This work aims to determine the real-life anthropometric and analytical benefits of adding subcutaneous semaglutide to previous insulin treatment in patients with type 2 diabetes.
MethodsThis is a descriptive, retrospective, open-label study describing the clinical and anthropometric characteristics of 117 patients diagnosed with type 2 diabetes followed-up on in the Endocrinology and Nutrition outpatient clinic of the Hospital Universitario Central de Asturias for 53 weeks after starting treatment with subcutaneous semaglutide (October–December 2019). All patients were on previous insulin treatment with or without oral antidiabetics.
ResultsOf the 117 initial patients, 17 did not complete the study due to adverse effects (nausea, vomiting), the physician’s decision, or loss to follow-up. Twelve months (week 53) after starting semaglutide, there was a decrease in HbA1c of 0.74% (95% CI 0.59–1.14, p < 0.05) as well as 3.61 kg of weight loss (95% CI 2.30–4.92, p < 0.05) and a decline in total insulin of 15.88 IU (95% CI 10.98–20.74, p < 0.05) from baseline figures. In patients without prior GLP-1 receptor analogs (GLP-1ra), the effect in terms of a reduction in HbA1c, weight, and the total insulin dose was statistically significant. However, in patients pre-treated with GLP-1ra only had improvements in terms of weight loss. No serious adverse events were observed.
ConclusionsThe addition of subcutaneous semaglutide to prior insulin treatment with or without oral antidiabetics safely led to a decrease in HbA1c, weight, and the insulin dose. This effect is greater in GLP-1ra naive patients.
Determinar en la vida real los beneficios antropométricos y analíticos de la adición de semaglutida por vía subcutánea al tratamiento previo con insulina en pacientes con diabetes tipo 2.
MétodosEstudio descriptivo, retrospectivo y abierto en el que se describen características clínicas y antropométricas de 117 pacientes diagnosticados de diabetes tipo 2 seguidos en las consultas externas de Endocrinología y Nutrición del Hospital Universitario Central de Asturias a lo largo de 53 semanas tras el inicio de tratamiento con semaglutida subcutánea (octubre–diciembre 2019). Todos los pacientes estaban en tratamiento previo con insulina, con o sin antidiabéticos orales.
ResultadosDe los 117 pacientes iniciales, 17 no completaron el estudio debido a efectos adversos (náuseas, vómitos), decisión clínica y pérdida de seguimiento. A los 12 meses (semana 53) del inicio de la semaglutida se obtuvo un descenso de HbA1c de 0,74% (IC 95% 0,59–1,14, p < 0,05), así como de 3,61 kg de peso (IC 95% 2,30–4,92, p < 0,05), y de 15,88 UI de insulina total (IC 95% 10,98–20,74, p < 0,05) respecto a las cifras basales. En pacientes sin análogo del receptor de GLP-1 (arGLP-1) previo, el efecto en la disminución de HbA1c, el peso y la dosis total de insulina fue estadísticamente significativo; sin embargo, los pacientes pretratados con arGLP-1 solo tuvieron mejoría en la reducción de peso. No se observaron eventos adversos graves.
ConclusionesLa adición de semaglutida subcutánea al tratamiento previo con insulina con o sin antidiabéticos orales induce una disminución de HbA1c, peso y dosis de insulina de forma segura. Este efecto es mayor en pacientes naïve para tratamiento con arGLP-1.
Type 2 diabetes mellitus (DM2) is a slowly progressive disease. On occasion, due to the gradual decline in pancreatic beta cell function, it is necessary to intensify the initial treatment to maintain adequate glycemic control.
Treatment with insulin (initially mostly basal) is frequently used following the failure of several oral lipid-lowering drugs maintain HbA1c at target levels1. However, data published in various randomized clinical trials (RCTs)2,3 and cohort studies4,5 show that a significant proportion of patients who receive insulin therapy do not reach glycemic control targets either.
When basal insulin requirements exceed 0.5 IU/kg of weight, subsequent dose increases do not result in an improvement in glycemic control, but do result in an increase in weight and hypoglycemic episodes6. At present, when faced with this situation, the American Diabetes Association recommends associating treatment with a GLP-1 receptor agonist (GLP-1ra) in patients in which atherosclerotic cardiovascular disease and/or obesity are predominant1.
The benefit of combined insulin and GLP-1ra treatment lies in the sum of their components’ synergic and complementary effects7,8.
GLP-1ras improve insulin secretion in a glucose-dependent manner, suppress glucagon secretion, act on post-prandial blood glucose by slowing stomach emptying, and promote weight loss. All of this improves glycemic control with a low risk of hypoglycemia and weight gain7,8.
Semaglutide (Novo Nordisk A/S, Denmark) is a GLP-1ra. To improve glycemic control in adults with DM2, in addition to diet and exercise, semaglutide can be administered in both subcutaneous (SC) formulations once per week9 or oral formulations once per day10.
In the SUSTAIN clinical trials, semaglutide SC was demonstrated to be superior by reducing HbA1c and body weight compared to the placebo and a wide variety of active comparators, including SGLT-2is, other GLP-1ras, and insulin glargine, with a safety profile similar to that of other GLP-1ras11–16.
However, the population represented in RCTs is not the same as patients in clinical practice, given that the former must meet strict inclusion and exclusion criteria. Real-world studies aim to verify the findings of these RCTs and are important for knowing the efficacy and safety of the drug in clinical practice settings.
Various studies have compared the combination of a GLP-1ra and insulin to an intensification of the insulin regimen17,18. Meta-analyses indicate that glycemic control is similar or better with the addition of GLP-1ras with benefits over insulin intensification, including a reduction in body weight and lower incidence of hypoglycemic episodes.
ObjectivesThe purpose of this study is to determine the benefits in clinical practice of starting semaglutide SC in patients who have previously been treated with subcutaneous insulin.
The primary objective was the change in HbA1c values at six and 12 months from the baseline value (%-point). Secondary aims included the change in body weight (kg), total insulin dose (IU), and insulin dose per weight (IU/kg) from baseline to the two time points.
In addition, the study seeks to confirm the drug’s safety in this subgroup of the population with diabetes by determining the rate of treatment suspensions and patient dropouts as well as severe hypoglycemic episodes.
Materials and methodsStudy design and locationThis retrospective, open-label, 53-week, observational study evaluated the real-world clinical repercussions of the addition of weekly semaglutide SC in adults with DM2 previously treated with insulin followed-up on in the Endocrinology and Nutrition Department outpatient clinic of the Central University Hospital of Asturias.
Study populationInclusion criteria: patients ≥18 years with DM2 with more 12 weeks since onset who were already in treatment with insulin (basal, basal-bolus, or premixed).
Exclusion criteria: treatment with any investigational drug (a drug not currently approved for treatment) within 90 days prior to inclusion, hypersensitivity to semaglutide or any of its excipients, and an estimated glomerular filtration rate <15 mL/min/1.73 m2 on the most recent blood test available.
The decision to start treatment with semaglutide SC, prescribe other hypoglycemic treatments, or prescribe a diet and exercise regimen was the responsibility of each physician.
All DPP4 inhibitor drug treatment was suspended when semaglutide was started.
The recruitment period was from October to December 2019. The data (up to January 2021) were retrospectively collected by the physicians.
The patients started treatment with semaglutide SC in a preloaded pen according to routine clinical practice. The medication was provided by the sponsor. The physician determined the initial dose, the escalation dose, and the maintenance dose as well as any later changes to the maintenance dose.
After visit 1 (0 weeks), the patients attended visits only if appropriate according to routine clinical practice (at six months and at one year from the start). Given that the study took place during the COVID-19 pandemic period, body measurement data are not available for all initially recruited individuals, given that many of the intermediate visits were online. Nevertheless, we consider glycated hemoglobin values as the primordial data set as they report on the principal objective (glycemic control). Patients (n = 17) who did not have follow-up blood tests performed between weeks 28 and 38 were considered lost to follow-up.
The main study variables were quantitative: HbA1c was measured using high-performance liquid chromatography (HPLC) (Jokoh HS-10 Analyzer). The initial values were standardizing according to IFCC criteria (JDS/JSCC HbA1c = 0.927 [IFCC HbA1c] + 1.73). Height, weight, and BMI (weight in kg divided by height in meters squared) were measured with the subject wearing light clothing and without shoes. Data on insulin units were gathered by the physician during the medical consultation according to data provided by the patient and subsequent prescriptions.
Statistical analysisWe used descriptive statistics (mean ± standard deviation (SD) for quantitative variables and percentages for categorical variables) to describe patients’ characteristics upon starting semaglutide.
The baseline characteristics were analyzed in the analysis of the total group, which included all patients who started treatment with semaglutide (n = 117).
The primary objective (change in HbA1c) was calculated with 95% confidence intervals and two-tailed p values. The secondary and exploratory objectives were calculated in the same manner as the primary objective.
The population for the outcomes analysis was individuals who started treatment with semaglutide and who had data on HbA1c (n = 100), weight (n = 55), and insulin units (n = 87) from the end of the study.
The repeated measures ANOVA comparison of means test was used for the statistical analysis, performed using the GraphPad program.
A subgroup analysis was performed based on whether the patient was already in treatment with another GLP-1ra or if it was started at that time (GLP-1ra naive vs. GLP-1ra experienced).
ResultsPatients’ baseline characteristicsA total of 117 patients were evaluated in the study period, of which 100 (85%) completed follow-up to week 53 (Table 1). The reasons the 17 patients did not complete follow-up were: lack of adherence to medical visits (n = 9), adverse effects (n = 6), and clinical decision (n = 2). The adverse effects which led to the suspension of the drug were gastrointestinal in nature (nausea and vomiting). The mean age of patients was 64.4 ± 9.4 years and 57% were male.
Patients’ baseline characteristics.
| Total | GLP-1ra naive | GLP-1ra experienced | |
|---|---|---|---|
| N (%) | 117 | 80 (69) | 37 (32) |
| Time since onset (years) | 16.64 ± 7.99 | 16.08 ± 8.52 | 17.86 ± 6.65 |
| Age (years) | 64.23 ± 9.43 | 63.51 ± 9.47 | 65.78 ± 9.30 |
| No. of OAD | 1.62 ± 1.02 | 1.63 ± 1.09 | 1.62 ± 0.86 |
| SGLT2i (%) | 54 (46) | 33 (41) | 21 (57) |
| Weight (kg) | 97.28 ± 18.43 | 95.54 ± 19.21 | 100.95 ± 16.34 |
| HbA1c (%) | 8.31 ± 1.33 | 8.39 ± 1.33 | 8.13 ± 1.32 |
| HbA1c <7% | 17 (15) | 8 (10) | 9 (24) |
| Total insulin dose (IU) | 63.78 ± 39.69 | 63.00 ± 38.64 | 65.46 ± 42.36 |
| Basal dose (IU) | 44.29 ± 21.33 | 42.85 ± 20.23 | 47.44 ± 23.52 |
| Prandial dose (IU) | 25.69 ± 19.49 | 20.15 ± 25.27 | 18.05 ± 25.55 |
| IU insulin/kg weight | 0.67 ± 0.44 | 0.69 ± 0.46 | 0.64 ± 0.39 |
Data are expressed as frequencies (percentages) or means ± SD.
IU: insulin dose; OAD: oral antidiabetics; SGLT2i: SGLT2 inhibitors.
The mean baseline HbA1c was 8.31%, the mean duration of diabetes was 16.7 years, and the mean body weight was 97.3 kg.
Patients were prescribed an initial dose of 0.25 mg of semaglutide weekly. At six months from starting treatment, 99 (85%) patients were on a dose of 0.5 mg/week, 11 (9%) were on a dose of 1 mg/week, and 7 (6%) continued with the initial dose of 0.25 mg/week. At the end of follow-up, 55 (47%) patients received a dose of 1 mg/week, 39 (33%) patients received 0.5 mg/week, and 6 (5%) maintained the dose of 0.25 mg/week. At 12 months from starting medical treatment, 6 (6%) continued with the initial dose of 0.25 mg/week, 39 (39%) were on 0.5 mg/week, and 55 (55%) had progressed to the maximum dose of 1 mg/week.
Before starting treatment with semaglutide, 37 (32%) patients were in treatment with another GLP-1ra: 17 patients with liraglutide, 17 with dulaglutide, and 7 with exenatide or lixisenatide.
The mean number of oral antidiabetics (OAD)/patient/day was 1.6 ± 1.02. A total of 20 patients (17%) were not taking any OAD, 26 patients (23%) took 1 OAD, 45 (39%) took 2 OAD, 23 (20%) took 3 OAD, and 2 (1%) took 4 OAD. At the start of the study, 91 (78%) patients were taking metformin.
In regard to the insulin regimen prior to starting semaglutide, 49 (42%) patients were only taking basal insulin, 42 (36%) were taking basal-bolus insulin, and 26 (22%) were taking premixed insulin.
HbA1cOne hundred of the initial 117 patients completed the study (data at baseline, at six months, and at one year). At six months from starting weekly semaglutide SC, the mean change in HbA1c was –0.86% (95% CI: –0.46; –1.15; p < 0.0001) (Fig. 1). At 12 months, the decrease in HbA1c was maintained. Therefore, the mean estimated change compared to the baseline figure was –0.74% (95% CI: –0.59; –1.14; p < 0.0001). There were no statistically significant differences between data at six and 12 months.
In the subgroup of GLP-1ra naive patients, the change in HbA1c from the value at baseline to the value at 12 months was slightly less (–0.94%; 95% CI: –1.36; –0.53; p < 0.0001) than the change in the value at six months (–1.21%). Regarding the GLP-1ra experienced group, the difference of –0.44% (95% CI: –0.90; 0.02; p = 0.0616) at 12 months compared to the initial value was not statistically significant; the value at six months was not significant either.
WeightThe mean baseline weight was 97.3 ± 18.43 kg. Only 55 (47.4%) patients have data available from the study’s three time points. In general, the mean change in body weight at six months compared to the initial value was –5.33 kg (95% CI: –4.25; –6.42; p < 0.0001). The value at 12 months was –3.61 kg (95% CI: –2.30; –4.92; p < 0.0001) (Fig. 2).
In the subgroup analysis, the initial baseline weight in the GLP-1ra naive group was 95.54 kg (SD 19.21). At six months from starting semaglutide, a mean decrease of –5.68 kg was achieved (95% CI: –4.57; –7.18; p < 0.0001) and at 12 months, a mean decrease of –4.54 kg was achieved (95% CI: –2.80; –6.28; p < 0.0001). However, the mean baseline weight was higher in the GLP-1ra experienced group: 100.95 kg (SD 16.34). At six months, a reduction of –1.01 kg was achieved (95% CI: –1.21; 3.24; NS) and at 12 months, a reduction of –2.70 kg was achieved (95% CI: –0.80; –4.60; p = 0.004).
Total insulin doseThe initial total insulin dose was 63.78 ± 39.69 IU. Data are available on 87 (75%) patients throughout the study (Fig. 3). Compared to the initial value, the mean change in the total population at six months was –15.18 IU (95% CI: –10.67; –19.68; p < 0.0001) and at 12 months was –15.88 IU (95% CI: –10.98; –20.74; p < 0.0001).
In the subgroup analysis, the initial total dose of insulin in the GLP-1ra naive group was 63.00 (SD 38.64). At six months from starting semaglutide, a mean reduction of –18.06 IU was achieved (95% CI: –12.33; –23.78; p < 0.0001) and at 12 months, a mean reduction of 22.21 IU was achieved (95% CI: –9.32; –35.10; p < 0.0001).
On the contrary, this decrease was less noteworthy in the GLP-1ra experienced subgroup. The total initial dose was 65.46 IU (SD 42.36). At six months from starting semaglutide, a mean reduction of 8.99 IU was achieved (95% CI: –2.21; 15.77; p = 0.008) and at 12 months, a mean reduction of 4.79 IU was achieved (95% CI: –2.76; 12.24; p = 0.2935).
Units of insulin/kg of weightOnly 52 (50%) of the 117 patients have data available on IU/kg of weight from the three evaluations. The mean initial figure was 0.67 ± 0.44 IU/kg. The mean change at six months compared to the initial value was –0.14 IU/kg (95% CI: –0.09; –0.20; p < 0.0001). At 12 months from the start of the study, a mean reduction of –0.15 IU/kg was achieved (95% CI: –0.10; –0.21; p < 0.0001).
In the subgroup analysis, the initial figure of IU/kg of weight in the GLP-1ra naive group (n = 35) was 0.69 IU/kg (SD 0.46). At six months, a reduction of –0.19 IU/kg was achieved (95% CI: –0.12; –0.25; p < 0.0001) and at 12 months, this reduction was maintained and increased, reaching –0.23 IU/kg (95% CI: –0.16; –0.30; p < 0.0001).
In the GLP-1ra experienced subgroup, the initial mean dose was 0.64 IU/kg of weight (n = 17). At six months, a reduction of just –0.01 IU/kg was achieved (95% CI: –0.06; 0.08; p > 0.9999) and at 12 months, a reduction of 0.03 IU was achieved (95% CI: –0.06; 0.08; p = 0.4051) (Table 2).
No patients reported grade 3 or severe hypoglycemic events during the course of this study. The number of mild or moderate hypoglycemic events was not quantified given that they were not considered clinically relevant.
DiscussionThis work has shown that weekly subcutaneous semaglutide in patients treated with insulin produces a significant reduction in HbA1c, weight, and insulin needs. This reduction was greater in patients who had not previously been treated with another GLP-1ra. The drug was well tolerated and just six (5%) patients needed to withdraw from the study due to adverse effects.
Compared to the Canadian SURE study19, published with real-world data on semaglutide added to insulin, the reduction in HbA1c was very similar (–0.86% vs. –0.90%), as was the reduction in body weight (–5.33 kg vs. –4.5 kg). In addition, patients started from an initial weight that was also similar (99.2 kg SURE vs. 97.29 kg in our study). It is important to note that all patients in this study were in treatment with insulin (during and prior to the study), but only 41% of patients were treated with insulin in the SURE study (other subgroups were only treated with oral hypoglycemic agents).
If we compare our data to those obtained in other non-real-world studies17,18, the conclusion is the same: combined insulin and semaglutide treatment significantly improved glycemic control in terms of a reduction in HbA1c and the number of patients within the target (HbA1c <7%), with associated weight loss.
Overall, our results support the role of semaglutide combined with insulin as a valuable intensification strategy for producing a clinical benefit in the management of DM2. Along these lines, the updated consensus document from the American Diabetes Association and the European Association for the Study of Diabetes (EASD) suggests intensifying treatment with GLP-1ra, among other drugs including SGLT-2is or prandial insulin, in patients who cannot meet glycemic targets at the start of insulin treatment in combination with oral medication1.
Due to the disease’s natural history, many patients with DM2 need insulin to reach their glycemic control target. A common approach for starting insulin therapy is a single injection of basal insulin, which is eventually titrated to optimize treatment1. If necessary, treatment intensification to reach the glycemic target can include the addition of one or more rapid-acting insulin doses before meals (basal-plus or basal-bolus regimen) or a change to a premixed insulin regimen in which there is a fixed amount of intermediate-acting or short- or rapid-acting insulin.
Although these regimens lead to a similar decrease in HbA1c levels20, there is a certain degree of reluctance to intensifying insulin therapy due to possible adverse effects (i.e., an increase in the number of hypoglycemic events and weight gain) and practical issues (i.e., difficulties in educating patients on injection therapy). In this context, the addition of semaglutide to insulin can be a valid therapeutic option for improving glycemic control and minimizing adverse events associated with insulin intensification.
The subgroup analysis shows that the patients who benefited most from treatment with semaglutide are those not previously treated with another GLP-1ra, i.e., those who had not been treated with another drug from the same family. Those who had previously been treated with another GLP-1ra had favorable results, especially in terms of weight loss, but not to the same extent in terms of glycemic control.
In conclusion, combined insulin and semaglutide therapy is a valuable treatment strategy for improving metabolic control in the management of DM221.
FundingThe authors declare that they have not received funding for conducting this study.
Conflicts of interestThe authors declare no conflicts of interest.
We would like to thank the entire Endocrinology and Nutrition Department of the Central University Hospital of Asturias, including physicians and nursing department, administrative, and auxiliary staff, for their collaboration.
Please cite this article as: Ares-Blanco J, Pujante-Alarcón P, Lambert C, Morales-Sánchez P, Delgado-Álvarez E, Menéndez-Torre EL. Efectos en vida real de la adición de semaglutida subcutánea semanal al tratamiento con insulina en diabetes mellitus tipo 2. Rev Clin Esp. 2022;222:516–522.
