SGLT-2 inhibitors in non-alcoholic fatty liver disease patients with type 2 diabetes mellitus: A systematic review.

BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) is a common comorbidity with type 2 diabetes. The existing therapeutic options for NAFLD are not adequate. Hypocaloric diet and exercise is the cornerstone of therapy in NAFLD. Pioglitazone is the only drug recommended in diabetes patients with biopsy proven non-alcoholic steatohepatitis. The frequent coexistence of NAFLD and type 2 diabetes with their combined adverse health consequences and inadequate therapeutic options makes it necessary to search for newer alternatives. AIM: To assess the effect of sodium glucose cotransporter-2 (SGLT-2) inhibitors on liver enzymes in type 2 diabetes patients with NAFLD.
METHODS: We searched PubMed/MEDLINE, Cochrane library, Google scholar, and Clinicaltrials.gov for the relevant articles to be included in this systematic review. Human studies done in type 2 diabetes patients with NAFLD treated with SGLT-2 inhibitors for at least 12 wk were included. Data from eight studies (four randomised controlled trials and four observational studies) were extracted and a narrative synthesis was done. A total of 214 patients were treated with SGLT-2 inhibitors in these studies (94 in randomised controlled trials and 120 in observational studies).
RESULTS: The primary outcome measure was change in serum alanine aminotransferase level. Out of eight studies, seven studies showed a significant decrease in serum alanine aminotransferase level. Most of the studies revealed reduction in serum level of other liver enzymes like aspartate aminotransferase and gamma glutamyl transferase. Five studies that reported a change in hepatic fat exhibited a significant reduction in hepatic fat content in those treated with SGLT-2 inhibitors. Likewise, among the three studies that evaluated a change in indices of hepatic fibrosis, two studies revealed a significant improvement in liver fibrosis. Moreover, there was an improvement in obesity, insulin resistance, glycaemia, and lipid parameters in those subjects taking SGLT-2 inhibitors. The studies disclosed that about 17% (30/176) of the subjects taking SGLT-2 inhibitors developed adverse events and more than 40% (10/23) of them had genitourinary tract infections.
CONCLUSION: Based on low to moderate quality of evidence, SGLT-2 inhibitors improve the serum level of liver enzymes, decrease liver fat, and fibrosis with additional beneficial effects on various metabolic parameters in type 2 diabetes patients with NAFLD.

Core tip: The frequent coexistence of non-alcoholic fatty liver disease (NAFLD) and type 2 diabetes, their adverse health consequences, and lack of adequate therapeutic options makes it necessary to search for newer alternatives. Currently, pioglitazone and vitamin E are recommended in addition to lifestyle modifications for the management of NAFLD. Animal studies have shown that sodium glucose cotransporter-2 inhibitors might be beneficial in NAFLD present in diabetes patients. The current systematic review shows that sodium glucose cotransporter-2 inhibitors improve the serum level of liver enzymes, liver fat, and liver fibrosis with additional beneficial effects on various metabolic parameters in type 2 diabetes patients with NAFLD.

INTRODUCTION

Non-alcoholic fatty liver disease (NAFLD) is an emerging public health issue worldwide. The prevalence of NAFLD in type 2 diabetes mellitus patients is three times greater as compared to the general population. Its prevalence in diabetic subjects ranges from 69%-87% depending upon the imaging modality used[1]. The spectrum of NAFLD includes simple steatosis, steatohepatitis, and cirrhosis[2]. Besides NAFLD is a risk factor for extrahepatic complications like cardiovascular disease, chronic kidney disease, and type 2 diabetes. In addition, the prevalence of both microvascular and macrovascular complications is increased in patients with NAFLD and type 2 diabetes[3].

The existing therapeutic options for NAFLD are not adequate. Hypocaloric diet and exercise is the cornerstone of therapy in NAFLD. Pioglitazone and vitamin E are recommended only in biopsy-proven non-alcoholic steatohepatitis (NASH), but vitamin E is not recommended in diabetic patients due to inadequate evidence[4]. The frequent coexistence of NAFLD and type 2 diabetes with their combined adverse health consequences and inadequate therapeutic options makes it necessary to search for newer alternatives. Based on the information from animal studies, sodium glucose cotransporter-2 (SGLT-2) inhibitors appear promising in the management of NAFLD[5-7]. This systematic review is an effort to review the available literature on the effect of SGLT-2 inhibitors on NAFLD in type 2 diabetes patients.

MATERIALS AND METHODS

Protocol and registration

This systematic review was performed according to the predefined protocol registered in PROSPERO (Registration ID: CRD42018104572). The protocol can be accessed at the website address https://www.crd.york.ac.uk/prospero. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis 2009 guidelines for reporting this systematic review[8]. Ethics committee approval was not required for this systematic review because it was done using published data found in the public domain.

Eligibility criteria

All observational and randomised controlled trials (RCTs) done using SGLT-2 inhibitors among type 2 diabetes patients with NAFLD having both baseline and post-treatment serum alanine aminotransferase (ALT) level data with a minimum follow-up duration of 12 wk were included in this systematic review. The studies with concomitant pharmacological therapy like pioglitazone or α-tocopherol (vitamin E) for treating NAFLD were excluded to avoid the confounding effects of these drugs on liver function tests. Only those studies that were done in humans and published in English were considered for inclusion. We excluded abstract-only articles, case reports, conference presentations, editorials, reviews, expert opinions, and studies with five participants and less.

Primary and secondary outcomes

The primary outcome was the change in serum ALT levels in type 2 diabetes patients with NAFLD treated with SGLT-2 inhibitors. The secondary outcomes were change in serum aspartate aminotransferase (AST) and gamma-glutamyl transferase (GGT) levels, hepatic fat, hepatic fibrosis, metabolic profile, anthropometric parameters, and the adverse effects of SGLT-2 inhibitors.

Information sources

PubMed/MEDLINE, Cochrane library, Google scholar, and Clinicaltrials.gov were searched from their date of inception until 31st August, 2018.

Literature search and study selection

The search terms/MeSH terms used were “NAFLD”, “Nonalcoholic fatty liver disease”, “Non-alcoholic fatty liver disease”, “Non alcoholic fatty liver disease”, “NASH”, “Non-alcoholic steatohepatitis”, “Nonalcoholic steatohepatitis”, “Non alcoholic steatohepatitis”, “Fatty liver”, “Type 2 diabetes mellitus”, “Type 2 diabetes”, “Diabetes mellitus type 2”, “Diabetes type 2”, “SGLT-2 inhibitors”, “Sodium glucose cotransporter-2 inhibitors”, “SGLT-2”, “SGLT2”, “SGLT 2”, “Canagliflozin”, “Dapagliflozin”, “Empagliflozin”, “Ipragliflozin”, “Luseogliflozin”, “Tofogliflozin”, “Sotagliflozin”, “Remogliflozin”, “Ertugliflozin”, and “Sergliflozin”(Table 1). The references of the search articles were scrutinised for relevant articles.

Table 1

Literature search strategy

S. No	Search terms
1	NAFLD
2	Nonalcoholic fatty liver disease
3	Non-alcoholic fatty liver disease
4	Non alcoholic fatty liver disease
5	NASH
6	Non-alcoholic steatohepatitis
7	Nonalcoholic steatohepatitis
8	Non alcoholic steatohepatitis
9	Fatty liver
10	1 OR 2 OR 3 OR 4 OR 5 OR 6 OR 7 OR 8 OR 9
11	Type 2 diabetes mellitus
12	Type 2 diabetes
13	Diabetes mellitus type 2
14	Diabetes type 2
15	11 OR 12 OR 13 OR 14
16	SGLT-2 inhibitors
17	Sodium glucose cotransporter-2 inhibitors
18	SGLT-2
19	SGLT2
20	SGLT 2
21	Canagliflozin
22	Dapagliflozin
23	Empagliflozin
24	Ipragliflozin
25	Luseogliflozin
26	Tofogliflozin
27	Sotagliflozin
28	Remogliflozin
29	Ertugliflozin
30	Sergliflozin
31	16 OR 17 OR 18 OR 19 OR 20 OR 21 OR 22 OR 23 OR 24 OR 25 OR 26 OR 27 OR 28 OR 29 OR 30
32	10 AND 15 AND 31

NAFLD: Non-alcoholic fatty liver disease; NASH: Non alcoholic steatohepatitis; SGLT-2: Sodium glucose cotransporter-2.

Data collection process

The titles and/or abstracts of studies were retrieved using the search strategy and those from additional sources were scrutinised independently by two review authors (HR and JPS) to identify studies that potentially met the inclusion criteria as outlined above. The full texts of these potentially eligible studies were retrieved and independently assessed for eligibility by three review team members (HD, SS, and RP). Any disagreements between the reviewers over the eligibility of particular studies were resolved through discussion with a fourth senior reviewer (SKK). A standardised, pre-formatted excel form was used to extract data from the included studies for the assessment of study quality.

Data items and synthesis of results

The extracted data included the author of the study with year, the study methodology, the recruitment and study completion rates, the types of population, the exposure/intervention (dose of SGLT-2 inhibitor, duration), the results (outcome measures like change in serum ALT, AST, GGT, hepatic fat, markers of liver fibrosis, fasting plasma glucose (FPG), glycosylated haemoglobin (HbA1c), lipid profile, homeostasis model assessment-estimated insulin resistance (HOMA-IR), body mass index (BMI), any adverse effects, information for the assessment of the risk of bias, and sources of funding/support.

The statistical review of the study was performed by a biomedical statistician (SSK). A narrative synthesis of the results of individual studies was done. The change in the difference in means and difference in proportions and the respective P values as mentioned in the original manuscripts were tabulated and explained in our study.

Risk of study bias

The risk of bias of the RCTs was done using Cochrane risk of bias tool[9]. The studies were graded as “good quality” or “fair quality” or “poor quality” according to the level of risk. Methodological Index for Non-Randomized Studies (MINORS) scale was used to assess the risk of bias of observational studies[10]. A study was considered to be an ideal study if the score was 16 for single arm and 24 for comparative studies.

RESULTS

Study selection

Our literature search from all the aforementioned databases yielded 73 articles (including references of the relevant articles). After eliminating duplicate articles, 55 articles were screened. Eight articles met all of the inclusion criteria (total 214 patients were on SGLT-2 inhibitors) (Figure 1).

Figure 1

Literature search and study selection.

Study characteristics

The summary of all studies included in this systematic review is given in Tables 2 and 3. Out of the eight studies, four are RCTs[11-14] and four are observational[15-18]. Five studies were conducted amongst the Japanese population. Ipragliflozin was used in three studies whereas canagliflozin and luseogliflozin were used in two studies each, but dapagliflozin and empagliflozin were used in one study each. All studies used one type of SGLT-2 inhibitor except the one authored by Seko et al[16], where both canagliflozin and ipragliflozin were used. The change in serum ALT was a secondary outcome while the effect of SGLT- 2 inhibitors on liver fat was the primary outcome in all RCTs.

Table 2

Randomised controlled trials

S. No	Ref.	Inclusion criteria	Age (yr)	Male gender	Intervention arm	Control arm	Follow-up duration	Primary outcome
1	Kuchay et al[11], 2018	Age > 20 yr, hepatic steatosis (MRI-PDFF > 6%), HbA1c > 7.0% to < 10.0%	Intervention arm: 50.7 (12.8)	Intervention arm: 16 (64%)	Standard treatment + Empagliflozin 10 mg daily (n = 25)	Standard treatment (n = 25)	20 wk	Change in liver fat content by MRI-PDFF
			Control arm: 49.1 (10.3)	Control arm: 17 (68%)
2	Ito et al[12], 2017	Age 20-75 yr, HbA1c 7.0–11.0%, BMI < 45 kg/m2, On diet and exercise therapy alone or with oral hypoglycaemic agents other than SGLT-2 inhibitors and thiazolidinediones and/or insulin, NAFLD, findings suggesting hepatic steatosis and hepatic dysfunction on clinical laboratory tests or on imaging studies (e.g., computed tomography or ultrasound)	Pioglitazone arm: 59.1 (9.8)	Pioglitazone arm: 18 (53%)	Ipragliflozin 50 mg daily (n = 32)	Pioglitazone 15-30 mg daily (n = 34)	24 wk	Change in L/S attenuation ratio
			Ipragliflozin arm: 57.3 (12.1)	Ipragliflozin arm: 14 (44%)
3	Shibuya et al[13], 2018	Fatty liver diagnosed on the basis of computed tomography or abdominal sonography, HbA1c 6.0%–10.0%, age 20–70 yr	Luseogliflozin arm: 51 (47-62)	Luseogliflozin arm: 10 (62.5%)	Luseogliflozin 2.5 mg daily (n = 16)	Metformin 1.5 g daily (n = 16)	24 wk	Change in L/S attenuation ratio
			Metformin arm: 60 (53-66)	Metformin arm: 8 (50%)
4	Eriksson et al[14], 2018	Age 40–75 yr, treated with a stable dose of metformin or sulfonylurea alone or in combination for at least 3 mo, MRI-PDFF > 5.5%, BMI 25–40 kg/m2	Dapagliflozin arm: 65 (6.5)	Dapagliflozin arm: 16 (76.2%)	Dapagliflozin 10 mg daily (n = 21) or Omega 3-carboxylic acid 4 g daily (n = 20) or Combination (n = 22)	Placebo (n = 21)	12 wk	Change in liver fat content by MRI-PDFF
			Omega 3-carboxylic acid arm: 66.2 (5.9)	Omega 3-carboxylic acid arm: 11 (55%)
			O + D arm: 65(5.4)	O + D arm: 15 (68.2%)
			Placebo arm: 65.6 (6.1)	Placebo arm: 17 (81%)

MRI-PDFF: Magnetic resonance imaging-derived proton density fat fraction; L/S: Liver/spleen; O + D: Omega 3-carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2; NAFLD: Non-alcoholic fatty liver disease.

Table 3

Observational studies

S. No	Ref.	Design	Inclusion criteria	Age (yr)	Male gender	Sample size	SGLT-2 inhibitor	Follow-up duration
1	Ohki et al[15], 2016	Prospective study	Type 2 diabetes with NAFLD treated with GLP-1 analogues or DPP-4 inhibitors and failed to normalise serum ALT levels	54.2 (49.3-60.1)	19 (79.2%)	24	Ipragliflozin 25-50 mg daily	320 d (302-329)
2	Seko et al[16], 2016	Retrospective cohort study	Type 2 diabetes with NAFLD	SGLT-2 inhibitor arm: 60.3 (1.8)	SGLT-2 inhibitor arm: 9 (37.5%)	24 (SGLT-2 inhibitor); 21 (Sitagliptin )	Canagliflozin 100 mg (n = 18) or Ipragliflozin 50 mg daily (n = 6)	24 wk
				Sitagliptin arm: 59.4 (3.7)	Sitagliptin arm: 8 (38.1%)
3	Gautam et al[17], 2018	Prospective study	Type 2 diabetes with NAFLD	-	-	32	Canagliflozin 100 mg daily	24 wk
4	Sumida et al[18], 2018	Prospective study	Age > 20 yr, HbA1c > 6.5% to < 8.5%, NAFLD	55.4 (13.6)	28 (70%)	40	Luseogliflozin 2.5 mg daily	24 wk

NAFLD: Non-alcoholic fatty liver disease; SGLT-2: Sodium glucose cotransporter-2; GLP-1: Glucagon like peptide-1; DPP-4: Dipeptidyl peptidase-4.

Risk of bias within studies

The risk of bias of RCTs was assessed using the Cochrane risk of bias tool. Among the four RCTs, the studies done by Kuchay et al[11] and Eriksson et al[14] were of good quality however those done by Ito et al[12] and Shibuya et al[13] were of fair quality (Table 4). The risk of bias of observational studies was assessed using the MINORS scale. All the observational studies were of less than ideal quality (Table 5).

Table 4

Assessment of study quality of randomised controlled trials

Study	Criteria	Risk of bias	Study quality
Kuchay et al[11]	Random sequence generation	Low risk	Good quality
	Allocation concealment	Low risk
	Selective reporting	Low risk
	Other bias	Low risk
	Blinding of participants and personnel	Low risk
	Blinding of outcome assessment	Low risk
	Incomplete outcome data	Low risk
Ito et al[12]	Random sequence generation	Low risk	Fair quality
	Allocation concealment	Unclear risk
	Selective reporting	Low risk
	Other bias	Low risk
	Blinding of participants and personnel	Low risk
	Blinding of outcome assessment	Low risk
	Incomplete outcome data	Low risk
Shibuya et al[13]	Random sequence generation	Unclear risk	Fair quality
	Allocation concealment	Unclear risk
	Selective reporting	Low risk
	Other bias	Low risk
	Blinding of participants and personnel	Low risk
	Blinding of outcome assessment	Low risk
	Incomplete outcome data	Low risk
Eriksson et al[14]	Random sequence generation	Low risk	Good quality
	Allocation concealment	Low risk
	Selective reporting	Low risk
	Other bias	Low risk
	Blinding of participants and personnel	Low risk
	Blinding of outcome assessment	Low risk
	Incomplete outcome data	Low risk

Table 5

Assessment of study quality of observational studies

S. No	Criteria	Ohki et al[15]	Seko et al[16]	Gautam et al[17]	Sumida et al[18]
1	A clearly stated aim	2	2	2	2
2	Inclusion of consecutive patients	0	2	2	1
3	Prospective collection of data	2	0	2	2
4	Endpoints appropriate to the aim of the study	2	2	2	2
5	Unbiased assessment of the study endpoint	0	0	0	0
6	Follow-up period appropriate to the aim of the study	2	2	2	2
7	Loss to follow up less than 5%	2	2	2	2
8	Prospective calculation of the study size	0	0	0	0
9	An adequate control group	NA	0	NA	NA
10	Contemporary groups	NA	2	NA	NA
11	Baseline equivalence of groups	NA	2	NA	NA
12	Adequate statistical analyses	NA	2	NA	NA
13	Total score	10/16	16/24	12/16	11/16

Primary outcome

Change in serum ALT levels: In all of the studies, there was a decrease in serum ALT levels from the baseline in those treated with SGLT-2 inhibitors (Table 6) but in the study done by Shibuya et al[13] it did not reach statistical significance.

Table 6

Change in serum alanine aminotransferase levels in individual studies

Study	Serum ALT level (IU/L)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	64.3 (20.2)	49.7 (25.8)	0.001	0.005
	Control	65.3 (40.3)	61.6 (38.4)	0.422
Ito et al[12]	Ipragliflozin	57.4 (27.3)	38.2 (20.5)	< 0.05	0.642
	Pioglitazone	53.1 (26.6)	36.8 (15.1)	< 0.05
Shibuya et al[13]	Luseogliflozin	49.5 (31.0, 70.0)	31 (26.0, 55.0)	0.057	0.064
	Metformin	39 (23.0, 56.0)	39 (27.0, 51.0)	0.518
Eriksson et al[14]	Placebo	33.53 (12.4)	-0.2 (8.8)1	-	-
	Omega-3 CA	37.65 (14.7)	+5.9 (16.5)1	-	Non-significant2
	Dapagliflozin	39.41 (14.7)	-8.2 (8.2)1	-	< 0.052
	O + D	35.88 (17.1)	+0.1 (12.9)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	62 (43.0-75.0)	38.0 (31.0-65.0)	0.01	-
Seko et al[16]	SGLT-2 inhibitor	70.8 (8.1)	48.8 (5.5)	0.002	0.039
	Sitagliptin	92.4 (11.2)	71.1 (10.0)	0.012
Gautam et al[17]	Canagliflozin	96 (18.7)	60.0 (17.6)	< 0.00001	-
Sumida et al[18]	Luseogliflozin	54.7 (28.2)	42.4 (26.5)	< 0.001	-

Change from baseline.

Compared to placebo.

ALT: Alanine aminotransferase; CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Kuchay et al[11] found a significant decrease in serum ALT levels in the empagliflozin arm compared to the control arm at the end of the study (difference between the two arms was -10.9 IU/L, P = 0.005). In the study done by Ito et al[12] ALT levels decreased equally in both the groups [Change from baseline in ipragliflozin group: -17.5 (4) and pioglitazone group: -20 (3.4), P = 0.642]. Similar results were found in the study by Shibuya et al[13] [ΔALT in luseogliflozin arm was 9 (-20, 1) and in metformin arm was 4.5 (-5, 9), P = 0.064]. Eriksson et al[14] found that the ALT reduction in the dapagliflozin arm was more compared to placebo [ΔALT in dapagliflozin arm was -8.24 (8.24) and in the placebo arm was -0.18 (8.82), P < 0.05]. Seko et al[16] demonstrated that the serum ALT levels in SGLT-2 inhibitor arm was lower compared to the sitagliptin arm at the end of the study [48.8 (5.5) vs 71.1 (10), P = 0.039]

Secondary outcomes

Change in serum AST levels: Seven of the included studies had data regarding change in serum AST levels (Table 7). The study done by Shibuya et al[13] did not have data on AST levels. All the studies showed a significant reduction in serum AST levels in those treated with SGLT-2 inhibitors. The decrease in AST with empagliflozin and ipragliflozin was similar compared to placebo and pioglitazone respectively whereas dapagliflozin was better than placebo.

Table 7

Change in serum aspartate aminotransferase levels in individual studies

Study	Serum AST levels (IU/L)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	44.6 (23.5)	36.2 (9.0)	0.04	0.212
	Control	45.3 (24.3)	44.6 (23.8)	0.931
Ito et al[12]	Ipragliflozin	39.7 (16.7)	27.3 (8.9)	< 0.05	0.802
	Pioglitazone	43.3 (20.5)	32.4 (15.4)	< 0.05
Eriksson et al[14]	Placebo	29.4 (13.2)	-1.2 (7.2)1	-	-
	Omega-3 CA	30.6 (10.2)	+4.8 (9.0)1	-	Non-significant2
	Dapagliflozin	31.2 (11.4)	-4.2 (5.4)1	-	< 0.052
	O + D	30 (10.2)	+1.2 (5.4)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	37 (29.0-52.0)	28 (23.0-31.0)	0.03	-
Seko et al[16]	SGLT-2 inhibitor	54.4 (5.6)	38 (3.1)	0.001	-
	Sitagliptin	67 (7.7)	52.5 (7.7)	0.016
Gautam et al[17]	Canagliflozin	72 (16.7)	53 (10.3)	< 0.00001	-
Sumida et al[18]	Luseogliflozin	40.7 (22.2)	31.9 (18.2)	< 0.001	-

Change from baseline.

Compared to placebo.

AST: Aspartate aminotransferase; CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Change in serum GGT levels: Seven studies had data regarding GGT levels. Six studies reported a significant decrease in serum GGT levels in those treated with SGLT-2 inhibitors (Table 8). In the study done by Seko et al[16], there was an insignificant decrease in both the SGLT-2 inhibitor and DPP-4 inhibitor groups. The decrease in GGT with empagliflozin and ipragliflozin was similar compared to placebo and pioglitazone correspondingly while dapagliflozin was better than placebo.

Table 8

Change in serum gamma-glutamyl transferase levels in individual studies

Study	Serum GGT (IU/L )			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	65.8 (36.1)	50.9 (24.6)	0.002	0.057
	Control	63.9 (45.3)	60.0 (39.0)	0.421
Ito et al[12]	Ipragliflozin	62.8 (58.3)	44.0 (38.3)	< 0.05	0.642
	Pioglitazone	71.6 (54.1)	48.8 (61.2)	< 0.05
Eriksson et al[14]	Placebo	32.4 (17.4)	+2.4 (9.6)1	-	-
	Omega-3 CA	54.0 (57.6)	+2.4 (12.0)1	-	Non-significant2
	Dapagliflozin	58.2 (43.2)	-4.8 (13.8)1	-	< 0.052
	O + D	40.2 (14.4)	-0.6 (13.8)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	75.0 (47.0-105.0)	60.0 (40.0-101.0)	0.03	-
Seko et al[16]	SGLT-2 inhibitor	61.7 (9.1)	58.7 (11.5)	0.051	-
	Sitagliptin	89.2 (11.8)	82.4 (11.9)	0.36
Gautam et al[17]	Canagliflozin	75.1 (31.8)	69.2 (26.2)	0.003	-
Sumida et al[18]	Luseogliflozin	62.4 (77.1)	48.2 (56.3)	0.003	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; GGT: Gamma-glutamyl transferase; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Change in hepatic fat: Kuchay et al[11] and Eriksson et al[14] evaluated hepatic fat using magnetic resonance imaging- derived proton density fat fraction (Table 9). It was found that there was a significant reduction in hepatic fat in the empagliflozin arm compared to the control arm in the study done by Kuchay et al[11]. In the study done by Eriksson et al[14], dapagliflozin or omega-3 carboxylic acid when administered alone or in combination reduced hepatic fat fraction significantly. When compared with placebo, only the combination of both drugs reduced hepatic fat fraction significantly. Sumida et al[18] showed that luseogliflozin significantly reduced hepatic fat fraction using magnetic resonance imaging-hepatic fat fraction. Ito et al[12] and Shibuya et al[13] used liver/spleen attenuation ratio for measuring hepatic fat. They found that ipragliflozin was equivalent to pioglitazone in improving liver/spleen attenuation ratio while luseogliflozin was found to be superior to metformin in the same aspect.

Table 9

Change in hepatic fat in individual studies

Study	Parameter	Group	Baseline	Study completion	P value	P value between groups
Kuchay et al[11]	MRI-PDFF	Empagliflozin	16.2 (7)	11.3 (5.3)	< 0.0001	< 0.0001
		Control	16.4 (7.3)	15.5 (6.7)	0.054
Ito et al[12]	L/S ratio	Ipragliflozin	0.8 (0.2)	1.0 (0.2)	< 0.05	0.90
		Pioglitazone	0.8 (0.3)	1.0 (0.2)	< 0.05
Shibuya et al[13]	L/S ratio	Luseogliflozin	0.9 (0.6-1.0)	1.0 (0.8-1.2)	0.0008	0.00002
		Metformin	1.0 (0.8-1.1)	0.9 (0.7-1.0)	0.017
Eriksson et al[14]	MRI-PDFF	Placebo	15.1 (6.5)	-0.6 (1.9)1	-	-
		Omega-3 CA	22.2 (11.0)	-3.2 (2.9)1	-	Non-significant2
		Dapagliflozin	17.3 (9.1)	-2.2 (3.3)1	-	Non-significant2
		O + D	17.8 (9.2)	-3.2 (3.5)1	-	< 0.052
Sumida et al[18]	MRI-HFF	Luseogliflozin	21.5 (7.2)	15.7 (6.8)	< 0.001	-

Change from baseline.

Compared to placebo.

MRI-PDFF: Magnetic resonance imaging-derived proton density fat fraction; L/S ratio: Liver/spleen attenuation ratio; MRI-HFF: Magnetic resonance imaging-hepatic fat fraction; CA: Carboxylic acid; O + D: Omega-3 CA + Dapagliflozin.

Effect on liver fibrosis indices

Ito et al[12] and Ohki et al[15] evaluated liver fibrosis using the FIB-4 index (Table 10). There was a significant decrease in the FIB-4 index in the ipragliflozin arms compared to baseline. Ipragliflozin was similar to pioglitazone in decreasing the FIB-4 index. Sumida et al[18] used both the FIB-4 index and NAFLD fibrosis score. There was no significant change in either indices.

Table 10

Assessment of liver fibrosis in individual studies

Study	Parameter	Group	Baseline	Study completion	P value	P value between groups
Ito et al[12]	FIB-4 index	Ipragliflozin	1.44 (0.64)	1.22 (0.55)	< 0.05	0.596
		Pioglitazone	1.84 (1.13)	1.71 (1.19)	Non-significant
Ohki et al[15]	FIB-4 index	Ipragliflozin	1.75 (0.82-1.93)	1.39 (0.77-1.99)	0.04	-
Sumida et al[18]	FIB-4 index	Luseogliflozin	1.63 (1.19)	1.52 (0.92)	0.17	-
	NAFLD fibrosis score	Luseogliflozin	1.61 (0.71)	1.62 (0.88)	0.86	-

FIB: Fibrosis 4; NAFLD: Non-alcoholic fatty liver disease.

Change in metabolic and anthropometric parameters

Seven studies reported changes in FPG and HbA1c (Tables 11 and 12). The majority of the studies showed a decrease in FPG and HbA1c.

Table 11

Change in fasting plasma glucose in individual studies

Study	Fasting plasma glucose (mg/dL)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	173.0 (44.0)	124.0 (17.0)	< 0.001	0.85
	Control	176.0 (57.0)	120.0 (19.0)	< 0.0001
Ito et al[12]	Ipragliflozin	160.1 (38.7)	136.5 (26.7)	< 0.05	0.785
	Pioglitazone	169.4 (50.9)	139.0 (26.6)	< 0.05
Shibuya et al[13]	Luseogliflozin	127.0 (116.0, 136.0)	125.0 (113.0, 138.0)	0.87	0.583
	Metformin	147.0 (126.0, 161.0)	134.0 (122.0, 145.0)	0.32
Eriksson et al[14]	Placebo	169.2 (29.7)	+6.7 (14.8)1	-	-
	Omega-3 CA	162.4 (26.6)	+3.8 (19.3)1	-	Non-significant2
	Dapagliflozin	161.8 (33.3)	-17.6 (26.8)1	-	< 0.052
	O + D	168.8 (35.5)	-16.4 (36.0)1	-	< 0.052
Ohki et al[15]	Ipragliflozin	162.0 (135.0-189.0)	135.0 (120.0-166.0)	0.3	-
Seko et al[16]	SGLT-2 inhibitor	125.0 (6.0)	116.6 (4.2)	0.07	Non-significant
	Sitagliptin	114.6 (7.0)	134.0 (10.5)	0.067
Sumida et al[18]	Luseogliflozin	142.0 (30.3)	135.4 (25.6)	0.04	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Table 12

Change in glycosylated haemoglobin in individual studies

Study	Glycosylated haemoglobin (%)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	9.0 (1.0)	7.2 (0.6)	< 0.001	0.88
	Control	9.1 (1.4)	7.1 (0.9)	< 0.0001
Ito et al[12]	Ipragliflozin	8.5 (1.5)	7.6 (1.0)	< 0.05	0.522
	Pioglitazone	8.3 (1.4)	7.1 (0.9)	< 0.05
Shibuya et al[13]	Luseogliflozin	7.8 (7.2, 7.9)	6.5 (6.4, 7.0)	0.002	0.023
	Metformin	7.4 (6.9, 7.7)	7.3 (6.7, 7.6)	0.362
Eriksson et al[14]	Placebo	7.4 (0.8)	-0.1 (0.4)1	-	-
	Omega-3 CA	7.4 (0.7)	+0.1 (0.4)1	-	Non-significant2
	Dapagliflozin	7.4 (0.6)	-0.6 (0.7)1	-	< 0.052
	O + D	7.5 (0.8)	-0.5 (0.5)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	8.4 (7.8-8.9)	7.6 (6.9-8.2)	< 0.01	-
Seko et al[16]	SGLT-2 inhibitor	6.7 (0.1)	6.5 (0.1)	0.055	Non-significant
	Sitagliptin	7.0 (0.3)	6.9 (0.3)	0.331
Sumida et al[18]	Luseogliflozin	7.3 (0.7)	7.0 (0.7)	0.002	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

In the study done by Ito et al[12] there was no difference in the change in HOMA-IR in those treated with either ipragliflozin or pioglitazone (P = 0.401) (Table 13). There was a significant decrease in HOMA-IR in those treated with dapagliflozin compared to placebo in the study done by Eriksson et al[14]. Surprisingly there was an insignificant increase in HOMA-IR in those treated with either a SGLT-2 inhibitor or a gliptin in the study done by Seko et al[16].

Table 13

Change in homeostasis model assessment-estimated insulin resistance in individual studies

Study	HOMA-IR			P value	P value between groups
	Group	Baseline	Study completion
Ito et al[12]	Ipragliflozin	5.2 (2.5)	4.8 (5.5)	Non-significant	0.401
	Pioglitazone	5.7 (3.4)	4.5 (2.7)	< 0.05
Eriksson et al[14]	Placebo	4.2 (2.4)	-0.2 (1.4)1	-	-
	Omega 3-CA	5.4 (2.9)	+0.3 (2.4)1	-	Non-significant2
	Dapagliflozin	4.3 (1.9)	-1.1 (1.4)1	-	< 0.052
	O + D	4.4 (1.7)	-0.9 (1.6)1	-	< 0.052
Seko et al[16]	SGLT-2 inhibitor	4.5 (0.5)	7.9 (2.3)	0.955	-
	Sitagliptin	4.4 (0.5)	6.5 (0.8)	0.163

Change from baseline.

Compared to placebo.

HOMA-IR: Homeostasis model assessment-estimated insulin resistance; CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Six studies included data on the changes in lipid profile (Tables 14, 15, and 16). There was a significant decrease in serum triglycerides in two studies (Kuchay et al[11] and Ito et al[12]). Three studies exhibited an increase in high-density lipoprotein cholesterol levels (Ito et al[12], Ohki et al[15], and Seko et al[16]). Most of the studies (Ito et al[12], Eriksson et al[14], Ohki et al[15], Seko et al[16], and Sumida et al[18]) showed no change in serum LDL levels.

Table 14

Change in serum triglycerides in individual studies

Study	Serum triglycerides (mg/dL)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	201.0 (124.0)	155.0 (52.0)	0.01	0.678
	Control	212.0 (115.0)	175.0 (43.0)	0.019
Ito et al[12]	Ipragliflozin	166.9 (76.4)	143.4 (81.4)	< 0.05	0.938
	Pioglitazone	188.4 (148.8)	169.3 (131.3)	Non-significant
Eriksson et al[14]	Placebo	169.2 (84.1)	-11.5 (45.6)1	-	-
	Omega-3 CA	186.9 (81.5)	-15.9 (47.4)1	-	Non-significant2
	Dapagliflozin	178.0 (103.6)	+14.2 (40.5)1	-	Non-significant2
	O + D	168.3 (72.6)	-25.7 (57.1)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	148.0 (107.0, 222.)	145.0 (114.0, 172.0)	0.75	-
Seko et al[16]	SGLT-2 inhibitor	153.8 (15.9)	137.8 (10.5)	0.236	-
	Sitagliptin	193.4 (25.2)	191.1 (23.8)	0.986
Sumida et al[18]	Luseogliflozin	158.1 (110.5)	129.4 (59.5)	0.062	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Table 15

Change in serum low-density lipoprotein cholesterol in individual studies

Study	Serum low-density lipoprotein cholesterol (mg/dL)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	112.0 (35.0)	95.0 (22.0)	0.018	0.512
	Control	114.0 (30.0)	96.0 (17.0)	0.001
Ito et al[12]	Ipragliflozin	108.3 (36.2)	110.7 (40.1)	Non-significant	0.057
	Pioglitazone	104.0 (27.9)	114.6 (29.5)	< 0.05
Eriksson et al[14]	Placebo	98.2 (34.4)	+1.6 (15.5)1	-	-
	Omega-3 CA	111.8 (34.4)	+2.3 (17.4)1	-	Non-significant2
	Dapagliflozin	109.4 (34.8)	+7.7 (20.5)1	-	Non-significant2
	O + D	88.9 (23.2)	+5.8 (21.7)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	113.0 (89.0-142.0)	103.0 (92.0-122.0)	0.08	-
Seko et al[16]	SGLT-2 inhibitor	119.2 (5.8)	119.8 (5.7)	0.943	-
	Sitagliptin	112.9 (4.9)	127.1 (8.8)	0.063
Sumida et al[18]	Luseogliflozin	101.0 (22.4)	105.0 (24.4)	0.11	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Table 16

Change in serum high-density lipoprotein cholesterol in individual studies

Study	Serum high-density lipoprotein cholesterol (mg/dL)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	42.0 (12.0)	45.0 (12.0)	0.087	0.752
	Control	45.0 (15.0)	47.0 (12.0)	0.097
Ito et al[12]	Ipragliflozin	48.9 (9.3)	54.7 (10.4)	< 0.05	0.82
	Pioglitazone	47.4 (11.6)	52.7 (13.5)	< 0.05
Eriksson et al[14]	Placebo	51.4 (14.9)	-0.4 (5.0)1	-	-
	Omega-3 CA	49.9 (14.1)	+0.4 (3.2)1	-	Non-significant2
	Dapagliflozin	49.9 (9.5)	+0.4 (4.8)1	-	Non-significant2
	O + D	51.4 (10.2)	+1.6 (5.0)1	-	Non-significant2
Ohki et al[15]	Ipragliflozin	42.0 (40.0-50.0)	44.0 (42.0-59.0)	0.01	-
Seko et al[16]	SGLT-2 inhibitor	53.9 (2.5)	55.4 (2.6)	0.043	-
	Sitagliptin	54.8 (3.3)	55.6 (2.3)	0.531
Sumida et al[18]	Luseogliflozin	55.6 (11.7)	57.5 (13.4)	0.062	-

Change from baseline.

Compared to placebo.

CA: Carboxylic acid; O + D: Omega-3 carboxylic acid + Dapagliflozin; SGLT-2: Sodium glucose cotransporter-2.

Five studies included BMI change (Table 17). There was a reduction in BMI in the SGLT-2 inhibitor arms in all the studies. Empagliflozin was similar to placebo in reducing BMI whereas luseogliflozin was superior to metformin in reducing BMI.

Table 17

Change in body mass index in individual studies

Study	Body mass index (kg/m2)			P value	P value between groups
	Group	Baseline	Study completion
Kuchay et al[11]	Empagliflozin	30.0 (3.8)	28.7 (3.5)	0.001	0.124
	Control	29.4 (3.1)	28.8 (2.8)	0.019
Shibuya et al[13]	Luseogliflozin	27.9 (26.2, 28.7)	27.0 (25.6, 28.3)	0.002	0.031
	Metformin	27.2 (24.8, 32.1)	27.3 (24.3, 31.6)	0.646
Ohki et al[15]	Ipragliflozin	30.1 (26.1-31.4)	27.6 (25.3-30.2)	< 0.01	-
Seko et al[16]	SGLT-2 inhibitor	29.6 (0.7)	28.3 (0.7)	< 0.001	-
	Sitagliptin	29.2 (1.5)	28.9 (1.4)	0.295
Sumida et al[18]	Luseogliflozin	27.8 (3.6)	27.2 (1.0)	< 0.001	-

SGLT-2: Sodium glucose cotransporter-2.

Adverse effects of SGLT-2 inhibitors

Out of the eight studies, six studies reported the adverse effects of SGLT-2 inhibitors. There were a total of 30 reported adverse events in 176 patients taking SGLT-2 inhibitors (Table 18). The most common adverse event was genitourinary tract infection (10 events).

Table 18

Adverse effects of sodium glucose cotransporter-2 inhibitors in individual studies

Study	No. of adverse events	No. of patients	Types of adverse events
Kuchay et al[11]	3	25	Nonspecific fatigue: 1
			Arthralgia: 1
			Balanoposthitis: 1
Ito et al[12]	9	32	UTI: 3
			Increased appetite: 2
			Nausea: 1
			Headache: 1
			Diarrhoea: 1
			Vaginal candidiasis: 1
Eriksson et al[14]	7	21	-
Seko et al[16]	2	26	UTI: 2
Gautam et al[17]	1	32	Recurrent UTI with genital candidiasis: 1
Sumida et al[18]	8	40	Low blood pressure: 3
			Vaginal itching: 2
			Constipation: 1
			Vertigo: 1
			Dehydration: 1
Total	30	176	Most common adverse event: Genitourinary tract infections-10

UTI: Urinary tract infection.

DISCUSSION

Type 2 diabetes is commonly associated with NAFLD. Serum ALT levels are commonly above the upper limit of normal with AST levels lesser than ALT levels[19]. Animal studies have shown that SGLT-2 inhibitors decrease liver enzymes (ALT, AST), liver weight, and hepatic steatosis[20-23]. There are several mechanisms for improvement in serum liver enzymes in the patients taking SGLT-2 inhibitors. These drugs cause hyperglucagonemia by increasing glucagon secretion from the pancreatic α cells. Glucagon stimulates gluconeogenesis and β-oxidation of fatty acids in the liver via stimulation of peroxisome proliferator-activated receptor alpha and carnitine palmitoyl transferase-1[13]. Thus SGLT-2 inhibitors help to reduce hepatic fat. They reduce collagen deposition and inflammatory cytokine expression in liver[5,22]. They decrease liver enzymes by additionally improving glycaemic parameters and insulin resistance. Out of eight studies, seven showed a decrease in serum ALT and AST levels in our systematic review. Shibuya et al[13] observed a decrease in ALT that almost reached statistical significance, however data regarding AST was unavailable[13]. Out of seven studies, six illustrated a significant decrease in GGT levels while in the study by Seko et al[16] the change in serum GGT level almost reached statistical significance.

Liver enzymes are surrogate markers of liver histological response, but an improvement in liver histology is not always associated with a decrease in serum liver enzymes[11]. The five studies that evaluated changes in hepatic fat showed a decrease in hepatic fat. There was no correlation of a change in ALT with a change in hepatic fat in the study by Shibuya et al[13], however there was a correlation between these two parameters in the study by Sumida et al[18]. The decrease in hepatic fat in the SGLT-2 inhibitor arm was comparable to pioglitazone, which is an approved drug for treatment of NAFLD irrespective of the presence of diabetes. Eriksson et al[14] observed that although the hepatic fat content decreased in the dapagliflozin arm it did not reach statistical significance compared to placebo. The lesser duration of this study (12 wk) compared to other studies may have contributed to this difference.

The progression of NAFLD to cirrhosis is determined to a large extent by the liver histology. Studies with up to 20 years follow-up have shown that the risk of progression to cirrhosis for simple steatosis, NASH, and NASH with fibrosis are 0%-4%, 25%, and 38%, respectively[24]. The FIB-4 index is a non-invasive tool to assess liver fibrosis[25]. It is calculated from the patient’s age, platelet count, ALT levels, and AST levels. The FIB-4 index was decreased with SGLT-2 inhibitor therapy in two out of three studies. Sumida et al[18] used the NAFLD fibrosis score in addition to the FIB-4 index to assess liver fibrosis. The NAFLD fibrosis score is a composite score of six variables (age, BMI, hyperglycaemia, platelet count, albumin, and AST/ALT ratio)[26]. There was no significant change in either indices in this study.

It has been shown that NAFLD is more common in those with poor glycaemic control than those with good glycaemic control[27]. SGLT-2 inhibitors promote glycosuria by inhibiting SGLT-2 in the proximal convoluted tubule. Therefore their action is dependent on blood glucose levels but insulin independent[28]. They cause a significant reduction in FPG[29]. A meta-analysis of RCTs has concluded that the average HbA1c reduction at 52 wk of SGLT-2 inhibitor therapy to be 0.6%[30]. Another meta-analysis has shown that SGLT-2 inhibitor monotherapy is equivalent to metformin monotherapy in reducing HbA1c levels[31]. However, the decrease in HbA1c was more in the luseogliflozin arm compared to the metformin arm in the study by Shibuya et al[13]. Four out of seven studies and six out of seven studies showed a decrease in FPG and HbA1c, respectively, in the SGLT-2 inhibitor arm. Thus, the improved glycaemic status is one of the mechanisms by which SGLT-2 inhibitors ameliorate NAFLD.

SGLT-2 inhibitors ameliorate insulin resistance in numerous ways. SGLT-2 inhibitors improve obesity associated insulin resistance by regulating macrophage recruitment and altering the proportion of pro-inflammatory and anti-inflammatory macrophages. They enhance fat utilization by promoting β-oxidation of fatty acids and browning of white adipose tissue by inducing the expression of thermogenin leading to an improvement in the lipid profile. Similar to other antidiabetic drugs, SGLT-2 inhibitors reduce insulin resistance by decreasing glucotoxicity. Dapagliflozin has been shown to improve insulin sensitivity by increasing adiponectin and zinc-A2-glycoprotein levels[32]. Only dapagliflozin was shown to decrease insulin resistance in the study by Eriksson et al[14].

SGLT-2 inhibitors cause weight reduction. The major mechanism that causes weight reduction is the decrease in fat mass. The decrease in fat mass is due to the shift in substrate utilization to lipids instead of carbohydrates[33,34]. Ito et al[12] and Shibuya et al[13] demonstrated that SGLT-2 inhibitors caused a significant reduction in abdominal visceral and subcutaneous fat area as measured by computed tomography scan. Similarly, Eriksson et al[14] showed that dapagliflozin significantly reduced abdominal visceral and subcutaneous adipose tissue volume as assessed by magnetic resonance imaging. The other mechanisms of weight loss are the urinary glucose loss which amounts to approximately 200 Kcal/d and osmotic diuresis[33,35]. Unlike the other weight-reducing effects of SGLT-2 inhibitors, which are potentially beneficial, osmotic diuresis is clearly an adverse effect. Seko et al[16] showed that ipragliflozin and canagliflozin significantly reduced total body water in addition to body fat mass as measured by bioelectrical impedance analysis. Five studies showed a significant decrease in BMI in patients on SGLT-2 inhibitor therapy. Thus, the major beneficial effects of SGLT-2 inhibitors on NAFLD are exerted via reduction in hepatic fat and fibrosis, improved glycaemic control, decrease in insulin resistance, and weight loss.

The most common adverse effects of SGLT-2 inhibitors are genitourinary tract infections. In addition, they may cause diabetic ketoacidosis, dizziness, acute kidney injury, lower limb amputations, and bone fractures[36,37]. A meta-analysis concluded that there was no difference between placebo and SGLT-2 inhibitors for serious adverse events[38]. Among the 30 adverse events reported in all the studies, the most common was genitourinary tract infections (10 out of 23 characterised events).

The major strength of this systematic review was that the effect of five SGLT-2 inhibitors on NAFLD in patients with type 2 diabetes was evaluated in both RCTs and observational studies. Moreover, liver fat, liver fibrosis, metabolic, and anthropometric parameters in addition to liver enzymes were assessed as outcome variables following SGLT-2 inhibitor therapy. Yet this systematic review has a few limitations. First, most of the studies were done amongst the Japanese population. As a result, the study findings may not be applicable to patients from other ethnicities. Second, the sample size was considerably small and the duration of follow-up was of limited period in most of the studies. Third, the confounding effect of concomitant anti-diabetes drugs like metformin, DPP- 4 inhibitors, and glucagon like peptide-1 analogues on NAFLD cannot be ruled out, particularly in observational studies. Fourth, two studies (Eriksson et al[14] and Sumida et al[18]) were funded by pharmaceutical companies, which is a source of potential conflicts of interest.

Summary and conclusion

In conclusion based on the available evidence, SGLT-2 inhibitors were found to improve serum levels of liver enzymes, liver fibrosis indices, and liver fat without significant side effects in type 2 diabetes patients with NAFLD. They showed additional beneficial effects on obesity, glycaemic parameters, insulin resistance, and dyslipidaemia in these subjects. However, the quality of evidence was low to moderate. Prospective studies, preferably RCTs, comparing different SGLT-2 inhibitors with standard treatments of NAFLD in multi-ethnic populations with a longer follow-up period are needed in the future.

ARTICLE HIGHLIGHTS

Research background

Non-alcoholic fatty liver disease (NAFLD) is a common comorbidity with type 2 diabetes. The existing therapeutic options for NAFLD are not adequate. Hypocaloric diet and exercise is the cornerstone of therapy in NAFLD. Pioglitazone is the only drug recommended in diabetes patients with biopsy proven non-alcoholic steatohepatitis. The frequent coexistence of NAFLD and type 2 diabetes along with their combined adverse health consequences and inadequate therapeutic options makes it necessary to search for newer alternatives. This systematic review is an effort to review the available literature on the effect of sodium glucose cotransporter-2 (SGLT-2) inhibitors on NAFLD in type 2 diabetes patients.

Research motivation

Because the existing therapeutic options are not adequate for NAFLD patients, there is a need for finding newer alternatives. SGLT-2 inhibitors have shown promise in the management of NAFLD in animals. Hence, we reviewed the available literature on the effect of SGLT-2 inhibitors in NAFLD in type 2 diabetes patients. This will promote further high quality research on the effect of SGLT-2 inhibitors in NAFLD.

Research objectives

The primary outcome was the change in serum alanine aminotransferase levels in type 2 diabetes patients with NAFLD treated with SGLT-2 inhibitors. The secondary outcomes were change in serum aspartate aminotransferase and gamma-glutamyl transferase levels, hepatic fat, hepatic fibrosis, metabolic profile, anthropometric parameters, and the adverse effects of SGLT-2 inhibitors.

Research methods

This systematic review was registered in PROSPERO and performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. We searched PubMed/MEDLINE, Cochrane library, Google scholar, and Clinicaltrials.gov for the relevant articles to be included in this systematic review. A narrative synthesis of the results of individual studies was done. The change in the difference in means and difference in proportions and the respective P values as mentioned in the original manuscripts were tabulated and explained. The quality of the randomised controlled trials and observational studies was analysed using the Cochrane risk of bias tool and MINORS scale, respectively.

Research results

Eight articles (four randomised controlled trials and four observational studies) were included in this systematic review. A total of 214 patients were treated with SGLT-2 inhibitors. SGLT-2 inhibitors caused a significant improvement in liver enzymes, hepatic fat, hepatic fibrosis, glycaemia, insulin resistance, obesity, and lipid parameters with minimal adverse effects. However, the quality of evidence is low to moderate.

Research conclusions

We found that SGLT-2 inhibitors improved the serum levels of liver enzymes, liver fat, and liver fibrosis with additional beneficial effects on various metabolic and anthropometric parameters in type 2 diabetes patients with NAFLD. However, the number of patients treated with SGLT-2 inhibitors was small. The findings of this systematic review will have impact in choosing anti-diabetes medication like SGLT-2 inhibitors to treat NAFLD associated with type 2 diabetes.

Research perspectives

The studies included in this systematic review were heterogeneous with regard to study design and intervention drugs. Most of the studies were done amongst the Japanese population. Prospective studies, preferably randomised controlled trials, comparing different SGLT-2 inhibitors with standard treatments of NAFLD in multi-ethnic populations with a longer follow-up period are needed in the future.