Articles were included if they were written in English or translated into English and were not restricted to any study type or publication year.

A narrative synthesis was used to describe the findings from the included articles in this review.

Results

Key studies that have examined the use of PPIs in acute pancreatitis in humans are listed in Table 1.

Table 1: Study comparison

AP = acute pancreatitis; APACHE-2 = Acute Physiology and Chronic Health Evaluation Score; ASD = acid suppressing drugs; BD = bis in die (twice a day); GI = gastrointestinal; H2RA = histamine 2-receptor antagonists; h = hours; IV = intravenous; PO = per os (by mouth); PPIs = proton pump inhibitors; RCT = randomised controlled trial; SAP = severe acute pancreatitis; SIRS = systematic inflammatory response syndrome; vs = versus.

Table 1 also available for download

Study	Design	Number of patients	Population	Severity of AP	Intervention	Control	Primary outcome	Key findings
Yoo et al. (2012)13	Open label RCT	40	Patients with AP Korea	All severities included	Pantoprazole 40 mg BD IV until oral intake resumed then PO	No pantoprazole administered	APACHE-II score at admission, at 48 hours, and at discharge	No difference in primary outcomes for control and intervention (p>0.05)
Ma et al. (2017)14	Open label RCT	45	Patients with SAP China	SAP	Esomeprazole 40 mg IV daily for 7 days	No esomeprazole administered	APACHE-II score, SIRS score, and Marshall score at day 1, 3, and 7	No difference in primary outcomes for control and intervention (p>0.05)
Murata et al. (2015)15	Observational studies	10 400	SAP patients admitted to hospitals (n = 1021) Japan	SAP	PPIs (various strength, frequency, routes)	No PPIs administered	Hospital mortality within 7, 14, and 28 days	No difference in primary outcomes for control and intervention (p<0.001)
Demcsák et al. (2020)16	Cohort study	17 422	17 422 adult patients 59 centres in 23 countries	AP, all severities	ASD (PPIs and H2RA)	No ASD therapy	Hospital mortality Rate of GI bleeding Rate of GI infection	Mortality was higher in those receiving ASD (p<0.001) GI bleeds were more common within ASD treatment groups vs control (p<0.001) No significant differences in GI infection for ASD receiving patients vs control (p=0.276)

Yoo et al. was a Korean open label randomised controlled trial. ¹³ The study randomised 40 patients with acute pancreatitis who were divided into two groups. The intervention group received intravenous (IV) pantoprazole 40 mg twice daily during fasting time, then pantoprazole 40 mg orally twice daily until discharge. The control group received no PPI therapy. Both groups were treated with standard conservative pancreatitis cares (analgesics, IV fluids, withhold all oral intake, antiemetics if required). Average length of stay was 7.4 days for the intervention group and 7.55 days for the control group. The team used Acute Physiology and Chronic Health Evaluation Score (APACHE-2) to measure patient outcomes. The APACHE-2 score takes into consideration 12 individual variables to evaluate the severity of a patient’s health condition. These variables are age, temperature, mean arterial pressure, pH, heart rate, respiratory rate, sodium, potassium, creatinine, haematocrit, white blood cell count, and Glasgow Coma Scale. In the intervention group APACHE-2 score on admission was 3.15 and on discharge was 2.35. In the control group APACHE-2 score was 4.4 at admission and 2.85 at discharge. The intervention group showed no change in mean length of stay compared to the control group (7.4 days vs 7.6 days). The intervention group showed no change in time to 50% reduction in lipase compared to the control group (46.74 hours vs 40.8 hours). This study showed no clinically significant difference between the control versus intervention group. The main limitation in this study was the sample size (N = 40) and differences in baseline characteristics between the groups in relation to the cause and severity of acute pancreatitis.

Ma et al. was an open label randomised controlled trial which took place in China.¹⁴ The study reviewed the effect of intravenous esomeprazole on patients with severe acute pancreatitis. Forty-five patients with severe acute pancreatitis were randomly assigned into two groups, control or intervention. The control group (N = 21) received baseline pancreatitis treatment, excluding a PPI. The intervention group (N = 24) received baseline pancreatitis treatment in addition to esomeprazole 40 mg intravenously every 24 hours for 7 days. The following pathology was tested on day 1 and on day 7: CRP, interleukin (IL)-6, IL-8, tumour necrosis factor (TNF)-alpha, and procalcitonin. APACHE-2, systematic inflammatory response syndrome (SIRS), and modified Marshall scoring system were obtained at day 1, day 3, and day 7. Upper gastrointestinal (GI) pH, faecal occult blood tests, and peptic ulcer manifestations were checked on day 7. The study found no significant difference on pathology results (p > 0.05), no significant difference in APACHE-2, SIRS, and Marshall scores between the two groups (p > 0.05). Unsurprisingly, the intervention group had significantly increased pH at 5.02 vs the control group at 2.83. Neither group showed any statistically significant difference between peptic ulcer rate or positive faecal occult blood (p = 0.592 and p = 0.752 respectively). The authors concluded that PPI therapy did not show any benefit in severe acute pancreatitis. Although this study had a larger sample size compared to Yoo et al (45 vs 40 respectively), it was still not large enough to draw a statistically significant conclusion as shown by the p-values. Interestingly, neither lipase or amylase were discussed in this study which are key markers in determining the effect of PPIs in the progression of acinar cell injury, as CRP alone is not selective to pancreatic inflammation.

Murata et al. was a retrospective study which took place over 1021 hospitals in Japan.¹⁵ The study reviewed the effect of PPI administration (various types and strengths) on patients with severe acute pancreatitis. Ten thousand four hundred patients with severe acute pancreatitis who presented to hospitals between 2010–2012 in Japan were included in the study. Patients were divided into the intervention group (n = 3879) who used proton pump inhibitors during their treatment, and the control group who were not given proton pump inhibitors during their treatment (n = 6521). Mortality within 7, 14, and 28 days was used as the primary outcome. The study found that there was a higher risk of mortality at 7, 14, and 28 days for the intervention group vs control (5.1%, 5.8%, 7.0% vs 3.5%, 3.9%, 4.4% respectively). The observed increase in mortality with PPIs is likely attributable to confounding factors such as increased severity of pancreatitis for intervention group. The authors adjusted the data for patient demographics and clinical factors via multiple logistic regression and generated an odds ratio for the intervention group mortality at 7, 14, and 28 days. This was 1.14 (95% confidence interval [CI] 0.91–1.42, p = 0.236), 1.10 (95% CI 0.89–1.35, p = 0.349), and 1.12 (95% CI 0.92–1.37, p = 0.224) respectively. Thus, showed that there was likely no reduction, or increase in mortality with PPI therapy. Although this retrospective study had a considerable number of participants, the greatest limitation was that it did not address any outcomes other than mortality.

Demcsák et al. was a multi-centre cohort study which took place over fifty-nine centres across twenty-three countries and included 17 422 patients.¹⁶ The study evaluated the effectiveness of acid suppressing drugs (ASD) on all severities of acute pancreatitis. Acid suppressing drugs in this study included PPIs and histamine 2-receptor antagonists (H₂RA). Information was collected for patients who presented with acute pancreatitis between January 2013–December 2018 diagnosed according to International Association of Pancreatology (IAP) and the American Pancreatic Association (APA) guidelines. The 17 422 patients included in the study were separated based on pancreatitis severity as determined by the revised Atlanta classification. Each centre provided data on patient mortality, signs of GI bleeding, and GI infection. Signs of GI bleeding was defined as positive rectal examination, macroscopic blood in the stool, vomit or gastric juice, positive faecal blood test, or bleeding verified by imaging. Bleeding associated with ERCP was excluded. The results of this retrospective study showed that 23.3% of patients were already taking an acid suppressing drug on admission, and during hospitalisation 86.6% of patients received an acid suppressing drug. The vast majority (81.8%) of these patients received a PPI. When divided into ASD and No ASD groups, based on whether patients received any ASD therapy during their admission, mortality was significantly higher in the ASD group (4.9% vs. 2.4%), as was the rate of GI bleeding (5.1% vs. 2.2%). There was no statistically significant difference in rates of GI infection between the two groups (p = 0.276). While there was a found association between ASD therapy on mortality and risk of GI bleed, this is likely due to ASDs being prescribed more in the moderate and severe pancreatitis groups. This study does, however, substantiate the suggestion that PPIs are routinely prescribed for patients with acute pancreatitis and are likely overused when there is little to no indication for them.

Discussion

There is no clear evidence to show the benefit of PPIs in acute pancreatitis. All the studies reviewed showed that administration of a PPI had little to no benefit when compared to placebo in relation to the most common primary outcomes; mortality rate, APACHE-II scores, and SIRS scores.

No study reviewed measured separate outcomes for moderate acute pancreatitis verse severe acute pancreatitis. This comparison may be an area for future investigation. There was no consistent intervention used in the studies reviewed, however they all evaluated a proton pump inhibitor of therapeutic strength administered as either oral or IV. Given PPI’s similar efficacy, it is unlikely that a consistent PPI intervention between all the studies would have altered the accumulated outcome.¹⁷ 

Adverse effects from PPIs

Demcsák et al. was the only study which actively investigated the incidence of infections following the administration of a PPI.¹⁶ Interestingly, the study found that PPIs increased the risk of GI infections (most notably Clostridium difficile), however this was deemed not clinically significant. Although this large-scale study strengthens the argument against the routine use of PPIs in acute pancreatitis, it’s limited primary outcomes (mortality, GI bleed, infection rate) may not address the qualitative benefits of PPIs such as patient comfort and reduction in pain scores.

Whilst the adverse effects from short term PPIs are relatively rare (e.g. gastric discomfort, headache, vomiting, rash), there are reported incidents of more severe reactions such as allergic reactions, acute interstitial nephritis and hypersensitivity reactions such as Stevens-Johnson Syndrome.¹⁸ Additional risks of PPI use include increased polypharmacy, increased cost to healthcare systems, and the risk of patients continuing on therapy longer than intended (i.e. medicine not being ceased on discharge).¹⁹ Whilst each of these negative aspects of PPIs are relatively minor, when combined there is a reasonable case to promote the deprescribing of these medicines where there is insufficient evidence of benefit.

Guidelines for the management of acute pancreatitis

Currently there are no Australian guidelines that recommend PPI initiation for acute pancreatitis alone. Management guidelines published by the IAP, in collaboration with the APA, recommend immediate fluid resuscitation, ICU management for severe pancreatitis and nutritional support, however, they make no recommendations for initiating a PPI.^20,21 Furthermore, the guidelines published in Gut in the United Kingdom endorse the recommendations of the IAP and state that there is no specific drug therapy that is effective for treating pancreatitis.²²

Factors contributing to PPI use in acute pancreatitis

This then begs the question, why are PPIs so routinely prescribed on admission? A possible justification for this is that PPIs are commenced while a diagnosis is being made, and the clinician is still suspicious of an upper GI ulcer or flare of gastro-oesophageal reflux disease (GORD). Epigastric pain alone may lead the clinician to begin a PPI while the diagnosis is still in question. The recommendation for patients with peptic ulcer disease or gastroesophageal reflux is to commence a PPI acutely.³ While this would be appropriate as initial management of a suspected ulcer or GORD flare, it does not justify continuation of the PPI once a diagnosis of acute pancreatitis is made. There may be an increased risk of developing upper GI ulcers in patients with severe acute pancreatitis which could substantiate prophylactic PPI use in this cohort of patients, but it does not explain their widespread use when upper GI bleeds affect only a small proportion of patients.²³

Primary outcomes of data

The most common primary outcome measured between the studies were APACHE-II used by Yoo et al.¹³ and Ma et al.¹⁴ and mortality used by Murata et al.¹⁵ and Demcsák.¹⁶ The differences in primary and secondary outcomes measured in these studies makes it difficult to draw consistent conclusions regarding the use of PPIs in acute pancreatitis. Although, as discussed APACHE-II is a universally accepted metric for severity of disease classification,²⁴ it is not a specific measurement for the progression of acute pancreatitis. Whilst scoring systems are useful metrics for quantitative data, they do not supersede clinical assessment in acute pancreatitis.²⁵ Measuring mortality outcomes may appear to be a comprehensive metric; however, it is more likely that the benefit of PPIs in acute pancreatitis is found from qualitative outcomes, rather than a reduction in mortality.

Conclusion

The two randomised controlled trials included in this review had a sound methodology for evaluating the primary outcomes, however, both are limited by their small sample sizes. Only 40 patients were included in Yoo et al.¹³ and 45 patients in Ma et al.,¹⁴ so they are likely to be underpowered. There is an opportunity for a large, randomised controlled trial to be conducted to improve the quality of the evidence for PPI use in acute pancreatitis. The major limitation of the observational study included in this review is that it is retrospective and may be confounded by patients who are taking PPIs being more prone to severe pancreatitis.¹⁵ The final study included is an international cohort study, and is also limited by being a retrospective study,¹⁶ so no causative conclusions can be drawn as the patients were not randomised. This review shows that there is a gap in our evidence for PPI use in acute pancreatitis, particularly when it is a mild or moderate severity.

There is a protocol for a systematic review underway in China. This study plans to undertake a systematic review of the literature to summarize previous evidence in order to clarify the effectiveness and safety of PPIs in acute pancreatitis.²⁶ The systematic review will analyse hospital mortality and duration of hospital stays as the primary outcomes. There are various additional studies which could be conducted to gain a deeper understanding about the roles of PPI in pancreatitis. Each study reviewed used different PPIs of differing strengths, administration routes and treated various levels of severity of pancreatitis. Existing evidence suggests that the clinical differences between therapeutically dosed PPIs of any type are quite minor and unlikely to have a significant effect of clinical outcomes,¹⁷ there is evidence to suggest that IV administration of PPIs is faster at achieving stomach pH > 6 (2–3 hours for IV vs 3–4 hours for oral).²⁷ Another systematic review found that oral PPIs may be as effective as IV PPIs in the management of bleeding ulcers,²⁸ however, there is no evidence to date regarding the use of oral vs IV PPIs in acute pancreatitis. Future studies could also compare the outcomes of patients taking high doses of PPIs compared to standard dosages of PPIs and compare the outcomes within acute pancreatitis. A meta-analysis conducted in 2010 found that high dose IV PPI was not superior to low dose IV PPI in reducing rebleeding odds.²⁹