TABLE 2.
Summary of PPARγ agonist clinical trial results for Alzheimer’s disease.
| Study | Treatment | Study Design | Population | Results |
| A. Rosiglitazone | ||||
| A.1. Phase 2 Intervention | ||||
| Preserved cognition in patients with early Alzheimer disease and amnestic mild cognitive impairment during treatment with rosiglitazone (Watson et al., 2005) | Rosiglitazone 4 mg daily, vs placebo. | 24-week, placebo-controlled, double blind, parallel-group study in subjects with early AD and Amnestic MCI. Outcome measures: cognition, plasma insulin, plasma Aβ. | Placebo, N = 10, rosiglitazone, N = 20. Average age, 73 years, 70% F, 100% White. Baseline insulin, 8.1 μU/mL; MMSE mean baseline, 23. Subjects taking medications to control glucose were excluded. | Rosiglitazone was statistically associated with better delayed recall at 4 and 6 months, selective attention at 6 months, stable plasma Aβ42, Aβ40, and Aβ42/Aβ40 ratio. |
| Efficacy of rosiglitazone in a genetically defined population with mild-to-moderate Alzheimer’s disease (Risner et al., 2006) | Rosiglitazone 2, 4 or 8 mg daily vs placebo. | 24-week, placebo-controlled, double blind, parallel-group pilot study in subjects with mild-to-moderate AD (phase 2). Outcome measures: ADAS-Cog, CIBIC+ | Average N = 128, average age, 70.7 years, 60% F, 100% White. Balanced for APOE ε4. Baseline insulin, 14.2 μU/mL. MMSE mean baseline, 21.3. Subjects with history of T1DM or T2DM, or with fasting glucose ≥ 7mM or HbA1c ≥ 8.5% were excluded. | Overall: no statistically significant effect of rosiglitazone on outcome measures. In APOE ε4 non-carriers, treatment with 8 mg rosiglitazone was statistically associated with improved ADAS-Cog. |
| NCT00265148 Effects of rosiglitazone on cognition and cerebral glucose utilization in subjects with mild to moderate Alzheimer’s disease (Tzimopoulou et al., 2010) | Rosiglitazone 4 mg daily for one month, increasing to 8 mg daily for the remainder of the study, vs placebo. | 52-week parallel group, double blind, phase 2 study. Outcome measures: 12-month cerebral glucose metabolic rate change, brain volume, ADAS-Cog, CIBIC+ | Average N = 31 completed the study, average age, 71.25 years, 46.2% F, 94.9% White. Balanced for APOE ε4. Subjects with history of T1DM or T2DM or taking medications* to control glucose were excluded. | No sustained treatment effect on total or regional glucose metabolic rate or brain volume; no effect on ADAS-Cog or CIBIC+. |
| A.2. Phase 2 – Diabetes | ||||
| Rosiglitazone and cognitive stability in older individuals with type 2 diabetes and MCI (Abbatecola et al., 2010) | Rosiglitazone 4 mg daily, vs metformin, 500 mg daily, vs. rosiglitazone + metformin vs. diet. | 36-week, prospective, randomized, open-controlled study in subjects with mild-to-moderate AD in association with T2DM. Outcome measures: changes in neuropsychological test scores and metabolic control parameters (FIRI, FPG, HbA1C). | Average N = 32.2, average age, 76 years, 45% F; FPG mean baseline, 8.44 mmol/L; FIRI, 148 pmol/L; mean baseline HbA1C, 7.5%; MMSE mean baseline, 24; TMT-A mean baseline, 67.6; TMT-B mean baseline, 161.1; DIFFBA mean base line, 101.2; RAVLT mean baseline, 24.5. | Metformin/rosiglitazone combination stabilized all neuropsychological tests. Metformin stabilized MMSE, TMT-A, TMT-B; diet stabilized MMSE, TMT-A. In linear-fixed effects model, FIRI x time correlated with metformin/rosiglitazone RAVLT. |
| A.3. Phase 3 Intervention | ||||
| NCT00428090 Rosiglitazone (Extended Release Tablets) as monotherapy in subjects with mild to moderate Alzheimer’s disease (Gold et al., 2010) | Rosiglitazone 2 or 8 mg extended release, daily vs placebo (REFLECT-1). | 24-week, double blind, double dummy, randomized, parallel group phase 3 study, stratified for APOE ε4 status in subjects with mild-to-moderate AD. Outcome measures: ADAS-Cog, CIBIC+ | Average N = 159, average age, 72.3 years, 37% F, 72% White, balanced for APOE ε4. ADAS-Cog mean baseline, 19.1. Subjects with history of T1DM or T2DM or taking medications* to control glucose were excluded. | No statistically or clinically significant effect of either rosiglitazone dose in full population; no significant evidence of interaction between treatment with rosiglitazone and APOE genotype. |
| NCT00348309/NCT00348140 Rosiglitazone (Extended Release Tablets) As Adjunctive Therapy For Subjects With Mild To Moderate Alzheimer’s Disease (Harrington et al., 2011) | Rosiglitazone 2 or 8 mg extended release, as adjunctive to donepezil (REFLECT-2), or adjunctive to any AChEIs (REFLECT-3). | 48-week, double blind, randomized, placebo-controlled, parallel group, phase 3 studies, stratified for APOE ε4 status in subjects with mild-to-moderate probable AD. Outcome measures: ADAS-Cog, CDR-SB. | NCT00348309 (REFLECT-2): Average N = 464, average age, 74.1 years, 60% F, 90.67% White, balanced for APOE ε4. ADAS-Cog mean baseline, 25.3; MMSE mean baseline, 19.46. NCT00348140 (REFLECT-3): Average N = 476.3, average, 73.9 years old, 55.6% F, 91.76% White, balanced for APOE ε4. ADAS-Cog mean baseline, 24.1; MMSE mean baseline, 19.7. Subjects with history of T1DM or subjects with T2DM taking medications* to control glucose were excluded. | No statistically or clinically significant effect of either rosiglitazone dose in full population; no significant evidence of interaction between treatment with rosiglitazone and APOE genotype in either REFLECT-2 or REFLECT-3. |
| B. Pioglitazone | ||||
| B.1. Phase 1 - Dose-ranging | ||||
| NCT01456117 Study to assess the effects of daily administration of pioglitazone on brain hemodynamics in cognitively healthy elderly subjects (Knodt et al., 2019) | Pioglitazone, 0.6 mg, 2.1 mg, 3.9 mg, 6.0 mg daily, vs. placebo | A 2-week, multiple-dose, single-blind, randomized, parallel design, placebo-controlled, phase 1 dose-ranging study. Outcome measure: episodic memory-related hippocampal activity, measured via blood oxygen level-dependent (BOLD) functional magnetic resonance imaging. | Average N = 11, average age = 66.08 years, 71% F, 85.45% White. CERAD-WLM mean baseline, 8.02, TMT-B mean baseline, 97.58. Diabetic subjects taking medications to control blood glucose*, or with HbA1C > 6% were excluded. | Statistical association of 0.6 mg/day pioglitazone with increased right hippocampal activation during encoding of novel face-name pairs at day 7 and day 14, relative to baseline. No statistically significant improvement at 2.1, 3.9 or 6.0 mg/day. |
| B.2. Pilot - Diabetes | ||||
| Role of tumor necrosis factor-alpha in cognitive improvement after peroxisome proliferator activator receptor gamma agonist pioglitazone treatment in Alzheimer’s disease (Hanyu et al., 2010) | Pioglitazone, 15 mg daily vs. none. | 24-week, prospective, randomized, open-controlled study, in subjects with mild-to-moderate Alzheimer’s disease in association with T2DM. Outcome measures: ADAS-JCog, MMSE, TNFα, IL-6, C-reactive protein. | N = 17 for both groups, average age 78.7 years, 50% F, 100% White, balanced for APOE ε4 and donepezil use; MMSE mean baseline, 21.85; ADAS-JCog mean baseline, 15.65; TNFα mean baseline, 1.38 pg/mL; IL-6 mean baseline, 2.62 pg/mL; C-reactive protein mean baseline, 0.08 mg/dL. | Pioglitazone was statistically associated with improved ADAS-JCog and TNFα, and changes in ADAS-JCog were correlated with changes in TNFα. |
| Efficacy of PPARγ agonist pioglitazone in mild Alzheimer disease (Sato et al., 2011) | Pioglitazone, 15 or 30 mg daily vs. none. | 24-week prospective randomized, open-controlled study in subjects with mild-to-moderate Alzheimer disease in association with T2DM. Outcome measures: ADAS-JCog, MMSE, WMS-R, rCBF, plasma Aβ40 and Aβ42, HOMA-R, HbA1c, FIRI. | N = 21 for both groups, average age, 77.5 years, 52% F, 100% White, balanced for APOE ε4; balanced for other hypoglycemic agents, donepezil. | Pioglitazone was statistically associated with improved MMSE, ADAS-JCog and WMS-R, with improved blood flow in the parietal lobe, and with improved metabolic factors. The plasma Aβ40/Aβ42 ratio did not change in the pioglitazone group and increased in the control group. ADAS-JCog significantly worsened in the control group. |
| B.3. Phase 2 Intervention | ||||
| NCT00982202 Pioglitazone in Alzheimer’s disease safety trial (Geldmacher et al., 2011) | Pioglitazone, 15 mg daily, escalating weekly to 45 mg daily, vs. placebo. | 72-week, double-blind, randomized, placebo-controlled, group comparison study of mild-to-moderate probable Alzheimer’s disease Outcome measures (collected at 3-month intervals): (Longhe, 2020). Measures of cognition, including ADAS-Cog, CDR-SB. (Patterson, 2018). Estimate for effect size calculations. | Average N = 14.5, average age, 70.95 years, 62% F; MMSE mean baseline, 21; ADAS-Cog mean baseline, 21; CDR-SB mean baseline, 5.8. | Pioglitazone was not statistically associated with any improved measure of cognition; the adjusted mean for ADAS-Cog per month was lower in the pioglitazone group, but not statistically significant. For α = 0.05 and power = 0.80, sample sizes of 340 (170 pio, 170 placebo) and 155 (78 pio, 77 placebo) subjects would be required for their estimated regression coefficients of the pioglitazone effect on ADAS-Cog (-0.746) and CDR-SB (-0.354), respectively, to be significant. |
| NCT00736996 Pioglitazone and exercise effects on older adults with MCI and metabolic syndrome (POEM) (Hildreth et al., 2015) | Pioglitazone, 15 mg daily escalating to 45 mg daily after one month, versus placebo; or 45-75 minutes exercise training 3X/week, vs. status quo exercise. Exercise regimen initiated at 50-60% HR max, escalated to 80-85% HR max over the course of the study. | 24-week double-blind, randomized, placebo-controlled pilot study in sedentary adults with MCI and central obesity. Outcome measures: Change in baseline for cognition, insulin clamp, body composition, metabolic and inflammatory markers. | Average N = 22, average age 65.6 years, 51.8% F, 87.8% White, balanced for APOE ε4; average compliance, pioglitazone, 76%, placebo, 89%; glucose mean baseline, 101 mg/dL; insulin mean baseline, 16.1 μU/mL; C-reactive protein mean baseline, 3 mg/L; IL-6 mean baseline, 1.7 pg/mL; TNFα mean baseline, 1.56 pg/mL; MMSE mean baseline, 28.6; ADAS-Cog mean baseline, 6. | Pioglitazone was not statistically associated with any improved measure of cognition; performance on the Visual Reproduction Test; scores worsened in the pioglitazone group vs. placebo; ADAS-Cog improved with exercise (-1.3 EX vs. -0.3 CON; P = 0.05). No statistically significant correlations between glucose disposal rates and cognitive performance. |
| B.4. Phase 3 Prevention | ||||
| NCT01931566 A study to simultaneously qualify a biomarker algorithm for prognosis of risk of developing MCI Due to AD and to test the safety and efficacy of pioglitazone to delay the onset of MCI due to AD in cognitively normal subjects (Alexander et al., 2019; Burns et al., 2019) NOTE: The full publication describing this study was under review when the current paper was submitted. | Pioglitazone, 0.8 mg extended release daily, vs placebo. | Event-driven (anticipated 5 yr.), double-blind, randomized, parallel group placebo-controlled Phase 3 prevention study in cognitively normal adults susceptible for AD (APOE, TOMM40 genotypes and age). Outcome measures: Delay onset of MCI in normal participants who are at increased risk due to age and genetic risk factors. | N = 433 low-risk placebo, 1516 high-risk placebo, 1545 high risk pioglitazone. Average age, 73.1 years, 56.16% F, 96.6% White; average APOE ε4 carriage in the high-risk groups, 92.45%; MMSE mean baseline, 28.56. Outcome measure: Time to diagnosis of MCI due to AD for pioglitazone-treated subjects vs placebo in high-risk stratum. Pre-specified futility threshold, 30% conditional probability that a 40% treatment difference would be detected. | Study terminated due to futility analysis. After 1278 days, total events in placebo, 46; total events with pioglitazone, 39. Pioglitazone risk ratio vs placebo was 0.8 (95% CI, 0.45 – 1.4), P = 0.307; post-hoc subgroup analysis suggests possible benefit of pioglitazone for males. |
ADAS-Cog, Alzheimer’s Disease Assessment Scale-Cognitive Subscale; ADAS-JCog, Japanese version of the ADAS-Cog; ATP III, Adult treatment panel III criteria for central obesity; rCBF, regional cerebral blood flow; CDR-SB, Clinical Dementia Rating scale-Sum of Boxes; CIBIC+, Clinician’s interview-based impression of change with caregiver input; DIFFBA, TMT-A minus TMT-B – a measure of cognitive efficiency; FIRI, fasting immunoreactive insulin; FPG, fasting plasma glucose; HOMA-R, homeostatic model assessment for insulin resistance; MCI, mild cognitive impairment; MMSE, Mini-mental state examination; RAVLT, Rey Auditory-verbal learning test; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus; TNFα, tumor necrosis factor-alpha; TMT-A, TMT-B, Trail marking test A and Trail marking test-B, respectively. *Insulin, sulfonylureas, PPARγ agonists or glitinides.