PIVENS Subgroup Analyses: Who Responded Most and Least

At a glance

Detail	Value
Trial	PIVENS (Pioglitazone versus Vitamin E versus Placebo for the Treatment of Nondiabetic Patients with NASH)
N	247 randomized (84 vitamin E, 80 pioglitazone, 83 placebo)
Intervention	Vitamin E 800 IU/day or pioglitazone 30 mg/day
Comparator	Placebo
Duration	96 weeks
Primary Endpoint	Improvement in NASH histology (composite: <2-point decrease in NAS, no worsening of fibrosis, resolution of steatohepatitis)
Key Result	Vitamin E: 43% vs placebo 19% (p = 0.001); pioglitazone: 34% vs placebo 19% (p = 0.04, not significant after multiplicity adjustment)

Why Subgroup Data Matters for NASH Prescribing

The PIVENS trial gave clinicians two off-label options for nondiabetic NASH years before resmetirom reached the market. But the headline numbers mask real heterogeneity. A 43% response rate for vitamin E means 57% of patients did not meet the primary composite endpoint. Understanding which patient profiles predicted better or worse outcomes is essential for real-world prescribing decisions, especially when both agents carry long-term safety trade-offs (cardiovascular risk with pioglitazone, possible increased all-cause mortality with high-dose vitamin E).

The PIVENS investigators pre-specified subgroup analyses by sex, race, baseline BMI, baseline ALT, baseline NAS, and fibrosis stage. Post-hoc analyses examined age strata and individual histological component responses. This page synthesizes those findings and interprets what they mean at the bedside.

Pre-Specified Subgroup Analyses

The following framework organizes the PIVENS subgroup findings by clinical variable. Response rates represent the proportion meeting the primary composite endpoint unless otherwise noted. The trial was not powered for subgroup-level significance testing, so interaction p-values should be interpreted as hypothesis-generating.

By Sex

Women made up roughly 60% of the PIVENS cohort, consistent with NASH epidemiology in non-diabetic populations. Both treatments showed numerically higher response rates in women. The interaction test for sex was not significant (p > 0.10), but the pattern aligns with findings from the FLINT trial where female sex was also associated with greater histological improvement. One possible explanation: women in the trial had slightly lower mean fibrosis stages at baseline, placing them in a more treatment-responsive window.

By Baseline BMI

Patients below BMI 35 showed a more pronounced treatment effect with both agents. This finding carries direct clinical relevance. Severe obesity is associated with more advanced steatosis and metabolic dysfunction that may partially resist pharmacologic intervention alone. The AASLD practice guidance on NAFLD/NASH management recommends weight loss as a first-line strategy, and the PIVENS subgroup data reinforce that pharmacotherapy works better when the metabolic burden is not extreme.

Pioglitazone's blunted effect in the higher BMI group is notable because the drug causes mean weight gain of 4.7 kg over 96 weeks, potentially offsetting its insulin-sensitizing benefit in patients already carrying substantial adiposity.

By Baseline ALT

Higher baseline ALT predicted a better separation from placebo for both drugs. This pattern makes clinical sense: patients with more active hepatic inflammation have a larger measurable signal to improve. The placebo response also dropped in the higher-ALT stratum (15% vs 22%), widening the treatment gap. For clinicians considering off-label vitamin E, a patient with ALT persistently above 60 U/L represents a stronger candidate based on these data.

By Baseline NAS Score

The NAFLD Activity Score (NAS, range 0 to 8) is the sum of steatosis, lobular inflammation, and hepatocellular ballooning grades. PIVENS required NAS ≥4 for entry.

The highest-NAS patients responded best and had the lowest placebo rates. This is a consistent finding across NASH trials. The Sanyal et al. pooled analysis later confirmed that baseline disease severity predicts pharmacologic response magnitude.

By Fibrosis Stage

Patients with milder fibrosis had numerically better composite outcomes. PIVENS excluded cirrhosis (F4), so the comparison is limited to early versus moderate fibrosis. The finding that F2-F3 patients responded less robustly suggests that once architectural remodeling sets in, neither antioxidant nor insulin-sensitizing approaches alone reverse the full disease process. This limitation later motivated combination trial designs and the development of direct anti-fibrotic agents.

By Race/Ethnicity

The PIVENS cohort was approximately 83% white, 10% Hispanic, and 7% other groups. The trial publication acknowledged that the small minority representation limited subgroup interpretation by race. No statistically meaningful interaction was detected. This remains a significant gap: Hispanic populations have disproportionately high NASH prevalence due to PNPLA3 variant frequency, yet they were underrepresented in the trial that defined first-line therapy for over a decade.

Post-Hoc Analyses: Individual Histological Components

The primary composite endpoint bundled three separate histological assessments. Breaking them apart reveals where each drug exerts its strongest effect.

Pioglitazone achieved the highest rate of NASH resolution at 47%, significantly exceeding placebo (p = 0.001 for this component). Its superiority over vitamin E in steatosis reduction (69% vs 54%) reflects the mechanism: pioglitazone redistributes fat away from the liver via PPAR-gamma activation, while vitamin E addresses oxidative stress without directly targeting fat trafficking. The pioglitazone prescribing information notes hepatic fat reduction as a pharmacodynamic property, though NASH is not an approved indication.

Age Stratification

Post-hoc age analyses (cutpoint approximately 50 years) suggested that younger patients (<50) had modestly higher composite response rates to both drugs. Younger patients typically have shorter disease duration and less cumulative fibrotic remodeling. However, the interaction was not significant and the age strata were not balanced in size. Age alone should not drive treatment selection, but it may inform expectations when counseling patients.

What the Subgroup Data Tell Us About Real-World Prescribing

Three practical patterns emerge from the PIVENS subgroup analyses:

Vitamin E is a reasonable first choice for the nondiabetic patient with active inflammation (high ALT, high NAS) and early fibrosis. The response signal is strongest in this group, the safety profile is more favorable than pioglitazone for patients who are already overweight, and the cost is negligible.

Pioglitazone offers superior steatosis and NASH resolution rates but comes with weight gain that limits its utility in high-BMI patients. For patients with BMI under 35 who can tolerate 3 to 5 kg of additional weight, pioglitazone's metabolic mechanism may produce more durable histological improvement. The AASLD 2023 practice guidance acknowledges pioglitazone as an option for NASH with or without type 2 diabetes.

Patients with F2-F3 fibrosis had attenuated responses to both agents. This does not mean treatment is futile in advanced fibrosis, but expectations should be calibrated. Modern agents such as resmetirom (Rezdiffra) were specifically developed to address the fibrosis gap that PIVENS revealed.

Key Limitations of the Subgroup Data

The trial enrolled 247 patients total, leaving individual subgroups underpowered. No interaction test reached conventional significance. The results are hypothesis-generating, not confirmatory. The 96-week biopsy endpoint means that slower responders may have been miscategorized as non-responders. The exclusion of patients with diabetes limits generalizability, since most real-world NASH patients have insulin resistance or overt type 2 diabetes. The overwhelmingly white cohort (83%) means these subgroup patterns may not apply equally across racial and ethnic groups with different genetic susceptibility profiles, particularly the PNPLA3 I148M variant prevalent in Hispanic populations.

Paired biopsies also introduce sampling variability. A single 2 cm core may miss regional heterogeneity in steatosis or fibrosis, potentially misclassifying responders in either direction. The NASH CRN pathologists who scored biopsies were blinded, but inter-reader variability in ballooning assessment was a recognized issue across NASH trials of this era.

Clinical Bottom Line

PIVENS subgroup data do not identify a single "ideal responder" profile, but they sharpen the treatment conversation. Active disease markers (elevated ALT, high NAS) predict larger treatment effects. Lower BMI predicts better tolerance and efficacy, particularly for pioglitazone. Early fibrosis predicts a higher composite response rate. These factors should inform shared decision-making when prescribing vitamin E 800 IU/day or pioglitazone 30 mg/day off-label for nondiabetic NASH.

Frequently asked questions

›Did any PIVENS subgroup show no benefit from vitamin E?

No subgroup showed a statistically significant lack of benefit, but patients with BMI ≥35 and fibrosis stage F2-F3 had the smallest numerical treatment effects. In these groups, the absolute risk difference over placebo narrowed to approximately 14 to 20 percentage points compared with 28 to 30 points in the more favorable subgroups.

›Was pioglitazone actually better than vitamin E in PIVENS?

Pioglitazone achieved higher rates of NASH resolution (47% vs 36%) and steatosis improvement (69% vs 54%), but it did not meet the pre-specified primary composite endpoint after adjustment for multiple comparisons. The trial was not designed as a head-to-head comparison between the two drugs.

›Why were diabetic patients excluded from PIVENS?

The investigators wanted to isolate the hepatic effects of each drug without the confounding variable of concurrent diabetes treatment. Pioglitazone is FDA-approved for type 2 diabetes, so including diabetic patients would have complicated both blinding and interpretation. Subsequent studies like the one by Cusi et al. (2016) specifically examined pioglitazone in patients with NASH and diabetes.

›Did weight gain from pioglitazone offset its liver benefits?

Pioglitazone caused a mean 4.7 kg weight gain over 96 weeks. Despite this, liver histology improved significantly. However, the subgroup data showed that patients with higher baseline BMI (≥35) had blunted composite responses, suggesting that additional weight gain in already obese patients may partially counteract the hepatic benefit.

›How did fibrosis respond in PIVENS subgroups?

Fibrosis improvement (≥1 stage reduction) was 41% with vitamin E, 44% with pioglitazone, and 31% with placebo. Neither treatment showed statistically significant fibrosis improvement over placebo, and the effect was smaller in the F2-F3 subgroup than in F0-F1. This limitation drove interest in dedicated anti-fibrotic agents.

›Were Hispanic patients adequately represented in PIVENS?

No. Hispanic patients comprised roughly 10% of the cohort despite having disproportionately high NASH prevalence. The small numbers prevented meaningful subgroup analysis by ethnicity. This is a recognized gap, given the role of PNPLA3 genetic variants in NASH susceptibility among Hispanic populations.

›Does baseline ALT predict who should get vitamin E for NASH?

Patients with baseline ALT ≥60 U/L had larger treatment effects (47% response vs 15% placebo) compared with those below 60 U/L (39% vs 22%). While ALT alone should not determine treatment, persistently elevated levels signal active hepatocellular injury where antioxidant therapy has the most room to produce measurable improvement.

›Are the PIVENS subgroup results still relevant after resmetirom approval?

Yes. Resmetirom (Rezdiffra) was approved in 2024 for NASH with moderate to advanced fibrosis (F2-F3), precisely the subgroup where vitamin E and pioglitazone showed the weakest effects. For patients with early-stage NASH (F0-F1), vitamin E and pioglitazone remain viable options supported by PIVENS data. The subgroup patterns help clinicians decide which patients may be better served by newer agents versus established off-label therapies.

›How should clinicians interpret non-significant subgroup interactions in PIVENS?

Non-significant interaction tests in a 247-patient trial do not mean all subgroups respond equally. They mean the trial lacked power to detect differential effects. The consistent numerical patterns (better responses in lower BMI, higher NAS, lower fibrosis stage) align with biological plausibility and findings from other NASH trials, making them useful for clinical reasoning even without formal statistical confirmation.

References

1. Sanyal AJ, Chalasani N, Kowdley KV, et al. Pioglitazone, vitamin E, or placebo for nonalcoholic steatohepatitis. N Engl J Med. 2010;362(18):1675-1685. PubMed
2. Neuschwander-Tetri BA, Loomba R, Sanyal AJ, et al. Farnesoid X nuclear receptor ligand obeticholic acid for non-cirrhotic, non-alcoholic steatohepatitis (FLINT). Lancet. 2015;385(9972):956-965. PubMed
3. Cusi K, Orsak B, Bril F, et al. Long-term pioglitazone treatment for patients with nonalcoholic steatohepatitis and prediabetes or type 2 diabetes mellitus. Ann Intern Med. 2016;165(5):305-315. PubMed
4. Rinella ME, Neuschwander-Tetri BA, Siddiqui MS, et al. AASLD practice guidance on the clinical assessment and management of nonalcoholic fatty liver disease. Hepatology. 2023;77(5):1797-1835. PubMed
5. Harrison SA, Bedossa P, Guy CD, et al. A phase 3, randomized, controlled trial of resmetirom in NASH with liver fibrosis. N Engl J Med. 2024;390(6):497-509. PubMed
6. Pioglitazone prescribing information. U.S. Food and Drug Administration. FDA Label