Tag Archives: Proton pump inhibitors

PPI Side Effects: “Dissecting the Evidence”

Posted on by

While proton pump inhibitors (PPIs) are used extensively for acid-related diseases and have been around for nearly 25 years, there have been a number of reports about potential side effects.  As a drug class, PPIs have a very good safety profile.  A recent article reviews some controversial adverse effects and summarizes the evidence for and against (Clin Gastroenterol Hepatol 2013; 11: 458-64).

I. Calcium/bone effects.  After reviewing a number of studies, the authors conclude: “There is no good evidence to establish that PPI use has a significant risk for bone density loss or osteroporotic-related fractures….Supplemental calcium is not recommended or justified solely because of PPI use.”

II. Iron. “Although it is conceivable that PPI therapy may reduce absorption of nonheme iron and retard iron pool replenishment, this effect has not been well-studied or evident from widespread use in clinical practice.

III.  Magnesium.  “The FDA recommendation to consider checking magnesium levels before starting is not practical, in particular for the over-the-counter market. In patients who may be predisposed to …ongoing magnesium loss…it may be reasonable to follow…Given the rarity of the reports and no controlled studies to delineate the mechanisms, it is important for health care providers to be aware of this” (rare reports of profound hypomagnesemia).

IV. Pneumonia. “Small relative risk associated with short-term and high-dose PPI use.  These relationships, however, do not offer a definitive explanation for the relative risk” due to the studies and confounding factors.

V. Clostridium difficile.  “To date, there is insufficient evidence to conclude that there is a definitive relationship between PPI use and C difficile infection…clinicians should be aware of this potential relationship.”

VI. Traveler’s diarrhea.  “The data…were overall supportive of no associated risk, albeit there were a few specific case reports suggesting a remote causal association.”

VII. Small intestinal bacterial overgrowth. “The relationship between PPI use and the development of SIBO is still not understood.”

VIII. Interstitial nephritis.  Extremely rare. “Investigators…did not find enough evidence to support a causative relationship.”

IX. Methotrexate.  “Coadministration of PPIs with high-dose methotrexate appears to be correlated with delayed methotrexate elimination.”

Also discussed: Vitamin B12, Clopidogrel, Spontaneous bacterial peritonitis

The authors conclude that the above reported associations have received considerable attention.  “Because PPIs are overprescribed in many patients, …the clinical effects always should be reviewed and attempts should be justified to stop any therapy that may not be needed.”

Related blog entries:

Do medicines work for GERD infants?

Posted on by

“Absence of evidence is not evidence of absence.” Carl Sagan

If medicines work for infantile GERD, it is difficult to prove (Winter H, et al. JPGN 2012; 55: 14-20).  The cited study is the latest having difficulty proving that proton pump inhibitors are effective in infants.  In this randomized, double-blind, placebo-controlled multinational study from 33 centers, 98 infants (1-11 months) were enrolled.  My colleague, Dr. Benjamin Gold, was one of the researchers.  Initially, a 2-week open-label treatment was given which was followed by a 4-week randomized phase.   Study participants had to have a clinical diagnosis of GERD with at least one GERD symptom  –at least twice per week in a 4-week period:

  • vomiting/regurgitation
  • irritability
  • supraesophageal manifestations (cough, wheeze, stridor)
  • respiratory symptoms triggered by feedings
  • feeding difficulties

The treatment administered was weight-based dosing of esomeprazole:

  • 3-5 kg:     2.5mg
  • 5-7.5kg    5mg
  • 7.5-12kg  10mg

Daily symptoms were captured with an interactive voice response system.  Among the initial 98 patients enrolled, 80 reached the randomization phase.  During the initial 2-week period, 81 (82.7%) had symptom improvement based on physician global assessment.  During the double-blind phase, 48.8% of placebo-treated patients and 38.5% of esomeprazole-treated patients discontinued therapy due to symptom worsening.  While the time for discontinuing esomeprazole was longer in a posthoc analysis, the primary outcome, discontinuation rate, was not significantly different.

So what is the reason that this was a negative study?  While the reasons are unclear, all of the following are possible:

  • Patient selection/lack of accurate diagnosis.  Mixed-population was recruited for the study –though this type of population is similar to clinical practice.
  • Dose of esomeprazole.
  • Inadequately powered study.  If the effectiveness of PPIs is small, a much larger population is needed.
  • Maybe these agents don’t work in infants.  Infants secrete less acid than children and adults; thus, acid blockers may not work as well. (The Medical Pendulum and Gastroesophageal Reflux)

Why not give PPIs even if they don’t work?  The previous link discusses many of the potential adverse problems that are possible with medical treatment of GERD. However, even if a medicine does not harm does not mean you should do something because it ‘might’ do some good.  An example of this is the apocryphal tale of the famous pianist who died one day in the middle of a recital. (I saw this in a journal article but cannot remember the reference.)  The manager came out to announce his death.  A man in the audience shouts, ‘Give him an enema.’ Initially, the manager tries to ignore him. After the man yells this three times, the manager responds, ‘the poor man is dead…what good will an enema do?’ The voice replies, ‘What harm will it do?

Additional references:

  • -JPGN 2010; 50: 609-18. Pantoprazole helped improve symptoms but there were no significant differences compared to placebo in withdrawal rates due to lack of efficacy. n=128.
  • -NASPGHAN 2009, Abstract#21. Meds/Rx of NICU pts did not shorten hospital stay or promote wt gain, n=1149. Abstract#26. prevacid more effective than ranitidine in infants.
  • -J Pediatr 2009; 155: 601 (letter). Should not be used to treat symptoms unless proven to be reflux.
  • -JPGN 2009; 49: 498. GERD guidelines.  “In infants and toddlers, there is no symptom or symptom complex that is diagnostic of GERD or predicts a response to therapy.” Identical response to placebo (vs prevacid) in largest double-blind randomized study (54% at 4 weeks) (J Pediatr 2009; 154: 514-20.)  Reflux is “not a common cause of unexplained crying. irritability..in otherwise healthy infants.” “There is no evidence to support the empiric use of acid suppression for the treatment of irritable infants.”

Looking better or feeling better in EoE?

Posted on by

When seeing a new diagnosis of eosinophilic esophagitis (EoE), I often try to explain that there are two potential goals of treatment: clinical remission (improvement in symptoms) and histologic remission (improvement in appearance of esophagus with microscope).  Unfortunately, these two outcomes are not always synchronous; more proof of this comes from a recent study (Clin Gastroenterol Hepatol 2012; 10: 742-49, 750-52 [editorial]).

In this double-blind, randomized, placebo-controlled study of fluticasone in adult patients with a new diagnosis of EoE, 19 patients were treated with fluticasone (880 μg BID) and 15 patients were treated with placebo inhaler –for six weeks.  Initially, 21 patients were assigned to each group; 2 dropped out of treatment group and 6 dropped out of placebo group before completion of followup EGD.   The average age in the treatment group was 37 years versus 38 years in the placebo group.  A complete histologic response was defined as >90% reduction in mean eosinophil count; this occurred in 62% of fluticasone patients and in none of the placebo group, based on an intention-to-treat analysis.  Another measure of eosinophil activity, eosinophil-derived neurotoxin (EDN), was reduced by 81% on intracellular staining in the treatment group compared with 8% in the placebo group.  Figures 1 through 3 show this staining –it’s pretty cool!

Yet, the clinical response was not statistically different.  Dysphagia was reduced by 57% in the treated subjects compared to 33% in the placebo subjects in an intention-to-treat analysis.  Results were improved modestly in those who actually were treated: 63% (12 of 19) compared to 47% of placebo patients.  A complete response for dysphagia was noted in 42.9% of fluticasone group compared with 28.6% of control group based on an intention-to-treat analysis.  A fairly high rate of candidiasis was noted in treated patients 26%;  no placebo patients developed candida.

Another interesting finding was that among those who continued PPIs for heartburn symptoms the response to fluticasone was not improved.  40% of PPI users had a complete histologic response compared with 79% of non-PPI users.

So what are the reasons for the discrepancy between clinical and histologic response?

  • Established strictures and small-caliber esophagus may require dilation rather than medicines to relieve dysphagia
  • Esophageal fibrosis and subsequent esophageal compliance may not respond to topical therapy or take a lot longer to improve
  • Secondary candidiasis may reduce clinical response –though in this study, 5 of 6 patients with candida did in fact have symptom resolution
  • Compensatory behaviors may improve clinical symptoms –chewing food, cutting up food better, drinking more fluids, and avoiding some foods.  This may make it harder to detect important differences.

Patient information link: (Eosinophilic esophagitis – CCDHC Home)

Related posts:

Look of improvement on an EoE diet

Guidelines for Eosinophilic Esophagitis

Eosinophilic Esophagitis -Six Food Group Diet

The undiscovered country

MicroRNA signature for eosinophilic esophagitis

Proton pump inhibitors–infection risk with cirrhosis

Posted on by

In a previous post (The Medical Pendulum and Gastroesophageal Reflux), I note that enthusiasm for proton pump inhibitors has started to wane.  In addition, a significant number of reported of potential side effects were referenced.  Another potential adverse effect is increasing the rate of spontaneous bacterial peritonitis (SBP) in patients with cirrhosis (Clin Gastroenterol Hepatol 2012; 10: 422-27).

This retrospective study examined 65 hospitalized cirrhotic patients with paracentesis-proven SBP between 2006-2009 and compared them to 65 contemporaneous hospitalized cirrhotic patients without SBP.   Patients with SBP had a higher incidence of use of PPI within previous 7 days: 71% versus 42%.  Of patients with SBP receiving PPI, the authors state that 68% did not have a documented indication for PPI use.

Additional references/previous posts:

  • –Treating reflux does not help asthma
  • -Risk of Hypomagnesemmia -2011. http://www.fda.gov/drugs/drugsafety/ucm245011.htm
  • –Gastroenterology 2010; 139: 1115.  Review of safety of PPIs.
  • –Gastroenterology 2010; 139: 93. n=167,000. PPIs associated with hip fracture risk, OR 1.3, in patients with other risk factors.
  • –Gastroenterology 2010; 138: 896-904. 5 yrs of PPI -no increase risk in hip/spine fx.
  • –Arch Intern Med 2010; 170: 765-71, 747 (ed). PPI not related to hip fx (n=161,806) women 50-79. INCREASE risk of spine fx, hazard risk 1.47
  • –Arch Intern Med 2010; 170: 772-8. PPIs increase risk of Clostridium difficile infection (hazard ratio 1.42 –42% increase in risk), n=1166.
  • –Arch Intern Med 2010; 170: 784-90. n=101,796. OR 1.74 for daily PPI, OR 2.36 if BID Rx; thus ~70% increase risk of nosocomial infection.
  • –Clin Gastro & Hep 2010; 8: 504. Increased bacterial overgrowth with PPI use.
  • -JAMA 2009; 301: 2120-2128. Use of PPIs associated with INCREASED hospital acquired pneumonia by ~30%. Could result in 180,000 HAP cases/yr with ~33,000 deaths. n+ 63,878 admissions, 52% on PPIs or H2RAs (83% PPIs, 17% H2RAs). H2RAs NOT associated with HAP cases.

All bleeding stops

Posted on by

Although the title is technically true, there are preferred ways to stop GI bleeding (Clinical Gastroenterol & Hepatol 2012; 10: 234-39).  This article summarizes the approach for nonvariceal upper GI bleeding.

Recommendations:

  • Endoscopy within 24 hours following ABCs/adequate resuscitation.  Use of a promotility agent prior to endoscopy may be helpful (in pediatric patients: erythromycin ~3 mg/kg)
  • Correct coagulopathy but do not delay endoscopy.
  • Consider nasogastric tube placement.
  • Do not use somatostatin or octreotide.
  • High-risk endoscopic stigmata should receive endoscopic hemostasis.  These lesions include actively spurting, oozing blood, nonbleeding visible vessel, and an adherent clot.
  • Pigmented dots or clean base ulcers do not require endoscopic hemostasis.
  • Endoscopic management includes clips, thermocoagulation, or sclerosant injection alone or in combination with epinephrine injection.  Epinephrine alone is not recommended for high risk lesions.
  • If a clot is found, attempts to remove it should be made to visualize underlying lesion.  If clot is adherent, intensive IV PPI therapy may be sufficient.  A typical dose would be esomeprazole 80mg bolus (for an adult) followed by 8 mg/hour for 72 hours.
  • Stable patients can be fed within 24 hours.
• IV PPI dose: 1mg/kg bolus followed by 0.1mg/kg/hr infusion.

Additional references:

  • -Ann Intern Med 2010; 152: 101-113. Consensus recommendations on UGI bleeding. Early endoscopy (<24hrs), data support attempts to dislodge clots, consider clips or thermocoagulation for Rx. Preendosocpy PPI can be helpful.
  • -Clin Gastro & Hep 2010; 8: 651. Article suggests second look only if difficult visualization on initial endoscopy (eg unable to remove clot).
  • -Ann Intern Med 2010; 152: 101-13. Systematic review of on UGI bleeding. Use IV PPI bolus, then continuous PPI if high risk stigmata after endoscopic Rx. Hospitalize for at least 72hrs.
  • -Gastroenterology 2010; 138: 1252. Review of upper GI bleeding.
  • -Clin Gastro & Hep 2009; 7: 828. Review of recurrent GI bleeding with negative initial evaluation.
  • -Gastroenterology 2008; 134: 1836. Frequent high dose oral PPI also effective with bleeding ulcers: prevacid 120mg x1, then 30mg q3hrs compared favorably with 90mg IV followed by 9mg/hr. n=66. intragastric pH >6 for 68% of study in IV PPI vs. 65% in oral PPI. 1st hour -more rapid onset with IV PPI.
  • -Gastroenterology 2007; 133: 1694. Position statement & review on obscure bleeding.
  • -Ann Intern Med 2003; 139: 843-857. Consensus on nonvariceal bleeding. Rec: lansoprazole 90mg bolus, then 6mg/hr x 72hrs or pantoprazole 80mg then 8mg/hr in high risk lesions
  • -Clin Gastro & Hep 2006; 4: 1459. Trends in non-variceal bleeding between 1993-2003 do NOT show improved outcomes with PPI. Overall mortality fairly steady @3.5%
  • -Clin Gastro & Hep 2005; 3: 959. WCE should be 2nd step in obscure bleed, p egd/col.
  • -NEJM 2004; 351: 488. case review.
  • -Gastroenterology 2002; 123: 17-23. IV erythromycin, 20 minutes before endoscopy, helped clear stomach (82% clear vs. 33% c placebo). Adults in this study received 250 mg. (thus, children probably need 3-4 mg/kg)
  • -Gastro Endosc 2002; 56: 174. erythromycin helpful-3mg/kg IV over 30 min
  • -Gastroenterology 2002; 123: 407-13. Endoscopic Rx of adherent clots c PUDz helpful (epinephrine injection, cold guillotining of clot, then coagulation cautery); Editorial on 632-635 emphasizes vigourous washing BUT NOT to remove adherent clot unless in centers with low rebleeding rates. Additionally, PPIs very helpful in preventing rebleeding in this situation (NEJM 1997; 336: 1054-8).
  • -NEJM 1999; 341: 38. Occult bleeding
  • -Gastroenterology 2000; 118: 197. AGA position statement.
  • -Gastro Endosc 2001; 53: 853 & 859. ASGE guidelines/algorithm for upper & lower GI bleeding

The Medical Pendulum and Gastroesophageal Reflux

Posted on by

In so many areas of pediatric gastroenterology, there is a gradual development of enthusiasm for a medical treatment.  In the vast majority, the enthusiasm goes too far and closer scrutiny often determines a more limited role for this medical treatment or potential adverse effects that were not initially appreciated.  The latest example of this may well be with the use of proton pump inhibitors (PPIs) for gastroesophageal reflux disease, particularly in infants and individuals with asthma.  Although these medications may not have reached their apogee, more and more their effectiveness for so many ailments has been questioned.  In this month’s issue of JPGN, this is highlighted (JPGN 2012; 54: 8-14).  The article which emanates from the offices of the FDA discusses the fact that the usage of PPIs has increased 11-fold from 2002-2009 in infants <12months of age; 404,000 prescriptions were dispensed to 145,000 infants in the U.S. in 2009.    At the same time, althougth there have been four randomized controlled trials of PPIs in infants, NO studies have demonstrated the effectiveness of these drugs in this population.  As a consequence, the authors recommend that these drugs be restricted to infants with endoscopically-proven GERD/erosive esophagitis.  No other tools are sufficient to identify infants who are likely to respond.  Perhaps the reason why these agents work less well in infants is due to the fact that acid secretion is much less in infants than in children and adults.  For example, at 4months of life, average acid secretion rate in infants is ~27-fold lower than in adults (Am J Dig Dis 1969; 14: 400-14). As a consequence, their symptoms may not be responsive to acid reduction treatments.

Other related references on GERD in infancy:

–JPGN 2010; 50: 609-18. Pantoprazole helped improve symptoms but there were no significant differences compared to placeblo in withdrawal rates due to lack of efficacy. n=128.
-NASPGHAN 2009, Abstract#21. Meds/Rx of NICU pts did not shorten hospital stay or promote wt gain, n=1149.
–JPGN 2009; 49: 498. NASPGHAN GERD guidelines. “In infants and toddlers, there is no symptom or symptom complex that is diagnositc of GERD or predicts response to therapy.” Identical response to placebo (vs prevacid) in largest double-blind randomized study (54% at 4 weeks) (J Pediatrics 2009; 154: 514-20.)-Reflux is “not a common cause of unexplained crying. irritability..in otherwise healthy infants.” “There is no evidence to support the empiric use of acid suppression for the treatment of irritable infants.”

GERD and respiratory/ENT issues:

–Gastroenterology 2010; 139: 1887. PPIs decreased postnasal drainage compared to placebo. n=75. (50% vs 5%) age discrepancy in patient populations.
–Clin Gastro & Hep 2010; 8: 741 (excellent editorial), 770 (article on rabeprazole improving heartburn Sx in pts with laryngitis), n=82. Editorial suggests 1-2month trial of BID PPI and if not effective, then little to offer. May change when studies looking at surgery (after impedance) outcomes.
–Gastroenterology 2010; 139: 754. 716 (editorial). Acoustic cough & reflux. Study recorded cough during pH measurement. n=71. ‘causality cannot be established until effective treatment’ available.
–Gastroenterology 2009; 137: 1844. Critical review of below NEJM article. ‘a subset of asthmatics will have objective detection of GERD without typical symptoms…work by Amer Lung Assn suggests that twice daily PPI will not be helpful’..however, ‘perhaps 3-6months of PPI may still be reasonable until we can accurately identify subgroups of pts who may respond.’ –Gary Falk, Cleveland Clinic
–NEJM 2009; 360: 1487, 1551. Use of PPIs (nexium 40mg bid) in poorly-controlled asthma with no symptoms of GER –did not help w asthma control & pH studies were not predictive of response. n=412 adults. 40% c abnl pH studies in each group (nexium vs. placebo).
–Clin Gastro & Hep 2007; 5: 1379. Review of ENT findings and reflux.
–Am J Gastro 2007; 102: 716. Poor specificity of ENT findings for diagnosis of laryngopharyngeal reflux.
–Aliment Pharm Ther 2007; 25: 385-92. meta-analysis. Rx c PPIs not more effective than placebo in resolving ENT symptoms presumed to be due to GER. Editorial suggests some patients may benefit, but better tools are needed to identify them.

GERD and surgery:

–Gastroenterology 2011; 141: 1938.  LOTUS study in JAMA summarized in this review. (JAMA 2011; 305: 1969) Medical treatment outperformed surgery. 92% under control (remission) with long term medical Rx vs 85% with surgery & fewer side effects of medical treatment.
–Clin Gastro & Hepatology 2009; 7: 1292, 1264 (editorial). 12yr outcomes for surgery vs PPI. n=154 omeprazole, n=144 surgery. Similar long term outcome ~50% with long term remission.

–Pediatrics 2006; 118:1828. 48,665 antireflux surgeries done from 1996-2003 (~7000/yr) in US

–Clin Gastro & Hep 2006; 4: 299. Frequent complications post-op and frequent need for GERD meds.  Dysphagia in 19%, dilatation in 6%, repeat surgery in 2%, mortality in 0.8% (n=3145). 50% required GERD meds.

–Clin Gastro & Hep 2004; 2: 978-984. Pediatric study.  n=198.  63% required post-op treatment for recurrent GERD -retrospective review 1996-99.

Proton Pump Inhibitors and reported adverse effects:

-Risk of Hypomagnesemmia -2011. http://www.fda.gov/drugs/drugsafety/ucm245011.htm
–NEJM 2010; 363: 2114. large Denmark study. 5082 fetuses with PPI exposure (out of 840,968 live births). Risk of birth defects NOT increased with exposure during 1st trimester. Possible slight increase risk with preconception use except with omeprazole.
–Gastroenterology 2010; 139: 1115. Review of safety of PPIs.
–Gastroenterology 2010; 139: 93. n=167,000. PPIs associated with hip fracture risk, OR 1.3, in patients with other risk factors.
–Gastroenterology 2010; 138: 896-904. 5yrs of PPI -no increase risk in hip/spine fx.
–Arch Intern Med 2010; 170: 765-71, 747 (ed). PP not related to hip fx (n=161,806) women 50-79. INCREASE risk of spine fx, hazard risk 1.47
–Arch Intern Med 2010; 170: 772-8. PPIs increase risk of Clostridium difficile infection (hazard ratio 1.42 –42% increase in risk), n=1166.
–Arch Intern Med 2010; 170: 784-90. n=101,796. OR 1.74 for daily PPI, OR 2.36 if BID Rx; thus ~70% increase risk of nosocomial infection.
–Clin Gastro & Hep 2010; 8: 504. Increased bacterial overgrowth with PPI use.

-JAMA 2009; 301: 2120-2128. Use of PPIs associated with INCREASED hospital acquired pneumonia by ~30%. Could result in 180,000 HAP cases/yr with ~33,000 deaths. n+ 63,878 admissions, 52% on PPIs or H2RAs (83% PPIs, 17% H2RAs). H2RAs NOT associated with HAP cases.
–Gastroenterology 2009; 137: 80. PPIs induce acid-related symptoms in ~22% vs 7% of placebo in healthy volunteers.
–Ann Intern Med 2008; 149: 391-398. Risk for pneumonia associated with short-term PPI use, not long term
–Clin Gastro & Hep 2007; 5: 1418. Increases risk of bacterial gastroenteritis.
–JPGN 2007; 45: 395, 421. Increasing use of PPIs-4-fold from 2000-2003; 0.5% of all infants. No safety/efficacy data.
–J Pediatrics 2007; 150: 262. Long term use (up to 11yrs of usage) of PPIs in 166 children; minimal problems: 2 c nausea, 2 c skin rash, 1 c diarrhea, 1 c agitation.
–JAMA 2006; 296: 2947-53. Risk of bone fracture –odds ratio 1.44-2.65 with long-term PPI treatment (>1yr); UK study looked at 1.8million