The effects of empagliflozin on cardiometabolic risk factors in patients with type 2 diabetes: A short-term observational study

Abstract

The aim of this study was to assess the short-term cardiometabolic outcomes in type 2 diabetes patients receiving empagliflozin in a tertiary referral center. Three hundred and fifteen consecutive patients started on empagliflozin were followed for a 4-month period after local ethics committee approval for a range of outcomes. Data were recorded on Microsoft Excel and transposed to SPSS for further analysis. Empagliflozin treatment resulted in statistically significant reductions in weight, glycosylated hemoglobin, and systolic and diastolic blood pressures along with favorable lipid profile outcomes over a 4-month period. The rates of discontinuation of the medications due to genomycotic infections were extremely low at 0.6% with no episodes of severe hypoglycemia or euglycemic diabetic ketoacidosis. Empagliflozin therapy, either in addition to other oral agents or insulin, seems to result in favorable outcomes in cardiometabolic risk factors in the immediate short term. Long-term follow-up of this cohort will shed light on cardiovascular outcomes and adverse effects in our population in real-world clinical practice.

Introduction

Diabetes is a vascular disease with huge public health and economic implications worldwide, more so in South Asia.[1,2] In an ideal world, one would expect a diabetes drug to reduce blood glucose and concurrently reduce the risk of cardiovascular (CV) deaths. Paradoxically, there has been more concern about the detrimental effects of diabetes drugs on CV outcomes, leading to the US Food and Drug Administration (FDA) issuing guidance and new regulations on CV safety of diabetes drugs from 2008. The US FDA was responding to concerns after two meta-analysis, one on muraglitazar and the other on rosiglitazone, both of which showed worsening CV complications versus comparators.[3,4] Since then, there have been a number of prospective CV outcome trials with dipeptidyl peptidase IV (DPP-IV) inhibitors, glucagon-like peptide-1 receptor analogs, and sodium-glucose cotransporter-2 (SGLT-2) inhibitors. The DPP-IV inhibitors led CV outcome (CVOT) trials showed only neutral effects on CV outcomes and, if anything, raised concerns regarding heart failure with certain gliptins.[5]

Empagliflozin was the first oral antidiabetic drug to show significant reductions in the rates of CV deaths and hospitalization for heart failure in high-risk CV patients with type 2 diabetes over a mean follow-up of 3.1 years. The reduction in CV deaths was large and clinically important (38% risk reduction), but the most interesting aspect was the observation that the reduction in CV deaths was seen as early as 3 months and sustained throughout the course of the trial. This led the US FDA to approve empagliflozin to reduce CV deaths in adults with type 2 diabetes. These reductions in CV outcomes could not be explained by the modest glucose-lowering effects of empagliflozin, suggesting that alternate mechanisms may be the reasons for such impressive reductions in CV outcomes within a short period of time. Empagliflozin was associated with small reductions in weight, waist circumference, uric acid level, systolic and diastolic blood pressures, and small increases in low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C). The rates of genital infections were higher in the empagliflozin group in the absence of any increases in diabetic ketoacidosis, fractures, or lower-limb amputations in this study.

On a background of significant reductions in CV deaths even within 3 months of treatment with empagliflozin, we set out to study the real-world short-term observational outcomes with empagliflozin in combination with other oral agents and insulin in a tertiary referral center in relation to a range of clinically important variables and adverse events.

Aims

The aim of this study was to assess the short-term outcomes (4 months) of empagliflozin use in patients with type 2 diabetes on variables including glycemic control, blood pressure, lipid parameters, body weight, and adverse events.

Methods

The study was carried out at the Diabetes and Endocrine Centre at Kovai Medical Centre and Hospital, Coimbatore, India, from February 10, 2018, for a period of 6 months. Local ethics committee approval was obtained (Ethical approval no. EC/AP/494/02/2018).

Inclusion criteria

• Type 2 diabetes confirmed by the treating diabetologist

• Patients on SGLT-2 inhibitor therapy in addition to other oral agents or insulin

• Age >20 years

• Estimated glomerular filtration rate (eGFR) >60 ml/min.

Exclusion criteria

• Patients with type 1 diabetes

• Age <20 years

• Pregnant women

• Those with eGFR <60 ml/min.

Variables including age, sex, past medical history, blood pressure, fasting glucose, glycosylated hemoglobin, fasting lipid profile, and creatinine were collected using a dedicated questionnaire and data sheet. Results were tabulated onto Microsoft Excel and transposed to SPSS for statistical analysis (SPSS) (IBM Corporation, Armonk, New York, USA) version 20 for analysis. Mean (standard deviation) was calculated and paired t-test was used to assess the differences between the two groups, and P < 0.05 was defined as statistically significant.

Results

A total of 315 consecutive patients on empagliflozin were followed for a period of 4 months [Table 1]. The mean age of the study population was 55.6 years (11.2) with a range from 25 to 85 years of age. The male:Female ratio was 60:40. Nearly 82% of patients were in the age range of 40–70 years. Fifty-six percent of patients had either hypertension or coronary artery disease. Empagliflozin was predominantly used in overweight or obese patients (95%). The distribution of empagliflozin use based on diabetes duration was 17% in <5 years of diabetes, 53% in those between 6 and 10 years of diabetes, 15% between 11 and 15 years of duration, and 15% more than 15 years of duration. Empagliflozin was used in combination with insulin ± oral agents in 50% of the study population and the rest 50% in combination with various oral agents.

Table 1.

Baseline characteristics (n=315)

Factors	Mean (SD)
Male:female	59.7:40.3
Duration of diabetes (years)	9.8 (5.2)
Mean weight (kg)	80.8 (12.3)
BMI	29.4 (3.4)
SBP (mmHg)	127.6 (14.4)
DBP (mmHg)	81.1 (8.4)
FBS (mg/dl)	213 (64.6)
HbA1c, n (%)	9.0 (1.3)
Total cholesterol	231 (79.7)
Triglycerides	243.5 (83.5)
HDL-C	34.8 (5.4)
LDL-C	116.4 (47.9)
Creatinine	0.8 (0.1)

BMI=Body mass index, SBP=Systolic blood pressure, DBP=Diastolic blood pressure, FBS=Fasting blood sugar, LDL-C=Low-density lipoprotein cholesterol, HDL-C=High-density lipoprotein cholesterol, HbA1c=Hemoglobin A1c, SD=Standard deviation

The study group was arbitrarily classified into those with duration of diabetes < or >5-year duration, and the variables were analyzed in two categories: Empagliflozin with insulin (±oral agents) and empagliflozin with other oral agents [Table 2]. Empagliflozin significantly reduced body weight at 4 months in combination with insulin or with other oral agents. The mean weight loss at the end of 4 months of treatment with empagliflozin in combination with insulin and duration of diabetes < and >5 years was −2.3 and −2.1 kg, respectively (P < 0.001). Similarly, in combination with oral agents, the mean weight loss was −2.7 kg and −2.1 kg at the end of 4 months in those with duration < or >5 years, respectively (P < 0.001). The fasting blood glucose reduced by 60 mg/dl in those with empagliflozin in combination with insulin and <5 years of duration of diabetes at the end of 4 months (P = 0.008) and 88 mg/dl in those with a duration of diabetes more than 5 years and on insulin + empagliflozin (P < 0.001). It is possible that a larger reduction in fasting glucoses in those with duration >5 years may be due to chance alone as this does not corroborate with HbA1c reductions in both the groups. Fasting glucoses in our study group are more prone for confounding factors as many patients travel long distances before arrival to hospital and it is possible that the fasting glucoses may not be a true fasting sample in a significant minority. Similar reductions were observed in those with empagliflozin and oral agents at < or >5-year duration of diabetes (68 and 76 mg/dl, respectively, P < 0.001). We did not have sufficient number of postprandial results to make a valid conclusion on the effect of empagliflozin on postprandial glucose levels. The mean HbA1c reduction was 2.8% from baseline after 4 months of treatment with empagliflozin in the insulin group with duration <5 years (P < 0.001) and 1.6% in this group with duration >5 years. The reductions in HbA1c in the empagliflozin + oral agents' group at the end of 4 months in those with duration < and >5 years were 1.2% and 1.1%, respectively (P < 0.001).

Table 2.

Comparison between baseline and 4-month data on empagliflozin with insulin group and 4-month data on empagliflozin with oral agents

	Empagliflozin with insulin <5 years			Empagliflozin with insulin >5 years			Empagliflozin with OHA <5 years			Empagliflozin with OHA >5 years
	Baseline	4 months	P	Baseline	4 months	P	Baseline	4 months	P	Baseline	4 months	P
Body weight	79.5 (14.3)	77.2 (13.5)	<0.01	81.02 (11.2)	78.9 (10.7)	<0.01	86.8 (16)	84.1 (14.7)	<0.01	78.9 (11.3)	76.8 (11.3)	<0.01
FBS	246.4 (93.3)	180.8 (77.9)	0.008	228.3 (71.0)	140.2 (45.4)	<0.01	184.9 (41.9)	116.6 (25.1)	<0.01	205.7 (57.3)	129.5 (38.3)	<0.01
HbA1c	11.3 (1.6)	8.57 (1.0)	<0.01	9.4 (0.9)	7.8 (0.8)	<0.01	8.3 (0.8)	7.19 (0.5)	<0.01	8.5 (1.1)	7.4 (0.8)	<0.01
Total cholesterol	200.7 (58.5)	194.6 (44.1)	0.3	225.9 (65.5)	209.8 (62.5)	0.06	221.6 (65.3)	200.6 (54.6)	0.06	241.2 (87.7)	219.5 (71)	0.06
Triglycerides	216 (60.5)	208 (75.6)	0.29	238.8 (78.4)	219.2 (68.4)	0.09	223.5 (77.7)	199.7 (62.7)	0.06	256 (62.7)	233.2 (79.7)	0.06
HDL	37.4 (3.1)	38 (3.9)	0.11	34.3 (4.2)	35.2 (3.6)	0.1	35.6 (4.4)	36.1 (4)	0.1	34.9 (6.5)	35.8 (5.6)	0.09
LDL	98.4 (53.1)	96.2 (46.4)	0.64	116.6 (41.7)	102.5 (36.1)	0.09	117.7 (49.5)	109.1 (41.8)	0.09	117.2 (52)	107.6 (42.1)	0.07

FBS=Fasting blood sugar, HbA1c=Hemoglobin A1c, LDL=Low-density lipoprotein, HDL=High-density lipoprotein

There were minor reductions in total cholesterol, triglycerides (TG), and LDL-C at the end of 4 months with a small increase in HDL-C across both insulin and other oral agent's combination groups [Table 2]. There were 50 episodes of self-reported hypoglycemia at the end of 4-month duration with no severe hypoglycemic events in any categories. Forty episodes were reported in the empagliflozin insulin combination group with duration of diabetes >5 years. Two patients had to stop treatment due to genomycotic infections, and there were no cases of diabetic ketoacidosis or lower-limb amputations in any category during the 4-month follow-up.

Discussion

We report statistically significant reductions in glycosylated hemoglobin, fasting blood glucoses, body weight, and favorable lipid profile changes with empagliflozin, both in combination with insulin and other oral agents in a short-term 4-month observational study. The rates of genomycotic infections were extremely low with no serious hypoglycemic events or diabetic ketoacidosis reported within this time period.

SGLT-2 inhibitors are thought to be relatively weak glucose-lowering agents with meta-analysis showing modest 0.5%–0.7% HbA1c reductions with SGLT-2 inhibitors compared to placebo. In our study, the reductions in glycosylated hemoglobin within 4 months were impressive ranging from around 1.2% in combination with oral agents and 2.8% in combination with insulin. However, this finding should be tempered as the baseline HbA1c in the insulin-treated group was high and one would expect a steeper fall in glycosylated hemoglobin at higher baseline levels. In addition, there was no comparator arm to know whether similar reductions would have been obtained with a comparator drug.

Weight loss was consistently between 2.1 and 2.7 kg at the end of 4 months in our study and was similar in both insulin-treated and oral agents' group. This is consistent with evidence from 12-week trials with all three SGLT-2 inhibitors showing a weight loss of 2–3 kg. Studies have shown that this weight loss is sustained over time with one meta-analysis showing a sustained 2.9 kg weight loss for up to 2 years with SGLT-2 inhibitors. Weight loss with SGLT-2 inhibitors is characterized by early body water and fat loss till 8 weeks followed by a slower rate of sustained fat loss. The fat loss is attributed to an approximately 200 kcal/day glucose loss in urine with a shift toward lipolysis and ketogenesis. Whether this early weight loss would be a major player in CV benefit with SGLT-2 inhibitors is open to debate, as this looks likely to be multifactorial centering around “superfuel” effect of ketones. However, this weight loss may play a significant role in the long-term beneficial effects of SGLT-2 inhibitors on CV outcomes.

Our study showed minor reductions in TG and LDL-C with a small increase in HDL-C [Table 2]. This is contrary to studies showing a 10% increase in LDL-C, which seems to be a class effect of this agent. Intuitively, one would expect the LDL-C to reduce with weight loss. Therefore, an increase in LDL-C raises concerns, as this marker is the most powerful determinant of CV outcomes in type 2 diabetes. However, randomized controlled trials with SGLT-2 inhibitors have shown beneficial CV outcomes, suggesting that the increase LDL-C with SGLT-2 inhibitors may not have detrimental CV outcomes. The possible mechanisms for increased LDL-C with SGLT-2 inhibitors include increased LDL-C production by enhanced lipoprotein lipase activity, impaired LDL-C catabolism, but more importantly suppressed conversion from large buoyant (lb) LDL-C to small dense (sd) LDL-C. LDL-C can be fractionated into lb-LDL-C and sd-LDL-C. sd-LDL-C is extremely atherogenic compared to lb fraction, with studies suggesting a threefold risk of coronary artery disease with sd-LDLC. The Quebec CV study revealed that lb-LDL-C was a negative risk factor for CV death. There is evidence to suggest that SGLT-2 inhibitors may preferentially increase the less atherogenic lb-LDL-C and reduce the highly atherogenic sd-LDL-C, thereby reducing the atherogenic burden. Considering the above discussions, the reductions in LDL-C in our study were unexpected, whether it was due to other confounding factors such as statin doses, compliance, and use in an observational study is not entirely clear. The reductions in TG and the increase in HDL-C were along expected lines, consistent with other studies showing similar changes, predominantly related to weight loss with SGLT-2 inhibitors.

Empagliflozin treatment, in our study, was associated with an approximately 4–5 mmHg systolic blood pressure reduction and 1 mmHg diastolic blood pressure reduction in all categories with the exception of the insulin arm with <5-year duration of diabetes. In a large meta-analysis of nearly 21,000 patients from six studies, SGLT-2 inhibitors were consistently associated with a 3.7 mmHg systolic BP reduction and 1.4 mmHg diastolic BP reduction. This reduction was based on 24-h ambulatory blood pressure values, which is a better predictor of CV risk than seated clinic blood pressure measurements.

The rates of discontinuation with genomycotic infections were extremely low in our study (0.6%), which is surprising given the data that up to 10% of women may develop vulvovaginal candidiasis with SGLT-2 inhibitors. Forty percent of our study population were women and only two discontinued treatment due to genomycotic infections. It is possible that women who had a single episode of genomycotic infections did not report if they had treatment locally and became better. There were no cases of pyelonephritis or Fournier's gangrene in our study population. Hypoglycemias were reported more with the insulin-treated group, however, there were no episodes of severe hypoglycemia needing hospitalization, both in the insulin-treated and oral agent's combination groups. We did not record any case of euglycemic diabetic ketoacidosis in this cohort.

There are major limitations in our study. Our study was a short-term observational study with no comparator arm, therefore, the possibility of confounding factors altering the interpretation of study is possible. There was a paucity of data on postprandial glucose levels due to logistical issues. The rates of discontinuation of the drug due to genomycotic infections were extremely low raising the possibility that a single episode of vulvovaginitis could have been missed in the data collection. Finally, empagliflozin was the only SGLT-2 inhibitor used in this study due to personal preferences by the diabetes team based on the current evidence, the slight concern with lower-limb amputations with canagliflozin, and the yet to be published CVOT with dapagliflozin at the time of the study.

In summary, we have shown short-term reductions in cardiometabolic risk factors in patients with type 2 diabetes in the absence of major adverse events in a real-world observational study. Further long-term follow-up of this cohort will shed light on CV outcomes in an Indian population.

Financial support and sponsorship

Nil.

Conflicts of interest

Krishnan Swaminathan reports receiving honoraria from Boehringer for speaking assignments on empagliflozin. Other authors report no conflict of interest.