2022 年 69 巻 8 号 p. 959-969
Recent studies have found compared with insulin glargine (IGlar), insulin degludec/aspart (IDeg/Asp) may provide adequate glycemic control and prevent hypoglycemia events in type 2 diabetes mellitus (T2DM). Consequently, we performed a meta-analysis to appraise and compare the efficiency and safety of IDeg/Asp and IGlar in the treatment of T2DM. We sought the databases including PubMed, Embase, Scopus, Cochrane library to confirm related articles which inspected the effect of IDeg/Asp versus IGlar for the treatment of T2DM until May 2021. Finally, six randomized controlled trials (RCTs) of 1,346 patients were included. The results showed that IDeg/Asp significantly decreased the mean hemoglobin A1c (HbA1c) level but was prone to serious adverse events, and IGlar increased the nocturnal confirmed hypoglycemia events. Besides, there were no significant changes in other indicators, including mean fasting plasma glucose (FPG) level, nine-point self-measured plasma glucose (SMPG) level, and adverse events. What’s more, we found that there was no significant difference in the occurrence of hypoglycemia overall, but our subgroup analysis of confirmed hypoglycemia revealed the population in this subgroup (duration of diabetes ≤11 years) might has its particularity effecting the hypoglycemia outcome. Concerning efficiency, IDeg/Asp may have advantages in controlling the mean HbA1c level. Regarding safety, IGlar might increase the risk of nocturnal confirmed hypoglycemia. Further evidence is needed to compare better the efficiency and safety of IDeg/Asp versus IGlar therapy.
DIABETES is a chronic metabolic disease with an increasing prevalence worldwide and now becomes a severe global health issue due to the high-carbohydrate diet and obesity [1]. The International Diabetes Federation (IDF) conservatively evaluated that in 2015, 12% of global health expenditure was spent on the treatment of diabetes and its related complications [2]. In addition, IDF has estimated that by 2040 the total number of people with diabetes worldwide will reach 642 million [3].
The treatment of type 2 diabetes mellitus (T2DM) includes lifestyle intervention, blood glucose monitoring, diabetes education, and hypoglycemic drugs therapy (oral hypoglycemic drugs, insulin therapy et al.) [4]. It is worth noting that insulin is not required for T2DM to maintain life, while it may be the most effective measure to control blood glucose at certain times, especially during long courses of disease [5-7]. In clinical practice, basal insulin glargine improves fasting plasma glucose (FPG) and hemoglobin A1c (HbA1c) levels. Basal insulin alone is the most expedient initial insulin regimen, which can be added to metformin and other oral medications. The starting dose can be estimated based on blood glucose levels, the degree of hyperglycemia, and individualized titrations over a range of days to weeks, as needed [8]. Insulin glargine (a basal insulin) provides a 22–24 hour duration and a ‘no peak’ concentrations, commonly used as the beginning of insulin therapy for many diabetics. Insulin degludec/aspart (IDeg/Asp) covers basal and meal-time insulin control, and it’s an alternative for patients to have a convenient and flexible regimen. A previous multinational consensus on the usage of IDeg/Asp recommended that IDeg/Asp for insulin initiation suits for the circumstance of symptoms of hyperglycemia, high carbohydrate diet, high HbA1c, high postprandial excursion and so on [9]. Nevertheless, compared with insulin degludec/aspart (IDeg/Asp), insulin glargine (IGlar) is associated with a potential risk of hypoglycemia. IDeg/Asp combines IDeg and IAsp (7:3), which provides effective glycemic control, reduces hypoglycemia rate (particularly nocturnal hypoglycemia), allows a flexible injection time [10]. As the first co-formulation of long-acting and short-acting insulin, IDeg/Asp retains its unique pharmacokinetic (PK) /pharmacodynamic (PD) properties [11]. Considering the potential of IDeg/Asp to facilitate patients’ adherence to therapy, provide adequate glycemic control, and prevents hypoglycemia events, further researches for evaluating the efficacy and safety of IDeg/Asp are needed.
In this context, the purpose of this randomized controlled trial meta-analysis was to appraise the efficiency and safety of the IDeg/Asp combination compared with basal insulin glargine in T2DM.
Two researchers independently retrieved the databases including PubMed, Embase, Scopus, Cochrane library for studies by using the combination of Mesh words and Entry Terms and selected the eligible articles on the basis of the inclusion and exclusion criteria by May 16, 2021. Moreover, the reference list that seemed to accord with the inclusion criteria was manually retrieved to determine whether it could be included. The Mesh words included ‘Diabetes Mellitus’, ‘insulin degludec, insulin aspart drug combination’, ‘Insulin Aspart Injection’, ‘Insulin Glargine’, ‘clinical trials’.
Inclusion and exclusion criteriaThe inclusion criteria were as below: (1) Randomized controlled trials (RCTs); (2) IDeg/Asp intervention, IGlar (100 units/mL or 300 units/mL) control; (3) Type 2 diabetes participants (adults with an HbA1c of 7–11%); (4) The sample size of the study shall not be less than 20.
The exclusion criteria were as below: (1) Non-randomized controlled trials (NRCTs); (2) Duplicate articles, individual case reports, and studies with no available data; (3) The efficacy of glycemic control unable to determine from the trials; (4) Diabetes with other diseases, such as type 1 diabetes, secondary diabetes cardiac disease, impaired liver or severe renal disease.
Data collectionTwo reviewers worked on the data collection process on their own. The necessary data and information were extracted and organized into tables by using Microsoft Excel. The following data and information were extracted: (1) Efficiency assessments: mean HbA1c end-of-trail, mean FPG end-of-trial, nine-point SMPG; (2) Safety assessments: confirmed hypoglycemia, nocturnal confirmed hypoglycemia, adverse events (AEs), and serious adverse events; (3) Other variables: participant, gender, country, age, weight, body mass index (BMI), duration of diabetes, intervention characteristics, sources of funding.
If there were any discrepancies in the data or disagreements between the two investigators, the third-party investigator would check the study and decide the general policy after proper discussion.
Missing dataIf any necessary data were missing, we contacted the corresponding authors or sponsors for the missing data or information via email. When the average change in the continuous data was not reported with a standard deviation (SD), but 95% confidential interval (95% CI) or p-value of both sets were reported simultaneously, these data were included in the meta-analysis by conversion. When binary data were not available, a sensitivity analysis was conducted for evaluation [12].
Quality assessmentThe quality of the inclusive articles was evaluated by two researchers respectively in conformity with the Cochrane risk of bias tool and Jadad scale. The Cochrane risk of bias tool was applied to assess the risk, including selection bias, performance bias, detection bias, attrition bias, and reporting bias as high, unclear, or low. Review Manager V.5.4 was used to appraise the risk of bias. Furthermore, we used the Jadad scale to appraise the methodological quality of each trail in line with the description of randomization, blinding, and dropouts, accounting for 2 points, 2 points, and 1 point, respectively. Finally, studies that scored 3 points or more were considered high-quality trials. Any divergences generated from this process were passed on to the third investigator, obtaining a consensus.
Data synthesisThe data of designated studies were analyzed by Review Manager V.5.4. We computed the mean difference (MD) with a 95% CI and obtained the means and SD from continuous results, and random-effects models were used to measure these results. The I2 value revealed the heterogeneity of all statistical tests in the forest plots, which was accounted for as follows: 0–40%: exhibiting low heterogeneity; 50–70%: indicating moderate heterogeneity, and >70%: showing significant heterogeneity [13]. Galbraith’s diagram reveals that the heterogeneity created by the circle beyond the lower and upper lines. To comprehend the effect on the results, sensitivity analysis was performed by excluding one literature to evaluate the effect on the result. STATA V.14.0 was applied to subsequent operations. Visual funnel plots and Egger’s test were used to assess the probability of publication bias, and the two-tailed p-value <0.05 was indicated statistically significant. Moreover, this meta-analysis was conducted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses PRISMA guidelines.
Totally, 511 studies were selected from PubMed (n = 47), Embase (n = 44), Cochrane Library (n = 108), Scopus (n = 161) and the other resources (n = 151). Two hundred forty-eight references remained after removing duplicate allusions. Through the primary inspecting of titles and abstracts, we excluded 227 articles, including case report (n = 14), reviews (n = 39), irrelevant interventions (n = 99), lack of comparison (n = 31), NRCTs (n = 35) and empirical perspective (n = 9). After remained 21 articles, we supplementary removed 15 articles, consisting of NRCT (n = 1), irrelevant interventions (n = 3), not original published (n = 3), conference abstract (n = 1), registered trial (n = 4), and three that had no reported outcome of interest. Finally, six references were retained. The specific selecting process flowchart was shown in the following Fig. 1.
The flowchart of selection process
In the six clinical trials of 12 to 52 weeks’ time range (mean = 24.7 weeks, median = 21 weeks), 1346 diabetes patients were randomized to IDeg/Asp group (n = 658) or IGlar group (n = 688). The trial population enrolled were insulin-Naïve people with T2DM, except one trial in Kumar 2017, and all studies accepted IDeg/Asp or IGlar once-daily, and four articles were combined with metformin or other medicine. The mean age varied from 49.0 years to 64.0 years, and the male/female ratio varied from 91/101 to 40/15. Moreover, the mean duration of diabetes ranged from 5.0 years to 12.4 years, and the mean BMI varied from 25.0 kg/m2 to 30.9 kg/m2. Furthermore, the mean HbA1c in the six trials ranged from 7.0% to 9.7%, and the mean FPG ranged from 7.8 mmol/L to 12.1 mmol/L. The main characteristic of each study was summarized in the following Table 1.
| Trials | Country | NCT No. | Patients | Time Range (weeks) | Sample size | Intervention measure | Injection time | Age (Years) | Gender (Male/Female) | Duration of diabetes (Years) | BMI (Kg/m2) | HbA1C ( %) | FPG (mmol/L) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IGlar | IDegAsp | Total | IGlar | IDegAsp | IGlar | IDegAsp | IGlar | IDeg/Asp | IGlar | IDeg/Asp | IGlar | IDeg/Asp | IGlar | IDeg/Asp | IGlar | IDeg/Asp | IGlar | IDeg/Asp | |||||
| Heise (2011) [14] | Europe | NCT00614055 | T2D (insulin naïve) | 16 | 55 | 55 | 110 | once-daily* | once-daily* | before the evening meal | before the evening meal | 58.4 (±8.4) | 58.7 (±8.8) | 40/15 | 35/20 | 8.5 (±4.8) | 9.1 (±8.0) | 30.5 (±3.5) | 30.2 (±3.4) | 8.4 (±1.3) | 8.3 (±1.2) | 12.1 (±3.5) | 11.1 (±3.3) |
| Kumar (2016) [15] | Countrya | NCT01169766 | T2D (insulin naïve) | 52 | 209 | 179 | 388 | once-daily* | once-daily* | either with breakfast or with the largest meal of the day | either with breakfast or with the largest meal of the day | 56.5 (±8.6) | 57.4 (±9.2) | 118/103 | 91/101 | 9.6 (±6.1) | 8.7 (±6.2) | 30.4 (±5.2) | 30.9 (±4.9) | 8.9 (±1.0) | 8.9 (±1.0) | 10.2 (±2.8) | 10.2 (±2.7) |
| Kumar (2017) [16] | Countryb | NCT01045447 | T2D | 26 | 205 | 196 | 401 | once-daily# | once- daily# |
the largest meal of the day | the largest meal of the day | 58.4 (±10.1) | 57.8 (±9.5) | 127/106 | 135/95 | 11.4 (±7.3) | 11.6 (±6.8) | 30.1 (±5.3) | 30.1 (±5.1) | 8.4 (±1.0) | 8.3 (±0.8) | 7.8 (±2.8) | 8.0 (±2.5) |
| Liebl (2013) [17] | Europe | NA | T2D (insulin naïve) | 16 | 57 | 55 | 112 | once-daily* | once-daily* | before the evening meal | before the evening meal | — | — | 36/21 | 40/15 | — | — | 25–37 (–) | 25–37 (–) | 7–11 (–) | 7–11 (–) | — | — |
| Onishi (2013) [18] | Japan | NCT01272193 | T2D (insulin naïve) | 26 | 149 | 147 | 296 | once-daily | once- daily |
either before breakfast or at bedtime at the discretion of each subject | before the largest meal of the day | 61 (±9.6) | 60 (±10.0) | 99/50 | 90/57 | 12.4 (±8.6) | 10.9 (±7.3) | 25.0 (±3.8) | 25.2 (±3.8) | 8.5 (±0.8) | 8.3 (±0.8) | 9.1 (±1.9) | 9.0 (±1.6) |
| Seiya (2019) [19] | Japan | NA | T2D (insulin naïve) | 12 | 13 | 26 | 39 | once-daily | once- daily |
before breakfast or dinner | before breakfast or dinner | 49 (–) | 64 (–) | 7/6 | 17/9 | 5.0 (–) | 11.0 (–) | 25.9 (–) | 26.0 (–) | 9.7 (–) | 8.9 (–) | — | — |
Countrya: Austria, India, Korea, Poland, Russia, Spain, Turkey, and America; Countryb: includes Croatia, France, India, Poland, South Africa, South Korea, Sweden, Turkey, and America; T2D, Type 2 diabetes; Values are mean (±SD); (—) devotes unclear; (*) means all in combination with metformin; (#) means combination with metformin ± pioglitazone ± dipeptidyl peptidase-4 inhibitors
The Cochrane risk of bias tool showed that five articles were evaluated as low risk of bias while one was unclear in two types of selection bias. In the risk of performance bias, one trial was regarded as low risk, one article was concerned with high risk, and the rest were considered an unclear risk. In terms of the detection bias, one trial was concerned as low risk, and others were unclear. The attrition bias was at low risk in three articles, one study was at unclear risk, and two trials were at high risk. Reporting bias was at low risk in 6 trials, and the risk of other bias was appraised as unclear (Fig. 2).
The assessment of the quality of clinical trials is shown. (a) Risk of bias graph. (b) Risk of bias
summary. (+) denotes a low risk of bias; (−) denotes a high risk of bias; (?) denotes an unclear risk of bias.
Based on the Jadad Scale, the scores ranged from 1 to 5, and trials scored 3 points or more were considered as high-quality trials. In our meta-analysis, four trials were scoring 3 points or more assessed as high quality, and two studies were 2 points as low quality (Table 2).
| Trials | Randomization mentioned | Concealment of randomization | Blinding | Appropriate blinding method | Reporting of withdrawals | Jadad score |
|---|---|---|---|---|---|---|
| Heise (2011) [14] | Yes | Yes | Unclear | Unclear | No | 2 |
| Kumar (2016) [15] | Yes | Yes | Unclear | Unclear | Yes | 3 |
| Kumar (2017) [16] | Yes | Yes | Unclear | Unclear | Yes | 3 |
| Liebl (2013) [17] | Yes | Unclear | Yes | Unclear | No | 2 |
| Onishi (2013) [18] | Yes | Yes | Yes | Unclear | Yes | 4 |
| Seiya (2019) [19] | Yes | Yes | Unclear | Unclear | Yes | 3 |
The pooled analysis showed that there was a pronounced significant difference in mean HbA1c end-of-trail between IDeg/Asp and IGlar [MD = –0.16; 95% CI (–0.31, –0.00); p = 0.04], with low heterogeneity (I2 = 0%). Egger’s test showed no publication bias, suggesting that IDeg/Asp has better control in mean HbA1c level than IGlar (Fig. 3).
Impact of IDeg/Asp on mean HbA1c end-of-trail
The pooled analysis showed that mean FPG end-of-trail was similar in patients between IDeg/Asp and IGlar, with no significant difference [MD = –0.01; 95% CI (–0.36, 0.34); p = 0.95]. The heterogeneity was very low (I2 = 0%). No publication bias was found after analysis (Table 3).
| Outcomes | Number of studies | Number of patients | Mean Difference IV, Random, 95% CI | Test for overall effect (p-value) | Heterogeneity I2 (%) |
| mean FPG end-of-trail |
4 | 1,173 | –0.01 [–0.36, 0.34] | 0.95 | 0 |
| nine-point SMPG | 3 | 899 | –0.12 [–0.31, 0.08] | 0.25 | 0 |
| Outcomes | Number of studies | Number of patients | Odds Ratio M-H, Random, 95% CI | Test for overall effect (p-value) | Heterogeneity I2 (%) |
| adverse events | 6 | 1,346 | 0.98 [0.68, 1.41] | 0.92 | 51 |
| 5* | 1,236 | 0.87 [0.68, 1.13] | 0.3 | 0 |
(*) means after removed Heise 2011
The pooled analysis showed that there was no significant difference in patients among IDeg/Asp to IGlar [MD = –0.12; 95% CI (–0.31, 0.08); p = 0.25], with low heterogeneity (I2 = 0%). No publication bias was found after analysis (Table 3).
Safety outcome Meta-analysis of confirmed hypoglycemiaThe pooled analysis showed that confirmed hypoglycemia was similar in patients between IDeg/Asp and IGlar that was no significant difference [OR = 1.59; 95% CI (0.97, 2.61); p = 0.07], with moderate heterogeneity (I2 = 66%). Throughout the subgroup analysis, we tend to found that the group of IDeg/Asp (duration of diabetes ≤11 years) considerably had a higher risk than IGlar (p < 0.0001) in confirmed hypoglycemia, with lower heterogeneity (I2 = 0%). No publication bias was found after analysis (Fig. 4).
Impact of IDeg/Asp on confirmed hypoglycemia
There was no significant distinction in nocturnal confirmed hypoglycemia [OR = 0.54; 95% CI (0.31, 0.94); p = 0.49], with moderate heterogeneity (I2 = 57%). Compared with alternative studies, the research by Kumar revealed in 2017 manifested prominent heterogeneity. After deleting this study, heterogeneity was low (I2 = 0%), and the comprehensive effects revealed that the nocturnal confirmed hypoglycemia of IDeg/Asp group was lower than IGlar group [OR = 0.40; 95% CI (0.26, 0.62); p < 0.0001]. No publication bias was found after analysis (Fig. 5).
Impact of IDeg/Asp on nocturnal confirmed hypoglycemia after deleting Kumar 2017
There was no significant distinction in adverse events [OR = 0.98; 95% CI (0.68, 1.41); p = 0.92], with moderate heterogeneity. Compared with alternative studies, the research by Heise revealed in 2011 manifested prominent heterogeneity. After deleting this study, heterogeneity was low (I2 = 0%), but the comprehensive effects revealed that adverse events still had no significant difference in the two groups [OR = 0.87; 95% CI (0.68, 1.13); p = 0.30]. It may be that the experimental cycle of the study by Heise is relatively short, and the adverse events of insulin may be more likely to occur in the late stage when it is used for a long time. No publication bias was found after analysis (Table 3).
Meta-analysis of serious AEsThe pooled analysis indicated that there was a pronounced significant difference in serious AEs among IDeg/Asp to IGlar [OR = 2.09; 95% CI (1.27, 3.45); p = 0.004], with low heterogeneity (I2 = 0%). Egger’s test showed no publication bias was detected, suggesting that IDeg/Asp has a higher risk of serious AEs than IGlar (Fig. 6).
Impact of IDeg/Asp on serious adverse events
T2DM is one of the most severe health matters. At present, the first-line drug recommended by the clinical guidelines is metformin, but as the gradual decline of β-cell function in patients with type 2 diabetes, exogenous insulin will be necessary to achieve the desired glucose levels [20, 21]. IDeg/Asp both has the characteristics of long-acting insulin and fast-acting insulin. Studies have found that compared with IGlar, IDeg/Asp could better control blood glucose and reduce the frequency of hypoglycemia events [10, 18, 22], but there have been few reports comparing efficacy and safety between IDeg/Asp and IGlar. Therefore, we conducted a meta-analysis of trials to evaluate and compare the efficacy and safety of IDeg/Asp and IGlar in the treatment of T2DM to provide a more beneficial comprehending and planning for clinical decision-making and future studies.
Our study included six RCTs of 1,346 patients, four of which were high quality and two of which were low quality. Concerning efficiency, our results showed that compared with IGlar, IDeg/Asp might have better control in mean HbA1c level, further highlighting that IDeg/Asp may have a more long-lasting effect in controlling HbA1c for each patient. However, for FPG and nine-point SMPG levels, the effect of IDeg/Asp and IGlar on their results were not significant, suggesting that the advantage of IDeg/Asp in controlling blood glucose was not very obvious, which has been only reflected in the mean HbA1c level so far. Additionally, a comparison with other similar studies reveals several similar results and some notable points. Kisioglu et al. [23] performed a real-world experience that included 174 participants, revealing that both IDeg/Asp treatment and IGlarU300 treatment showed a significant decrease in FPG and HbA1c. Nevertheless, Kawaguchi et al. [24] compared the efficacy and safety of twice-daily IDeg/Asp and IGlar300/Glu, found that IDeg/Asp ended up poor in efficacy and safety to IGlar300/Glu. Therefore, many multicenter randomized controlled trials are required to determine the efficiency of IDeg/Asp versus IGlar.
In terms of safety, our pooled analysis found that after deleting the research of the research by Kumar revealed in 2017, the nocturnal confirmed hypoglycemia of IDeg/Asp was significantly lower than IGlar. It may be that the baseline of fasting plasma glucose of the patients included in the group of IDeg/Asp was lower than those included in the group of IGlar in the study of Kumar. Besides, it is also probable that the mean age baseline of the patients included in the group of IDeg/Asp was higher than those included in the group of IGlar. Therefore, the patients in IDeg/Asp group were more prone to nocturnal hypoglycemia, while other studies did not have such significant differences in a baseline of mean age and FPG. However, there was no significant difference in the occurrence of hypoglycemia overall, our subgroup analysis of confirmed hypoglycemia indicated that the population in this subgroup (duration of diabetes ≤11 years) might has its particularity effecting the hypoglycemia outcome, which emphasized the significance of personalized treatment for different patients and identified the appropriate diet of the day with which to administer IDeg/Asp, further studies are needed.
The low incidence of nocturnal hypoglycemia in IDeg/Asp may be due to the flat and stable pharmacodynamic characteristics of essential IDeg components and long half-life [25]. It has been reported that the half-life of IDeg is about 25 hours, twice as much as IGlar [26], so that IDeg/Asp taken with the evening main meal decreases post-dinner glucose excursions and maintains more steady nocturnal glucose standards than IGlar [17]. As for hypoglycemia, a great number of studies have pointed out that hypoglycemia is most common in elderly patients with multiple or late comorbidities, patients with prolonged diabetes, or patients with a previous history of hypoglycemia [27, 28], and that a vital and potentially modifiable risk factor for hypoglycemia is the choice of intensive glucose therapy, with insulin or insulin secretagogues at the highest risk [29, 30]. Three Phase III studies (the Onish study, the Step by Step study, and the China-registered study) have confirmed that compared with the control group (the control group of the Onish study was insulin glargine once daily with or without oral hypoglycemic agents, the control group of the Step by Step study was insulin glargine once daily + insulin aspart 1–3 times daily with or without oral hypoglycemic agents, and the control group of the China-registered study was premix insulin twice daily), there was no significant difference or significant reduction in the incidence of confirmatory hypoglycemia and night confirmatory hypoglycemia of IDegAsp [18, 22, 31], which is consistent with our study. Kumar [15] revealed that the rate of overall confirmed hypoglycemic episodes was distinguishably higher with IDeg/Asp than IGlar. In comparison, the IDeg/Asp group had a significantly lower rate of nocturnal hypoglycemic episodes than IGlar, suggesting that IDeg/Asp was mostly used in the elderly with a higher risk of nocturnal hypoglycemia. IGlar was more suitable for those who had higher overall hypoglycemia risk, which is consistent with the findings of our study. Notably, Hanne et al. [32] mentioned that once-daily IDeg/Asp might achieve a low risk of postprandial hypoglycemia if the dose and injection time are adjusted according to the individual differences of patients and dietary habits, which indicated that the flexibility of injection time and dose changes should be considered when using IDeg/Asp to match any significant changes in daily dietary pattern. Additionally, from the Multinational Consensus of Kalra et al. [9], the efficacy of IDeg/Asp might be relative to the consumption of high carbohydrate meals. Therefore, the study of Onishi [16] contributed high heterogeneity in our result of confirmed hypoglycemia might be that the people who consume high carbohydrate meals have tolerance for IDeg/Asp.
Furthermore, studies have also illustrated that IDeg/Asp and IGlar used in TD2M treatment had no significant difference in the incidence of general adverse events, while IDeg/Asp had a higher risk of severe adverse events. Of note, in this analysis, we found that one patient was predisposed to depression or experienced mood changes after taking IDeg/Asp [14], and another two patients had a transient ischemic attack [14] and a cardiac failure [16], leading to the subject’s discontinuation respectively. Studies have also illustrated that IDeg/Asp may increase polymyalgia rheumatica, hemorrhagic stroke, and thrombophlebitis [15]. This suggests that IDeg/Asp used in elderly patients with multiple or severe complications, especially elderly patients with a high risk of hypoglycemia, should be used with caution, and individualized treatment programs should be implemented in light of specific conditions.
When considering these results, several limitations should be aware of. First, the factors that cannot be manipulated constantly may influence, such as different ethnicity, ages at the initiation of intervention, duration of intervention, formulation and dose of IDeg/Asp and IGlar, and dietary and exercise guidelines. In addition, the inclusive literature may be potentially biased during the review process. Finally, we included a small number of articles and only incorporated the published studies, which may impact the results. Further research is needed on this aspect.
Our pooled analysis suggested that compared with IGlar, IDeg/Asp had a tremendous advantage in controlling mean HbA1c level but had no significant difference in controlling mean FPG level and nine-point SMPG. Furthermore, we found that there was no significant difference in the occurrence of hypoglycemia overall, but our subgroup analysis of confirmed hypoglycemia revealed the population in this subgroup (duration of diabetes ≤11 years) might has its particularity effecting the hypoglycemia outcome, which emphasized the significance of personalized treatment for different patients. In contrast, IGlar may increase the risk of hypoglycemia confirmed at night, which suggested that in the future use of IDeg/Asp. It is worth noting that the vast majority of the studies we included were patients who received insulin treatment for the first time based on oral hypoglycemic drugs. We reached conclusions through data collection and analysis. Therefore, it is necessary for us to limit our conclusions to patients treated with insulin for the first time. As for whether it can be applied to non insulin primary treatment and long-term efficacy, further research is needed.
AEs, Adverse events; BMI, Body mass index; FPG, Fasting plasma glucose; HbA1c, Hemoglobin A1c; IDeg/Asp, Insulin degludec/aspart; IGlar, Insulin glargine; MD, Mean difference; NRCT, Non-randomized Controlled Trial; NCT No., National clinical trial number; OR, Odds ratio; PD, Pharmacodynamic; PK, Pharmacokinetic; RCT, Randomized Controlled Trial; SD, Standard deviation; SMPG, Self-measured plasma glucose; T2DM, Type 2 diabetes mellitus; 95% CI, 95% confidential interval
The authors would like to express special gratitude to all the personnel who supported or helped with this study.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
None of the authors have any potential conflicts of interest associated with this research.
