INTRODUCTION — The thiazolidinediones increase insulin sensitivity by acting on adipose, muscle, and, to a lesser extent, liver to increase glucose utilization and decrease glucose production. Two thiazolidinediones (rosiglitazone and pioglitazone) are currently available in the United States. In 2010, the European Medicines Agency suspended sales of rosiglitazone, and in June 2011, the French and German Medicines Agencies also suspended the use of pioglitazone, owing to concerns that the overall risks of rosiglitazone and pioglitazone exceed their benefits. A third thiazolidinedione, troglitazone, was the first drug in this class and was removed from markets because it caused liver dysfunction and, in some patients, liver failure.

The pharmacology and use of thiazolidinediones will be reviewed here. Other oral anti-hyperglycemic medications are discussed separately:

●(See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus".)

●(See "Management of persistent hyperglycemia in type 2 diabetes mellitus".)

●(See "Metformin in the treatment of adults with type 2 diabetes mellitus".)

MECHANISM OF ACTION — The thiazolidinediones increase insulin sensitivity by acting on adipose tissue and muscle to increase glucose utilization. To a lesser degree, they decrease glucose production by the liver [1,2]. The mechanism by which the thiazolidinediones exert their effect is not fully understood. They bind to and activate one or more peroxisome proliferator-activated receptors (PPARs), mostly PPAR-gamma, which alters the transcription of multiple genes involved in glucose and lipid metabolism [1,3].

PPAR-gamma is found predominantly in adipose tissue, pancreatic beta cells, vascular endothelium, macrophages, and the central nervous system (CNS) [4]. PPAR-alpha is expressed mostly in liver, heart, skeletal muscle, and vascular walls. The various thiazolidinediones have differential effects on PPAR-gamma and PPAR-alpha. Rosiglitazone is a pure PPAR-gamma agonist, while pioglitazone also exerts some PPAR-alpha effects (similar to fibrates). This may account for the different effects that pioglitazone and rosiglitazone have on lipids and cardiovascular ischemic outcomes (see 'Cardiovascular effects' below). Partial PPAR-gamma agonists are under investigation, with the hope that partial activation of the nuclear receptor will reduce adverse events [5].

In human adipocytes, treatment with thiazolidinediones has been reported to increase expression of genes involved in promoting lipid storage and decrease expression of genes associated with inflammation, such as interleukin (IL)-6 [6]. Thus, the insulin-sensitizing effect of thiazolidinediones may be related in part due to regulation of adipose tissue production of adipokines via PPAR-gamma activation [7].

In the CNS, PPAR-gamma activation may mediate weight gain by promoting increased feeding, as demonstrated in studies of knockout murine models [8-10]. This finding may account, in part, for the weight gain associated with thiazolidinedione treatment, although a substantial component is likely related to fluid retention that has been documented in patients. (See 'Weight gain' below.)

CLINICAL USE

Potential indications — Generally, thiazolidinediones are not prescribed as initial therapy in patients with type 2 diabetes. Pioglitazone may have a role as initial therapy in patients who have contraindications to other oral agents (eg, metformin, sulfonylureas), who decline injectable therapies (eg, insulin, glucagon-like peptide 1 [GLP-1] receptor agonists), and in whom medications like dipeptidyl peptidase 4 (DPP-4) inhibitors or sodium-glucose co-transporter 2 (SGLT2) inhibitors cannot be afforded or are found to be less effective. It is more often prescribed as second-line or, more commonly, as third-line therapy, when other oral agents in combination are not resulting in adequate glycemic management, especially when cost is a major concern (pioglitazone is available as a low-cost generic) [11]. (See "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Monotherapy failure'.)

Pioglitazone may still offer an advantage in certain clinical settings; for example, in the context of high risk for hypoglycemia, severe insulin resistance, coexisting nonalcoholic steatohepatitis (NASH), or recent stroke. However, risks of weight gain, heart failure, fractures, and lingering concerns about a possible small increase in risk of bladder cancer require careful risk-benefit discussion with the patient. Lower doses should be used when possible to mitigate these potential adverse consequences, particularly weight gain and edema. As a fundamental principle, the choice and doses of type 2 diabetes mellitus therapies should be individualized. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus", section on 'Initial pharmacologic therapy' and "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Our approach' and 'Safety' below.)

In contrast, rosiglitazone is rarely used and is no longer widely available because of the concern about a potential increased risk for cardiovascular events suggested by a controversial meta-analysis of phase III clinical trials [12,13]. In addition, potentially adverse atherogenic effects of rosiglitazone as a consequence of altered lipid levels in comparison with the more widely available pioglitazone make it a less desirable choice. (See 'Major cardiovascular events' below.)

Contraindications — Thiazolidinediones should not be used in patients with:

●Heart failure or any evidence of fluid overload

●History of fracture or at high risk for fracture (eg, postmenopausal women with low bone mass)

●Active liver disease (liver transaminases >2.5 times above the upper reference limit)

●Active or history of bladder cancer

●Type 1 diabetes

●Pregnancy

●Macular edema

Combination therapy with thiazolidinediones and insulin may be particularly problematic due to increased risk of edema and heart failure [14-16]. Thiazolidinediones should be discontinued in patients who develop signs and symptoms of heart failure.

Dosing and monitoring — Liver function tests should be measured prior to initiating pioglitazone or rosiglitazone and periodically thereafter. (See 'Contraindications' above.)

Pioglitazone is usually initiated at a dose of 15 mg daily. If, after a few weeks, there is inadequate glycemic management based upon fasting blood glucose values, the daily dose can be increased in increments of 15 mg to the maximum dose of 45 mg daily. However, side effects of weight gain and edema increase at higher dosing levels (see 'Safety' below). Most clinicians still prescribing this medication keep patients in the 15 to 30 mg dosing range. Dose escalation should be performed gradually (eg, every 4 to 12 weeks) while reassessing the patient for weight gain, edema, and other signs of heart failure.

Although pioglitazone has not been associated with liver toxicity, periodic liver function testing is recommended because of hepatotoxicity demonstrated with other thiazolidinediones (see 'Hepatotoxicity' below). Glycated hemoglobin (A1C) should be measured at least twice yearly in patients meeting glycemic goals and more frequently (quarterly) in patients whose therapy has changed or who are not meeting goals. Monitoring glycemia is discussed in more detail elsewhere. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus", section on 'Monitoring' and "Glucose monitoring in the management of nonpregnant adults with diabetes mellitus", section on 'Type 2 diabetes'.)

GLYCEMIC EFFICACY

Monotherapy — When thiazolidinediones are used as monotherapy, the expected decrease in A1C is approximately 0.5 to 1.4 percent [17-19]. In a randomized trial comparing rosiglitazone, metformin, and glyburide as initial therapy for type 2 diabetes, rosiglitazone had a lower incidence of monotherapy failure at five years [20]. However, the difference in A1C levels between the rosiglitazone- and metformin-treated groups was small (approximately 0.1 percent). Treatment with the thiazolidinedione was associated with greater weight gain, peripheral edema, and increases in low-density lipoprotein (LDL) concentrations. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus", section on 'Glycemic efficacy'.)

Combination therapy — In addition to use as monotherapy, the thiazolidinediones have been studied in combination with metformin, sulfonylureas, insulin, dipeptidyl peptidase 4 (DPP-4) inhibitors, glucagon-like peptide 1 (GLP-1) receptor agonists, and sodium-glucose co-transporter 2 (SGLT2) inhibitors [21]. Combination therapy is discussed in detail in separately. (See "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Our approach'.)

Although the addition of a thiazolidinedione to insulin improves glycemic management compared with insulin monotherapy [22,23], the combination of insulin plus a thiazolidinedione is associated with an increased incidence of weight gain, edema, and heart failure and should be avoided in patients where these may be concerns [14-16].

CARDIOVASCULAR EFFECTS — The cardiovascular effects of thiazolidinediones may depend upon the specificity of peroxisome proliferator-activated receptor (PPAR) activation [24]. Rosiglitazone (a pure PPAR-gamma agonist) and pioglitazone (PPAR-gamma with some PPAR-alpha agonist activity) have different effects on lipid concentrations, with pioglitazone exerting a less unfavorable effect (see 'Lipids' below). Although both drugs have a similar effect on the incidence of heart failure (both increasing the risk), they appear to have disparate effects on ischemic outcomes [12,25]. (See 'Major cardiovascular events' below.)

The results of the meta-analyses and observational studies suggest increased caution with use of rosiglitazone in particular. Any potential cardiovascular benefits of pioglitazone must be weighed against the increased risk of heart failure, weight gain, fluid retention, and fractures [26,27]. (See 'Major cardiovascular events' below and 'Safety' below.)

Lipids — Rosiglitazone and pioglitazone have similar effects on glycemia, but their effects on serum lipid concentrations are different. Most randomized trials have shown a more favorable lipid profile with pioglitazone [1,28,29]. In a review of six randomized trials, low-density lipoprotein (LDL) cholesterol levels typically remained constant when monotherapy or combination therapy with pioglitazone was used, while increases in LDL cholesterol levels ranging from 8 to 16 percent were observed in studies of rosiglitazone [1]. High-density lipoprotein (HDL) cholesterol levels increased by approximately 10 percent with both drugs. Decreases in triglyceride levels were observed more often with pioglitazone than with rosiglitazone.

Major cardiovascular events — In 2010, the US Food and Drug Administration (FDA) imposed marked restrictions on the prescribing of rosiglitazone because of concerns about increased risk of acute myocardial infarction (MI) and cardiovascular deaths [30]. These restrictions were largely removed by the FDA in 2013 after a reevaluation of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) study [31,32], and the Risk Evaluation and Mitigation Strategy (REMS; aimed at educating clinicians about the cardiovascular side effects of rosiglitazone) was eliminated in 2015 [33]. The absence of an adverse effect of rosiglitazone on cardiovascular risk, however, has not been excluded. Based on this controversy, the drug is no longer widely used. The only thiazolidinedione still used to any degree is pioglitazone.

Heart failure — Rosiglitazone and pioglitazone increase the risk of heart failure. In meta-analyses of randomized trials of thiazolidinediones for the treatment or prevention of type 2 diabetes, the estimated relative risk (RR) of heart failure in patients randomly assigned to thiazolidinediones compared with placebo ranged from 1.5 to 2.1 (95% CIs 1.2-2.4 and 1.1-4.1, respectively) [34-36]. In the subsequently published RECORD study, the addition of rosiglitazone to metformin or sulfonylurea in people with type 2 diabetes increased the risk of heart failure compared with combination metformin and sulfonylurea (2.7 versus 1.3 percent, hazard ratio [HR] 2.10, 95% CI 1.35-3.27) [32].

In retrospective analyses of prescription data obtained from national databases, rosiglitazone was associated with a greater risk of adverse heart failure events than pioglitazone [37-39].

Atherosclerotic cardiovascular events

●Rosiglitazone – The effect of rosiglitazone on the risk of MI is uncertain. Meta-analyses of trials evaluating rosiglitazone for the treatment of diabetes show increases in [12,13,36,40-42] or no evidence of effects on [43,44] MI and other adverse outcomes (morbidity, mortality, quality of life). In the RECORD study mentioned above (see 'Heart failure' above), the effect of rosiglitazone on MI was inconclusive (HR for rosiglitazone compared with the active comparators for fatal and nonfatal MI 1.14, 95% CI 0.80-1.63) due to the small number of events and possibly affected by greater statin use in the rosiglitazone group. Rosiglitazone did not increase the risk of overall cardiovascular morbidity or mortality compared with standard glucose-lowering drugs [32].

●Pioglitazone – Pioglitazone does not appear to induce the same atherosclerotic cardiovascular risk profile as rosiglitazone. In fact, it likely decreases the risk of major adverse atherosclerotic cardiovascular events [25,45]. This was illustrated by a meta-analysis of 19 trials of pioglitazone for the treatment of type 2 diabetes [25]. MI occurred in 131 patients (1.5 percent) in the pioglitazone group and 159 (2 percent) in the comparator group (placebo, metformin, sulfonylurea, rosiglitazone; HR 0.81, 95% CI 0.64-1.02). The primary composite endpoint of death, nonfatal MI, or nonfatal stroke occurred in 4.4 and 5.7 percent of patients in the pioglitazone and control groups, respectively (HR 0.82, 95% CI 0.72-0.94). As in the meta-analyses described above, the majority of trials were not originally intended to assess cardiovascular endpoints, which were collected as adverse events and not uniformly adjudicated.

The largest of the trials included in the meta-analysis, however, was specifically designed to evaluate cardiovascular outcomes. The Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive) trial evaluated the effect of pioglitazone on cardiovascular events and mortality in 5238 patients with type 2 diabetes and established macrovascular complications (prior MI, stroke, coronary artery bypass graft [CABG] surgery, acute coronary syndrome, or symptomatic peripheral artery disease) [46]. There was a nonsignificant effect on the predefined primary outcome of the study (composite of all-cause mortality, nonfatal MI and silent MI, stroke, acute coronary syndrome, surgical intervention on coronary or leg arteries, or leg amputation; 19.7 versus 21.7 percent, HR 0.90, 95% CI 0.80-1.02). There was, however, a significant decrease in the main secondary endpoint (composite of all-cause mortality, nonfatal MI, or nonfatal stroke) in the pioglitazone group (11.6 versus 13.6 percent, HR 0.84, 95% CI 0.72-0.98). Hospitalization for heart failure was more frequent in the pioglitazone group, with the number of episodes counterbalancing the decreased number of major adverse cardiovascular events. In subgroup analyses, pioglitazone reduced the risk of subsequent MI in patients with prior MI (5.3 versus 7.2 percent with placebo, HR 0.72, 95% CI 0.52-0.99) [47] and subsequent stroke in those with prior stroke (5.6 versus 10.2 percent with placebo, HR 0.53, 95% CI 0.34-0.85) [48].

In other trials in patients with insulin resistance and recent stroke or transient ischemic attack, pioglitazone reduced the incidence of acute coronary syndrome and stroke, but increased risk of weight gain, edema, and fracture [49-51]. The risk of heart failure was not increased, although investigators were allowed to decrease the study drug dose for edema or weight gain during the trial.

HEPATIC BENEFITS — In patients with type 2 diabetes mellitus and biopsy-proven nonalcoholic steatohepatitis (NASH), pioglitazone improves fibrosis as well as inflammation and steatosis. This topic is reviewed in detail separately. (See "Management of nonalcoholic fatty liver disease in adults", section on 'Patients with NASH and diabetes'.)

SAFETY — Thiazolidinediones are not considered first-choice agents for patients with type 2 diabetes, due to adverse effects including increased risk of weight gain, fluid retention, heart failure, fractures, and the potential small increased risk of bladder cancer (pioglitazone). In 2010, the European Medicines Agency suspended sales of rosiglitazone [52], and in June 2011, the French and German Medicines Agencies also suspended the use of pioglitazone because of the potential increased risk of bladder cancer and the concern that the overall risks of pioglitazone exceed its benefits. The European Medicines Agency has not suspended the sales of pioglitazone [53].

Hypoglycemia — The thiazolidinediones improve blood glucose primarily by increasing insulin sensitivity. Therefore, thiazolidinedione monotherapy is much less likely to cause hypoglycemia than sulfonylureas or insulin [17,20]. However, hypoglycemic events can occur when these agents are used in conjunction with sulfonylureas or insulin.

Weight gain — All of the thiazolidinediones cause weight gain. This effect is both dose- and time-dependent and can be substantial [17,54-59]. In a 26-week trial, the improvement in A1C with rosiglitazone was accompanied by a dose-related weight gain of up to 4.5 kg (9.9 pounds) compared with placebo [18]. In a cohort of patients taking pioglitazone continuously for over 30 months, the average weight gain increased steadily up to 30 months but then plateaued (at around 5.3 kg, or 12 pounds) by 36 months [60].

The weight gain is caused, in part, by fluid retention. A small study (n = 8) found that 12 weeks of therapy with pioglitazone 45 mg resulted in 2.4 liters of water accumulation, which accounted for 75 percent of the 3.1 kg weight gain [61]. (See 'Fluid retention/heart failure' below.)

In addition, weight gain may be due to thiazolidinedione-induced activation of peroxisome proliferator-activated receptor (PPAR)-gamma in the central nervous system (CNS) [10]. In animal studies, activation of CNS PPAR-gamma by rosiglitazone increased feeding and expanded adipose tissue mass [8,9].

In humans, upregulation of genes that facilitate adipocyte lipid storage has been described [62]. Weight gain may also result from the proliferation of new adipocytes. In 20 patients with type 2 diabetes treated with the previously used thiazolidinedione troglitazone, the increase in fat deposition occurred primarily in subcutaneous fat stores, resulting in a decreased visceral-to-subcutaneous fat ratio [59].

Fluid retention/heart failure — Fluid retention, which is more prominent with concomitant insulin therapy, can occur with all the thiazolidinediones but does not account for all of the weight gain. (See 'Weight gain' above.)

Peripheral edema occurs in 4 to 6 percent of patients treated with thiazolidinediones (compared with 1 to 2 percent with placebo) and in a higher percentage of patients with a history of heart failure [1]. This fluid retention may lead to the precipitation or worsening of heart failure. (See 'Heart failure' above.)

As mentioned above, the thiazolidinediones act by binding to and activating PPAR-gamma [1,63]. Along the nephron, PPAR-gamma is most abundant in the collecting tubules, and the fluid retention with thiazolidinediones appears to result from PPAR-gamma stimulation of sodium reabsorption by sodium channels (called the epithelial sodium channel) in the luminal membrane of the collecting tubule cells [64]. This effect is mediated by increased expression of the gamma subunit of the sodium channel gene mRNA.

Skeletal fractures — There is an increasing body of evidence suggesting that thiazolidinediones decrease bone density and increase fracture risk, particularly in women [65-73]. In a large population-based study using the United Kingdom-based General Practice Research Database, pioglitazone (odds ratio [OR] 2.59, 95% CI 0.96-7.01) and rosiglitazone (OR 2.38, 95% CI 1.39-4.09) were associated with low-trauma fracture in both females and males [69]. The absolute increase in fracture risk associated with thiazolidinedione treatment appears to be small. Nevertheless, these drugs should probably not be used in women with low bone density or other risk factors for fracture (see 'Contraindications' above). The results of observational studies in men are conflicting and suggest that additional studies are required to determine if there is a similar increase in fracture risk.

The adverse impact of thiazolidinediones on bone in patients with type 2 diabetes is illustrated by the following studies:

●In the Diabetes Outcome Progression Trial (ADOPT), fracture rates per 100 patient-years were 2.7, 1.5, and 1.3 for rosiglitazone, metformin, and glyburide, respectively, for women, whereas for men, fracture rates did not differ significantly by treatment (1.2, 1.0, and 1.1, respectively) [20]. Risk curves began to diverge after one year of follow-up for women.

●In the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial, the incidence of fractures was higher in the rosiglitazone group (8.3 versus 5.3 percent in the control group, relative risk [RR] 1.57, 95% CI 1.26-1.97) [32]. Fractures occurred mainly in the upper and distal lower limbs and were more common in women than in men.

●In a retrospective review of adverse events in the Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive) study, there was a higher rate of fractures in women receiving pioglitazone (5.1 versus 2.5 percent with placebo), but not in men [46,72].

●In patients with insulin resistance and recent stroke or transient ischemic attack, there was an increased risk of fracture in patients receiving pioglitazone (13.6 versus 8.8 percent with placebo, hazard ratio [HR] 1.53, 95% CI 1.24-1.89) [73]. Increased risk for any fracture was observed in both men (9.4 versus 5.2 percent, HR 1.83, 95% CI 1.36-2.48) and women (14.9 versus 11.6 percent, HR 1.32, 95% CI 0.98-1.78). The increased risk of fracture appeared to emerge after two years of pioglitazone therapy.

In animal models, treatment with rosiglitazone resulted in bone loss by suppression of osteoblast differentiation and formation [74]. The bone loss was associated with an increase in marrow adipocytes. In vivo studies suggest that this finding is mediated by rosiglitazone-induced activation of PPAR-gamma. PPAR-gamma2 isoform is an important regulator of adipocyte differentiation. Activation of PPAR-gamma2 results in diversion of bone marrow stromal cells from the osteoblast lineage into the adipocyte lineage, which subsequently leads to a decrease in bone formation rates and increase in adipogenesis [75]. This process, in turn, may be regulated by the insulin-like growth factor (IGF) system as rosiglitazone-induced activation of PPAR-gamma has been demonstrated to downregulate components of the IGF system both in vivo and in vitro, and IGF-1 is an important regulator of osteoblast proliferation and differentiation [76].

Bladder cancer — Although there are no data linking rosiglitazone to bladder cancer, there is concern about increased risk with pioglitazone. The association is controversial, and different reports have yielded conflicting results. Pioglitazone should not be used in patients with active bladder cancer [77]. In patients with a history of bladder cancer, the benefits of glycemic management versus the unknown risk for cancer recurrence with pioglitazone should be considered. (See 'Contraindications' above.)

In preclinical studies, pioglitazone was associated with bladder tumors in male rats. In the PROactive trial described above, there were more cases of bladder cancer (14 versus 5) in the pioglitazone than placebo group [72]. In an interim analysis of an ongoing, 10-year observational study of pioglitazone use in patients with diabetes, there was not a significant association between pioglitazone exposure and bladder cancer among patients with median duration of therapy of two years (HR for bladder cancer 1.2, 95% CI 0.9-1.5) [78,79]. In the 10-year full report from this study, there was no increased risk of bladder cancer (HR 1.06, 95% CI 0.89-1.26) even when accounting for dose and duration of exposure [80].

Other observational studies have shown conflicting results. A meta-analysis involving 26 separate studies concluded that there is no increased risk (HR 1.07, 95% CI 0.96-1.18) [81]. This has resulted in uncertainty about the association between pioglitazone and bladder cancer. If there is increased risk, this likely is small but should nonetheless be discussed with patients prior to initiation of therapy.

Hepatotoxicity — The thiazolidinedione troglitazone was removed from the market in the United States and United Kingdom because of reports of severe hepatocellular injury [82], resulting in death or the need for liver transplantation in some patients [83,84]. Similar hepatoxicity has not been observed with rosiglitazone or pioglitazone [85]. Nonetheless, based on the troglitazone experience, the US Food and Drug Administration (FDA) currently recommends that patients receiving pioglitazone or rosiglitazone undergo baseline liver biochemical testing followed by periodic monitoring of liver function based upon clinical judgement. In patients with nonalcoholic fatty liver disease, transaminase levels often improve with use of thiazolidinediones, which correlates with their known effects to decrease liver fat content. (See "Management of nonalcoholic fatty liver disease in adults", section on 'Patients with NASH and diabetes'.)

Macular edema — Macular edema has been reported in patients taking thiazolidinediones, though the frequency of occurrence is unknown [86-90]. In a prospective cohort study of 172,006 individuals with diabetic macular edema (DME), the use of thiazolidinediones (98 percent with pioglitazone) was associated with a 2.6-fold increased risk of developing DME, and this relationship remained even after correcting for potential confounders, such as age or poor glycemic management [87]. However, in the randomized clinical trial Action to Control Cardiovascular Risk in Diabetes (ACCORD) eye substudy, data from a subgroup of 3473 participants who had fundus photographs that were centrally evaluated in a standard fashion did not demonstrate a conclusive link between thiazolidinedione use and DME [87,91]. Thus, data suggest that DME can occur at least sporadically in patients with diabetes mellitus treated with thiazolidinedione, but this is relatively rare and may be related to underlying confounders, such as duration or severity of diabetes. Patients at greatest risk seem to be those who are also at risk for peripheral edema. (See 'Fluid retention/heart failure' above.)

DIABETES PREVENTION — We do not suggest using thiazolidinediones for diabetes prevention. (See "Prevention of type 2 diabetes mellitus", section on 'Our approach'.)

Several studies have suggested efficacy in preventing type 2 diabetes in patients at risk, at least while thiazolidinedione therapy is continued [92-95]. Although thiazolidinediones delay the onset of diagnosis of diabetes and, therefore, reduce the length of exposure of hyperglycemia, the benefit or harm of the intervention (independent of the effect on hyperglycemia) must be considered. Thiazolidinediones are limited by adverse effects (see 'Safety' above). Thus, the risks may outweigh the benefits.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diabetes mellitus in adults".)

SUMMARY AND RECOMMENDATIONS

●The thiazolidinediones increase insulin sensitivity by acting on adipose tissue and muscle to increase glucose utilization. To a lesser degree, they decrease glucose production by the liver. (See 'Mechanism of action' above.)

●When used as monotherapy, thiazolidinediones reduce glycated hemoglobin (A1C) values by approximately 0.5 to 1.4 percentage points. (See 'Glycemic efficacy' above.)

●In general, thiazolidinediones are not prescribed as initial therapy in patients with type 2 diabetes. Pioglitazone may have a role as initial therapy in patients who have contraindications to other oral agents (eg, metformin, sulfonylureas), who decline injectable therapies (eg, insulin, glucagon-like peptide 1 [GLP-1] receptor agonists), and in whom medications like dipeptidyl peptidase 4 (DPP-4) inhibitors or sodium-glucose co-transporter 2 (SGLT2) inhibitors cannot be afforded or are found to be less effective. Pioglitazone is more often prescribed as second-line or, more commonly as third-line therapy, when other oral agents in combination are not resulting in adequate glycemic management, especially when cost is a major concern (pioglitazone is available as a low-cost generic). Pioglitazone may also play a role in certain clinical settings; for example, in the context of high risk for hypoglycemia, severe insulin resistance, coexisting nonalcoholic steatohepatitis (NASH), or potentially, recent stroke. However, the risk of weight gain, heart failure, and fractures, and the concerns regarding bladder cancer need to be considered. (See 'Potential indications' above and 'Safety' above and "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Monotherapy failure'.)

●Rosiglitazone is rarely used (and is no longer widely available) because of the increased risk of heart failure and concern for increased risk of atherosclerotic cardiovascular events. (See 'Major cardiovascular events' above.)

●Pioglitazone increases the risk of heart failure but likely decreases the risk of major adverse atherosclerotic cardiovascular events. (See 'Major cardiovascular events' above.)

●Thiazolidinediones have been evaluated as a therapy to prevent diabetes. Due to concerns about side effects, we suggest not using thiazolidinediones for type 2 diabetes prevention (Grade 2B). (See 'Diabetes prevention' above and "Prevention of type 2 diabetes mellitus".)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges David McCulloch, MD, who contributed to earlier versions of this topic review.