Pioglitazone for sympathetic overactivity in CKD: the biomarker the trial was built on was not there

What was tried

The University of Texas at Arlington ran NCT03471117, "Targeting ADMA With Pioglitazone to Reduce Sympathetic Overactivity in CKD Patients," with UT Southwestern at Dallas as collaborator. The study started on 2018-04-01 and both primary completion and completion are recorded as 2020-01-14. The record was updated on 2026-07-27 as terminated.

The design was a randomized, placebo-controlled crossover with single masking. Each participant received pioglitazone 15 mg daily for one month and matching placebo for one month in randomized order. ClinicalTrials.gov classifies the trial as Phase 4, which normalizes to Unknown under the Claidex scheme. The stated purpose was basic science.

Eligibility required CKD at stage 3 or 4, with estimated glomerular filtration rate between 15 and 59 mL/min/1.73 m2 by the MDRD formula, in adults aged 35 to 70. The primary endpoint was muscle sympathetic nerve activity in bursts per minute, by microneurography of the peroneal nerve, with mean arterial pressure and flow-mediated dilation as secondary. Actual enrollment reached 5 participants. ChEMBL records pioglitazone as CHEMBL595, a small molecule, maximum phase 4, first approved in 1999.

The biological hypothesis

The chain of reasoning had four links. Chronic kidney disease raises circulating asymmetric dimethylarginine, an endogenous inhibitor of nitric oxide synthase. Elevated ADMA reduces nitric oxide availability. That disinhibits central sympathetic outflow. Pioglitazone, a PPAR-gamma agonist acting on PPARG, was expected to lower ADMA and therefore lower sympathetic nerve traffic.

Each link had published support. Sympathetic overactivity in CKD is well characterized (International Journal of Molecular Sciences 2017), arginine analogs suppressing nitric oxide production in CKD were reviewed two decades ago (Nature Clinical Practice Nephrology 2006), and the clinical behavior of ADMA and SDMA in CKD has been mapped since (International Journal of Molecular Sciences 2019). The specific link between sympathetic nerve traffic and ADMA in CKD was reported in 2011 (Clinical Journal of the American Society of Nephrology). Pioglitazone itself had been studied against muscle sympathetic nerve activity in type 2 diabetes (Autonomic Neuroscience 2010) and in obese participants (Journal of Clinical Endocrinology and Metabolism 2014).

Open Targets scores the PPARG association with chronic kidney disease at 0.3491, from a literature datatype score of 0.7423 and a clinical score of 0.5372, and lists 40 distinct clinical candidates against PPARG.

What actually happened

The investigators terminated the study and posted an unusually specific reason: the CKD participants they recruited did not have elevated ADMA, the main premise for the intervention. They noticed this in the first cohort, then collected blood from their other ongoing CKD studies, which they report further solidified the finding.

Results were posted. Four participants were analyzed, two randomized to pioglitazone first and two to placebo first, and all four completed both periods. Baseline age was 66 (standard deviation 3), all 4 were female, and baseline muscle sympathetic nerve activity was 41 bursts per minute (standard deviation 10).

On the primary endpoint, muscle sympathetic nerve activity was 43 bursts per minute (standard deviation 8) on pioglitazone against 41 (10) on placebo. Mean arterial pressure was 112 mmHg (26) against 110 (24). Flow-mediated dilation was 3.5 percent (1.9) against 3.7 percent (1.5). No between-group statistical test was posted. Both arms recorded 0 serious adverse events and 0 deaths among 4 participants at risk.

Failure mechanism, best guess

This is a clean biomarker failure. The intervention was never refuted, because it was never given a population in which it could act. Every downstream link was contingent on link one, elevated ADMA, which did not hold in the recruited cohort.

Recent independent work supports the investigators' reading. A 2025 study reported impaired vascular function in CKD patients with elevated symmetric dimethylarginine but not asymmetric dimethylarginine (American Journal of Physiology-Renal Physiology). The same observation from a different group: in contemporary CKD cohorts, the dimethylarginine that moves is often SDMA. If SDMA carries the vascular signal, an ADMA-lowering strategy targets the wrong methylarginine.

Cohort composition is a second contributor. All 4 analyzed participants were female, and sex differences in sympathetic activity in CKD have since been documented (American Journal of Physiology-Renal Physiology 2024). A baseline of 41 bursts per minute is not low, so these participants did have sympathetic overactivity, just not the proposed cause of it. That points toward renal afferent signaling rather than nitric oxide deficiency.

How to prevent this next time

Endpoint-level data exist, so part of this section is computed.

Computed from the posted means: 43 minus 41 is a difference of plus 2 bursts per minute on pioglitazone, opposite to the hypothesis. Pooled standard deviation, from the posted values 8 and 10 as the root mean square, is 9.06 bursts per minute, so the standardized difference is 2 divided by 9.06, or 0.22, in the wrong direction.

Illustrative: treating the two periods as independent groups of 4, at two-sided alpha 0.05 and 80 percent power, the smallest detectable difference is (1.96 plus 0.8416) times 9.06 times the square root of 2 divided by 4, or 17.9 bursts per minute. Assumptions: normality, independence between periods, pooled standard deviation 9.06 from the posted values, and no adjustment for crossover correlation, which a paired analysis would improve on. Against a baseline of 41, that threshold is a 44 percent reduction. No plausible one-month effect of pioglitazone 15 mg is that large, so this design could not have detected a real effect even in an ADMA-elevated population.

Biomarker enrichment is the decisive lever. ADMA is a plasma assay. Screening for it before randomization, rather than assuming it from stage 3 to 4 CKD, would have surfaced the problem at screening. The investigators ran exactly this check retrospectively, on blood from their other CKD studies.

The single highest leverage change would have been to make documented elevated ADMA an inclusion criterion measured at screening, converting a 5-participant terminated trial into a one-page feasibility finding.

What this means for similar programs

The Claidex Mechanism Risk Score for PPARG in this indication stands at 30, in the yellow band, from contributions of 4.70 for phase burden, 8.52 for archetype severity, 4.25 for recency, 9.76 for genetic deficit, and 2.50 for saturation. Genetic deficit is the largest term, reflecting an Open Targets association score of 0.3491 against a target that already carries 40 clinical candidates.

Two warnings transfer. First, repurposing trials built on a mechanistic biomarker should verify it in the intended population before randomizing anyone, especially when the supporting evidence is a decade old. Second, teams here should treat SDMA as a candidate mediator alongside ADMA. Programs pursuing renal denervation or exercise, where sympathetic modulation has been shown in CKD (JCI Insight 2023), do not depend on the methylarginine hypothesis.

PPARG agonism carries its own liability profile a cardiovascular indication must price in: Open Targets lists 19 PPARG safety liabilities including heart failure, water retention, and increased body weight.

Open questions

What were the measured ADMA concentrations in the first cohort and in the blood from the other CKD studies, and how did they compare with the reference values used to power the trial? Were SDMA levels measured in the same samples? Did the 2011 finding rest on a cohort with different dialysis status or stage distribution?

Sources

1. • ClinicalTrials.gov API v2, record NCT03471117 including posted results section, accessed 2026-07-30. https://clinicaltrials.gov/study/NCT03471117 - Open Targets Platform GraphQL API v4, target ENSG00000132170 (PPARG), association against MONDO_0005300, accessed 2026-07-30. https://platform.opentargets.org/target/ENSG00000132170 - ChEMBL API, CHEMBL595 (pioglitazone), accessed 2026-07-30. - openFDA drug event API (FAERS), pioglitazone queries, accessed 2026-07-30. - Sympathetic Nerve Traffic and Asymmetric Dimethylarginine in Chronic Kidney Disease. Clinical Journal of the American Society of Nephrology, 2011.- Impaired vascular function in patients with chronic kidney disease who have elevated symmetric dimethylarginine but not asymmetric dimethylarginine. American Journal of Physiology-Renal Physiology, 2025.- Sympathetic Overactivity in Chronic Kidney Disease: Consequences and Mechanisms. International Journal of Molecular Sciences, 2017.- Asymmetric (ADMA) and Symmetric (SDMA) Dimethylarginines in Chronic Kidney Disease: A Clinical Approach. International Journal of Molecular Sciences, 2019.- Arginine, arginine analogs and nitric oxide production in chronic kidney disease. Nature Clinical Practice Nephrology, 2006.- Effect of pioglitazone on muscle sympathetic nerve activity in type 2 diabetes mellitus with alpha-glucosidase inhibitor. Autonomic Neuroscience, 2010.- A Randomized Controlled Trial of the Effects of Pioglitazone Treatment on Sympathetic Nervous System Activity and Cardiovascular Function in Obese Subjects. Journal of Clinical Endocrinology and Metabolism, 2014.- Sex differences in sympathetic activity and pulse wave velocity in adults with chronic kidney disease. American Journal of Physiology-Renal Physiology, 2024.- Exercise modulates sympathetic and vascular function in chronic kidney disease. JCI Insight, 2023.