It is well recognized that changes in albumin excretion are predictors of cardiovascular and renal outcomes and can be used as surrogate endpoints [1-3]. Although the focus of treatment in patients with type 2 diabetes mellitus (T2DM) has shifted to new drugs such as SGLT2 inhibitors and GLP-1 receptor agonists [4,5], metformin is still a mainstay of therapy in these patients and is recommended as first-line treatment in most present guidelines [6-8]. Many trials have shown that intensive glucose control has a favorable effect on microvascular diabetic complications such as diabetic nephropathy [9-11]. Metformin has been shown to reduce albuminuria in patients with T2DM [10,12,13], and nephroprotective effects of metformin beyond its well-known glucose-lowering effects have been described [14-17].

Metformin is effective, inexpensive and largely devoid of severe side effects [18]. However, gastrointestinal disturbances, such as diarrhea, bloating, meteorism, nausea, and abdominal pain are very frequent [18]. In most patients they subside after a few weeks or months, but in a sizable minority of the patients (typically between 10 and 25% in different trials) [18,19] they remain indefinitely, disturbing the patients’ well-being and quality of life [20], reducing adherence and compliance [18], and leading to poorer outcomes [21].

Conventional, immediate-release metformin, has a short duration of action, with a marked plasmatic peak 2.5 hours after ingestion, but since 2003 extended-release formulations of metformin are widely available, with a delayed and much less marked peak of action about 7 hours after ingestion and an effective duration of action nearing 24 hours [22-24]. The use of this extended-release metformin (XRM) in associated in several trials to a marked improvement in gastrointestinal tolerance vs. conventional metformin [24-28], even in patients previously labeled as metformin-intolerant [25]: it can improve the patients’ adherence and compliance [26-28], enhance their quality of life [28] and the achievement of glycemic targets [26-28]. Since 2005 the British NICE guidelines for treatment of type 2 diabetes in adults (NG28) recommend offering XRM to patients who complain of gastrointestinal disturbances with the use of conventional metformin [6].

However, in the Spanish market, due to economic issues, XRM is still not available as a monotherapy. Only since 2022, when the patent of sitagliptin expired, many fixed combinations of generic metformin plus sitagliptin suddenly became available, most of them based on conventional metformin but some on XRM, in pills containing 1000 mg of XRM + 50 mg of sitagliptin. Thus we were able to prescribe XRM in patients previously labeled as metformin-intolerant, provided that the use of sitagliptin was also adequate. Favorable experiences in our first patients induced us to design an open-label, investigator-promoted assay in order to assess the tolerability and effectiveness the XRM/sitagliptin single-pill combination in patients with T2DM labeled as metformin-intolerant and treated with a DPP4 inhibitor but insufficiently controlled, and if the use of this combination can result in an improvement in renal function, particularly in reduction of albumin excretion.

Methods

For this study, consecutive unselected patients were recruited except for the following inclusion and exclusion criteria:

Inclusion

?      Both sexes

?      Aged 40 - 69 years

?      Clinically diagnosed T2DM

?      Insufficient blood glucose control (HbA1c > 7%, in the last 2 available visits)

?      Gastrointestinal intolerance to metformin

?      Pharmacologic treatment of T2DM including full-dose of a DPP4 inhibitor

?      Informed consent

Exclusion

?      Clinical suspicion of T1DM

?      Very poor blood glucose control (HbA1c >9%)

?      Concomitant severe disease

?      Cardiovascular event in the 3 previous months

?      Alcoholism or addictions

?      Known poor adherence to medication or to scheduled clinical visits

?      Present or planned pregnancy

?      Stage III obesity (BMI >40 kg/m²)

?      Chronic kidney disease stage  IIIb, IV o V (estimated GFR < 45 mL/min/1.73 m²)

?      Metformin intolerance or contraindication due to non-gastrointestinal causes (allergy, etc.)

Protocol

In the first visit and after having checked the compliance with the inclusion and exclusion criteria and obtained informed consent, weight and height were measured by standard clinical procedures; blood pressure and heart rate were obtained according to the present European Society of Hypertension guidelines [29]. Fasting blood glucose, HbA1c, creatinine, lipids (total cholesterol, HDL-cholesterol, non-HDL cholesterol and triglycerides), and the albumin/creatinine ratio in a morning urine sample were measured by routine laboratory methods. The GFR was estimated by the CKD-EPI equation [30], and the risk of progression or renal disease was classified according to the 2012 KDIGO categories [31]. The health-related quality of life was estimated by the well-validated EuroQol-5D-3L questionnaire [32] which includes five dimensions (“Mobility”, “Self-Care”, “Usual Activities”, “Pain/Discomfort”, and “Anxiety/Depression”), and three levels for each dimension (good, mediocre or bad), and also an analog visual scale (AVS) or “health thermometer” [32].

The DPP4 inhibitor which the patient was taking was withdrawn and switched to the single-pill combination of XRM/sitagliptin (1000/50 mg), one pill once a day (preferably with dinner) during the first month. After one month, if the tolerance was adequate, the dose was increased to two pills (preferably taken together). In case that the patient was taking additional medication for diabetes control it was not modified, but further recommendations on lifestyle and  drug treatment for control of hypertension and dyslipidemia were given, according to the present Guidelines for Cardiovascular Prevention in Clinical Practice of the European Society of Cardiology [33].

The follow-up visit was scheduled after 3-4 months; the body weight, blood pressure and heart rate were measured again, the labs tests and the EuroQol-5D-3L questionnaire were repeated, and a simple Likert scale [34] questionnaire (selection of one of five smiley-type icons conventionally associated to “Very High”, “High”, “Indifferent”, “Low” or “Very Low” satisfaction; with the additional option of “No Opinion”) was offered. In order to avoid bias, the questionnaires were never administered to the patients by their responsible physician.  Data on medication compliance and tolerance were obtained by an ad-hoc questionnaire.

Statistical Analysis

The normality of the studied parametric variables was assessed by the Kolmogorov-Smirnov test. Those normally distributed (Gaussian) were compared by the Student’s paired t-test, and the non-Gaussian by the Mann-Whitney’s U-test. In order to compare non-parametric variables in 2 x 2 tables, the Fischer’s exact test was used, and the chi-square test in larger tables. For bivariate correlation Pearson’s R was used, and for multivariate correlation analysis, a backwards logistic step-by-step analysis was performed. All analyses were done by intention to treat, except where specifically indicated. Calculations were performed by the IBM SPSS 28.0.1.1 for MS-Windows software package. For normally distributed parametric variables, values are given as mean ± standard deviation; median (interquartile range) is used for non-normal ones.

Ethical aspects

The present study has been performed according to the local applicable regulations (European Community Directive 2001/83/EC) and with the ethical principles established by the Helsinki Declaration, updated in Fortaleza, Brazil (2013) [35]. It was submitted for approval in the Ethics Institutional Committee of the Hospital Universitario San Roque of Las Palmas de Gran Canaria, Spain, and for registry in the webpage htpps://clinicaltrials.gov.

Results

A total 72 patients were recruited, 45 (62%) were women, aged 55 ± 8 years, and the known duration of T2DM was 7.5 ± 3.2 years. The main data on the patients at baseline and at the follow-up visit are shown in Table 1. All the analyzed parametric variables were normally distributed except for plasma triglycerides and the albumin/creatinine quotient.

Of 72 patients, 60 (83%) tolerated 1 pill of XRM/sitagliptin 1000/50 mg, although 17 of them (28%) reported mild gastrointestinal symptoms that subsided along the first months and did not cause withdrawal. 52 (72%) of the patients tolerated 2 pills, although 11 of them (21%) reported mild gastrointestinal symptoms after the dose increase. 12 patients (17%) did not tolerate the combination.

Of the 20 patients that did not tolerate 1 or 2 pills, 14 (19.4%) reported diarrhea, 7 (9.7%) flatulence, 4 (5.6%) nausea, 2 (2.8%) dyspepsia, y 1 (1.4%) abdominal pain.

In the 8 patients who tolerated only 1 pill of XRM/sitagliptin, the self-reported compliance by questionnaire was 89.2 ± 6.3% and never < 75%. In the 52 patients who tolerated 2 pills, compliance was 82.3 ± 7.9%, and < 75% in only one patient who reported 72.5% compliance and was not excluded from analysis.

There were no changes in the weight or BMI of the patients. Fasting glycemia dropped from 175 ± 34 mg/dL to 129 ± 23 mg/dL (p < 0.05) in the patients who tolerated 1 pill, and to 121 ± 21 (p < 0.01) in those who tolerated 2 pills. HbA1c dropped from 8.3 ± 1.0% to 7.7 ± 0.8% (p < 0.01) with one pill, and to 7.4 ± 0.7% (p < 0.01) with 2 pills. Data on lipid profile, blood pressure and heart rate of the patients were similar in patients who tolerated 1, 2 or no pills and are shown in Table 1.

There were no significant changes in the estimated GFR (CKD-EPI, 57±16 mL/min/1.73 m² in the first visit and 59±16 mL/min/1.73 m² in the follow-up visit), but the albumin/creatinine quotient (morning urine sample) dropped from 39 (17-78) mg/g to 29 (14-66) mg/g (p = 0.029) (Figure 1). In none of the 7 patients who had a baseline quotient ? 300 mg/g did it drop to < 300 mg/g, but in 3 of 11 patients (27.3%) with a baseline quotient ? 30 mg/g it dropped to < 30 mg/g (regression to normal albuminuria). No patient had an increase in the quotient from < 30 mg/g to ? 30 mg/g, or from < 300 mg/g to ? 300 mg/g. No patients had their estimated GFR reduced to < 45 mL/min/1.73 m² in the follow-up visit. The distribution of the risk of renal disease progression (2012 KDIGO categories) was: low in 27 patients (38%), moderate in 31 (43%), high in 9 (12%) and very high in 5 (7%) at baseline, which changed to low in 29 patients (40%), moderate in 30 (42%), high in 8 (11%) and very high in 5 (7%) at the follow-up visit. Although there seemed to be a favorable trend, this change was not statistically significant (p = 0.147).