Efficacy and safety of fasiglifam (TAK-875), a G protein-coupled receptor 40 agonist, in Japanese patients with type 2 diabetes inadequately controlled by diet and exercise: a randomized, double-blind, placebo-controlled, phase III trial.

AIM: To assess the efficacy and safety of fasiglifam 25 and 50 mg in Japanese patients with type 2 diabetes inadequately controlled by diet and exercise.
METHODS: This phase III, double-blind, placebo-controlled, multicentre study included 192 patients randomized to once-daily treatment with fasiglifam 25 mg (n = 63) or 50 mg (n = 62) or placebo (n = 67) for 24 weeks. The primary efficacy endpoint was the change from baseline in glycated haemoglobin (HbA1c) at week 24.
RESULTS: At week 24, both fasiglifam groups had significantly reduced HbA1c levels compared with the placebo group (p < 0.0001). The least squares mean change from baseline in HbA1c was 0.16% with placebo, -0.57% with fasiglifam 25 mg and -0.83% with fasiglifam 50 mg. The percentage of patients who achieved an HbA1c target of <6.9% at week 24 was also significantly higher (p < 0.05) for fasiglifam 25 mg (30.2%) and 50 mg (54.8%) compared with placebo (13.8%). Fasiglifam significantly reduced fasting plasma glucose levels at all assessment points, starting from week 2. The incidence and types of treatment-emergent adverse events in each fasiglifam group were similar to those in the placebo group, and hypoglycaemia was reported in 1 patient receiving fasiglifam 50 mg. There were no clinically meaningful changes in body weight in any treatment group.
CONCLUSIONS: Fasiglifam significantly improved glycaemic control and was well tolerated, with a low risk of hypoglycaemia in Japanese patients with type 2 diabetes inadequately controlled by diet and exercise; however, in a recent review of data from overall fasiglifam global clinical trials, concerns about liver safety arose and the clinical development of fasiglifam was terminated after this trial was completed.

RCT Entities:   Population Interventions Outcomes

Introduction

Type 2 diabetes is fast becoming a worldwide epidemic which is placing an enormous burden on global healthcare systems and resources 1,2. Persistent uncontrolled hyperglycaemia is the hallmark metabolic abnormality of type 2 diabetes and is attributable largely to ongoing deterioration of insulin-secreting pancreatic β-cell function, which is characterized by both insulin resistance and/or secretory failure 3,4. There is now abundant scientific evidence supporting the notion that achievement of tight glycaemic control will substantially reduce hyperglycaemia-related morbidity 5–8. This would be relatively straightforward if sufficient insulin could be administered to reduce glucose levels to within the ‘normal’ range, but this ignores the devastating effects of hypoglycaemia, the most feared adverse consequence of glucose-lowering therapy 9–11. A wide range of antidiabetic agents have been developed to help improve glycaemic control in patients with type 2 diabetes mellitus and, in Japan, insulin secretagogues, such as sulphonylureas, are commonly used, either as monotherapy or in combination regimens 12–14. Sulphonylureas stimulate insulin secretion by selectively targeting ATP-regulated K+ channels in the plasma membrane of pancreatic β cells. This mechanism is independent of plasma glucose levels, however, and thus increases the risk of hypoglycaemia 15,16. This has led to a search for agents that can augment insulin secretion depending on the circulating glucose level, so as to minimize the chances of hypoglycaemia. In this regard, recent interest has centred on long-chain fatty acids which have been shown to amplify glucose-stimulated insulin secretion from pancreatic β-cells; this effect is mediated through activation of the G protein-coupled receptor 40 (GPR40), also known as free fatty acid receptor 1 17,18. It has been suggested that selective GPR40 agonists may potentiate nutrient-induced insulin secretion and this will favour enhanced prandial secretion and reduce the risk of interprandial hypoglycaemia 19.

Fasiglifam, an orally bioavailable GPR40 agonist, has been shown to produce marked dose-dependent glucose-lowering effects and improvements in other indices of glycaemic control 20–22. In exploratory and dose-finding studies published to date in the USA/Central America and Japan, it has been well tolerated and did not increase the risk of hypoglycaemia 23–25. Indeed, in the US study it was associated with a significantly lower rate of hypoglycaemia than was glimepiride 25.

Based on the antihyperglycaemic effects and overall tolerability of fasiglifam reported in the pilot and dose-ranging studies, it was decided in the present phase III placebo-controlled trial to evaluate the glycaemic effects and safety of fasiglifam 25 and 50 mg administered once daily for the longer duration of 24 weeks in Japanese patients with type 2 diabetes inadequately controlled by diet/exercise. Concerns relating to liver safety, however, after a review of data obtained during the clinical trials’ programme, resulted in termination of the development of fasiglifam after the present trial was completed.

Methods

This was a phase III, multicentre (21 clinical institutions in Japan), randomized, double-blind, parallel group, placebo-controlled 24-week comparative study in Japanese patients with type 2 diabetes inadequately controlled by diet and exercise. After a 4-week screening period, a 24-week treatment period in which patients were randomly assigned to receive either placebo or fasiglifam (25 or 50 mg) followed by a 1-week follow-up period was initiated. The study was performed in accordance with the Declaration of Helsinki and the International Conference on Harmonisation Harmonized Tripartite Guideline for Good Clinical Practice. All participants provided written informed consent. Central or local institutional review boards approved the protocol at each study site.

Patients eligible for the study were aged ≥20 years and had type 2 diabetes mellitus with a glycated haemoglobin (HbA1c) level of between 6.9 and 10.4% at the start of the screening period. All patients had received specific dietary and exercise therapy for at least 8 weeks before the start of screening. Patients were excluded if they had been treated with any antidiabetic drug within 8 weeks before the start of screening or during the screening period, or if they had systolic blood pressure of ≥180 mmHg or diastolic blood pressure of ≥110 mmHg. Additionally, patients with requirement for insulin, those with hepatic or renal impairment, and those with serious cardiovascular, pancreatic or haematological disorders, or with any malignancy were also excluded. Pregnant or lactating women and those of child-bearing age who were not using contraception were not eligible.

Each participant visited the institution at the start and completion of the screening period (weeks −4 and 0) and on weeks 2, 4, 8, 12, 16, 20, 24 and 25. Provided they met the eligibility criteria during week 0, they were randomly allocated to receive double-blind therapy with fasiglifam 25 or 50 mg, or placebo. All treatments were administered as identical tablets once daily before breakfast.

During the 4-week screening period, the following were recorded: demographics/patient characteristics, medical history, concomitant medications, physical examination details, body weight, vital signs, 12-lead ECG (repeated on weeks 12 and 24), clinical laboratory test results, pregnancy test (repeated on week 24), measures of glycaemic control [HbA1c, fasting plasma glucose (FPG), fasting insulin, fasting C-peptide, fasting glucagon and fasting proinsulin levels], adiponectin, compliance with diet/exercise, and pretreatment-emergent adverse events. On weeks 2, 4, 8, 12, 16, 20 and 24 and after an additional 1 week’s follow-up, the following measurements were performed: physical examination, body weight, vital signs, concomitant medications, clinical laboratory tests, compliance with diet/exercise, measures of glycaemic control, such as HbA1c, FPG, fasting insulin, fasting C-peptide, fasting glucagon and fasting proinsulin levels (but not at the follow-up visit), and treatment-emergent adverse events (TEAEs). All assessments were performed at the end of the treatment period at week 24 or at the point of discontinuation for patients who prematurely withdrew from the study.

The study was registered under clinical trials number: JapicCTI-111604.

Efficacy and Safety Outcome Measures

The primary endpoint was the change from baseline in HbA1c at week 24 [last observation carried forward (LOCF)]. Secondary endpoints included changes in HbA1c, FPG, fasting insulin, fasting C-peptide and fasting glucagon levels. Safety was assessed by recording all TEAEs using the Medical Dictionary for Regulatory Activities terminology, vital signs, 12-lead ECG, clinical laboratory test results and body weight. TEAE reporting included investigator assessments for severity, time of onset and relationship to study drug.

All clinical laboratory tests were performed at an independent central laboratory (Mitsubishi Chemical Medience Corporation, Tokyo, Japan). Values for HbA1c (%) were estimated using the National Glycohaemoglobin Standardization Program (NGSP) equivalent values (%), which were calculated using the formula: HbA1c (NGSP) (%) = 1.02 × HbA1c (Japan Diabetes Society) (%) + 0.25%. This equation takes into consideration the relationship between HbA1c (Japan Diabetes Society) (%) calculated using the previous Japanese standard measurement methods and HbA1c (NGSP) 26.

Exploratory endpoints included fasting proinsulin, homeostatic model assessment of insulin resistance (HOMA-IR), homeostatic model assessment of β-cell function (HOMA-β), proinsulin/insulin ratio, adiponectin and high-molecular-weight adiponectin. HOMA-IR and HOMA-β were calculated using the following formulas: HOMA-IR = fasting insulin (μIU/ml) × fasting glucose (mg/dl)/405, HOMA-IR = 360 × [fasting insulin (μIU/ml)/(fasting glucose (mg/dl) − 63)], respectively.

Statistical Methods

Sample size was calculated based on the results of two dose-ranging studies performed in the USA/Central America and Japan 24,25. A total of 60 participants per group (180 participants in all) was assumed to have >90% of power using two sample t-tests to detect differences in the primary endpoint between both fasiglifam groups and placebo [−1.10% in 50 mg and −0.70% in 25 mg vs placebo with a common standard deviation (s.d.) per group of 1.10%] at a significance level of 5%.

The full analysis set was used to evaluate efficacy and included all patients randomly assigned to treatment groups who received at least one dose of double-blind study drug. Missing values of week 24 were imputed by LOCF using any post-baseline values. No evaluable data were excluded for any reason. For the primary endpoint, an analysis of covariance model with study treatment as a fixed effect and baseline HbA1c as a covariate was performed. Based on this model, least squares (LS) means, standard error (s.e.) values, and corresponding two-sided 95% confidence intervals (CIs) were calculated for each treatment group. Both fasiglifam groups were compared with placebo based on a stepwise testing procedure. Two-sided hypothesis testing was performed at a significance level of 5%. Other efficacy endpoints were summarized at each visit by treatment group using descriptive statistics or frequencies, and exploratory two-sided 95% CIs for differences between treatment groups were also calculated.

Safety assessments included all patients who received at least one dose of study medication. Safety findings were summarized using descriptive statistics or frequency distributions.

Results

Details of patient disposition are shown in Figure 1. Of 230 patients who signed a consent form, 192 eligible patients were randomized to placebo (n = 67), fasiglifam 25 mg (n = 63) or fasiglifam 50 mg (n = 62), and 181 (94.3%) completed the study. In total, there were 136 men (70.8%) and 56 women (29.2%) with a mean age of 60.4 years, a mean body mass index of 24.97 kg/m2, and a mean duration of diabetes of 5.6 years. In the randomized population the mean (s.d.) baseline HbA1c and FPG were 7.79 (0.835)% and 8.63 (1.73) mmol/l, respectively. Baseline characteristics were generally similar among all treatment groups (Table1). Compliance with treatment, diet and exercise was high and very similar in all treatment groups.

Figure 1

Participant disposition.

Table 1

Baseline demographic and clinical characteristics

		Fasiglifam
Placebo		25 mg	50 mg	Total
No. of participants	67	63	62	192
Age (years)	61.1 (9.2)	60.2 (10.5)	59.8 (10.9)	60.4 (10.1)
Male	46 (68.7)	43 (68.3)	47 (75.8)	136 (70.8)
Female	21 (31.3)	20 (31.7)	15 (24.2)	56 (29.2)
BMI (kg/m2)	24.85 (3.70)	24.79 (3.90)	25.27 (3.60)	24.97 (3.72)
Duration of diabetes (years)	6.66 (6.50)	5.01 (4.41)	5.00 (4.43)	5.58 (5.27)
HbA1c (%)	7.83 (0.917)	7.84 (0.847)	7.69 (0.728)	7.79 (0.835)
FPG (mmol/l)	8.69 (1.69)	8.66 (1.73)	8.54 (1.78)	8.63 (1.73)
Fasting insulin (μIU/ml)	6.96 (3.837)	7.28 (7.807)	7.21 (4.669)	7.15 (5.644)

Data are mean (s.d.) or number of participants (%). BMI, body mass index; FPG, fasting plasma glucose; HbA1c, glycated haemoglobin.

Clinical Efficacy

At week 24, both fasiglifam groups had a significant reduction in HbA1c compared with the placebo group (p < 0.0001; Figure 2A). The LS mean change from baseline in HbA1c was 0.16% with placebo, −0.57% with fasiglifam 25 mg and −0.83% with fasiglifam 50 mg. The placebo-corrected LS mean (95% CI) reduction from baseline in HbA1c was −0.75% (−0.985, −0.521) with fasiglifam 25 mg and −1.01% (−1.244, −0.777) with fasiglifam 50 mg. The mean change from baseline in HbA1c showed a time-dependent decrease from weeks 2 to 24 in both fasiglifam groups and significant decreases occurred as early as week 4 compared with placebo (Figure 2B). The percentage of patients who achieved the clinical response target [HbA1c level <6.9% (based on the Japanese guideline)] 27 at week 24 was significantly higher in both fasiglifam groups than in the placebo group: 13.8% (9/65 participants) in the placebo group, 30.2% (19/63 participants) in the fasiglifam 25 mg group, 54.8% (34/62 participants) in the fasiglifam 50 mg group (2/67 participants in the placebo group had no post-baseline values of HbA1c and were excluded from the clinical response analysis). Differences (95% CI) in the percentages between the fasiglifam groups and placebo were 16.3% (2.208, 30.417) with fasiglifam 25 mg and 41.0% (26.028, 55.957) with fasiglifam 50 mg. Significant reductions from baseline were recorded for FPG in both fasiglifam groups compared with placebo at week 24 (Figure 3A). The mean change from baseline in FPG was −0.17 mmol/l with placebo, −1.40 mmol/l with fasiglifam 25 mg and −1.41 mmol/l with fasiglifam 50 mg. Differences (95% CI) in the mean changes from baseline in FPG between the fasiglifam groups and placebo were −1.23 mmol/l (−1.70, −0.76) for fasiglifam 25 mg and −1.24 mmol/l (−1.75, −0.73) for fasiglifam 50 mg. The reduction from baseline in FPG occurred from week 2 onwards in both fasiglifam groups, and the reduction was greater in these two groups compared with placebo at all assessment points during treatment (Figure 3B).

Figure 2

Change in glycated haemoglobin (HbA1c; %). (A) Primary endpoint least squares (LS) mean change from baseline at week 24 (LOCF). Error bars show standard error (s.e.) values. *p < 0.0001 versus placebo [analysis of covariance (ancova)]. (B) Mean ± standard deviation (s.d.) values over time.

Figure 3

Change in fasting plasma glucose (FPG, mmol/l). (A) Mean ± standard deviation (s.d.) change from baseline at week 24 (last observation carried forward). *p < 0.05 versus placebo. (B) Mean ± s.d. values overtime.

With regard to other glycaemic measures, such as change in fasting insulin, fasting C-peptide, fasting glucagon, fasting proinsulin, HOMA-IR, proinsulin/insulin ratio, adiponectin or high-molecular-weight adiponectin, no significant differences were recorded at week 24 in the fasiglifam groups compared with the placebo group. In contrast, HOMA-β was significantly increased by both dosages of fasiglifam compared with placebo at all assessment points except for week 24 (Figure S1).

Subgroup analyses were performed on the primary endpoint for baseline patient’s characteristics. HbA1c was decreased at week 24 in the fasiglifam groups compared with the placebo group in all subgroups. The reductions in HbA1c in the fasiglifam groups tended not to be affected by age, gender, BMI or duration of diabetes (data not shown), but were larger when baseline HbA1c levels were higher. Thus, in patients with a baseline HbA1c ≥8.4%, decreases were 1.37 and 1.40% in the fasiglifam 25 and 50 mg groups, respectively, compared with decreases of 0.41 and 0.62% in those with a baseline HbA1c <8.4%.

Safety and Tolerability

Safety and tolerability findings with fasiglifam 25 and 50 mg and placebo, including TEAEs with an incidence of 5% or higher, are summarized in Table2. The overall incidence of TEAEs was similar for the fasiglifam and placebo groups. All events with fasiglifam were mild or moderate in severity, and there was a slightly higher incidence of treatment-related TEAEs in the fasiglifam 50 mg group (14.5% vs ∼6% in the other two groups). The incidence of TEAEs leading to treatment discontinuation was similar between treatments. Serious TEAEs were reported in 1 patient on placebo (renal cell carcinoma) and 2 patients on 25 mg fasiglifam (one each of abnormal hepatic function and uterine cancer). The case of abnormal hepatic function was moderate in severity and led to discontinuation of fasiglifam 25 mg with full recovery. The most commonly reported TEAEs in the fasiglifam groups were nasopharyngitis and upper respiratory tract inflammation. Hypoglycaemia was reported in 1 participant receiving fasiglifam 50 mg, which was mild in severity and did not lead to discontinuation of study treatment.

Table 2

Summary of treatment-emergent adverse events

		Fasiglifam, n (%)
	Placebo, n (%)	25 mg	50 mg
Number of participants	67	63	62
Patients with any TEAEs	40 (59.7)	37 (58.7)	35 (56.5)
Patients with any TEAEs: mild	35 (52.2)	31 (49.2)	35 (56.5)
Patients with any TEAEs: moderate	5 (7.5)	6 (9.5)	0
Patients with any TEAEs: severe	0	0	0
Patients with any treatment-related TEAEs	4 (6.0)	4 (6.3)	9 (14.5)
Patients with any TEAEs leading to drug discontinuation	3 (4.5)	2 (3.2)	1 (1.6)
Patients with any treatment-emergent serious AEs	1 (1.5)	2 (3.2)	0
Deaths	0	0	0
Patients with any TEAEs and a frequency of at least 5% in any group
Nasopharyngitis	11 (16.4)	11 (17.5)	9 (14.5)
Upper respiratory tract inflammation	1 (1.5)	5 (7.9)	5 (8.1)
Patients with any treatment-related TEAEs
Lymphadenopathy	1 (1.5)	0	0
Constipation	1 (1.5)	0	0
Feeling abnormal	0	0	1 (1.6)
Hepatic function abnormal	0	1 (1.6)	1 (1.6)
Cholelithiasis	0	0	1 (1.6)
Bronchitis	0	0	1 (1.6)
Increased alanine aminotransferase	0	1 (1.6)	3 (4.8)
Increased aspartate aminotransferase	0	1 (1.6)	0
Blood creatine phosphokinase increased	0	1 (1.6)	0

TEAEs are listed in order of severity. Each participant was counted only once by maximum severity of Preferred Term or System Organ Class. AE, adverse event; TEAE, treatment-emergent AE.

The incidence of alanine aminotransferase (ALT) ≥3 times the upper limit of normal (ULN) was slightly higher in the fasiglifam 50 mg (6.5%, n = 4) and 25 mg (4.8%, n = 3) groups than in the placebo group (1.5%, n = 1). The incidence of ALT ≥5 times the ULN was similar among all treatment groups: fasiglifam 50 mg, 0% (n = 0); fasiglifam 25 mg, 1.6% (n = 1); and placebo, 1.5% (n = 1). The ALT abnormalities were observed at different time periods in the fasiglifam groups: at week 4 in 1 participant, weeks 8–12 in 3 participants, and weeks 16–24 in 3 participants. Follow-up of the 7 fasiglifam-treated participants with ALT >3× ULN revealed that in 5 participants, levels returned to normal or back to baseline within 4 weeks, either during continued treatment (3 participants) or after stopping study treatment (2 participants). In 2 participants, ALT increases occurred at week 24 and, in both these participants, ALT levels also decreased during the extended follow-up period. Five of 7 fasiglifam-treated participants had evidence of hepatic steatosis. The incidence of other abnormal liver tests [aspartate aminotransferase (AST), alkaline phosphatase or γ-glutamyltransferase ≥3× ULN or total bilirubin ≥2× ULN] were similar among all treatment groups. Study treatment was discontinued in only 1 participant each in the placebo, fasiglifam 25 mg and fasiglifam 50 mg groups, respectively. The participant in the fasiglifam 25 mg group experienced a serious TEAE of abnormal hepatic function (ALT 10× ULN and total bilirubin 3× ULN) which was accompanied by epigastric pain, and gallstones, biliary sludge and adenomyomatosis (confirmed by medical imaging). The participant in the fasiglifam 50 mg group had a TEAE of increased ALT (approximately 4× ULN), and the participant in the placebo group had TEAEs of increased ALT and AST levels, approximately 6× ULN and 3× ULN, respectively. No other participants discontinued study treatment as a result of a hepatic adverse event.

No clinically meaningful differences were found among the treatment groups for vital signs or 12-lead ECGs. There were no clinically meaningful changes in body weight in any treatment group: the mean (s.d.) change from baseline was −0.88 kg (2.84) with placebo, 0.42 kg (1.66) with fasiglifam 25 mg and 0.44 kg (2.09) with fasiglifam 50 mg.

Discussion

One of the main goals of the present phase III comfirmatory study was to ascertain the benefit-to-risk properties of the GPR40 agonist fasiglifam in Japanese patients with type 2 diabetes inadequately controlled by diet and exercise. Fasiglifam 25 and 50 mg administered orally once daily for 24 weeks produced clinically relevant and significant reductions in HbA1c and FPG versus placebo. The percentage of participants who achieved an HbA1c level of <6.9% was 30.2% in the fasiglifam 25 mg group, 54.8% in the fasiglifam 50 mg group and 13.8% in the placebo group, showing that more than half of participants achieved the glycaemic control target with the fasiglifam monotherapy. These results are consistent with previous 12-week dose-ranging studies performed in the USA/Central America and Japan in patients with type 2 diabetes 24,25. In a Japanese dose-ranging study 24, LS mean change from baseline in HbA1c at week 12 was 0.09% with placebo, −0.88% with fasiglifam 25 mg and −1.27% with fasiglifam .50 mg. While the reductions in HbA1c were slightly less in the present study, HbA1c levels at baseline were lower. Since the subgroup analysis showed that HbA1c tended to be reduced more in participants with higher HbA1c levels at baseline, the smaller reduction in HbA1c in the present study might be associated with the lower HbA1c baseline level.

Overall, the incidence of TEAEs in each group was similar for the fasiglifam and placebo groups with the majority of TEAEs being mild in severity. There was a slightly higher incidence of TEAEs in the fasiglifam 50 mg group, but this was not considered clinically relevant and there was no difference between fasiglifam and placebo regarding the number of participants discontinuing treatment as a result of a TEAE. As noted earlier, hypoglycaemia is the most feared TEAE associated with type 2 diabetes and agents that can reduce glucose levels without causing episodes of hypoglycaemia would be advantageous. The novel mode of action of fasiglifam, via GPR40-stimulated glucose-dependent increases in insulin secretion 17, has been proposed as a mechanism which would pose a low risk of causing hypoglycaemia 14,18. The US/Central American study reported an incidence rate of mild hypoglycaemia for fasiglifam 6.25–200 mg (2%) that was similar to that for placebo (3%), and these rates were significantly lower than the rate of 19% reported for glimepiride 22. Similarly, in a Japanese dose-ranging study the incidence of mild hypoglycaemia was 0.7% with fasiglifam compared with 4.1% for glimepiride, and no dose-dependent trends were observed 21. In the present study, which was twice the duration of the earlier studies, only one case of hypoglycaemia was recorded in a participant taking fasiglifam 50 mg and this highlights the low potential for fasiglifam to provoke episodes of hypoglycaemia.

The incidence of ALT ≥3 times the ULN was slightly higher in the fasiglifam groups than in the placebo group, while the incidence of ALT ≥5 times the ULN was no different between fasiglifam and placebo. The majority of participants who experienced elevation of aminotransferases had confounding factors such as hepatic steatosis and/or cholelithiasis. In some cases, however, drug-induced liver injury cannot be excluded completely. The incidence rates of other abnormal liver tests or of discontinuation of study treatment as a result of a hepatic adverse event were similar among all treatment groups. One participant treated with fasiglifam 25 mg developed a serious TEAE of abnormal hepatic function that was considered possibly drug-related by the investigator, but cholecystitis with gallstones and adenomyomatosis may have been a confounding factor in this participant. No participants treated with fasiglifam 50 mg had a serious hepatic-related TEAE. The earlier phase II studies also showed that no dose-dependent trends relating to tolerability and safety, including liver tests, were observed with doses of fasiglifam up to 200 mg once daily for 12 weeks 24,25; however, in a recent review of data from fasiglifam global clinical trials by an independent panel of experts overseeing the clinical development programme, concerns about liver safety arose and the sponsor has voluntarily terminated the clinical development of fasiglifam 28. The mechanisms causing the hepatic changes have not been identified so far. In contrast to most idiosyncratic drug-induced liver injuries, which commonly develop within 3 months after the initiation of a suspected drug, 3 out of 7 participants who experienced an ALT increase after the treatment with fasiglifam had a relatively long latency period in the present study, especially the 2 participants who experienced an ALT increase at the last visit during the treatment period. Although the meaning of this observation is currently unclear, further analysis using global phase III data may provide additional information for better understanding the fasiglifam drug-induced liver injury. Additional research is clearly essential to ascertain the mechanisms causing the hepatic changes and to assess whether they are potentially a class effect related to GPR40 agonism or findings specific to fasiglifam.

Whilst this phase III study was well designed in terms of controls to reduce bias and variation, the 24-week treatment period was short for a disease that will probably need life-long treatment. In addition, the number of participants treated was too small to be able to detect statistically significant differences in HbA1c between the various subgroups tested, or to detect less common TEAEs. These limitations may explain why we were not able to clearly identify any liver safety issues for fasiglifam in the present study.

In conclusion, fasiglifam is the first GPR40 agonist to undergo extensive clinical development. Fasiglifam 25 and 50 mg produced significant improvements in glycaemic control (HbA1c and FPG) with a low risk of causing hypoglycaemia in patients with type 2 diabetes inadequately controlled by diet and exercise.