Pioglitazone

What it is

Pioglitazone is a synthetic thiazolidinedione - an odourless white crystalline powder, practically insoluble in water, taken as a 15, 30 or 45 mg tablet. It is made and used as a racemic mixture whose two forms interconvert in the body. It is an insulin sensitiser, not an insulin or an insulin-releasing drug.

What the research says

Pioglitazone makes muscle, fat and liver tissue more responsive to the insulin a person already makes, which lowers blood glucose. Beyond glucose, the outcome trials are unusually informative and cut both ways. In the PROactive trial its primary composite endpoint was not met, while the secondary endpoint of death, heart attack or stroke was reduced. In IRIS, in people without diabetes who had insulin resistance after a stroke or TIA, it cut stroke or heart attack from 11.8% to 9.0% over 4.8 years. Against that sit a boxed warning for congestive heart failure, more fractures in women, weight gain and oedema, and a bladder-cancer signal that two large cohort studies answer in opposite directions. In biopsy-proven non-alcoholic steatohepatitis, an off-label use, an 18-month randomised trial showed clear histological improvement.

Evidence grade: Well established.

How it works

Drug class: Thiazolidinedione; agonist at the peroxisome proliferator-activated receptor gamma (PPAR-gamma)

Pioglitazone switches on PPAR-gamma, a nuclear receptor in fat, muscle and liver. That changes which insulin-responsive genes are read, and the result is that the body's own insulin works better: tissues take up more glucose and the liver releases less. It only works if the person still makes insulin - it does not make the pancreas release any. (Source 1)

Boxed warning

· Thiazolidinediones, including pioglitazone hydrochloride tablets, cause or exacerbate congestive heart failure in some patients [see Warnings and Precautions (5.1)].

(Source 2)

What it is used for

• Glucose-lowering is well established; what the trials dispute is whether that translates into fewer cardiovascular events. A meta-analysis of 19 trials (16,390 patients) found death, myocardial infarction or stroke in 4.4% on pioglitazone versus 5.7% on control, alongside more serious heart failure (2.3% vs 1.8%). The label's own position is that no study has established conclusive macrovascular risk reduction. Evidence: established. (Source 3)
• The IRIS trial (3,876 patients, NIH-funded) found a primary outcome of fatal or non-fatal stroke or MI in 9.0% on pioglitazone versus 11.8% on placebo over 4.8 years - an absolute difference of 2.8 percentage points - with no mortality difference and clear excesses of weight gain, oedema and fracture. One trial, in a narrowly selected population. Evidence: limited. (Source 4)
• In a single-centre randomised trial of 101 patients, 58% on pioglitazone reached the primary histological endpoint, a treatment difference of 41 percentage points (95% CI 23 to 59), and 51% had resolution of NASH. It is one single-centre trial with 101 patients and more weight gain in the treated group. Evidence: limited. (Source 5)
• In IRIS, diabetes developed in 3.8% on pioglitazone versus 7.7% on placebo over 4.8 years. This was a secondary outcome in a trial designed for cardiovascular events, in people already selected for insulin resistance. Evidence: limited. (Source 4)

Interactions

• Gemfibrozil and other strong CYP2C8 inhibitors (pharmacokinetic study): Gemfibrozil roughly triples pioglitazone exposure, so the label caps the dose at 15 mg daily when they are combined. (Source 6)
• Rifampicin and other CYP2C8 inducers (pharmacokinetic study): These can markedly reduce pioglitazone exposure, so glucose control may drift when one is started or stopped. (Source 7)
• Topiramate (pharmacokinetic study): Pioglitazone and its active metabolites fall when topiramate is added; the clinical importance is unknown. (Source 8)
• Insulin and sulfonylureas (insulin secretagogues) (label): Pioglitazone alone rarely causes low blood sugar, but combined with insulin or a sulfonylurea it can, and the other drug's dose may have to come down. Fluid retention is also worst in combination with insulin. (Source 9)
• Food and meals (label): No meaningful food interaction: the label states the tablet can be taken without regard to meals. No grapefruit interaction is described, which fits the drug being cleared mainly by CYP2C8 rather than CYP3A4. (Source 10)
• Alcohol (theoretical): No alcohol interaction study for pioglitazone was found, and the US label describes none. The relevant sourced point is indirect: the drug requires baseline liver tests and should be interrupted if ALT exceeds three times the upper limit, so anything else that loads the liver matters. (Source 11)
• Supplements generally (theoretical): No documented pharmacokinetic interaction between pioglitazone and any dietary supplement was found in the searches made. The mechanistically plausible route is CYP2C8, so a supplement that inhibited or induced CYP2C8 would be expected to behave like gemfibrozil or rifampicin - but that expectation is theory, not a study. (Source 12)

Stopping it

• There is no withdrawal syndrome and no tapering requirement in the literature searched. What the label specifies instead is stopping for harm: if heart failure develops, discontinuation or dose reduction must be considered. (Source 2)
• For liver signals, the label's position is interruption at an ALT above three times the upper limit of the reference range, and not restarting without another explanation for the abnormality. (Source 13)
• For macular oedema, the label notes that some patients improved after the thiazolidinedione was stopped, which is the closest thing to a described reversal on discontinuation. (Source 14)

What goes wrong

In PROactive, hospital admission for heart failure was half again as common on pioglitazone, in absolute terms 6% versus 4%. (Source 15)

• Randomized trial, High certainty.
• Size: 5,238 patients.
• Who: Type 2 diabetes with macrovascular disease.
• How long: Average 34.5 months.
• Result: Heart-failure hospital admission 6% (149 patients) on pioglitazone vs 4% (108 of 2633) on placebo; heart-failure mortality did not differ between groups.
• Funding: industry-funded.

Limit of this finding: The published abstract misprints the pioglitazone denominator here as "149 of 2065". Everywhere else in the same abstract the pioglitazone group numbers 2,605, and 149 of 2,605 is 5.7%, which is what the 6% refers to; 149 of 2,065 would be 7.2%. The quote is left exactly as the journal printed it, but the 2065 figure should not be used as a denominator.

6% (149 of 2065) and 4% (108 of 2633) of those in the pioglitazone and placebo groups, respectively, were admitted to hospital with heart failure; mortality rates from heart failure did not differ between groups.

The same trial quantified the harms: weight gain over 4.5 kg, oedema and fracture requiring surgery or hospitalisation were all significantly more common than on placebo. (Source 4)

• Randomized trial, High certainty.
• Size: 3,876 patients.
• Who: Non-diabetic insulin-resistant stroke and TIA patients.
• How long: 4.8 years.
• Result: Weight gain over 4.5 kg 52.2% vs 33.7% (P<0.001); oedema 35.6% vs 24.9% (P<0.001); fracture requiring surgery or hospitalisation 5.1% vs 3.2% (P=0.003) - an absolute excess of 1.9 percentage points.
• Funding: independent (NINDS)

Pioglitazone was associated with a greater frequency of weight gain exceeding 4.5 kg than was placebo (52.2% vs. 33.7%, P<0.001), edema (35.6% vs. 24.9%, P<0.001), and bone fracture requiring surgery or hospitalization (5.1% vs. 3.2%, P=0.003).

A systematic review and meta-analysis of 78 randomised trials reported an association between pioglitazone and a higher fracture risk overall, concentrated in women and in low-energy fractures from falls, but the overall result held only in the fixed-effect analysis. (Source 16)

• Meta-analysis, Low certainty.
• Size: 78 trials met inclusion criteria out of 860 identified; 34 at high risk of bias, 8 unclear, 36 low.
• Who: Trials of pioglitazone versus placebo or other antihyperglycaemic drugs reporting fractures.
• How long: Varied.
• Result: Any fracture RR 1.21 (95% CI 1.01-1.45), P=0.04; serious fractures RR 1.48 (1.10-1.98); low-energy fractures from falls RR 1.49 (1.20-1.87); women RR 1.56 (1.20-2.02); no difference versus other antihyperglycaemics RR 1.08 (0.73-1.59)
• Funding: not stated.

Limit of this finding: Two cautions, both from the review itself. Its headline risk ratio of 1.21 was significant in the fixed-effect model but not in the random-effects model, so the overall result is fragile. And the sentence in which it reports subgroups puts a null result under the word "exacerbated": in people with insulin resistance the risk ratio was 0.87 (95% CI 0.43-1.76, P = 0.69), which shows no increase in fracture risk in that subgroup. Nothing here supports an increased fracture risk in insulin-resistant people.

The meta-analysis revealed an association between pioglitazone and a significant increase in fracture risk (risk ratio [RR] 1.21; 95% CI 1.01-1.45; P = 0.04), including non-serious (RR 1.25; 95% CI 1.03-1.51; P = 0.02) and serious fractures (RR 1.48; 95% CI 1.10-1.98; P = 0.01).

The same cohort found unexpected associations with prostate and pancreatic cancer that the authors could not interpret. (Source 17)

• Cohort study, Very low certainty.
• Size: 236,507 people in the 10-additional-cancers cohort.
• Who: Kaiser Permanente Northern California members with diabetes.
• How long: 1997-2005 to June 2012.
• Result: Prostate cancer HR 1.13 (95% CI 1.02-1.26); pancreatic cancer HR 1.41 (1.16-1.71); crude pancreatic incidence 81.1 vs 48.4 per 100,000 person-years; no dose or duration pattern for any cancer.
• Funding: not stated.

ever use of pioglitazone was associated with increased risk of prostate cancer (HR, 1.13; 95% CI, 1.02-1.26) and pancreatic cancer (HR, 1.41; 95% CI, 1.16-1.71).

A UK population-based cohort of 145,806 people reached the opposite conclusion on bladder cancer, with a dose and duration response, and found no such signal for rosiglitazone. (Source 18)

• Cohort study, Moderate certainty.
• Size: 145,806 patients newly treated with antidiabetic drugs; 689,616 person-years; 622 incident bladder cancers.
• Who: UK Clinical Practice Research Datalink, type 2 diabetes.
• How long: 2000-2013 with follow-up to July 2014, exposure lagged one year.
• Result: Pioglitazone 121.0 vs 88.9 per 100,000 person-years, HR 1.63 (95% CI 1.22-2.19); rosiglitazone 86.2 vs 88.9, HR 1.10 (0.83-1.47)
• Funding: independent (the paper reports non-commercial research support)

Compared with other antidiabetic drugs, pioglitazone was associated with an increased risk of bladder cancer (121.0 v 88.9 per 100,000 person years; hazard ratio 1.63, 95% confidence interval 1.22 to 2.19).

Fractures in PROactive were a women's harm specifically: 5.1% of women on pioglitazone versus 2.5% on placebo, with no excess in men. (Source 19)

• Official position, Moderate certainty.
• Size: 5,238 patients (1,775 women)
• Who: PROactive participants with type 2 diabetes and macrovascular disease.
• How long: Mean 34.5 months.
• Result: Bone fracture in women 5.1% (44/870) vs 2.5% (23/905); men 1.7% vs 2.1%; mostly non-vertebral lower-limb and distal upper-limb fractures.
• Funding: industry (sponsor label summary)

During a mean follow-up of 34.5 months, the incidence of bone fracture in females was 5.1% (44/870) for pioglitazone tablets versus 2.5% (23/905) for placebo.

Congestive heart failure was the commonest serious adverse event causing withdrawal in PROactive, roughly doubling the placebo rate. (Source 20)

• Official position, Moderate certainty.
• Size: Over 8,500 patients across the randomised programme, including 2,605 in PROactive.
• Who: Adults with type 2 diabetes.
• How long: 16 weeks to over 2 years depending on trial.
• Result: In PROactive, withdrawals for adverse events 9.0% vs 7.7%; congestive heart failure causing withdrawal 1.3% vs 0.6%.
• Funding: industry (sponsor label)

Congestive heart failure was the most common serious adverse event leading to withdrawal occurring in 1.3% of patients treated with pioglitazone hydrochloride and 0.6% of patients treated with placebo.

Fatal and non-fatal liver failure have been reported after marketing, although the controlled trials showed no drug-induced hepatotoxicity. (Source 11)

• Case series, Very low certainty.
• Size: Spontaneous postmarketing reports, denominator unknown.
• Who: People taking pioglitazone.
• How long: Not applicable.
• Result: No rate can be calculated; the label states the reports contain insufficient information to establish the probable cause.
• Funding: industry (sponsor label)

There have been postmarketing reports of fatal and non-fatal hepatic failure in patients taking pioglitazone tables, although the reports contain insufficient information necessary to establish the probable cause.

What the evidence supports

In people without diabetes who had insulin resistance after a stroke or TIA, pioglitazone cut fatal or non-fatal stroke or myocardial infarction from 11.8% to 9.0% over 4.8 years, with no change in all-cause mortality. (Source 4)

• Randomized trial, High certainty.
• Size: 3,876 patients (1,939 pioglitazone, 1,937 placebo)
• Who: Recent ischaemic stroke or TIA, no diabetes, HOMA-IR above 3.0.
• How long: 4.8 years.
• Result: Primary outcome 175/1939 (9.0%) vs 228/1937 (11.8%), HR 0.76 (95% CI 0.62-0.93), P=0.007 - an absolute difference of 2.8 percentage points; all-cause mortality HR 0.93 (0.73-1.17), P=0.52.
• Funding: independent (National Institute of Neurological Disorders and Stroke)

By 4.8 years, a primary outcome had occurred in 175 of 1939 patients (9.0%) in the pioglitazone group and in 228 of 1937 (11.8%) in the placebo group (hazard ratio in the pioglitazone group, 0.76; 95% confidence interval [CI], 0.62 to 0.93; P=0.007).

In biopsy-proven NASH with prediabetes or type 2 diabetes, 18 months of pioglitazone produced a 41 percentage point absolute improvement in the primary histological endpoint. (Source 5)

• Randomized trial, Moderate certainty.
• Size: 101 patients.
• Who: Prediabetes or type 2 diabetes with biopsy-proven NASH, single university hospital.
• How long: 18 months double-blind, then 18 months open-label.
• Result: Primary outcome reached by 58% on pioglitazone, treatment difference 41 percentage points (95% CI 23 to 59); NASH resolution 51%, difference 32 points (13 to 51), P < 0.001 for each; fibrosis score difference -0.5 (-0.9 to 0.0), P=0.039; weight gain 2.5 kg greater.
• Funding: independent (Burroughs Wellcome Fund and American Diabetes Association)

Limit of this finding: One number in this abstract needs care: the fibrosis-score difference is printed as -0.5 with a confidence interval of -0.9 to 0.0 alongside P = 0.039. That is the journal's rounding of an interval that only just excludes zero, not a misprint - so the fibrosis improvement should be read as borderline, not as a clean result. The primary histological endpoint and the NASH-resolution figures are not affected.

Among patients randomly assigned to pioglitazone, 58% achieved the primary outcome (treatment difference, 41 percentage points [95% CI, 23 to 59 percentage points]) and 51% had resolution of NASH (treatment difference, 32 percentage points [CI, 13 to 51 percentage points]) (P < 0.001 for each).

What the evidence does not support

PROactive did not meet its primary endpoint: the broad composite of death, MI, stroke, acute coronary syndrome, revascularisation and amputation was not significantly reduced. (Source 15)

• Randomized trial, High certainty.
• Size: 5,238 patients (2,605 pioglitazone, 2,633 placebo)
• Who: Type 2 diabetes with evidence of macrovascular disease.
• How long: Average observation 34.5 months.
• Result: Primary composite 514/2605 vs 572/2633, HR 0.90 (95% CI 0.80-1.02), p=0.095; main secondary composite (death, non-fatal MI, stroke) 301 vs 358, HR 0.84 (0.72-0.98), p=0.027.
• Funding: industry-funded (manufacturer-sponsored trial)

Limit of this finding: The same abstract contains a printing error a little further on - it gives the heart-failure numerator and denominator as "149 of 2065" when the pioglitazone group numbered 2,605 throughout. It does not affect the endpoint figures quoted here.

514 of 2605 patients in the pioglitazone group and 572 of 2633 patients in the placebo group had at least one event in the primary composite endpoint (HR 0.90, 95% CI 0.80-1.02, p=0.095).

A 10-year prospective cohort of 193,099 people found no significant association between ever using pioglitazone and bladder cancer. (Source 17)

• Cohort study, Moderate certainty.
• Size: 193,099 people aged 40 or over in the bladder-cancer cohort; 34,181 (18%) pioglitazone users; 1,261 incident bladder cancers.
• Who: Kaiser Permanente Northern California members with diabetes.
• How long: Followed from 1997-2002 to December 2012; median pioglitazone duration 2.8 years.
• Result: Crude incidence 89.8 vs 75.9 per 100,000 person-years; adjusted HR 1.06 (95% CI 0.89-1.26); nested case-control adjusted OR 1.18 (0.78-1.80)
• Funding: not stated.

Ever use of pioglitazone was not associated with bladder cancer risk (adjusted hazard ratio [HR], 1.06; 95% CI, 0.89-1.26).

Where the evidence is mixed

In a meta-analysis of the manufacturer's individual-patient trial data, pioglitazone lowered the composite of death, myocardial infarction or stroke in absolute terms by 1.3 percentage points, while increasing serious heart failure by 0.5 points. (Source 3)

• Meta-analysis, Moderate certainty.
• Size: 16,390 patients across 19 randomised trials.
• Who: Adults with type 2 diabetes, with and without established vascular disease.
• How long: Treatment 4 months to 3.5 years.
• Result: Death, MI or stroke 375/8554 (4.4%) vs 450/7836 (5.7%), HR 0.82 (95% CI 0.72-0.94), P = .005; serious heart failure 200 (2.3%) vs 139 (1.8%), HR 1.41 (95% CI 1.14-1.76), P = .002; no heterogeneity (I-squared = 0%)
• Funding: industry data, independently analysed - the patient-level database was transferred from the manufacturer.

Death, myocardial infarction, or stroke occurred in 375 of 8554 patients (4.4%) receiving pioglitazone and 450 of 7836 patients (5.7%) receiving control therapy (hazard ratio [HR], 0.82; 95% confidence interval [CI], 0.72-0.94; P = .005). Progressive separation of time-to-event curves became apparent after approximately 1 year of therapy. Individual components of the primary end point were all reduced by a similar magnitude with pioglitazone treatment, with HRs ranging from 0.80 to 0.92. Serious heart failure was reported in 200 (2.3%) of the pioglitazone-treated patients and 139 (1.8%) of the control patients (HR, 1.41; 95% CI, 1.14-1.76; P = .002).

The source of that meta-analysis was the manufacturer's own trial database, which is a quality limitation worth stating. (Source 21)

• Meta-analysis, Moderate certainty.
• Size: 19 randomised, double-blinded, placebo- or active-controlled trials.
• Who: Pioglitazone clinical trial programme.
• How long: Varied.
• Result: Not an effect estimate - a provenance statement.
• Funding: industry-supplied data.

A database containing individual patient-level time-to-event data collected during pioglitazone clinical trials was transferred from the drug's manufacturer for independent analysis.

That fracture meta-analysis was itself fragile: the overall result held in the fixed-effect model but not in the random-effects model, and half the included trials were at high risk of bias. (Source 16)

• Meta-analysis, Very low certainty.
• Size: 78 trials.
• Who: As above.
• How long: Varied.
• Result: The review states the fracture result was significant in the fixed-effect model but not in the random-effects model.
• Funding: not stated.

Limit of this finding: Two cautions, both from the review itself. Its headline risk ratio of 1.21 was significant in the fixed-effect model but not in the random-effects model, so the overall result is fragile. And the sentence in which it reports subgroups puts a null result under the word "exacerbated": in people with insulin resistance the risk ratio was 0.87 (95% CI 0.43-1.76, P = 0.69), which shows no increase in fracture risk in that subgroup. Nothing here supports an increased fracture risk in insulin-resistant people.

Fracture risk was significant in the fixed-effect model but not in the random-effects model.

Within PROactive itself, bladder cancer was diagnosed in 0.54% on pioglitazone versus 0.19% on placebo during the trial, but over 13 years of trial plus observational follow-up the rates converged exactly. (Source 22)

• Official position, Low certainty.
• Size: 5,238 randomised, with a large subset followed up to 10 additional years.
• Who: PROactive participants.
• How long: 3-year trial plus up to 10 years of observation.
• Result: In-trial 14/2605 (0.54%) vs 5/2633 (0.19%); excluding exposure under one year, 6 (0.23%) vs 2 (0.08%); over 13 years HR 1.00 (95% CI 0.59 to 1.72)
• Funding: industry (sponsor label summary of sponsor trial)

during the three year PROactive clinical trial, 14 patients out of 2605 (0.54%) randomized to pioglitazone hydrochloride and 5 out of 2633 (0.19%) randomized to placebo were diagnosed with bladder cancer.

Where the research disagrees

Whether pioglitazone increases the risk of bladder cancer

• Lewis and colleagues, JAMA (2015), Kaiser Permanente cohort, 10-year prospective cohort of 193,099 people plus a nested case-control analysis: Pioglitazone use was not associated with a statistically significant increased risk of bladder cancer, although an increased risk, as previously observed, could not be excluded. (Source 23)
• Tuccori and colleagues, BMJ (2016), UK CPRD cohort, Population-based cohort of 145,806 patients, 689,616 person-years, with duration and dose-response analyses: Compared with other antidiabetic drugs, pioglitazone was associated with an increased risk of bladder cancer (121.0 v 88.9 per 100,000 person years; hazard ratio 1.63, 95% confidence interval 1.22 to 2.19). (Source 18)
• The US label (2026), summarising both, Regulatory position citing the opposing observational studies: Inconsistent findings and limitations inherent in these and other studies preclude conclusive interpretations of the observational data. (Source 24)

Whether pioglitazone reduces cardiovascular events in type 2 diabetes

• Lincoff and colleagues, JAMA (2007), Meta-analysis of 19 trials, 16,390 patients, manufacturer-supplied individual patient data: Death, myocardial infarction, or stroke occurred in 375 of 8554 patients (4.4%) receiving pioglitazone and 450 of 7836 patients (5.7%) receiving control therapy (hazard ratio [HR], 0.82; 95% confidence interval [CI], 0.72-0.94; P = .005). (Source 3)
• The US label (2026), Regulatory position; PROactive's primary endpoint was not met (HR 0.90, 95% CI 0.80-1.02, p=0.095): There have been no clinical studies establishing conclusive evidence of macrovascular risk reduction with pioglitazone hydrochloride. (Source 25)

How much

• Reference intake: Dosing is set by the prescriber, not the reader. As a position, the US pioglitazone tablet label (version effective 2026-08-27) states the recommended starting dose is 15 mg or 30 mg once daily without congestive heart failure, and 15 mg once daily in NYHA Class I or II heart failure, taken with or without food. (Source 10)
• Upper limit: The same label gives a maximum of 45 mg once daily, reduced to a maximum of 15 mg daily when taken with gemfibrozil or another strong CYP2C8 inhibitor. These are regulatory ceilings, not targets. (Source 6)
• Studied: PROactive titrated oral pioglitazone from 15 mg to 45 mg daily in 2,605 patients on top of their existing glucose-lowering drugs, against matching placebo in 2,633. (Source 26)
• Studied: IRIS used a target dose of 45 mg daily in 1,939 non-diabetic patients with insulin resistance after stroke or TIA. (Source 4)
• Studied: The NASH trial gave pioglitazone 45 mg/d or placebo for 18 months on top of a 500-kcal/d deficit diet. (Source 27)

A common belief, and what the research shows

The belief: Pioglitazone was withdrawn or banned because it causes bladder cancer.

What the research shows: It carries a boxed warning for congestive heart failure, not bladder cancer, and it remains licensed. The bladder-cancer evidence is genuinely split: a 193,099-person US cohort found "Ever use of pioglitazone was not associated with bladder cancer risk (adjusted hazard ratio [HR], 1.06; 95% CI, 0.89-1.26)." while a 145,806-person UK cohort found a hazard ratio of 1.63 with a dose-response. The US label's own summary is that "Inconsistent findings and limitations inherent in these and other studies preclude conclusive interpretations of the observational data." The harms that are not in doubt are heart failure, oedema, weight gain and fractures in women.

Questions and answers

What is it?

Pioglitazone is a synthetic thiazolidinedione drug - an odourless white crystalline powder that is practically insoluble in water, pressed into 15, 30 and 45 mg tablets. It is made and used as a racemic mixture whose two mirror-image forms interconvert in the body with no difference in activity. It is a manufactured medicine, not a nutrient. (Source 28)

What does it do in the body?

It reduces insulin resistance in muscle, fat and liver by switching on the nuclear receptor PPAR-gamma, which changes the reading of insulin-responsive genes. The result is more insulin-driven glucose disposal and less glucose output from the liver. It depends on the person still making their own insulin and does not stimulate insulin release. (Source 1)

Is it good or bad for you?

Both, and the balance depends on the person. In IRIS, 4.8 years of pioglitazone in insulin-resistant stroke survivors without diabetes cut stroke or heart attack from 11.8% to 9.0% - but in the same trial 52.2% gained more than 4.5 kg versus 33.7% on placebo, oedema hit 35.6% versus 24.9%, and fractures needing surgery or hospital care 5.1% versus 3.2%. It carries a boxed warning for congestive heart failure and is contraindicated in NYHA Class III or IV heart failure. (Source 29)

How do you get more of it?

Pioglitazone is prescription-only; the amount in the body is set by the prescribed dose. Blood levels are raised about three-fold by gemfibrozil and other strong CYP2C8 inhibitors, which is why the label caps the dose at 15 mg daily in that combination, and lowered by CYP2C8 inducers such as rifampicin. Food makes no difference. (Source 6)

If it is harmful, what reduces it?

The drug is stopped or the dose reduced - there is no antidote and no reported withdrawal syndrome. The label requires exactly this for its two main harms: discontinuation or dose reduction must be considered if heart failure develops, and treatment should be interrupted if ALT rises above three times the upper limit of normal. Macular oedema improved in some patients after the drug was stopped. (Source 2)

Why might someone be low in it or missing it?

This does not apply in the nutrient sense - nobody is naturally low in pioglitazone. Levels fall if the drug is stopped, if a CYP2C8 inducer such as rifampicin or an anticonvulsant is added, or if topiramate is co-prescribed. It is also contraindicated rather than merely reduced in people with NYHA Class III or IV heart failure, and the label says it is not recommended in symptomatic heart failure at all. (Source 2)

Which whole foods contain it or feed it?

No food contains pioglitazone; it is a synthesised thiazolidinedione formulated with pharmaceutical excipients. Diet is relevant in a different way: the NASH trial that showed histological benefit gave every participant a 500-kcal/day deficit diet alongside the drug or placebo, so the trial effect sits on top of calorie restriction, not instead of it. (Source 27)

What happens if you do not have it?

Nothing happens from not having pioglitazone - it is one option among several for type 2 diabetes, not something the body requires. Its approved role is as an adjunct to diet and exercise to improve glycaemic control in adults with type 2 diabetes, and it does not work at all in type 1 diabetes or diabetic ketoacidosis because it needs the person's own insulin to act. (Source 30)

How can you test for it?

There is no blood level of pioglitazone that is measured. What is monitored is response and harm: HbA1c guides whether the dose is titrated, a liver test panel (ALT, AST, alkaline phosphatase, bilirubin) is obtained before starting and measured promptly if liver symptoms appear, and weight, oedema and heart-failure symptoms are watched after starting and after every dose increase. Routine periodic liver testing is explicitly not recommended in people without liver disease. (Source 10)

References

1. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
2. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
3. JAMA. Pioglitazone and risk of cardiovascular events in patients with type 2 diabetes mellitus: a meta-analysis of randomized trials.. 2007. PMID 17848652, DOI 10.1001/jama.298.10.1180. Read the source
4. New England Journal of Medicine. Pioglitazone after Ischemic Stroke or Transient Ischemic Attack.. 2016. PMID 26886418, DOI 10.1056/NEJMoa1506930. Read the source
5. Annals of Internal Medicine. Long-Term Pioglitazone Treatment for Patients With Nonalcoholic Steatohepatitis and Prediabetes or Type 2 Diabetes Mellitus: A Randomized Trial.. 2016. PMID 27322798, DOI 10.7326/M15-1774. Read the source
6. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
7. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.) - section 7.2 CYP2C8 Inducers. 2026. Read the source
8. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.) - section 7.3 Topiramate. 2026. Read the source
9. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
10. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
11. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
12. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.) - section 7.1 Strong CYP2C8 Inhibitors. 2026. Read the source
13. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
14. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.) - section 5.7 Macular Edema. 2026. Read the source
15. The Lancet. Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial.. 2005. PMID 16214598, DOI 10.1016/S0140-6736(05)67528-9. Read the source
16. Frontiers in Pharmacology. Clinical implications of fracture severity risk with pioglitazone: a systematic review and meta-analysis of clinical randomized trials.. 2025. PMID 40115256, DOI 10.3389/fphar.2025.1357309. Read the source
17. JAMA. Pioglitazone Use and Risk of Bladder Cancer and Other Common Cancers in Persons With Diabetes.. 2015. PMID 26197187, DOI 10.1001/jama.2015.7996. Read the source
18. BMJ. Pioglitazone use and risk of bladder cancer: population based cohort study.. 2016. PMID 27029385, DOI 10.1136/bmj.i1541. Read the source
19. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
20. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
21. JAMA. Pioglitazone and risk of cardiovascular events in patients with type 2 diabetes mellitus: a meta-analysis of randomized trials - Data sources and study selection section of the published abstract. 2007. PMID 17848652, DOI 10.1001/jama.298.10.1180. Read the source
22. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
23. JAMA. Pioglitazone Use and Risk of Bladder Cancer and Other Common Cancers in Persons With Diabetes - Conclusions and relevance section of the published abstract. 2015. PMID 26197187, DOI 10.1001/jama.2015.7996. Read the source
24. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
25. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.) - section 5.8 Macrovascular Outcomes. 2026. Read the source
26. The Lancet. Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial - Methods section of the published abstract. 2005. PMID 16214598, DOI 10.1016/S0140-6736(05)67528-9. Read the source
27. Annals of Internal Medicine. Long-Term Pioglitazone Treatment for Patients With Nonalcoholic Steatohepatitis and Prediabetes or Type 2 Diabetes Mellitus: A Randomized Trial. - Intervention section of the published abstract. 2016. PMID 27322798, DOI 10.7326/M15-1774. Read the source
28. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source
29. New England Journal of Medicine. Pioglitazone after Ischemic Stroke or Transient Ischemic Attack - Conclusions section of the published abstract. 2016. PMID 26886418, DOI 10.1056/NEJMoa1506930. Read the source
30. DailyMed / US FDA Structured Product Label, label version effective 2026-08-27. PIOGLITAZONE TABLET - prescribing information (Preferred Pharmaceuticals Inc.). 2026. Read the source