Inappropriate Discharge on Proton Pump Inhibitors After Intensive Care Unit Hospitalization

Abstract

Proton pump inhibitors (PPIs) are recommended for stress ulcer prophylaxis in critically ill patients at high risk for upper gastrointestinal bleeding. We hypothesized that many PPIs started in the intensive care unit (ICU) for prophylaxis are continued at discharge without a documented indication and are therefore inappropriate. We aimed to identify risk factors associated with discharge on inappropriate PPIs after ICU stay. We performed a retrospective cohort study from 2014-2018 of patients ≥ 18 years old who were initiated on a PPI in the ICU and survived to discharge. Patients were excluded if they were prescribed a PPI prior to ICU admission or had an appropriate indication for continuing a PPI after ICU hospitalization, including ulcer, esophagitis, or erosions on upper endoscopy; diagnostic codes for gastroesophageal reflux disease, Barrett’s esophagus, peptic ulcer disease, Helicobacter pylori; were a lung transplant recipient; or were discharged on NSAIDs, dual antiplatelet therapy, or an antiplatelet medication with an anticoagulant. A logistic regression model was used to identify characteristics associated with discharge on an inappropriate PPI after hospitalization, compared to those whose PPI was discontinued. We identified 24,751 patients admitted to one of 9 ICUs who survived to discharge, with 4,127 initiated on a new PPI. Of all PPIs initiated in the ICU, 2,467/4,127 (60%) lacked documentation of one of the above indications and were classified as inappropriate. Of these 2,467, 1,122 (45%) were continued on a PPI after transfer to the floor and 668 (27%) were discharged on a PPI (Figure 1). Of patients discharged on a PPI who had primary care records available, 18/28 (64%) were continued on the PPI at follow-up. On multivariable logistic regression, after excluding patients with appropriate PPI indications and adjusting for significant covariates, risk factors for inappropriate discharge on PPI included having an upper endoscopy (odds ratio [OR] 1.70, 95% confidence interval [CI] 1.08-2.66), admission to the surgical compared to the medical ICU (OR 2.03, 95% CI 1.32-3.10), and discharge to a nursing (OR 1.43, 95% CI 1.04-1.96) or rehabilitation facility (OR 2.29, 95% CI 1.62-3.24) compared to home (Table 1). Of all patients who were initiated on a PPI in the ICU, 16% were discharged on a PPI without an appropriately documented indication. Among all patients without an indication for outpatient PPI use, 27% were nonetheless discharged on a PPI. Risk factors for discharge on an inappropriate PPI included admission to the surgical ICU, receiving an upper endoscopy, and discharge to a nursing or rehabilitation facility. Medically complex ICU patients are at increased risk of receiving PPIs without appropriate documented indications, and careful review of medication lists at discharge should occur in these high-risk groups.Table-1Multivariable Model of Predictors of Discharge on a PPI, among Patients Initiated on one in the ICU without Long-Term IndicationVariableN discharged on Inappropriate PPI/ N at risk (%)Adjusted odds ratio (95% confidence interval)p-valueICU where PPI started Medical ICU84/299 (28%)1 (reference) Cardiac ICU99/273 (36%)1.31 (0.90-1.92)0.16 Surgical ICU191/426 (45%)2.03 (1.32-3.10)<0.01 Cardiothoracic ICU98/909 (11%)0.33 (0.21-0.51)<0.01 Neuro ICU196/560 (35%)1.49 (0.79-2.83)0.22Discharge destination Home186/918 (20%)1 (reference) Home with services223/830 (27%)1.27 (0.98-1.63)0.07 Rehabilitation facility115/262 (44%)2.29 (1.62-3.24)<0.01 Skilled nursing facility133/406 (33%)1.43 (1.04-1.96)0.03 Other6/24 (25%)0.97 (0.37-2.59)0.96 Hospice5/27 (19%)0.60 (0.22-1.68)0.33Admit Service Medicine216/756 (29%)1 (reference) Neurology74/267 (28%)0.51 (0.27-0.96)0.04 Surgery255/1095 (23%)1.22 (0.87-1.71)0.26 Neurosurgery108/261 (41%)1.03 (0.54-1.98)0.93 Other15/88 (17%)0.32 (0.17-0.60)<0.01Underwent upper endoscopy No621/2,369 (26%)1 (reference) Yes47/98 (48%)1.70 (1.08-2.66)0.02Number of Discharge Medications 0 to 7265/1,257 (21%)1 (reference) 8 to 10155/579 (27%)1.66 (1.29-2.14)<0.01 >10248/631 (39%)2.40 (1.88-3.06)<0.01Number of Locations 1 to 3352/1,210 (29%)1 (reference) 4155/589 (26%)0.83 (0.64-1.07)0.15 >4161/668 (24%)0.70 (0.54-0.91)<0.01Length of Stay (days) 1 to 7198/931 (21%)1 (reference) 8 to 15201/788 (26%)1.07 (0.82-1.39)0.64 >15269/748 (36%)1.56 (1.15-2.12)<0.01Intubated during admission No282/752 (38%)1 (reference) Yes386/1,715 (23%)0.57 (0.44-0.73)<0.01 Open table in a new tab