Cohort studycited by the Tayebat archivefull text checked★ key study A-PPI-01 in-depth review

PPI use and chronic kidney disease

Proton Pump Inhibitor Use and the Risk of Chronic Kidney Disease.

Overview

Lazarus B et al.. JAMA Intern Med 2016 · PMID 26752337 · DOI 10.1001/jamainternmed.2015.7193

PPI users had about 1.2–1.5 times the rate of new chronic kidney disease compared with non-users or H2-blocker users.

Does the study support the specific claim? Partly · Methodological quality: moderate

Original publication

Plain-language summary

Study explained simply

What did the researchers want to find out?: Test whether PPI use is linked to new CKD.
How was the study done?: Follow-up of two large US cohorts.
How many people, animals or samples were studied?: 10,482 plus 248,751 adults with normal kidney function.
What was actually found?: PPI use was associated with more new CKD.
How large was the observed effect?: Adjusted HR 1.50 (ARIC), 1.24 (Geisinger).
What do the results mean in everyday life?: Long-term PPI users may have a slightly higher kidney risk, or may simply be sicker.
What are the study's limitations?: Observational; confounding; outcome definitions differ.
What does it say about Elawady's specific claim?: Supports an association, not a proven harm.
Do the results apply to Germany and Europe?: Applicable as association.
What can explicitly not be concluded from this study?: Cannot show that PPIs cause CKD.

Scientific in-depth review

1. Full reference

Authors: Lazarus B, Chen Y, Wilson FP, Sang Y, Chang AR, Coresh J, Grams ME
Journal: JAMA Intern Med
Year: 2016
Identifier: PMID 26752337 DOI 10.1001/jamainternmed.2015.7193 · PMC4772730
Verification: verified – PubMed E-utilities efetch, 2026-10-25
Retracted: no
Corrections: none
Extent checked: full text
Conflicts of interest and funding: NHLBI contracts (ARIC), NIDDK K08DK092287 to one author; USRDS data; no conflict-of-interest statement included in the available author manuscript text.

2. Research question

Is PPI use associated with incident CKD?

3. Study design and population

Study design: Cohort study – Prospective population cohort (ARIC) with replication in an administrative cohort (Geisinger)
Population: ARIC: community-based adults (visit 4, mean age ~62.5 y) from four US communities with eGFR >=60; Geisinger: patients of a rural Pennsylvania health system with outpatient eGFR >=60 (mean age ~50 y)
Country: USA
Period: ARIC baseline 1996-1999, follow-up to 31 Dec 2011 (median 13.9 years); Geisinger 13 Feb 1997 - 9 Oct 2014 (median 6.2 years)
Sample size: ARIC 10,482 (322 baseline PPI users, 956 H2RA users); Geisinger 248,751 (16,900 PPI users, 6640 H2RA users)
Inclusion and exclusion criteria: eGFR >=60 mL/min/1.73 m2; ARIC complete-case (excluded 215 missing eGFR/ACR, 725 eGFR <60, 234 missing covariates)

4. Methods and endpoints

Exposure / intervention: ARIC: PPI use from visual inspection of medication bottles at baseline (updated by telephone from 2006); Geisinger: outpatient PPI prescription within 90 days before baseline, once vs twice daily
Comparator: Non-users; H2RA users (active comparator/negative control); propensity-matched non-users
Endpoints: Incident CKD: ARIC hospital discharge/death ICD codes or ESRD (USRDS) - codes have 35.5% sensitivity, 95.7% specificity; Geisinger sustained outpatient eGFR <60 or ESRD. Secondary: AKI
Statistical methods: Cox models adjusted for demographics, socioeconomic status, eGFR, ACR, smoking, BP, BMI, comorbidities, medications; time-varying ever-use; H2RA active comparator and negative control; propensity-score matching; new-user design (Geisinger); exclusion of albuminuria; 10-year absolute risk differences

5. Actual results

Results: ARIC: 56 CKD events in 322 PPI users (14.2/1000 person-years) vs 1382 in 10,160 non-users (10.7/1000 py); unadjusted HR 1.45 (1.11-1.90); adjusted HR 1.50 (1.14-1.96); time-varying 1.35 (1.17-1.55); vs H2RA 1.39 (1.01-1.91); propensity-matched 1.76 (1.13-2.74). Geisinger: 20.1 vs 18.3/1000 py; adjusted HR 1.17 (1.12-1.23); time-varying 1.22 (1.19-1.25); new-user 1.24 (1.20-1.28); vs H2RA 1.29 (1.19-1.40); twice-daily 1.46 (1.28-1.67) vs once-daily 1.15 (1.09-1.21). H2RA use not associated with CKD (ARIC 1.15, 0.98-1.36; Geisinger 0.93, 0.88-0.99). AKI: ARIC adjusted HR 1.64 (1.22-2.21).
Absolute risks: 10-year CKD risk among baseline PPI users: ARIC 11.8% vs expected 8.5% without PPI (absolute difference 3.3%); Geisinger 15.6% vs 13.9% (1.7%)
Authors' conclusion: PPI use is associated with a higher risk of incident CKD; future research should evaluate whether limiting PPI use reduces CKD.

6. Strengths

Two independent cohorts with different outcome definitions
Medication exposure confirmed by pill-bottle inspection in ARIC
Active comparator (H2RA), negative control, propensity matching and new-user design
Absolute risk differences reported

7. Weaknesses

Observational; PPI users more obese, hypertensive and on more medications - residual confounding by indication
Only 322 baseline PPI users in ARIC; exposure largely single baseline (updates from 2006)
Low sensitivity (35.5%) of ARIC CKD codes; possible surveillance bias
OTC use may cause misclassification
Dose comparison (twice vs once daily) only in Geisinger

8. Risk of bias

Orienting tool: ROBINS-E
Type of assessment: structured assessment based on the full text by a single reviewer (no formal double review)
Full formal assessment: no
Overall risk of bias: some concerns
Rationale: Single-reviewer structured ROBINS-E judgement from full text; not formal (needs two independent reviewers and protocol).

8. Risk of bias
Domain	Judgement	Rationale
Confounding	some concerns	Extensive adjustment, active comparator and propensity matching, but confounding by indication and frailty cannot be excluded.
Exposure measurement	some concerns	Baseline exposure only until 2006 in ARIC; prescription-based in Geisinger; OTC use unrecorded.
Selection of participants	some concerns	Prevalent users in main analyses (new-user design only in Geisinger); complete-case analysis (multiple imputation reported as consistent).
Post-exposure interventions	low	Time-varying exposure analyses performed.
Missing data	low	Small exclusions; imputation did not change inference.
Outcome measurement	some concerns	Insensitive discharge codes in ARIC; differential testing frequency possible in Geisinger.
Selection of reported result	low	Prespecified sensitivity analyses reported consistently, including negative control.

Full-text review (25 October 2026): Added event rates, 10-year absolute risk differences (3.3% ARIC, 1.7% Geisinger), pill-bottle exposure assessment, low sensitivity of CKD codes (35.5%), H2RA negative-control results, AKI results, follow-up durations and funding; ROBINS-E domains (overall unchanged).

9. Assessment of the statistical interpretation

Modest relative associations (HR 1.17-1.50) with small absolute risk differences (1.7-3.3 percentage points over 10 years); robust across sensitivity analyses, but magnitudes in this range remain vulnerable to residual confounding. The study does not address mortality.

10. Relevance to the specific claim

Directly addresses CKD claim (C-PPI-01).

11. Independent contrary findings

COMPASS RCT (I-PPI-01) found no significant CKD excess over 3 years.

12. Transferability to Germany and Europe

High: European (Swedish, UK, Dutch, German) data show similar associations; PPIs are widely used in Germany.

13. Overall methodological judgement

Well-conducted cohort showing a modest association; does not prove causation.

Multidimensional assessment
Multidimensional assessment
Authenticity of the source	verified
Methodological quality	moderate
Risk of bias	some concerns
Precision of results	moderate
Directness	direct
Reproducibility	replicated
Consistency with the overall evidence	consistent
Regional transferability	high
Adequacy of the interpretation (by the archive)	partly correct but over-interpreted

14. Open questions

Would de-prescribing reduce CKD incidence?

15. Linked fact checks

Long-term PPI use causes chronic kidney disease and excess deaths (about 45 extra deaths per 1,000 users, mainly from cardiovascular disease, kidney disease and upper gastrointestinal cancer). – Studies supporting the claim

Use by the Tayebat archive

Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.

What the archive says about the study: ARIC cohort (n=10,482): PPI use linked to 35–50% higher incident CKD risk; twice-daily dosing nearly doubles risk versus once-daily (HR 1.46 vs 1.15).
A. What was actually studied?: Association of PPI use with incident CKD in two US cohorts.
B. What did the authors find?: Adjusted HR 1.50 (ARIC) and 1.17-1.24 (Geisinger); 10-year absolute risk difference 3.3% and 1.7%; twice-daily vs once-daily dosing HR 1.46 vs 1.15 in Geisinger only.
C. What conclusions do the authors themselves draw?: Association; called for research on whether limiting PPIs reduces CKD.
D. What does the archive derive from it?: The archive presents the association as a harm caused by PPIs and attributes the dosing result to ARIC.
E. Is this derivation scientifically justified?: Partly: the numbers are correct, but the dosing comparison comes from the Geisinger cohort, 1.46 vs 1.15 is not 'nearly double', the absolute risk differences (1.7-3.3% over 10 years) are omitted, and causation is not shown. The study provides no mortality data for the '45 extra deaths per 1,000' part of the claim.
Adequacy of the interpretation (by the archive): partly correct but over-interpreted