Ertugliflozin Delays Insulin Initiation and Reduces Insulin Dose Requirements in Patients With Type 2 Diabetes: Analyses From VERTIS CV

Abstract

Context

VERTIS CV evaluated the cardiovascular safety of ertugliflozin in patients with type 2 diabetes and atherosclerotic cardiovascular disease (ASCVD).

Objective

The aim of these analyses was to assess the insulin requirements of VERTIS CV patients over the trial duration.

Methods

Patients received ertugliflozin 5 mg, 15 mg, or placebo once daily; mean follow-up was 3.5 years. Time to insulin initiation in patients who were insulin naïve at baseline, change in insulin dose in patients receiving baseline insulin, and hypoglycemia incidence in both patient groups were assessed.

Results

In VERTIS CV, mean duration of type 2 diabetes was 13.0 years; glycated hemoglobin was 8.2%. Among 4348 (53%) insulin-naïve patients, the likelihood of insulin initiation was significantly reduced with ertugliflozin vs placebo (ertugliflozin 5 mg: hazard ratio [HR] 0.70, 95% CI 0.58-0.84; ertugliflozin 15 mg: HR 0.64, 95% CI 0.53-0.78). Time to insulin initiation was delayed with ertugliflozin; the estimated delay in reaching a 10% cumulative incidence of new insulin initiations vs placebo was 399 days with ertugliflozin 5 mg and 669 days with ertugliflozin 15 mg. Among 3898 (47%) patients receiving baseline insulin, the likelihood of requiring a ≥20% increase in insulin dose was significantly reduced with ertugliflozin vs placebo (ertugliflozin 5 mg: HR 0.62, 95% CI 0.52-0.75; ertugliflozin 15 mg: HR 0.51, 95% CI 0.41-0.62). The incidence of hypoglycemia events was not increased with ertugliflozin treatment.

Conclusion

In VERTIS CV patients, ertugliflozin reduced the likelihood of insulin initiation, delayed the time to insulin initiation by up to ∼1.8 years, and reduced insulin dose requirements vs placebo, without increasing hypoglycemia events.

ertugliflozin, type 2 diabetes mellitus, glycated hemoglobin, insulin sparing

Insulin therapy, usually in combination with other antihyperglycemic medications, remains an important treatment option for people with type 2 diabetes, particularly those with long-standing disease or severe hyperglycemia (1). However, adherence to insulin therapy can be affected by patients’ fear of injections, hypoglycemic events, and weight gain (2, 3). In addition to patients’ aversion to injections, and perhaps because of it, reluctance to initiate insulin therapy in patients who need optimization of glycemic control has been well documented among primary care providers (4-7).

Sodium–glucose cotransporter 2 (SGLT2) inhibitors are a class of antihyperglycemic medications that act by limiting re-uptake of glucose from urine filtrate in the proximal convoluted tubule of the kidney, thereby decreasing circulating levels of glucose leading to lowering of glycated hemoglobin (HbA_1c) levels in patients with hyperglycemia (8, 9). SGLT2 inhibitors have a favorable safety profile for hypoglycemia and can effect clinically meaningful and sustained reductions in body weight over time (10-12). The glycemic efficacy of SGLT2 inhibitors may facilitate insulin sparing through avoidance of or dose reductions in background insulin use (13).

Ertugliflozin is an SGLT2 inhibitor approved for use in adults with type 2 diabetes as an adjunct to diet and exercise to improve glycemic control (14, 15). The eValuation of ERTugliflozin effIcacy and Safety CardioVascular (VERTIS CV) trial evaluated the effects of ertugliflozin treatment on cardiorenal outcomes in patients with type 2 diabetes and established atherosclerotic cardiovascular disease (ASCVD) (16, 17). The present analyses investigated insulin use in VERTIS CV trial participants who were either insulin naïve or receiving insulin at baseline to assess whether treatment with ertugliflozin compared with placebo affected insulin requirements over the duration of the trial. The analyses also assessed change from baseline in HbA_1c and the incidence of hypoglycemic events for each patient subgroup.

Materials and Methods

Trial Design

The VERTIS CV trial (ClinicalTrials.gov identifier: NCT01986881) was a multicenter, randomized, double-blind, placebo-controlled, parallel-group, event-driven trial conducted in patients with type 2 diabetes and established ASCVD to assess the cardiovascular safety and efficacy of ertugliflozin vs placebo (16, 17). Patients were randomized (1:1:1) to once-daily ertugliflozin 5 mg, ertugliflozin 15 mg, or placebo in addition to background standard-of-care treatment, which could include the use of insulin therapy at study entry (16, 17). After week 18 in the trial, changes in background antihyperglycemic medications—including initiation of insulin—were encouraged to meet individualized treatment requirements for optimal glycemic control.

Trial Population

Patients were eligible for inclusion in the trial if they were ≥40 years of age with type 2 diabetes (HbA_1c 7.0-10.5%, inclusive) and had stable, established ASCVD involving the coronary, cerebrovascular, and/or peripheral arterial systems. Key exclusion criteria included a history of type 1 diabetes, a history of ketoacidosis, or an estimated glomerular filtration rate <30 mL/min/1.73 m². The trial was conducted in compliance with the ethics principles of the Declaration of Helsinki and with International Conference on Harmonisation Good Clinical Practice Guidelines. Written informed consent was obtained from all participants, and the final protocol and informed consent documentation were reviewed and approved by the institutional review board or independent ethics committee at each investigational center. For the present analyses, enrolled patients were categorized by baseline use of insulin.

Study Outcomes

Assessments of insulin use in VERTIS CV patients receiving ertugliflozin vs placebo were prespecified in the trial analysis plan and consisted of (1) time to initiation of insulin in patients not being treated with insulin at baseline; and (2) change from baseline in insulin dose over time in patients being treated with insulin at baseline. Assessments of the effect of ertugliflozin on HbA_1c change from baseline and the incidence of hypoglycemic events evaluated as part of these subgroup analyses in patients with and without insulin use at baseline, and the assessment of time to first ≥20% increase in insulin dose in patients with insulin use at baseline, were conducted in post hoc analyses.

Hypoglycemia Assessments

Hypoglycemia adverse event (AE) assessments covered the period from randomization to 14 days after the final dose of study medication, and from randomization to 2 days after the final dose of study medication for laboratory assessments. A hypoglycemia assessment log was used to collect information on potential episodes of hypoglycemia (including a concurrent fingerstick glucose value). Hypoglycemic events were categorized as documented symptomatic, asymptomatic, or severe. Documented symptomatic hypoglycemia was an event during which symptoms of hypoglycemia were accompanied by a measured fingerstick or plasma glucose concentration ≤70 mg/dL. Asymptomatic hypoglycemia was an event not accompanied by symptoms of hypoglycemia but with a measured fingerstick or plasma glucose concentration ≤70 mg/dL. Severe hypoglycemia was an event requiring the assistance of another person to actively administer carbohydrate, glucagon, or other resuscitative actions.

Statistical Methods

Analyses were conducted in the population of randomized patients who received at least 1 dose of study medication (ertugliflozin or placebo).

In the analyses of insulin use, for patients who were not being treated with insulin at baseline, time to insulin initiation was analyzed using a Cox proportional hazards model and included Kaplan–Meier estimates and hazard ratios (HRs), 95% CIs, and P values for likelihood of insulin initiation. Estimates of the time taken for patients to reach cumulative incidences of new insulin initiations of 5%, 10%, and 15% were also analyzed. For patients being treated with insulin at baseline, change in insulin dose over time from baseline to month 60 was calculated; change in insulin dose from baseline to last visit on insulin was also determined, providing a comprehensive analysis of dose adjustments by capturing all insulin users during the course of the study. The difference and 95% CI for the difference in change in insulin dose from baseline to last visit on insulin between ertugliflozin 5 mg and placebo, and ertugliflozin 15 mg and placebo, was calculated using the 2-sample t-test procedure. In addition, for patients being treated with insulin at baseline, the time to first ≥20% increase in average daily insulin dose (computed monthly) was assessed by Cox proportional hazards modeling.

Analyses of change in HbA_1c were conducted in the same population of patients who, in addition, had at least 1 postbaseline assessment of HbA_1c. Change in HbA_1c over time from baseline to month 60 was summarized. Change in HbA_1c at the week 18 time point was also calculated separately, as this was the timeframe for formal testing of glycemic efficacy in the VERTIS CV trial during which background antihyperglycemic medications were held constant to allow an assessment of HbA_1c with minimal confounding from changes in background antihyperglycemic medications during the first 18 weeks of the trial (16, 17).

For the analysis of hypoglycemia AEs, the estimated difference and 95% CI for the difference in percentage of patients with hypoglycemia events in the ertugliflozin groups vs the placebo group was calculated using the Miettinen and Nurminen method (18).

Results

Baseline Characteristics

Of 8246 patients randomized in the VERTIS CV trial, 53% (n = 4348) were not being treated with insulin at baseline and 47% (n = 3898) were being treated with insulin at baseline. Among patients with baseline insulin use, the majority (∼60%) were receiving a basal–bolus regimen (60.1% in the ertugliflozin group and 61.2% in the placebo group); ∼30% were using only basal insulin (32.6% and 32.3%, respectively); and fewer than 10% were using only bolus insulin (7.2% and 6.5%, respectively) (Table 1).

Table 1.

Baseline characteristics of patients with and without insulin use at baseline

	Patients with baseline insulin use			Patients without baseline insulin use
Characteristic	Placebo n = 1345	Ertugliflozin n = 2553	All n = 3898	Placebo n = 1402	Ertugliflozin n = 2946	All n = 4348
Age, y	64.5 (7.7)	64.7 (7.9)	64.6 (7.8)	64.3 (8.3)	64.1 (8.3)	64.2 (8.3)
Male, n (%)	900 (66.9)	1763 (69.1)	2663 (68.3)	1003 (71.5)	2103 (71.4)	3106 (71.4)
White race,a n (%)	1186 (88.2)	2241 (87.8)	3427 (87.9)	1228 (87.6)	2585 (87.7)	3813 (87.7)
BMI, kg/m2	32.8 (5.5)	32.6 (5.4)	32.7 (5.4)	31.1 (5.3)	31.3 (5.2)	31.3 (5.2)
Duration of T2D, y	16.2 (8.7)	16.2 (8.5)	16.2 (8.6)	10.1 (6.9)	10.1 (6.9)	10.1 (6.9)
HbA1c, %	8.4 (0.9)	8.4 (0.9)	8.4 (0.9)	8.1 (0.9)	8.1 (1.0)	8.1 (1.0)
Disease history, n (%)
Dyslipidemia	1108 (82.4)	2061 (80.7)	3169 (81.3)	967 (69.0)	2041 (69.3)	3008 (69.2)
CAD	1027 (76.4)	1945 (76.2)	2972 (76.2)	1085 (77.4)	2199 (74.6)	3284 (75.5)
DMD	676 (50.3)	1244 (48.7)	1920 (49.3)	413 (29.5)	828 (28.1)	1241 (28.5)
Diuretic use, n (%)	823 (61.2)	1430 (56.0)	2253 (57.8)	648 (46.2)	1337 (45.4)	1985 (45.7)
AHA use, n (%)
Insulin	1345 (100)	2553 (100)	3898 (100)	0 (0)	0 (0)	0 (0)
Basal + bolusb	823 (61.2)	1535 (60.1)	2358 (60.5)	—	—	—
Basal alonec	434 (32.3)	832 (32.6)	1266 (32.5)	—	—	—
Bolus aloned	87 (6.5)	183 (7.2)	270 (6.9)	—	—	—
Metformin	907 (67.4)	1667 (65.3)	2574 (66.0)	1215 (86.7)	2497 (84.8)	3712 (85.4)
Sulfonylureas	213 (15.8)	426 (16.7)	639 (16.4)	905 (64.6)	1839 (62.4)	2744 (63.1)
DPP4i	100 (7.4)	187 (7.3)	287 (7.4)	193 (13.8)	431 (14.6)	624 (14.4)

Values are mean (SD) unless otherwise stated.

Abbreviations: AHA, antihyperglycemic agent; BMI, body mass index; CAD, coronary artery disease; DMD, diabetic microvascular disease (includes diabetic complications such as retinopathy, nephropathy, and neuropathy); DPP4i, dipeptidyl peptidase-4 inhibitor; HbA_1c, glycated hemoglobin; T2D, type 2 diabetes.

Race was reported by the participant.

Basal + bolus: (intermediate/long-acting and short-acting insulin) OR premix insulin.

Basal alone: intermediate/long-acting insulin.

Bolus alone: short-acting insulin.

Baseline characteristics did not differ between ertugliflozin and placebo treatment groups within the subgroups of patients with and without baseline insulin use (Table 1). Although no significance testing was performed, patients who were receiving baseline insulin had a longer mean duration of type 2 diabetes, a higher mean HbA_1c level, more prevalent dyslipidemia and diabetic microvascular disease, more frequent use of diuretics, and less frequent use of other antihyperglycemic agents—particularly the use of sulfonylureas—compared with patients who were not receiving baseline insulin (Table 1). Mean age did not differ between the subgroups of patients with or without insulin use at baseline (Table 1).

Glycemic Efficacy

Treatment with ertugliflozin 5- and 15-mg doses reduced mean HbA_1c levels over the duration of the trial compared with placebo in patients with or without insulin treatment at baseline (Fig. 1). In patients receiving insulin at baseline, the mean (95% CI) HbA_1c change from baseline to week 18 (the primary glycemic endpoint for the VERTIS CV trial) was −0.70% (−0.74, −0.66) with ertugliflozin 5 mg, −0.70% (−0.74, −0.65) with ertugliflozin 15 mg, and −0.22% (−0.26, −0.17) with placebo. Similar reductions in HbA_1c from baseline to week 18 were observed in patients who were not receiving insulin at baseline: −0.75% (−0.79, −0.71) with both ertugliflozin 5 mg and ertugliflozin 15 mg, and −0.25% (−0.29, −0.20) with placebo.

Figure 1.

Mean (SE) HbA_1c change from baseline over time in (A) patients with insulin use at baseline and (B) patients without insulin use at baseline. Numbers are the number of patients with nonmissing assessments at both baseline and the specific timepoint. Abbreviations: BL, baseline; ERTU, ertugliflozin; HbA_1c, glycated hemoglobin; PBO, placebo; SE, standard error.

Insulin Initiations

In 4341 patients not being treated with insulin at baseline and who received treatment with ertugliflozin or placebo during the VERTIS CV trial, insulin was initiated in 184 patients (12.8%) in the ertugliflozin 5-mg group, 174 patients (11.6%) in the ertugliflozin 15-mg group, and 245 patients (17.5%) in the placebo group. Cox proportional hazard modeling demonstrated a significantly reduced likelihood of insulin initiation for both the ertugliflozin 5-mg and 15-mg doses compared with placebo: HR 0.70, 95% CI 0.58-0.84 (P < .001) for ertugliflozin 5 mg; and HR 0.64, 95% CI 0.53-0.78 (P < .001) for ertugliflozin 15 mg (Fig. 2).

Figure 2.

Time to insulin initiation in patients without insulin use at baseline. Time to insulin initiation was analyzed using a Cox proportional hazards model that included treatment as a covariate and randomized assignment as a stratification factor. The time delay in patients reaching cumulative incidences of new insulin initiations of 5%, 10%, and 15% with ertugliflozin 5 mg and ertugliflozin 15 mg vs placebo is also shown (arrows and table). Numbers are the number of patients at risk at the specific timepoint. Abbreviations: BL, baseline; ERTU, ertugliflozin; HR, hazard ratio; PBO, placebo.

An assessment of the time taken to reach cumulative incidences of new insulin initiations of 5%, 10%, and 15% for patients in the ertugliflozin groups vs the placebo group demonstrated that treatment with ertugliflozin delayed new initiations of insulin therapy (Fig. 2). Delays of up to 245 days to reach a 5% incidence, of up to 669 days to reach a 10% incidence, and of up to 619 days to reach a 15% incidence of new insulin initiations were observed in the ertugliflozin groups compared with placebo (Fig. 2).

Insulin Dose Adjustments

In 3897 patients being treated with insulin at baseline and who received treatment with ertugliflozin or placebo during the VERTIS CV trial, mean daily insulin doses were stable or slightly decreased in patients in the ertugliflozin treatment groups, whereas an increase in mean daily insulin dose was observed in patients in the placebo group (Fig. 3). The mean (95% CI) change in insulin dose from baseline to last visit on insulin was 1.02 IU/day (−0.14, 2.17) with ertugliflozin 5 mg, −1.70 IU/day (−3.03, −0.37) with ertugliflozin 15 mg, and 5.20 IU/day (3.79, 6.61) with placebo. The difference (95% CI) vs placebo was −4.18 IU/day (−6.35, −2.01) for ertugliflozin 5 mg and −6.90 IU/day (−9.21, −4.59) for ertugliflozin 15 mg.

Figure 3.

Mean (SE) insulin dose change over time and at last visit on insulin in patients with insulin use at baseline. Numbers are the number of patients with nonmissing assessments at both baseline and the specific timepoint. Abbreviations: BL, baseline; ERTU, ertugliflozin; PBO, placebo; SE, standard error.

An increase of ≥20% in the average daily insulin dose (computed monthly) occurred in 175 patients (13.4%) in the ertugliflozin 5-mg group, 140 patients (11.2%) in the ertugliflozin 15-mg group, and 273 patients (20.3%) in the placebo group. Cox proportional hazard modeling demonstrated a significantly reduced likelihood of requiring a ≥20% increase in daily insulin dose for both the ertugliflozin 5-mg and 15-mg doses compared with placebo: HR 0.62, 95% CI 0.52-0.75 (P < .001) for ertugliflozin 5 mg; and HR 0.51, 95% CI 0.41-0.62 (P < .001) for ertugliflozin 15 mg (Fig. 4). More patients in the ertugliflozin 15-mg arm (4.1%) had discontinued insulin therapy at end of study compared with the ertugliflozin 5-mg (2.4%) or placebo (2.2%) arms. A small and generally similar reduction in the proportion of patients on sulfonylureas was observed across treatment arms (3.2-3.8%) during the VERTIS CV study (Table 2).

Figure 4.

Time to ≥20% increase in average daily insulin dose (computed monthly) in patients with insulin use at baseline. Time to ≥20% increase in average daily insulin dose was analyzed using a Cox proportional hazards model that included treatment as a covariate and randomized assignment as a stratification factor. Average daily insulin dose was computed monthly. Numbers are the number of patients at risk at the specific timepoint. Abbreviations: BL, baseline; ERTU, ertugliflozin; HR, hazard ratio; PBO, placebo.

Table 2.

Change in antihyperglycemic medication use between baseline and end of study in VERTIS CV patients

Antihyperglycemic medication	Placebo	Ertugliflozin 5 mg	Ertugliflozin 15 mg	All
	n = 2745	n = 2746	n = 2747	n = 8238
Sulfonylureas
Baseline, n	1121	1120	1147	3388
End of study, n	1078	1084	1110	3272
Difference, n	−43	−36	−37	−116
Difference, %	−3.8	−3.2	−3.2	−3.4
Biguanides
Baseline, n	2122	2070	2097	6289
End of study, n	2095	2033	2036	6164
Difference, n	−27	−37	−61	−125
Difference, %	−1.3	−1.8	−2.9	−2.0
DPP4i
Baseline, n	290	300	318	908
End of study, n	392	386	370	1148
Difference, n	102	86	52	240
Difference, %	35.2	28.7	16.4	26.4
GLP-1 RA
Baseline, n	86	109	83	278
End of study, n	153	140	128	421
Difference, n	67	31	45	143
Difference, %	77.9	28.4	54.2	51.4
Thiazolidinediones
Baseline, n	59	50	49	158
End of study, n	93	55	56	204
Difference, n	34	5	7	46
Difference, %	57.6	10	14.3	29.1

Values are n or %.

Abbreviations: DPP4i, dipeptidyl peptidase-4 inhibitor; GLP-1 RA, glucagon-like peptide-1 receptor agonist.

Hypoglycemia Incidence

Patients receiving baseline insulin had a higher incidence of hypoglycemic events over the duration of the study than patients who were not receiving insulin at baseline (Table 3). The incidence of hypoglycemic events generally did not differ between the ertugliflozin groups and the placebo group within patient subgroups with or without baseline insulin use (Table 3). In patients receiving insulin at baseline, the total number of symptomatic and severe hypoglycemia episodes were lower, and asymptomatic hypoglycemia episodes were higher, in the ertugliflozin groups than the placebo group (Table 3).

Table 3.

Incidence of hypoglycemia events in patients with and without insulin use at baseline

	Patients with baseline insulin use			Patients without baseline insulin use
	Placebo	Ertugliflozin 5 mg	Ertugliflozin 15 mg	Placebo	Ertugliflozin 5 mg	Ertugliflozin 15 mg
Hypoglycemia eventsa	n = 1345	n = 1304	n = 1248	n = 1400	n = 1442	n = 1499
Documented symptomatic, n (%)	553 (41.1)	546 (41.9)	503 (40.3)	206 (14.7)	194 (13.5)	192 (12.8)
Difference in % (95% CI) vs placebob	—	0.8 (−3.0, 4.5)	−0.8 (−4.6, 3.0)	—	−1.3 (−3.8, 1.3)	−1.9 (−4.4, 0.6)
Total number of episodes	7712	6683	5637	1247	931	886
Patients with 1 episode, n (%)	106 (7.9)	121 (9.3)	101 (8.1)	71 (5.1)	72 (5.0)	81 (5.4)
Patients with 2 episodes, n (%)	70 (5.2)	59 (4.5)	75 (6.0)	32 (2.3)	36 (2.5)	30 (2.0)
Patients with ≥3 episodes, n (%)	377 (28.0)	366 (28.1)	327 (26.2)	103 (7.4)	86 (6.0)	81 (5.4)
Asymptomatic, n (%)	686 (51.0)	711 (54.5)	672 (53.8)	331 (23.6)	365 (25.3)	370 (24.7)
Difference in % (95% CI) vs placebob	—	3.5 (−0.3, 7.3)	2.8 (−1.0, 6.7)	—	1.7 (−1.5, 4.8)	1.0 (−2.1, 4.2)
Total number of episodes	7819	9555	8730	2357	2598	2537
Patients with 1 episode, n (%)	165 (12.3)	129 (9.9)	155 (12.4)	110 (7.9)	131 (9.1)	128 (8.5)
Patients with 2 episodes, n (%)	106 (7.9)	97 (7.4)	96 (7.7)	48 (3.4)	57 (4.0)	60 (4.0)
Patients with ≥3 episodes, n (%)	415 (30.9)	485 (37.2)	421 (33.7)	173 (12.4)	177 (12.3)	182 (12.1)
Documented (symptomatic/asymptomatic), n (%)	820 (61.0)	810 (62.1)	788 (63.1)	395 (28.2)	427 (29.6)	432 (28.8)
Difference in % (95% CI) vs placebob	—	1.2 (−2.6, 4.9)	2.2 (−1.6, 5.9)	—	1.4 (−1.9, 4.7)	0.6 (−2.7, 3.9)
Total number of episodes	15 531	16 238	14 367	3604	3529	3423
Patients with 1 episode, n (%)	151 (11.2)	118 (9.0)	143 (11.5)	117 (8.4)	134 (9.3)	136 (9.1)
Patients with 2 episodes, n (%)	78 (5.8)	77 (5.9)	83 (6.7)	52 (3.7)	61 (4.2)	65 (4.3)
Patients with ≥3 episodes, n (%)	591 (43.9)	615 (47.2)	562 (45.0)	226 (16.1)	232 (16.1)	231 (15.4)
Severe, n (%)	123 (9.1)	118 (9.0)	121 (9.7)	39 (2.8)	18 (1.2)	27 (1.8)
Difference in % (95% CI) vs placebob	—	−0.1 (−2.3, 2.1)	0.6 (−1.7, 2.8)	—	−1.5 (−2.6, −0.5)	−1.0 (−2.1, 0.1)
Total number of episodes	404	334	303	61	27	58
Patients with 1 episode, n (%)	63 (4.7)	62 (4.8)	70 (5.6)	28 (2.0)	12 (0.8)	17 (1.1)
Patients with 2 episodes, n (%)	24 (1.8)	19 (1.5)	27 (2.2)	6 (0.4)	4 (0.3)	4 (0.3)
Patients with ≥3 episodes, n (%)	36 (2.7)	37 (2.8)	24 (1.9)	5 (0.4)	2 (0.1)	6 (0.4)

Proportions of patients with ≥1 episode of a hypoglycemia event; patients are counted a single time for each applicable category. Hypoglycemia events were categorized as documented symptomatic hypoglycemia (an event during which symptoms of hypoglycemia are accompanied by a measured fingerstick or plasma glucose concentration ≤70 mg/dL); asymptomatic hypoglycemia (an event not accompanied by symptoms of hypoglycemia but with a measured fingerstick or plasma glucose concentration ≤70 mg/dL); and severe hypoglycemia (an event requiring the assistance of another person to actively administer carbohydrate, glucagon, or other resuscitative actions).

The estimated difference and 95% CI for the difference between ertugliflozin 5 mg vs placebo and ertugliflozin 15 mg vs placebo were calculated using the Miettinen and Nurminen method (18).

Discussion

In the VERTIS CV trial in patients with type 2 diabetes and ASCVD, addition of ertugliflozin to a background standard-of-care therapy resulted in a significant reduction in the overall insulin requirements relative to placebo. Patients who were not being treated with insulin at study entry and received ertugliflozin had a significantly reduced likelihood of initiating insulin during the trial compared with patients who received placebo, with reductions in the relative risk of a new insulin initiation with ertugliflozin 5 mg and 15 mg of 30% and 36%, respectively. Among those patients who required a new insulin initiation during the trial, time to insulin initiation was delayed by up to 1.8 years in those receiving ertugliflozin compared with placebo. In addition to reducing or delaying new insulin initiations, treatment with ertugliflozin also resulted in a stabilization of, or slight decrease in, daily insulin doses in those patients who were already being treated with insulin at baseline; conversely, a progressive increase in daily insulin dose was observed with placebo. In patients who were being treated with insulin at baseline and received ertugliflozin, a significantly reduced likelihood of requiring a ≥20% increase in the average daily insulin dose was observed compared with placebo, with reductions in the relative risk of such insulin dose increases with ertugliflozin 5 mg and 15 mg of 38% and 49%, respectively. The reduction in insulin requirements with ertugliflozin was achieved alongside robust and durable reductions in HbA_1c that were similarly observed in patients with and without insulin use at baseline. The results of these analyses add to the growing body of evidence that suggest SGLT2 inhibitors are a useful adjunct for the treatment of type 2 diabetes in patients with and without concurrent insulin use.

Our present findings are consistent with recent reports from clinical trials showing a delay in the need for initiation of insulin therapy among patients with type 2 diabetes treated with other SGLT2 inhibitors compared with placebo (19, 20). Most recently, an analysis of EMPA-REG OUTCOME trial data assessing sustained insulin initiation rates in insulin-naïve patients with type 2 diabetes and cardiovascular disease treated with empagliflozin or placebo (19) found that insulin initiations were significantly reduced in patients treated with empagliflozin compared with placebo (adjusted HR 0.40, 95% CI 0.32-0.49; P < .0001), and empagliflozin delayed the time to insulin initiation over the duration of the trial (19). Furthermore, an analysis of real-world data from primary care patients found that, compared with patients who had received only metformin monotherapy prior to initiating insulin therapy, time to insulin initiation was increased by 276 days with sulfonylureas, by 398 days with dipeptidyl peptidase 4 inhibitors, and by 637 days with SGLT2 inhibitors (all P < .001) (21). Together with our current analyses, these findings indicate that treatment with an SGLT2 inhibitor vs placebo can be insulin sparing.

As we observed in VERTIS CV, reductions in insulin dose requirements following treatment with SGLT2 inhibitors in patients with type 2 diabetes receiving insulin therapy have been reported in other clinical trials (19, 22-27). Of note, the above-mentioned analysis of EMPA-REG OUTCOME trial data (19) also assessed insulin dose changes in patients receiving insulin at baseline. Empagliflozin reduced the proportion of patients requiring a >20% increase in insulin dose compared with placebo (adjusted HR 0.42, 95% CI 0.36-0.49; P < .0001) (19). Empagliflozin treatment also significantly increased the proportion of patients achieving sustained and appropriate (ie, without increases in HbA_1c) insulin dose reductions of >20% from baseline compared with placebo (adjusted HR 1.87, 95% CI 1.39-2.51; P < .0001) (19).

Several mechanisms could have contributed to the decrease in insulin initiations observed following treatment with ertugliflozin. In addition to improvement in glycemic control and amelioration of glucotoxicity, ertugliflozin and other SGLT2 inhibitors are well known to decrease body weight (28). Furthermore, ertugliflozin treatment has been found to be associated with decreased indices of hepatic steatosis (29). The resultant improvement in insulin sensitivity following weight loss and reduction in liver fat can be expected to decrease insulin requirements. Additional mechanisms involving anti-inflammatory effects of SGLT2 inhibitors, by decreasing cytokine expression, can also improve insulin sensitivity and decrease the need for exogenous insulin (30).

We observed that the reductions in mean HbA_1c levels over the duration of the VERTIS CV trial in patients treated with ertugliflozin vs placebo were similar between the patient subgroups with and without baseline insulin use. However, patients receiving baseline insulin experienced a higher incidence of hypoglycemic events than those not receiving insulin at baseline. Nonetheless, the incidence of hypoglycemia events generally did not differ between ertugliflozin and placebo groups with or without baseline insulin use. Overall, our study indicates that SGLT2 inhibitors may offer a safe and effective add-on treatment option in patients with type 2 diabetes receiving insulin who need additional glycemic control, and may facilitate a reduction in daily insulin requirements in these patients.

This study should be considered alongside its limitations. This was a subgroup analysis of VERTIS CV trial data rather than a dedicated study to assess changes in insulin use following ertugliflozin treatment and, as such, is subject to the limitations inherent to this type of analysis. However, the assessment of treatment effects within each patient subgroup rather than between patient subgroups serves to minimize any potential bias. The VERTIS CV protocol stipulated that background antihyperglycemic medications be held constant to week 18 to allow the assessment of ertugliflozin glycemic efficacy without confounding changes in background antihyperglycemic therapy during that time. Given that protocol requirement, our findings may not accurately reflect typical clinical practice where insulin doses might be adjusted as needed to improve glycemic control. However, the present analyses of insulin use included data after week 18 through to the end of the study, when investigators were encouraged to optimize glycemic therapy to achieve individualized treatment goals. In this context, the delayed initiation of insulin and lower proportion of patients requiring a ≥20% increase in average daily insulin doses in ertugliflozin-treated patients were observed despite generally balanced changes in antihyperglycemic medications across treatment groups (Table 2). By that token, our data provide a valuable insight into the effects of ertugliflozin on insulin initiation and dose requirements in a manner that should be broadly applicable to patients in real-world settings. Another limitation is the lack of information in the present analyses on the different insulin species and regimens used at baseline or initiated during the VERTIS CV trial. Therefore, conclusions regarding the impact of ertugliflozin treatment on a specific insulin regimen should not be inferred. Finally, continuous glucose monitoring and quality-of-life measures were not included in the VERTIS CV study, and the inclusion of these assessments might have provided additional insights into the impact of ertugliflozin treatment on insulin use over time.

In summary, this study of VERTIS CV trial participants, who had type 2 diabetes and ASCVD, found that ertugliflozin treatment reduced the likelihood of insulin initiation during the trial (by up to 36%), delayed the time to new insulin initiations (by up to 1.8 years) in insulin-naïve patients, and reduced daily insulin doses in those patients who were already receiving insulin at study entry, compared with placebo. These data show that ertugliflozin can reduce the need for insulin therapy in both insulin-naïve and insulin-treated patients, and add to the body of evidence on SGLT2 inhibitors as a useful adjunct, with or without concurrent insulin use, for the treatment of type 2 diabetes.

Funding

This study was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and Pfizer Inc., New York, NY, USA. Editorial support was provided by Shirley Smith, PhD, of Engage Scientific Solutions (Horsham, UK) and was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA and Pfizer Inc., New York, NY, USA.

Author Contributions

S.D.-J., C.P.C., D.Z.I.C., R.E.P., F.C., R.F., J.L., H.S., U.M., and I.G. contributed to the conception, design, and/or planning of the analyses. R.F., J.L., H.S., U.M., and I.G. contributed to the acquisition of data for the analyses. All authors contributed to the conduct of the analyses and/or interpretation of the results. S.D.-J., R.F., J.L., and I.G. contributed to the drafting of the manuscript. All authors contributed to the critical review and revision of the manuscript. All authors approved the final version of the manuscript for publication.

Disclosures

S.D.-J. has led clinical trials for AstraZeneca, Boehringer Ingelheim, and Novo Nordisk Inc.; has received consulting fees from AstraZeneca, Boehringer Ingelheim, Janssen, Merck & Co., Inc., and Sanofi; and has equity interests in Jana Care Inc. and Aerami Therapeutics. C.P.C. has received research grants from Amgen, Better Therapeutics, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Janssen, Merck, Novo Nordisk, and Pfizer; has received consulting fees from Aegerion/Amryt, Alnylam, Amarin, Amgen, Applied Therapeutics, Ascendia, Boehringer Ingelheim, Bristol-Myers Squibb, Eli Lilly, Janssen, Lexicon, Merck, Pfizer, Rhoshan, and Sanofi; and has served on data and safety monitoring boards for the Veterans Administration, Applied Therapeutics, and Novo Nordisk. D.Z.I.C. has received consulting fees and/or speaking honoraria from AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Janssen, Merck & Co., Inc., Mitsubishi-Tanabe, Novo Nordisk, Prometic, and Sanofi; and has received operating funds from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck & Co., Inc., Novo Nordisk, and Sanofi. F.C. has received research grants from the Swedish Research Council, Swedish Heart & Lung Foundation, and King Gustav V and Queen Victoria Foundation; and has received consulting fees from Abbott, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Merck Sharp & Dohme, Novo Nordisk, and Pfizer Inc. R.E.P. has received grants (directed to his institution) from Hanmi Pharmaceutical Co., Ltd, Janssen, Metavention, Novo Nordisk, Poxel SA, and Sanofi; and has received consulting fees (directed to his institution) from Bayer, Corcept Therapeutics Incorporated, Dexcom, Hanmi Pharmaceutical Co., Ltd, Merck & Co., Inc., Novo Nordisk, Pfizer Inc., Sanofi, Scohia Pharma Inc., and Sun Pharmaceutical Industries. At the time these analyses were conducted, J.L. was an employee of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and may own stock and/or stock options in Merck & Co., Inc., Rahway, NJ, USA. I.G. is an employee of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and may own stock and/or stock options in Merck & Co., Inc., Rahway, NJ, USA. R.F., H.S., and U.M. are employees of, and may own shares/stock options in, Pfizer Inc., New York, NY, USA.

Pfizer Data Sharing Statement

Upon request, and subject to review, Pfizer will provide the data that support the findings of this study. Subject to certain criteria, conditions, and exceptions, Pfizer may also provide access to the related individual de-identified participant data. See https://www.pfizer.com/science/clinical-trials/trial-data-and-results for more information.

Clinical Trial Information

NCT01986881, clinicaltrials.gov (November 19, 2013).

Prior Presentation

Parts of this study were presented in abstract form at the American Diabetes Association 81st Scientific Sessions, held virtually June 25-29, 2021.

Abbreviations

 
• AE

  adverse event

 
• ASCVD

  atherosclerotic cardiovascular disease

 
• HbA_1c

  glycated hemoglobin

 
• HR

  hazard ratio

 
• SGLT2

  sodium–glucose cotransporter 2

Author notes

Affiliation at the time these analyses were conducted.