Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Comparisons between dipeptidyl peptidase-4 inhibitors and other classes of hypoglycemic drugs using two distinct biomarkers of pancreatic beta-cell function: A meta-analysis

Abstract

Background and objective

Dipeptidyl peptidase-4 (DPP-4) inhibitors have been suggested to have pancreatic beta-cell preserving effect according to studies using homeostatic model of assessment for beta-cell function (HOMA-β). However, whether HOMA-β is a suitable biomarker for comparisons between hypoglycemic drugs with different mechanisms of action remains unclear. Therefore, we conducted a meta-analysis to compare the effects of DPP-4 inhibitors and other classes of hypoglycemic drugs on HOMA-β and proinsulin-to-insulin ratio (PIR).

Methods

We searched MEDLINE, CENTRAL, and Ichushi-web for the period of 1966 to May 2020. We collected randomized, controlled clinical trials in patients with type 2 diabetes mellitus comparing DPP-4 inhibitors and other classes of hypoglycemic agents [α-glucosidase inhibitors (α-GIs), glucagon-like peptide-1 (GLP-1) analogues, metformin, sodium-glucose cotransporter 2 (SGLT2) inhibitors, sulfonylureas, or thiazolidinediones]. Weighted mean differences and 95% confidence intervals of changes in HOMA-β or PIR during study periods were calculated for pairwise comparisons.

Results

Thirty-seven and 21 relevant trials were retrieved for comparisons of HOMA-β and PIR, respectively. HOMA-β and PIR consistently showed superiority of DPP-4 inhibitors compared with α-GIs. Both biomarkers consistently supported inferiority of DPP-4 inhibitors compared with GLP-1 analogues. However, PIR showed inferiority of DPP-4 inhibitors compared with metformin, and superiority compared with SGLT2 inhibitors, whereas HOMA-β showed no significant differences between DPP-4 inhibitors and the two other agents.

Conclusion

DPP-4 inhibitors appear to be superior to α-GIs but inferior to GLP-1 analogues in preservation of beta-cell function assessed by either HOMA-β or PIR. DPP-4 inhibitors seem to be superior to SGLT2 inhibitors but inferior to metformin on islet function assessed only by PIR. Because HOMA-β and PIR may indicate different aspects of beta-cell function, results of beta-cell function preserving effects of hypoglycemic agents should be interpreted with caution.

Citation: Takahashi M, Shibasaki M, Echizen H, Kushiyama A (2020) Comparisons between dipeptidyl peptidase-4 inhibitors and other classes of hypoglycemic drugs using two distinct biomarkers of pancreatic beta-cell function: A meta-analysis. PLoS ONE 15(7): e0236603. https://doi.org/10.1371/journal.pone.0236603

Editor: Tatsuo Shimosawa, International University of Health and Welfare, School of Medicine, JAPAN

Received: May 6, 2020; Accepted: July 8, 2020; Published: July 24, 2020

Copyright: © 2020 Takahashi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: The authors received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Patients with type 2 diabetes mellitus (DM) may exhibit impaired insulin secretion and insulin resistance. While multifaceted pathophysiological abnormalities of pancreatic islet cells are known to be involved in progressive deterioration of beta-cell mass and function in DM patients [1–4], debates continue on whether some hypoglycemic drugs may have pancreatic beta-cell function preserving effects. There is a possibility that dipeptidyl peptidase-4 (DPP-4) inhibitors may slow deterioration of beta-cell function in patients with type 2 DM, since administration of these drugs increased or restored the pancreatic beta-cell mass in mice and rats [5–7]. At present, however, it remains unclear whether DPP-4 inhibitors have advantage over other hypoglycemic drugs by exhibiting pancreatic beta-cell function preserving effect.

While many biomarkers have been proposed for quantifying in vivo pancreatic beta-cell function, the homeostatic model of assessment of beta-cell function (HOMA-β) has been employed most frequently in epidemiological studies [8]. Recently, the proinsulin-to-insulin ratio (PIR) has also been used in many studies as a biomarker of beta-cells [9, 10]. While HOMA-β may be associated with insulin secretion or beta-cell mass [11], PIR may reflect the efficiency of proinsulin processing within beta-cells [12]. As a result, it is of interest to study whether the two biomarkers show similar or different changes over time after initiation of hypoglycemic drugs with different mechanisms of action in patients with type 2 DM.

Previous studies using mainly HOMA-β as biomarker demonstrated that DPP-4 inhibitors may have greater pancreatic beta-cell function preserving effect compared with placebo in patients with type 2 DM [10, 13–16]. Many pair-wise comparisons of DPP-4 inhibitors with α-glucosidase inhibitors (α-GIs), metformin or sulfonylureas using various biomarkers have been reported [17–20]. However, none of the previous studies in patients with type 2 DM compared the pancretic beta-cell function preserving effect between DPP-4 inhibitors and other classes of hypoglycemic drugs using both HOMA-β and PIR as biomarkers. Here, we present our results of a meta-analysis that cast doubt about interpretationg epidemiologic data based on assessment by HOMA-β alone.

Materials and methods

The present study was performed according to the Preferred Reporting Items of Systematic Reviews and Meta-Analyses (S1 Table) [21].

Curation of literature

We searched for relevant studies using MEDLINE (from 1966 to May 2020), Cochrane Central Register of Controlled Trials (CENTRAL, Issue 5 of 12, May 2020), and Ichushi-web (Japan Medical Abstracts Society, from 1983 to May 2020). The search formula consisted of generic names of the seven DPP-4 inhibitors (alogliptin, anagliptin, linagliptin, saxagliptin, sitagliptin, teneligliptin, and vildagliptin) AND (diabetes mellitus [all fields]) AND (Randomized Controlled Trial [publication type]). The detailed search strategies for each database were shown in S2 Table. Further search was performed using information available from the references of the retrieved studies, where necessary.

Inclusion criteria were: (1) studies were written in English or Japanese; (2) patients with type 2 DM were studied; (3) changes in HOMA-β and/or PIR from the beginning to the end of study were available; (4) comparisons of biomarkers were made between one of the seven DPP-4 inhibitors and another class of hypoglycemic agent [α-GIs, GLP-1 analogues, metformin, sodium-glucose cotransporter 2 (SGLT2) inhibitors, sulfonylureas, or thiazolidinediones]; and (5) studies had a prospective and randomized design. When biomarkers were measured more than once during the study period, or interim and final reports of a study were published independently, we only used the data having the longest follow-up period for that study. Inclusion of individual studies in the meta-analysis was decided by two independent reviewers (MT and MS) according to the above inclusion criteria. In the case of disagreement between two reviewers on the decision of inclusion, the difference was resolved by discussion.

Biomarkers

HOMA-β (%) was calculated according to the following equation: (20 × insulin [μU/mL])/(glucose [mmol/L]– 3.5) [22], or by computer program [23]. HOMA-β values are known to be reduced in patients with type 2 DM, and the reduction is considered to be associated with pancreatic beta-cell dysfunction. PIR was calculated as serum proinsulin level divided by insulin level. When PIR was given as (pmol/L)/(μU/mL), the value was divided by 7.175 [24] and converted to dimensionless value. Elevated PIR has also been known to be another biomarker of beta-cell dysfunction.

Data analysis

Mean and standard deviation (SD) of within-patient changes in HOMA-β and/or PIR from baseline to the end of study were computed from the curated data (numerical data in tables or text). When only graphical presentations of the data were available, relevant data points were converted to digital data using UnGraph ver. 5 digitizer program (BIOSOFT, Cambridge, United Kingdom), which gives X and Y coordinates of lines and points on scanned images. When multiple doses of a DPP-4 inhibitor were given to different groups of patients, the data obtained from patients given doses that are approved in Japan were used for analysis. When data of biomarkers at baseline and at the end of study were available only as group mean values (not as within-patient changes), we substituted within-group changes by within-patient changes. We also estimated SD of the within-patient changes according to the equation reported elsewhere [25]. where Nbefore and Nafter are the numbers of patients before and after intervention, and SDbefore and SDafter are the SD of the group before and after intervention, respectively. When no SD was given in the original study and only 95% CI was given, we estimated the corresponding SD values according to the equation given below [26]: For statistical analysis of meta-analysis, we calculated weighted mean differences (WMDs) of HOMA-β or PIR for DPP-4 inhibitor groups versus comparator drug groups, and the 95% confidence intervals (CIs).

We compared the pancreatic beta-cell function preserving effects of various hypoglycemic drugs using DPP-4 inhibitors as reference. Changes in mean HOMA-β during the study period obtained from hypoglycemic drugs other than DPP-4 inhibitors (such as GLP-1 analogues) were subtracted from the value obtained from DPP-4 inhibitors. Consequently, when the calculated value was positive, the DPP-4 inhibitor was considered to have greater beta-cell preserving effect than the comparator. For the comparisons of PIR, however, when the calculated value was negative, the DPP-4 inhibitor was considered to have greater effect than the comparator.

Besides HOMA-β and PIR, mean values of patients’ demographic data and clinical characteristics [including body mass index (BMI), glycosylated hemoglobin A1c (HbA1c) (%), and duration since diagnosis of type 2 DM] were curated for assessment of comparability of trials. Quality of study was assessed using the risk of bias tool provided by the Cochrane Handbook [27]. Two reviewers (MT and MS) independently evaluated all studies according to the following seven critical items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias (i.e., enough patients enrolled to reach the endpoint, adequate compliance). Publication bias was assessed visually using funnel plots and statistically using the method of Egger et al. [28]. Heterogeneity of curated data from individual studies was examined by Cochrane’s Q test. When an I2 value exceeded 50%, we considered that there was heterogeneity between the studies analyzed. We used fixed effects model when heterogeneity of curated data was rejected, and used a random effects model otherwise. For statistical analysis of meta-analysis, we calculated WMDs and their 95% CIs using Review Manager version 5.3 (The Nordic Cochrane Centre, the Cochrane Collaboration, Copenhagen, Denmark) based on an inverse variance method.

For sensitivity analyses, studies considered having high risk of bias in at least two out of seven domains (such as unblinded participants or personnel) were removed, and then analyses were repeated. In addition, we performed sensitivity tests when funnel plots of meta-analysis data suggested a possibility of publication bias. Specifically, we excluded suspected studies one by one from the corresponding meta-analysis and examined if there was a qualitative change in the result of statistical inference. When the statistical significance was lost by exclusion of a specific study, we considered that the result of the corresponding meta-analysis was conditional.

We presented the results of meta-analyses in a two-dimensional plot, in which we plotted the WMDs with 95% CIs obtained from comparisons of DPP-4 inhibitors versus other classes of hypoglycemia drugs (comparators: α-GIs, GLP-1 analogues, metformin, SGLT2 inhibitors, sulfonylureas, and thiazolidinediones) using HOMA-β (ordinate) and PIR (abscissa) as biomarkers. In addition, we depicted the approximate sample size of each comparison by the area of a circle in the plot.

Results

Characteristics of the included studies

In this meta-analysis, we retrieved 40 relevant studies from the three databases; MEDLINE, CENTRAL, and Ichushi-web, as shown in the flowchart (Fig 1). Among these studies, data for analysis were available for six of seven DPP-4 inhibitors: alogliptin (n = 2) [29, 30], anagliptin (n = 1) [31], linagliptin (n = 3) [32–34], saxagliptin (n = 4) [35–38], sitagliptin (n = 26) [39–64], and vildagliptin (n = 4) [65–68], but not for teneligliptin (n = 0). Comparators of DPP-4 inhibitors were α-GIs (n = 9) [30, 31, 34, 35, 41, 50, 63–65], GLP-1 analogues (n = 5) [47, 52, 58, 62, 68], metformin (n = 5) [29, 38–40, 56], SGLT2 inhibitors (n = 5) [36, 53, 54, 60, 61], sulfonylureas (n = 12) [32, 33, 37, 43–46, 48, 57, 59, 66, 67], and thiazolidinediones (n = 4) [42, 49, 51, 55].

thumbnail
Download:
Fig 1. Flowchart displaying the steps of selection of studies on seven dipeptidyl peptidase-4 inhibitors for meta-analysis.

Numbers of articles that met either the inclusion or exclusion criteria are shown in small tables. Abbreviations: CEN, CENTRAL; IW, Ichushi-web; MED, MEDLINE; MS, manual search.

https://doi.org/10.1371/journal.pone.0236603.g001

A concise summary of the studies included in the meta-analysis was shown in Table 1. The overall mean HOMA-β at baseline was 55.9% (range: 25.5% to 150.6%), and mean PIR was 0.35 (range: 0.20 to 0.60). The mean follow-up period was 34 weeks (range: 8 to 104 weeks). As for the quality of studies, 11 studies [35, 38, 50, 52, 55, 57, 59, 61, 63, 66, 68] were considered at high risk of bias according to the tool of Cochrane Handbook. All 11 were open label studies (S3 Table). Particularly, two studies were considered to be at high risk of bias in three domains [50, 63]. Both studies compared the clinical outcomes between sitagliptin and α-GIs, and adherence was significantly lower in the α-GIs group than in the sitagliptin group.

thumbnail
Download:
Table 1. Summary of studies included in the meta-analysis.

https://doi.org/10.1371/journal.pone.0236603.t001

Comparison of DPP-4 inhibitors with other hypoglycemic drugs using HOMA-β as biomarker

Relevant data for comparing the effect on HOMA-β between DPP-4 inhibitors and other hypoglycemic drugs were available from eight studies of α-GIs [30, 31, 35, 41, 50, 63–65], five studies of GLP-1 analogues [47, 52, 58, 62, 68], five studies of metformin [29, 38, 39, 40, 56], five studies of SGLT2 inhibitors [36, 53, 54, 60, 61], ten studies of sulfonylureas [37, 43, 44–46, 48, 57, 59, 66, 67], and four studies of thiazolidinediones [42, 49, 51, 55] (Table 1). According to the results of Cochrane’s Q test, the random effects model was employed for the meta-analysis of DPP-4 inhibitors versus α-GIs, SGLT2 inhibitors and sulfonylureas, but the fixed effects model was employed for DPP-4 inhibitors versus the remaining comparators. Analysis of all 37 studies showed that DPP-4 inhibitors increased HOMA-β to a significantly greater extent than α-GIs [WMD (95% CI): 7.54% (3.84, 11.24%)]. In contrast, DPP-4 inhibitors increased HOMA-β to a significantly less extent than GLP-1 analogues [-13.08% (-14.11, -12.05%)]. No significant differences in the change in HOMA-β were observed between DPP-4 inhibitors and metformin [-2.18% (-10.48, 6.12%)], between DPP-4 inhibitors and SGLT2 inhibitors [-4.08% (-11.65, 3.48%)], between DPP-4 inhibitors and sulfonylureas [-14.18% (-29.80, 1.45%)], and between DPP-4 inhibitors and thiazolidinediones [6.63% (-1.55, 14.81%)] (Fig 2).

thumbnail
Download:
Fig 2. Forest plots of meta-analyses for changes in homeostatic model of assessment for beta-cell function (HOMA-β).

Filled square represents weighted mean difference (WMD) of homeostatic model of assessment for beta-cell function (HOMA-β) for dipeptidyl peptidase-4 inhibitor versus other hypoglycemic agents. The two ends of the horizontal line passing through WMD represent the lower and upper limits of 95% confidence interval (CI). The size of each square corresponds to the weight assigned to the specific WMD data in the meta-analysis. Filled diamond at the bottom of each forest plot shows the combined WMD for individual comparison. Lateral tips of the diamond represent the lower and upper limits of 95% CI. A diamond located to the right of y-axis favors DPP-4 inhibitors over comparators. Abbreviations: CI, confidence interval; DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; IV, inverse variance; SD, standard deviation; SGLT2, sodium-glucose cotransporter 2; alo, alogliptin; ana, anagliptin; saxa, saxagliptin; sita, sitagliptin; vilda, vildagliptin.

https://doi.org/10.1371/journal.pone.0236603.g002

Comparison of DPP-4 inhibitors with other hypoglycemic drugs using PIR as biomarker

Relevant data for comparing the effect on PIR between DPP-4 inhibitors and other hypoglycemic drugs were available from four studies of α-GIs [31, 34, 50, 63], one study of a GLP-1 analogue [52], four studies of metformin [29, 39, 40, 56], two studies of SGLT2 inhibitors [54, 61], seven studies of sulfonylureas [32, 33, 44–46, 48, 57], and three studies of thiazolidinediones [42, 49, 51] (Table 1). In accordance with the results of Cochrane Q test, the random effects model was employed for the comparison with sulfonylureas, and the fixed effects model for comparisons with all other drugs. Analysis of all 21 studies showed that DPP-4 inhibitors reduced PIR to a significantly greater extent than α-GIs [WMD (95%CI): -0.04 (-0.07, -0.02)] and SGLT2 inhibitors [-0.08 (-0.14, -0.03)]. In contrast, GLP-1 analogues and metformin reduced PIR to a significantly greater extent than DPP-4 inhibitors [0.06 (0.00, 0.12) and 0.04 (0.02, 0.06), respectively]. No statistically significant differences were detected for the comparisons of DPP-4 inhibitors with sulfonylureas [-0.02 (-0.05, 0.01)], and with thiazolidinediones [-0.01 (-0.05, 0.02)] (Fig 3).

thumbnail
Download:
Fig 3. Forest plots of meta-analyses for changes in proinsulin-to-insulin ratio (PIR).

Filled square represents weighted mean difference (WMD) of proinsulin-to-insulin ratio (PIR) for dipeptidyl peptidase-4 inhibitor versus other hypoglycemic agents. The two ends of the horizontal line passing through WMD represent the lower and upper limits of 95% confidence interval (CI). The size of each square corresponds to the weight assigned to the specific WMD data in the meta-analysis. Filled diamond at the bottom of each forest plot shows the combined WMD for individual comparison. Lateral tips of the diamond represent the lower and upper limits of 95% CI. A diamond located to the left of y-axis favors DPP-4 inhibitors over comparators. Abbreviations: CI, confidence interval; DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; IV, inverse variance; SD, standard deviation; SGLT2, sodium-glucose cotransporter 2; alo, alogliptin; ana, anagliptin; lina, linagliptin; sita, sitagliptin.

https://doi.org/10.1371/journal.pone.0236603.g003

Two-dimensional display for the effects of DPP-4 inhibitors versus other hypoglycemic drugs using HOMA-β and PIR as biomarkers

Fig 4 showed a two-dimensional plot depicting WMDs with 95% CIs obtained from comparisons of DPP-4 inhibitors versus other classes of hypoglycemic drugs using HOMA-β and PIR as biomarkers. The data of each pair of comparison appeared as a cluster in one of the four quadrants. Numerical results of each comparison between DPP-4 inhibitors and other classes of hypoglycemia drugs already mentioned above. Fig 4 showed those results in an integrated manner to facilitate visual understanding. When both perpendicular and horizontal lines were present within the second quadrant, HOMA-β and PIR consistently showed superiority of DPP-4 inhibitors compared with other classes of hypoglycemic drugs. Whereas, when both perpendicular and horizontal lines were present within the fourth quadrant, HOMA-β and PIR consistently showed inferiority of DPP-4 inhibitors.

thumbnail
Download:
Fig 4. Two-dimensional display of comparisons of DPP-4 inhibitors versus other hypoglycemic drugs using HOMA-β and PIR.

Combined weighted mean differences (WMDs) of homeostatic model of assessment for beta-cell function (HOMA-β) for various comparisons are plotted on the ordinate, and combined WMDs of proinsulin-to-insulin ratio (PIR) are plotted on the abscissa. The center of each circle represents the combined WMD of HOMA-β and PIR. Perpendicular and horizontal lines represent 95% confident intervals for HOMA-β and PIR, respectively. Approximate number of patients in the sample is shown by the area of the circle. Note that an increase in HOMA-β and a decrease in PIR are associated with amelioration of beta-cell function in patients with type 2 DM. Abbreviations: DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; HOMA-β, homeostatic model of assessment for beta-cell function; PIR, proinsulin-to-insulin ratio; SGLT2, sodium-glucose cotransporter 2.

https://doi.org/10.1371/journal.pone.0236603.g004

Sensitivity analysis and evaluation of publication bias

We conducted a separate meta-analysis after excluding 11 studies [35, 38, 50, 52, 55, 57, 59, 61, 63, 66, 68], since they appeared to have high risk of bias according to Cochrane’s bias tool (S3 Table). Excluding those studies changed the significance of the difference in HOMA-β for DPP-4 inhibitors versus SGLT2 inhibitors [-7.82% (-11.89, -3.75%)] and versus sulfonylureas [-18.70% (-36.86, -0.53%)].

Funnel plots of the comparisons of DPP-4 inhibitors versus α-GIs, SGLT2 inhibitors, and sulfonylureas for HOMA-β (S1A, S1D and S1E Fig) as well as those of DPP-4 inhibitors versus α-GIs and sulfonylureas for PIR (S2A and S2D Fig) displayed apparently asymmetric distribution of data. However, the 95% CIs of the intercepts of Egger’s regression lines [28] for HOMA-β between DPP-4 inhibitors and α-GIs [(-1.50, 3.47), p = 0.37], SGLT2 inhibitors [(-2.33, 9.37), p = 0.15] or sulfonylureas [(-5.71, 4.67), p = 0.82]; as well as for PIR between DPP-4 inhibitors and α-GIs [(-6.77, 3.72), p = 0.34] or sulfonylureas [(-4.27, 1.50), p = 0.27] included the origin of the coordinates. As a result, we concluded that sensitivity analysis of meta-analysis to evaluate the effect of publication bias may not be required for either HOMA-β or PIR.

Discussion

To the best of our knowledge, the present study is the first to compare the pancreatic beta-cell preserving effect between DPP-4 inhibitors and other hypoglycemic drugs using two distinct biomarkers, HOMA-β and PIR. In this study, the results of HOMA-β and PIR consistently showed superiority of DPP-4 inhibitors over α-GIs and inferiority of DPP-4 inhibitors to GLP-1 analogues. However, PIR showed inferiority of DPP-4 inhibitors to metformin and superiority over SGLT2 inhibitors, while HOMA-β showed no significant differences in these comparisons. On the other hand, no significant differences in the change in both HOMA-β and PIR were observed between DPP-4 inhibitors and sulfonylureas as well as between DPP-4 inhibitors and thiazolidinediones. Recently, Wu et al. [69] reported a network meta-analysis on the comparisons of incretin-based therapies (i.e. DPP-4 inhibitors and GLP-1 analogues) with α-GIs, metformin, sulfonylureas, thiazolidinediones and placebo using HOMA-β as the sole biomarker for beta-cell function in patients with type 2 DM. Their results agree well with our present meta-analysis. The implication of our study is that results of comparison of beta-cell preserving effect between DPP-4 inhibitors and other hypoglycemic drugs using HOMA-β and PIR as biomarker should be interpreted with caution, because the two biomarkers may indicate different functions of beta-cell.

The changes observed in WMDs for HOMA-β and PIR yielded comparable results in the comparisons between DPP-4 inhibitors and two hypoglycemic drugs; α-GIs and GLP-1 analogues, in patients with type 2 DM. These results suggest that DPP-4 inhibitors are superior to α-GIs but inferior to GLP-1 analogues in terms of beta-cell preserving effect. In our meta-analysis between DPP-4 inhibitors and α-GIs using all the retrieved articles, we included a study in which all participants were given insulin [64]. Nevertheless, removal of that study [64] from meta-analysis did not change the statistical results.

On the other hand, the changes observed in WMDs for HOMA-β and PIR when comparing DPP-4 inhibitors and sulfonylureas showed opposite trends: the changes in HOMA-β appeared to favor sulfonylureas over DPP-4 inhibitors, whereas those in PIR appeared opposite. Hyperglycemia in patients with type 2 DM may be attributed to reduced insulin secretion from the pancreas and/or to peripheral insulin resistance. Attenuated basal secretion of insulin which is stimulated mainly by fasting serum glucose levels can be detected by HOMA-β [70]. However, insulin secretion in patients receiving sulfonylureas may be augmented even during euglycemia by the pharmacological effects of the drugs [71]. As a result, increases of HOMA-β observed after the initiation of sulfonylurea may not be attributed simply to the beta-cell preserving effect of the drug [72]. In our study, the comparison between DPP-4 inhibitors and sulfonylureas using HOMA-β as biomarker did not reach a significant level. One of the reasons for this result may be the selection of statistical model. We employed the random effect model, since the data reported by Derosa et al. [48] showed a heterogenously large effect on HOMA-β. Indeed, when their data were removed from the comparative integration analysis, heterogeneity disappeared and the overall comparison was statistically significant. The heterogenous data of Derosa et al. were probably due to markedly higher baseline HOMA-β in their patients (93.0%) compared to other studies used in comparison (range: 38.2% to 67.5%). In contrast, there was no significant difference in baseline PIR between the study of Derosa et al (0.25) and the other studies (range: 0.20 to 0.49). PIR is associated with the efficiency of proinsulin processing to insulin within beta-cells rather than the beta-cell mass [12]. In this respect, PIR may be a more suitable biomarker than HOMA-β for assessing the beta-cell function in patients with type 2 DM receiving sulfonylureas. In the present study, no significant differences were observed between DPP-4 inhibitors and sulfonylureas when using HOMA-β or PIR, although there was a trend of superiority for DPP-4 inhibitors over sulfonylureas when using PIR.

HOMA-β and PIR showed discrepant results in the assessment of beta-cell preserving effects of SGLT2 inhibitors. HOMA-β was proposed as a marker of beta-cell function on the assumption that blood glucose level depends only the rate of insulin secretion rate [22]. In patients receiving SGLT2 inhibitors, however, blood glucose levels are controlled not only by the insulin-dependent intracellular transport but also by the SGLT2 inhibitor-induced augmentation of urinary loss of glucose [73]. Consequently, the effect of SGLT2 inhibitors on beta-cell preservation may be overestimated by HOMA-β. The other biomarker examined in this study, PIR, was employed in a recent large cohort study searching for clinical factors that influence the progression of pancreatic beta-cell dysfunction (BetaDecline study) [74]. Taken together, PIR may be a useful biomarker of pancreatic beta-cell function for comparing DPP-4 inhibitors with SGLT2 inhibitors.

There are few studies on the effects of thiazolidinediones and metformin on beta-cell function in humans. Ishida et al. [75] reported that pioglitazone enhanced beta-cell function in diabetic mice through reduction of oxidative stress in pancreatic beta-cells. It has been reported that the reduction of oxidative stress could be due to increase the intrinsic activity of the glucose transporters [76] or augmented beta-cell pancreatic duodenal homeobox-1 expression [77]. Meanwhile, the favorable effects of metformin for beta-cell function are considered to be partially based on improving insulin processing insufficiency. Nagi et al. [78] reported that metformin decreased plasma proinsulin concentrations in subjects with type 2 diabetes mellitus. Their findings support the results of our meta-analysis that metformin reduced PIR to a significantly greater extent than DPP-4 inhibitors. However, precise mechanisms by which thiazolidinediones and metformin improve beta-cell function in clinical settings remain to be clarified.

In the present study, we made a new comprehensive attempt to compare the beta-cell preserving effects of hypoglycemic drugs using HOMA-β and PIR as biomarkers. We present a two-dimensional display of the beta-cell preserving effect of DPP-4 inhibitors versus other hypoglycemic drugs when assessed by HOMA-β and PIR (Fig 4). While an increase in HOMA-β may be associated with amelioration of the beta-cell function of insulin secretion [11], a reduction in PIR may be related to the improvement of proinsulin processing to insulin [79]. Therefore, data plotted in either the second or fourth quadrant imply qualitative similarity of HOMA-β and PIR in assessing the beta-cell function preserving effect of hypoglycemic drugs. We revealed that there is an agreement between HOMA-β and PIR regarding the pancreatic beta-cell preserving effects of hypoglycemic drugs compared with DPP-4 inhibitors, except for sulfonylureas and SGLT2 inhibitors for which the assumption of HOMA-β may not hold. In the plot of HOMA-β versus PIR in Fig 4, comparisons between DPP-4 inhibitors and sulfonylureas or SGLT2 inhibitors are plotted in the third quadrant. Therefore, evaluation using HOMA-β alone might not be suitable. Data presentation as Fig 4 may be useful for comprehensive interpretation of the pancreatic beta cell function preserving effects of hypoglycemic drugs with different mechanisms of action. Further studies using this approach to compare the performance of different biomarkers of pancreatic beta-cells are anticipated.

HOMA-β is most frequently used as a clinical and epidemiological biomarker for assessing pancreatic beta-cell function. However, our results of comparison between DPP-4 inhibitors and sulfonylureas or SGLT2 inhibitors illustrate that HOMA-β may not reflect beta-cell function accurately for drugs with hypoglycemic actions via direct insulin secretion or independent of insulin secretion. In addition, although HOMA-β assumes that all glucose-lowering action is provided by correctly processed insulin derived from beta-cells [80], impaired processing of proinsulin to insulin is observed in patients with type 2 DM [81].

One of the limitations of the present study is inherent to the methodology of meta-analysis. The biomarkers employed for evaluating the functional beta-cell reserve; HOMA-β and PIR, were secondary endpoints of the clinical trials included in the present meta-analysis. While the Cochrane risk of bias tool may be useful for assessing the quality of studies and the primary outcome, whether it is also valid for assessing the quality of the secondary outcome remains unclear. In particular, when the data of secondary outcomes are derived from a subgroup of patients, random allocation of patients may not be guaranteed. As a result, we cannot categorically eliminate selection bias or attrition bias for the assessment of the secondary outcomes analyzed in the present study. In addition, we cannot eliminate publication bias legitimately from meta-analysis despite the performance of funnel plot analyses. Furthermore, Egger’s regression analysis used for assessing bias and homogeneity of studies may be useful if more than ten publications are available for each integration analysis [82].

Conclusions

We performed comprehensive meta-analyses of the effects of DPP-4 inhibitors on pancreatic beta-cell function compared with other classes of hypoglycemic drugs using HOMA-β and PIR as biomarkers. DPP-4 inhibitors appear to be superior to α-GIs but inferior to GLP-1 analogues in terms of preservation of beta-cell function assessed by either HOMA-β or PIR. DPP-4 inhibitors seem to be superior to SGLT2 inhibitors but inferior to metformin in terms of islet function assessed only by PIR. Since HOMA-β and PIR may represent different aspects of beta-cell function, results of comparison of beta-cell function preserving effects among hypoglycemic agents should be interpreted with caution.

Supporting information

S1 Fig. Funnel plots of the data employed in the meta-analysis of homeostatic model of assessment for beta-cell function (HOMA-β).

This plot is effect size (weighted mean difference of homeostatic model of assessment for beta-cell function [HOMA-β] from individual articles) (X axis) versus standard error of effect size (Y axis). Dotted vertical line represents pooled estimate of effects. Abbreviations: DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; MD, mean difference; SE, standard error; SGLT2, sodium-glucose cotransporter 2.

https://doi.org/10.1371/journal.pone.0236603.s001

(TIF)

S2 Fig. Funnel plots of the data employed for the meta-analysis of proinsulin-to-insulin ratio (PIR).

This plot is effect size (weighted mean difference of proinsulin-to-insulin ratio [PIR] from individual articles) (X axis) versus standard error of effect size (Y axis). Dotted vertical line represents pooled estimate of effects. Abbreviations: DPP-4, dipeptidyl peptidase-4; MD, mean difference; SE, standard error; SGLT2, sodium-glucose cotransporter 2.

https://doi.org/10.1371/journal.pone.0236603.s002

(TIF)

S1 Table. PRISMA 2009 checklist.

https://doi.org/10.1371/journal.pone.0236603.s003

(DOC)

S2 Table. Detailed search strategies for the MEDLINE, CENTRAL and Ichushi-web databases.

https://doi.org/10.1371/journal.pone.0236603.s004

(DOCX)

S3 Table. Summary of the authors’ judgement on the risk of bias of the publications included in the present study.

Symbols: +, low risk of bias; -, high risk of bias; ?, unclear risk of bias.

https://doi.org/10.1371/journal.pone.0236603.s005

(DOCX)

References

  1. 1. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC. Beta-cell deficit and increased beta-cell apoptosis in humans with type 2 diabetes. Diabetes. 2003;52(1):102–110. pmid:12502499.
  2. 2. Inaishi J, Saisho Y, Sato S, Kou K, Murakami R, Watanabe Y, et al. Effects of Obesity and Diabetes on alpha- and beta-Cell Mass in Surgically Resected Human Pancreas. J Clin Endocrinol Metab. 2016;101(7):2874–2882. pmid:27070277; PubMed Central PMCID: PMC4929842.
  3. 3. Muller WA, Faloona GR, Aguilar-Parada E, Unger RH. Abnormal alpha-cell function in diabetes. Response to carbohydrate and protein ingestion. N Engl J Med. 1970;283(3):109–115. pmid:4912452.
  4. 4. Sato Y, Komatsu M, Katakura M, Ohfusa H, Yamada S, Yamauchi K, et al. Diminution of early insulin response to glucose in subjects with normal but minimally elevated fasting plasma glucose. Evidence for early beta-cell dysfunction. Diabet Med. 2002;19(7):566–571. pmid:12099959.
  5. 5. Mu J, Petrov A, Eiermann GJ, et al. Inhibition of DPP-4 with sitagliptin improves glycemic control and restores islet cell mass and function in a rodent model of type 2 diabetes. Eur J Pharmacol. 2009;623(1–3):148–154. pmid:19765579.
  6. 6. Mu J, Woods J, Zhou YP, et al. Chronic inhibition of dipeptidyl peptidase-4 with a sitagliptin analog preserves pancreatic beta-cell mass and function in a rodent model of type 2 diabetes. Diabetes. 2006;55(6):1695–1704. pmid:16731832.
  7. 7. Pospisilik JA, Martin J, Doty T, Ehses JA, Pamir N, Lynn FC, et al. Dipeptidyl peptidase IV inhibitor treatment stimulates beta-cell survival and islet neogenesis in streptozotocin-induced diabetic rats. Diabetes. 2003;52(3):741–750. pmid:12606516.
  8. 8. Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27(6):1487–1495. pmid:15161807.
  9. 9. Lu J, Zang J, Li H. Impact of three oral antidiabetic drugs on markers of beta-cell function in patients with type 2 diabetes: a meta-analysis. PLoS One. 2013;8(10):e76713. Epub 2013/11/10. pmid:24204660; PubMed Central PMCID: PMC3808360.
  10. 10. Riche DM, East HE, Riche KD. Impact of sitagliptin on markers of beta-cell function: a meta-analysis. Am J Med Sci. 2009;337(5):321–328. Epub 2009/03/27. pmid:19322069.
  11. 11. Hermans MP, Levy JC, Morris RJ, Turner RC. Comparison of tests of beta-cell function across a range of glucose tolerance from normal to diabetes. Diabetes. 1999;48(9):1779–1786. Epub 1999/09/10. pmid:10480608.
  12. 12. Rhodes CJ, Alarcon C. What beta-cell defect could lead to hyperproinsulinemia in NIDDM? Some clues from recent advances made in understanding the proinsulin-processing mechanism. Diabetes. 1994;43(4):511–517. Epub 1994/04/01. pmid:8138054.
  13. 13. Gao W, Wang Q, Yu S. Efficacy, safety and impact on beta-cell function of dipeptidyl peptidase-4 inhibitors plus metformin combination therapy in patients with type 2 diabetes and the difference between Asians and Caucasians: a meta-analysis. J Endocrinol Invest. 2016;39(9):1061–1074. Epub 2016/04/14. pmid:27072669.
  14. 14. Li X, Huang X, Bai C, Qin D, Cao S, Mei Q, et al. Efficacy and Safety of Teneligliptin in Patients With Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Front Pharmacol. 2018;9:449. Epub 2018/05/22. pmid:29780322; PubMed Central PMCID: PMC5946087.
  15. 15. Lyu X, Zhu X, Zhao B, Du L, Chen D, Wang C, et al. Effects of dipeptidyl peptidase-4 inhibitors on beta-cell function and insulin resistance in type 2 diabetes: meta-analysis of randomized controlled trials. Sci Rep. 2017;7:44865. Epub 2017/03/23. pmid:28322294; PubMed Central PMCID: PMC5359588.
  16. 16. van Genugten RE, van Raalte DH, Diamant M. Dipeptidyl peptidase-4 inhibitors and preservation of pancreatic islet-cell function: a critical appraisal of the evidence. Diabetes Obes Metab. 2012;14(2):101–111. Epub 2011/07/15. pmid:21752172.
  17. 17. Cai X, Yang W, Zhou L, Zhang S, Han X, Ji L. Comparisons of the efficacy of glucose control, lipid profile, and beta-cell function between DPP-4 inhibitors and AGI treatment in type 2 diabetes patients: a meta-analysis. Endocrine. 2015;50(3):590–597. pmid:26048437.
  18. 18. Du Q, Wu B, Wang YJ, Yang S, Zhao YY, Liang YY. Comparative effects of sitagliptin and metformin in patients with type 2 diabetes mellitus: a meta-analysis. Curr Med Res Opin. 2013;29(11):1487–1494. pmid:23927568.
  19. 19. Zhang Y, Hong J, Chi J, Gu W, Ning G, Wang W. Head-to-head comparison of dipeptidyl peptidase-IV inhibitors and sulfonylureas—a meta-analysis from randomized clinical trials. Diabetes Metab Res Rev. 2014;30(3):241–256. pmid:24123720.
  20. 20. Zhou JB, Bai L, Wang Y, Yang JK. The benefits and risks of DPP4-inhibitors vs. sulfonylureas for patients with type 2 diabetes: accumulated evidence from randomised controlled trial. Int J Clin Pract. 2016;70(2):132–141. pmid:26709610.
  21. 21. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6(7):e1000100. Epub 2009/07/22. pmid:19621070; PubMed Central PMCID: PMC2707010.
  22. 22. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–419. pmid:3899825.
  23. 23. Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21(12):2191–2192. pmid:9839117.
  24. 24. von Berghes C, Brabant G, Biebermann H, Krude H, Wiegand S. Proinsulin and the proinsulin/insulin ratio in overweight and obese children and adolescents: relation to clinical parameters, insulin resistance, and impaired glucose regulation. Pediatr Diabetes. 2011;12(3 Pt 2):242–249. Epub 2011/04/27. pmid:21518412
  25. 25. Borenstein M, Hedges L, Higgins J, Rothstein H. Effect sizes based on means. In: Introduction to Meta-Analysis. Chichester: Wiley; 2009. pp. 21–32.
    • 26. Higgins J, Deeks J. Selecting studies and collecting data. In: Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions. Chichester: Wiley; 2008. pp. 151–185.
      • 27. Higgins J, Altman D. Assessing risk of bias in included studies. In: Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions. Chichester: Wiley; 2008. pp. 187–241.
        • 28. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–634. Epub 1997/10/06. pmid:9310563; PubMed Central PMCID: PMC2127453.
        • 29. Pratley RE, Fleck P, Wilson C. Efficacy and safety of initial combination therapy with alogliptin plus metformin versus either as monotherapy in drug-naive patients with type 2 diabetes: a randomized, double-blind, 6-month study. Diabetes Obes Metab. 2014;16(7):613–621. pmid:24400655.
        • 30. Seino Y, Fujita T, Hiroi S, Hirayama M, Kaku K. Efficacy and safety of alogliptin in Japanese patients with type 2 diabetes mellitus: a randomized, double-blind, dose-ranging comparison with placebo, followed by a long-term extension study. Curr Med Res Opin. 2011;27(9):1781–1792. pmid:21806314.
        • 31. Kaku K. Efficacy and Safety of Anagliptin in Japanese Patients with Type 2 Diabetes—A Multi-centre, Randomized, Placebo-and Active Comparator-controlled, Double-blind, Parallel-group Study—. Jpn Pharmacol Ther. 2012;40(11):985–995. pmid:2013185801.
        • 32. Forst T, Uhlig-Laske B, Ring A, Graefe-Mody U, Friedrich C, Herbach K, et al. Linagliptin (BI 1356), a potent and selective DPP-4 inhibitor, is safe and efficacious in combination with metformin in patients with inadequately controlled Type 2 diabetes. Diabet Med. 2010;27(12):1409–1419. pmid:21059094.
        • 33. Gallwitz B, Rosenstock J, Rauch T, Bhattacharya S, Patel S, von Eynatten M, et al. 2-year efficacy and safety of linagliptin compared with glimepiride in patients with type 2 diabetes inadequately controlled on metformin: a randomised, double-blind, non-inferiority trial. Lancet. 2012;380(9840):475–483. pmid:22748821.
        • 34. Kawamori R, Inagaki N, Araki E, Watada H, Hayashi N, Horie Y, et al. Linagliptin monotherapy provides superior glycaemic control versus placebo or voglibose with comparable safety in Japanese patients with type 2 diabetes: a randomized, placebo and active comparator-controlled, double-blind study. Diabetes Obes Metab. 2012;14(4):348–357. pmid:22145698.
        • 35. Du J, Liang L, Fang H, Xu F, Li W, Shen L, et al. Efficacy and safety of saxagliptin compared with acarbose in Chinese patients with type 2 diabetes mellitus uncontrolled on metformin monotherapy: Results of a Phase IV open-label randomized controlled study (the SMART study). Diabetes Obes Metab. 2017;19(11):1513–1520. pmid:28296055.
        • 36. Ekholm E, Hansen L, Johnsson E, Iqbal N, Carlsson B, Chen H, et al. Combined treatment with saxagliptin plus dapagliflozin reduces insulin levels by increased insulin clearance and improves beta-cell function. Endocr Pract. 2017;23(3):258–265. pmid:27849380.
        • 37. Goke B, Gallwitz B, Eriksson J, Hellqvist A, Gause-Nilsson I, Investigators DC. Saxagliptin is non-inferior to glipizide in patients with type 2 diabetes mellitus inadequately controlled on metformin alone: a 52-week randomised controlled trial. Int J Clin Pract. 2010;64(12):1619–1631. pmid:20846286.
        • 38. Tao T, Wu P, Wang Y, Liu W. Comparison of glycemic control and beta-cell function in new onset T2DM patients with PCOS of metformin and saxagliptin monotherapy or combination treatment. BMC Endocr Disord. 2018;18(1):14. Epub 2018/02/28. pmid:29482528; PubMed Central PMCID: PMC5828487.
        • 39. Aschner P, Katzeff HL, Guo H, Sunga S, Williams-Herman D, Kaufman KD, et al. Efficacy and safety of monotherapy of sitagliptin compared with metformin in patients with type 2 diabetes. Diabetes Obes Metab. 2010;12(3):252–261. pmid:20070351.
        • 40. Derosa G, Maffioli P, Salvadeo SA, Ferrari I, Ragonesi PD, Querci F, et al. Effects of sitagliptin or metformin added to pioglitazone monotherapy in poorly controlled type 2 diabetes mellitus patients. Metabolism. 2010;59(6):887–895. pmid:20015525.
        • 41. Iwamoto Y, Tajima N, Kadowaki T, Nonaka K, Taniguchi T, Nishii M, et al. Efficacy and safety of sitagliptin monotherapy compared with voglibose in Japanese patients with type 2 diabetes: a randomized, double-blind trial. Diabetes Obes Metab. 2010;12(7):613–622. pmid:20590736.
        • 42. Scott R, Loeys T, Davies MJ, Engel SS, Sitagliptin Study G. Efficacy and safety of sitagliptin when added to ongoing metformin therapy in patients with type 2 diabetes. Diabetes Obes Metab. 2008;10(10):959–969. pmid:18201203.
        • 43. Scott R, Wu M, Sanchez M, Stein P. Efficacy and tolerability of the dipeptidyl peptidase-4 inhibitor sitagliptin as monotherapy over 12 weeks in patients with type 2 diabetes. Int J Clin Pract. 2007;61(1):171–180. pmid:17156104.
        • 44. Seck T, Nauck M, Sheng D, Sunga S, Davies MJ, Stein PP, et al. Safety and efficacy of treatment with sitagliptin or glipizide in patients with type 2 diabetes inadequately controlled on metformin: a 2-year study. Int J Clin Pract. 2010;64(5):562–576. pmid:20456211.
        • 45. Arjona Ferreira JC, Corry D, Mogensen CE, Sloan L, Xu L, Golm GT, et al. Efficacy and safety of sitagliptin in patients with type 2 diabetes and ESRD receiving dialysis: a 54-week randomized trial. Am J Kidney Dis. 2013;61(4):579–587. pmid:23352379.
        • 46. Arjona Ferreira JC, Marre M, Barzilai N, Guo H, Golm GT, Sisk CM, et al. Efficacy and safety of sitagliptin versus glipizide in patients with type 2 diabetes and moderate-to-severe chronic renal insufficiency. Diabetes Care. 2013;36(5):1067–1073. pmid:23248197; PubMed Central PMCID: PMC3631833.
        • 47. Berg JK, Shenouda SK, Heilmann CR, Gray AL, Holcombe JH. Effects of exenatide twice daily versus sitagliptin on 24-h glucose, glucoregulatory and hormonal measures: a randomized, double-blind, crossover study. Diabetes Obes Metab. 2011;13(11):982–989. pmid:21615670; PubMed Central PMCID: PMC3258427.
        • 48. Derosa G, Cicero AF, Franzetti IG, Querci F, Carbone A, Piccinni MN, et al. A comparison between sitagliptin or glibenclamide in addition to metformin + pioglitazone on glycaemic control and beta-cell function: the triple oral therapy. Diabet Med. 2013;30(7):846–854. pmid:23413771.
        • 49. Henry RR, Staels B, Fonseca VA, Chou MZ, Teng R, Golm GT, et al. Efficacy and safety of initial combination treatment with sitagliptin and pioglitazone—a factorial study. Diabetes Obes Metab. 2014;16(3):223–230. pmid:23909985.
        • 50. Kobayashi K, Yokoh H, Sato Y, Takemoto M, Uchida D, Kanatsuka A, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor sitagliptin compared with alpha-glucosidase inhibitor in Japanese patients with type 2 diabetes inadequately controlled on sulfonylurea alone (SUCCESS-2): a multicenter, randomized, open-label, non-inferiority trial. Diabetes Obes Metab. 2014;16(8):761–765. pmid:24447683.
        • 51. Perez-Monteverde A, Seck T, Xu L, Lee MA, Sisk CM, Williams-Herman DE, et al. Efficacy and safety of sitagliptin and the fixed-dose combination of sitagliptin and metformin vs. pioglitazone in drug-naive patients with type 2 diabetes. Int J Clin Pract. 2011;65(9):930–938. pmid:21849007.
        • 52. Pratley R, Nauck M, Bailey T, Montanya E, Cuddihy R, Filetti S, et al. One year of liraglutide treatment offers sustained and more effective glycaemic control and weight reduction compared with sitagliptin, both in combination with metformin, in patients with type 2 diabetes: a randomised, parallel-group, open-label trial. Int J Clin Pract. 2011;65(4):397–407. pmid:21355967; PubMed Central PMCID: PMC3085127.
        • 53. Rosenstock J, Aggarwal N, Polidori D, Zhao Y, Arbit D, Usiskin K, et al. Dose-ranging effects of canagliflozin, a sodium-glucose cotransporter 2 inhibitor, as add-on to metformin in subjects with type 2 diabetes. Diabetes Care. 2012;35(6):1232–1238. pmid:22492586; PubMed Central PMCID: PMC3357223.
        • 54. Schernthaner G, Gross JL, Rosenstock J, Guarisco M, Fu M, Yee J, et al. Canagliflozin compared with sitagliptin for patients with type 2 diabetes who do not have adequate glycemic control with metformin plus sulfonylurea: a 52-week randomized trial. Diabetes Care. 2013;36(9):2508–2515. pmid:23564919; PubMed Central PMCID: PMC3747923.
        • 55. Takihata M, Nakamura A, Tajima K, Inazumi T, Komatsu Y, Tamura H, et al. Comparative study of sitagliptin with pioglitazone in Japanese type 2 diabetic patients: the COMPASS randomized controlled trial. Diabetes Obes Metab. 2013;15(5):455–462. pmid:23279373.
        • 56. Williams-Herman D, Johnson J, Teng R, Golm G, Kaufman KD, Goldstein BJ, et al. Efficacy and safety of sitagliptin and metformin as initial combination therapy and as monotherapy over 2 years in patients with type 2 diabetes. Diabetes Obes Metab. 2010;12(5):442–451. pmid:20415693.
        • 57. Fukui K, Kawahito H, Wakana N, Kikai M, Terada K, Yamamoto K, et al. Dipeptidyl peptidase-4 inhibitor sitagliptin improves pancreatic beta-cell function in hypertensive diabetic patients treated with angiotensin receptor blockers. J Renin Angiotensin Aldosterone Syst. 2015;16(4):1001–1009. pmid:26195265.
        • 58. Gadde KM, Vetter ML, Iqbal N, Hardy E, Ohman P, investigators D-N-s. Efficacy and safety of autoinjected exenatide once-weekly suspension versus sitagliptin or placebo with metformin in patients with type 2 diabetes: The DURATION-NEO-2 randomized clinical study. Diabetes Obes Metab. 2017;19(7):979–988. pmid:28205322; PubMed Central PMCID: PMC5485171.
        • 59. Park SE, Lee BW, Kim JH, Lee WJ, Cho JH, Jung CH, et al. Effect of gemigliptin on glycaemic variability in patients with type 2 diabetes (STABLE study). Diabetes Obes Metab. 2017;19(6):892–896. pmid:28058753.
        • 60. Pratley RE, Eldor R, Raji A, Golm G, Huyck SB, Qiu Y, et al. Ertugliflozin plus sitagliptin versus either individual agent over 52 weeks in patients with type 2 diabetes mellitus inadequately controlled with metformin: The VERTIS FACTORIAL randomized trial. Diabetes Obes Metab. 2018;20(5):1111–1120. Epub 2017/12/22. pmid:29266675; PubMed Central PMCID: PMC5947297.
        • 61. Tsurutani Y, Nakai K, Inoue K, Azuma K, Mukai S, Maruyama S, et al. Comparative study of the effects of ipragliflozin and sitagliptin on multiple metabolic variables in Japanese patients with type 2 diabetes: A multicentre, randomized, prospective, open-label, active-controlled study. Diabetes Obes Metab. 2018;20(11):2675–2679. Epub 2018/06/13. pmid:29893003.
        • 62. Weinstock RS, Guerci B, Umpierrez G, Nauck MA, Skrivanek Z, Milicevic Z. Safety and efficacy of once-weekly dulaglutide versus sitagliptin after 2 years in metformin-treated patients with type 2 diabetes (AWARD-5): a randomized, phase III study. Diabetes Obes Metab. 2015;17(9):849–858. pmid:25912221; PubMed Central PMCID: PMC5008205.
        • 63. Yokoh H, Kobayashi K, Sato Y, Takemoto M, Uchida D, Kanatsuka A, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor sitagliptin compared with alpha-glucosidase inhibitor in Japanese patients with type 2 diabetes inadequately controlled on metformin or pioglitazone alone (Study for an Ultimate Combination Therapy to Control Diabetes with Sitagliptin-1): A multicenter, randomized, open-label, non-inferiority trial. J Diabetes Investig. 2015;6(2):182–191. pmid:25802726; PubMed Central PMCID: PMC4364853.
        • 64. Shi C, Zhang R, Bai R, Liu D, Wang Y, Zhang X, et al. Efficacy and safety of sitagliptin added to metformin and insulin compared with voglibose in patients with newly diagnosed type 2 diabetes. Clinics (Sao Paulo). 2019;74:e736. Epub 2019/04/29. pmid:31038562.
        • 65. Iwamoto Y, Kashiwagi A, Yamada N, Terao S, Mimori N, Suzuki M, et al. Efficacy and safety of vildagliptin and voglibose in Japanese patients with type 2 diabetes: a 12-week, randomized, double-blind, active-controlled study. Diabetes Obes Metab. 2010;12(8):700–708. pmid:20590747; PubMed Central PMCID: PMC2916214.
        • 66. Kim G, Oh S, Jin SM, Hur KY, Kim JH, Lee MK. The efficacy and safety of adding either vildagliptin or glimepiride to ongoing metformin therapy in patients with type 2 diabetes mellitus. Expert Opin Pharmacother. 2017;18(12):1179–1186. pmid:28714741.
        • 67. Sawayama Y, Hatajima A, Koga T. Effect of Vildagliptin on Type 2 Diabetic Patients with Non-Alcoholic Fatty Liver Disease. J New Rem & Clin. 2013;62(10):1804–1812. pmid:2013356889.
        • 68. Takeshita Y, Takamura T, Kita Y, Otoda T, Kato K, Wakakuri H, et al. Vildagliptin vs liraglutide as a second-line therapy switched from sitagliptin-based regimens in patients with type 2 diabetes: A randomized, parallel-group study. J Diabetes Investig. 2015;6(2):192–200. pmid:25802727; PubMed Central PMCID: PMC4364854.
        • 69. Wu S, Gao L, Cipriani A, Huang Y, Yang Z, Yang J, et al. The effects of incretin-based therapies on beta-cell function and insulin resistance in type 2 diabetes: a systematic review and network meta-analysis combining 360 trials. Diabetes Obes Metab. 2019;21(4):975–983. pmid:30536884.
        • 70. Turner RC, Holman RR, Matthews D, Hockaday TD, Peto J. Insulin deficiency and insulin resistance interaction in diabetes: estimation of their relative contribution by feedback analysis from basal plasma insulin and glucose concentrations. Metabolism. 1979;28(11):1086–1096. pmid:386029.
        • 71. Aguilar-Bryan L, Nichols CG, Wechsler SW, Clement JPt, Boyd AE 3rd, Gonzalez G, et al. Cloning of the beta cell high-affinity sulfonylurea receptor: a regulator of insulin secretion. Science. 1995;268(5209):423–426. Epub 1995/04/21. pmid:7716547.
        • 72. prospective diabetes study U.K. 16. Overview of 6 years' therapy of type II diabetes: a progressive disease. U.K. Prospective Diabetes Study Group. Diabetes. 1995;44(11):1249–1258. Epub 1995/11/01.
        • 73. Vick H, Diedrich DF, Baumann K. Reevaluation of renal tubular glucose transport inhibition by phlorizin analogs. Am J Physiol. 1973;224(3):552–557. Epub 1973/03/01. pmid:4691268.
        • 74. Russo GT, Giorda CB, Cercone S, Nicolucci A, Cucinotta D. Factors associated with beta-cell dysfunction in type 2 diabetes: the BETADECLINE study. PLoS One. 2014;9(10):e109702. Epub 2014/10/28. pmid:25347846; PubMed Central PMCID: PMC4210056.
        • 75. Ishida H, Takizawa M, Ozawa S, Nakamichi Y, Yamaguchi S, Katsuta H, et al. Pioglitazone improves insulin secretory capacity and prevents the loss of beta-cell mass in obese diabetic db/db mice: Possible protection of beta cells from oxidative stress. Metabolism. 2004;53(4):488–494. Epub 2004/03/27. pmid:15045697.
        • 76. Masuda K, Okamoto Y, Tsuura Y, Kato S, Miura T, Tsuda K, et al. Effects of Troglitazone (CS-045) on insulin secretion in isolated rat pancreatic islets and HIT cells: an insulinotropic mechanism distinct from glibenclamide. Diabetologia. 1995;38(1):24–30. pmid:7744225.
        • 77. Gupta D, Jetton TL, Mortensen RM, Duan SZ, Peshavaria M, Leahy JL. In vivo and in vitro studies of a functional peroxisome proliferator-activated receptor gamma response element in the mouse pdx-1 promoter. J Biol Chem. 2008;283(47):32462–32470. Epub 2008/08/21. pmid:18718916.
        • 78. Nagi DK, Ali VM, Yudkin JS. Effect of metformin on intact proinsulin and des 31,32 proinsulin concentrations in subjects with non-insulin-dependent (type 2) diabetes mellitus. Diabet Med. 1996;13(8):753–757. pmid:8862952.
        • 79. Kahn SE, Halban PA. Release of incompletely processed proinsulin is the cause of the disproportionate proinsulinemia of NIDDM. Diabetes. 1997;46(11):1725–1732. Epub 1997/11/14. pmid:9356018.
        • 80. Pfützner A, Derwahl M, Jacob S, et al. Limitations of the HOMA-B score for assessment of beta-cell functionality in interventional trials-results from the PIOglim study. Diabetes Technol Ther. 2010;12(8):599‐604. pmid:20615100.
        • 81. Yoshioka N, Kuzuya T, Matsuda A, Taniguchi M, Iwamoto Y. Serum proinsulin levels at fasting and after oral glucose load in patients with type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia. 1988;31(6):355‐360. pmid:3046976.
        • 82. Sterne JAC, Egger M, Moher D. Addressing reporting biases. In: Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions. Chichester: Wiley; 2008. pp. 297–333.