Discontinuation of Proton Pump Inhibitors in Patients With Chronic Kidney Disease

Abstract

Purpose: Proton pump inhibitors (PPIs) are commonly used medications and are historically well tolerated. Recent studies have linked PPI use to the development of chronic kidney disease (CKD) and end-stage renal disease. This study investigated the impact of discontinuing PPIs on renal function in patients with CKD. Methods: We conducted a retrospective chart review of patients with established CKD, defined as 2 eGFR (estimated glomerular filtration rate) measurements of less than 60 mL/min/1.73 m² at least 90 days apart, who were on a PPI from January 1, 2014 to December 31, 2014, with a medication possession ratio greater than or equal to 70%. We compared baseline eGFR to a final eGFR after at least 6 months of discontinuation or continuation of a PPI. After power analysis, we targeted an enrollment of 200 patients (100 in each group) to achieve a power of 0.80 and an alpha of 0.05. Summary: A total of 97 patients in the PPI discontinuation group and 100 patients in the PPI continuation group met the study inclusion criteria. Baseline eGFR in the PPI continuation group was 47.9 mL/min/1.73 m² and 50.7 mL/min/1.73 m² in the discontinuation group. Final eGFR in the PPI continuation group was significantly higher than baseline at 51.1 mL/min/1.73 m² (+3.25 ± 12.8, P = .01). Final eGFR in the PPI discontinuation group was 51.8 mL/min/1.73 m² (+1.09 ± 12.8, P = .3). The average time between baseline and final eGFRs was 270 days in the PPI continuation group and 301 days in the discontinuation group. There was no statistically significant difference in the change in eGFRs between groups (95% confidence interval [CI] = −5.48-2.03, P = .37). Conclusions: Proton pump inhibitor discontinuation after prolonged continuous use in patients with CKD was not associated with a significant change in renal function after 1 year.

Introduction

Proton pump inhibitors (PPIs) are commonly used medications in the United States, often with a duration of use that exceeds guideline-based recommendations. Although generally well tolerated, PPIs have been linked to a myriad of adverse effects including hip fracture, community-acquired pneumonia, hyponatremia, hypomagnesemia, and acute kidney injury resulting from acute interstitial nephritis (AIN).^1-5

Recently, in large database studies, an association between PPI usage and chronic kidney disease (CKD) has emerged.^6-8 In a retrospective study using the Veterans Affairs National Database, Xie and colleagues compared 173 321 PPI users with 20 270 Histamine 2 receptor antagonist (H2RA) users after 5 years of therapy. In this study, patients on PPIs were 22% more likely to drop below an estimated glomerular filtration rate (eGFR) of 60 mL/min/1.73 m² and 28% more likely to develop CKD compared with those on H2RAs. It was also shown that the AKI risk was increased in PPI users (hazard ratio [HR] = 2.15, 95% confidence interval [CI] = 2.00-2.32), supporting concerns for a link between PPI-induced AIN and development and progression of CKD. ⁶ In an ancillary study of the Atherosclerosis Risk in Communities Study population, Lazarus et al examined the association of self-reported PPI use with the development of CKD at the 2-year follow-up visit in 10 482 patients without baseline CKD. The study found that PPI users had greater risk of developing CKD compared with nonusers or H2RA users (adjusted HR = 1.45, 95% CI = 1.14-1.95, P = .006). The absolute risk for the development of CKD in PPI users was 3.3%. Proton pump inhibitor use was also associated with AKI risk (adjusted HR = 1.64, 95% CI = 1.22-2.21). ⁸ Peng et al performed a review of patients in Taiwan with established CKD on PPI therapy to analyze the risk of progression to end-stage renal disease (ESRD). After nearly 4 years of follow-up, CKD patients on a PPI had a significantly higher risk of progression to ESRD (adjusted odds ratio [OR] = 1.88, 95% CI = 1.71-2.06, P < .001). ⁹ Taken together, these retrospective and observational studies lend support to and heighten concern for the apparent increased risk of CKD and progression of CKD in patients using PPIs.

It is not clear what happens to renal function on cessation of chronic PPI use. Initial reports of PPI-induced AIN conveyed persistent renal dysfunction after PPI cessation.^10,11 The study by Xie and colleagues suggests that a longer duration off PPI use is correlated with worse renal outcomes. ⁶ It is unknown if extended time of PPI use leads to an improvement in renal function. In the present analysis of established CKD patients in a Veterans Affairs medical center, we aimed to assess the difference in renal function (eGFR) between those who have PPIs discontinued versus those with continued PPI therapy after a 1-year period.

Materials and Methods

Patient Selection

The VA Northeast Ohio Healthcare system (VANEOHS) services more than 100 000 veterans, with an estimated 12 751 of those veterans having CKD. It is estimated that about a third of these patients are on a PPI. We performed a retrospective, single-center study at the VANEOHS to assess the impact of discontinuing chronic PPI therapy on renal function in patients with stage 1 to 5 CKD not on dialysis. The study was reviewed and approved by the Institutional Review Board at the VANEOHS.

Patients were included for analysis if they had baseline CKD and at least 1 year of continuous PPI use from January 1, 2014 to December 31, 2014. Exclusion criteria included dialysis dependence during the screening period or if eGFR values were unavailable. Manual chart reviews were conducted using the VA’s electronic medical record system, Computerized Patient Record System (CPRS), to determine eligibility for study inclusion. Patients were then placed in either of 2 groups: PPI continuation group and PPI discontinuation group (Figure 1). Patients included in the PPI continuation group had an additional year of continued PPI use (entire calendar year of 2015). They were excluded if PPI therapy was changed during this time. Patients included in the PPI discontinuation group were those that stopped their PPI anytime in 2015 for a minimum of 180 days.

Figure 1.

Screen and selection.

Note. The most common reason for exclusion was not meeting the definition of PPI discontinuation, which accounted for 57% of excluded patients. The next most common reasons for exclusion were not meeting the definition of CKD (16%) and missing eGFR data (14%). PPI = proton pump inhibitor; CKD = chronic kidney disease; eGFR = estimated glomerular filtration rate.

Definitions

Continuous PPI use was defined as a medication possession ratio (MPR) greater than or equal to 70%, indicating that patients had access to their medication at least 70% of the time. Chronic kidney disease was defined as 2 eGFR values of less than 60 mL/min/1.73 m² at least 90 days apart using the Modification of Diet in Renal Disease (MDRD) equation. Demographic data and baseline and final eGFR values were collected for both groups. For patients in the discontinuation group, baseline eGFRs were defined as the value within 180 days before the date of discontinuation, using the value closest to the date of discontinuation. For the PPI continuation group, the baseline eGFR value was defined as the eGFR value within 180 days prior to January 1, 2015, using the value closest to January 1, 2015. The final eGFR values for the both groups were defined as the eGFR value at least 180 days and up to 365 days from the date of the baseline eGFR, using the value closest to 365 days (Figure 2).

Figure 2.

Group timelines.

Note. The figure shows the time point for data collection in both groups. The screening period was from January 1, 2014 to December 31, 2014. During this time, all patients had to meet inclusion criteria for enrollment. Patients were then divided into 2 groups: the discontinuation group and the continuation group. (a) Illustrates the timeline for eGFR value selection in the PPI discontinuation group. For each patient, baseline and final eGFR values were based on the discontinuation date of the PPI. If a patient stopped a PPI on July 14, 2015, the baseline eGFR collected was the value closest to the stop date within 6 months of July 14, 2015. The final eGFR was collected 1 year after the stop date. (b) The timeline for eGFR value selection in the PPI continuation group. All baseline eGFRs were collected on January 1, 2015 (value closest to this date, but had to be within 6 months of this date). The final eGFR was selected between July 2, 2015 and January 1, 2016, using the value closest to January 1, 2016. These dates were chosen to closely assess changes in eGFR in after 1 full calendar year of continuing the PPI. eGFR = estimated glomerular filtration rate; PPI = proton pump inhibitor.

Statistical Analysis

A 2-group by 2-time-point repeated-measures analysis of variance (ANOVA) was used to analyze eGFR values for both groups. With the repeated-measures ANOVA, 3 different analyses were done. First, the between-group comparisons were used to determine eGFR differences between the PPI discontinuation and PPI continuation groups. Second, the within-group analysis assessed the mean eGFR changes from baseline to final eGFR within each group (PPI discontinuation and PPI continuation groups). Finally, interaction effects between groups and time were examined to identify statistically significant differences in the eGFR trend between the PPI discontinuation and PPI continuation groups. Descriptive statistics were used for baseline characteristics. G power analysis v 3.1.92 was used for calculating power. Using repeated-measures ANOVA, within factors, power analysis was calculated using an alpha of 0.05, a power of 0.80, and n = 200, with 2 groups and 2 measurements, to get a moderately small effect size of 0.09. Using repeated-measures ANOVA, between factors, power analysis as defined could detect a moderately small effect size of 0.17. Using repeated-measures ANOVA, within-between interaction, power analysis as defined could detect a moderately small effect size of 0.09.

Results

Baseline Characteristics

From January 1, 2014 through December 31, 2014, there were 19 187 patients identified as having concomitant CKD and chronic PPI use. Of those patients, 1600 charts were assessed to enroll 97 patients in the PPI discontinuation group and 100 patients in the PPI continuation group who met eligibility criteria for inclusion (Figure 1). Baseline characteristics were collected for all patients (Table 1). Both groups largely consisted of men (95%). There was similar presence of comorbid diseases between groups. Laboratory baseline parameters showed high albumin/creatinine ratio and lower magnesium levels in the PPI continuation group. The most common documented indication for PPI use was gastroesophageal reflux disease (GERD) for both the PPI discontinuation group (57%) and the PPI continuation group (66%). Omeprazole was the most frequently prescribed PPI.

Table 1.

Baseline Characteristics.

Characteristic	PPI discontinuation (n = 97)	PPI continuation (n = 100)
Sex—male, n (%)	93 (96)	97 (97)
Race, n (%)
White	82 (85)	90 (90)
African American	7 (7)	3 (3)
Mean age, y	72.9	73.9
Past medical history, n (%)
Diabetes	54 (56)	48 (48)
Hypertension	81 (84)	85 (85)
Heart failure	16 (16)	9 (9)
CAD	41 (42)	37 (37)
Upper GIB	3 (3)	10 (10)
GERD	55 (57)	66 (66)
PVD	11 (11)	6 (6)
Barrett esophagus	0 (0)	3 (3)
Helicobacter pylori	0 (0)	1 (1)
Esophageal stricture	0 (0)	1 (1)
Esophageal carcinoma	0 (0)	0 (0)
Baseline medications, n (%)
ACEi or ARB	45 (46)	40 (40)
Diuretic	38 (39)	46 (46)
Aspirin	30 (31)	32 (32)
NSAIDs	4 (4)	14 (14)
Oral anticoagulants	17 (18)	16 (16)
Serum creatinine (SCr), mg/dL	1.52	1.52
Magnesium (Mg), mg/dL	3.11	1.91
Albumin/creatinine, mg/g	147.00	248.67
Baseline MPR, %	95	99
PPI, n (%)
Omeprazole	87 (90)	91 (91)
Other	10 (10)	9 (9)

Note. PPI = proton pump inhibitor; CAD = coronary artery disease; GIB = gastrointestinal bleeding; GERD = gastroesophageal reflux disease; PVD = peripheral vascular disease; ACEi = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; NSAIDs = nonsteroidal anti-inflammatory drugs; MPR = medication possession ratio.

Changes in eGFR

The average baseline eGFR was 50.7 (±13.6) mL/min/1.73 m² in the PPI discontinuation group and 47.9 (±11.5) mL/min/1.73 m² in the PPI continuation group. Final eGFR values for the PPI discontinuation and continuation groups were 51.8 (±15.5) and 51.1 (±17.3) mL/min/1.73 m², respectively. There was no statistically significant difference in baseline or final eGFR values between groups (P = .20; Table 2).

Table 2.

Renal Function Changes.

	PPI discontinuation (n = 97)	PPI continuation (n = 100)
Baseline eGFR (SD), mL/min/1.73 m2	50.7 (±13.6)	47.9 (±11.5)
Final eGFR (SD), mL/min/1.73 m2	51.8 (±15.5)	51.1(±17.3)
Change in eGFR (SD), mL/min/1.73 m2	1.09 (±10.7)*	3.25 (±12.8)**
Time between eGFRs, average days	301	270

Note. PPI = proton pump inhibitor; eGFR = estimated glomerular filtration rate.

^*P = .32. **P = .01.

The average change in eGFR was not statistically significant in the PPI discontinuation group (+1.09 mL/min/1.73 m², P = .32). There was a statistically significant increase in eGFR seen in the PPI continuation group (+3.25 mL/min/1.73 m², P = .013). However, there was no statistically significant difference seen in the change in eGFR between groups (95% CI = −5.48-2.03, P = .37; Figure 3).

Figure 3.

Estimated marginal difference Between eGFR Means.

Note. The figure shows the estimated marginal difference between eGFR means between the PPI continuation group and the PPI discontinuation group. No statistically significant difference was seen between groups. eGFR = estimated glomerular filtration rate; CI = confidence interval; PPI = proton pump inhibitor.

Discussion

Recently published studies highlight associations between PPI use and increased risk of developing CKD and progression of established CKD to ESRD.^6-9 These findings raise significant concern for populations that are often continued on PPI therapy for inappropriately long periods of time. Indeed, in this study, most of the patients were on extended-duration PPIs for GERD, which is not an established indication for long-term use.^1-5 Our study aimed to assess the effect of PPI discontinuation, after therapy for 1 full year, on renal function as measured by eGFR using the MDRD equation. ¹⁴ The results show that discontinuation of chronic PPI therapy did not have a significant impact on renal function over the subsequent year-long period compared with patients that continued PPIs. Both groups had an overall small increase in eGFR, which, although statistically significant in the PPI continuation group, would likely not be considered clinically significant. Based on our data and a relatively small sample size, there does not appear to be any immediate, hemodynamic effect from withdrawal of PPIs in patients with established CKD.

Several factors could have influenced this result. In the studied patient population, despite the high prevalence of diabetes and hypertension, there was effectively no change in renal function in either the PPI withdrawal or the control group during the follow-up period. This could certainly be a reflection of a relatively short follow-up period.

It is possible that certain patient populations may be at an increased risk of developing CKD and progression to ESRD from exposure to a PPI. However, given the retrospective nature of this study and lack of renal biopsy data, we were unable to parse out exact causes of CKD and thus could not ascertain whether any prestudy eGFR declines were solely due to PPI use. Moreover, the underlying comorbidities were similar between groups and withdrawal of the potential offending agent did not appear to impact renal function the first year after discontinuation.

Another possibility is that renal damage induced by chronic PPI use occurs relatively early in the course of therapy, where our study only collected PPI prescribing data from January 1, 2014 to December 31, 2016, and therefore did not take into account the exact duration of total PPI exposure in our patient population. In support of this concept, Xie et al ⁶ found that a longer duration of PPI exposure led to a higher risk of developing adverse renal outcomes, but their data suggested that this risk may diminish when PPI exposure exceeds 720 days. Finally, although the exact mechanism is not fully understood, PPI-induced AIN has been documented by case series, which, although certainly being a cause of AKI, may also contribute to the development and progression of CKD that may not be readily reversible with PPI withdrawal.^6-8,10,11-13

We identified several limitations to our study. Proton pump inhibitors are available over the counter, and therefore patients could have been taking additional PPIs apart from their prescribed medications. Moreover, patients included in the discontinuation group may have continued PPIs from outside sources, which was unable to be assessed in our study. Second, eGFR follow-up of up to 1 year may not be adequate to see a change in renal function after stopping PPIs. It is unclear how long it would take to see any impact from discontinuing the medication. Similarly, the sample size used to calculate power may not have been large enough to detect a difference. In addition, comorbid conditions were obtained through documentation in the patient problem list, and therefore the accuracy of this information is dependent on the accuracy of the provider. It is possible that a significant difference in confounding comorbid conditions could exist between groups that we did not capture. Also, patients could have been on long-term PPIs for appropriate indications that were not documented in the problem list.

Conclusions

In conclusion, this study showed that discontinuing chronic PPIs in patients with baseline CKD did not influence the change in renal function over 1 year compared with patients continued on PPIs. Larger studies would be warranted to further assess whether PPI discontinuation in patients with CKD impacts progression of CKD and development of ESRD.