Abstract
Achieving glycemic control and targets are challenging in type-1 diabetes management. To achieve this, intensive insulin therapy or multiple daily injections (MDI) and continuous subcutaneous insulin infusion (CSII) or pump therapy have been used in various health care settings. However, there has been a debate on their superiority. Some of researchers have recommended MDI, while others SCII. We compared MDI with CSII by a literature search. We have conducted mata-analysis for MDI and CSII on ten randomized controlled trials on 809 type-1 diabetics 809, MDI (N = 394) or CSII (N = 415). Heterogeneity between trials was quantified by conventional Q-statistic (Cochran’s heterogeneity statistic) and Higgins I2 statistic with 0-40% representing negligible heterogeneity, 30-60% moderate heterogeneity, 50-90% substantial heterogeneity and 75-100% considerable heterogeneity. tau-squared (τ2) was used to observe between-study random-effects variance. Meta Analyst software was used to analyze the data and to conduct meta-analysis. SPSS was used to analyze HbA1c student’s t-test for MDI and CSII. A random-effect analysis ((DerSimonian-Laird method) performed on ten studies found that the percentage of HbA1c was lower in patients receiving CSII compared with those receiving MDI; standardized mean difference (SMD) was 0.441, 95% confidence interval 0.267 to 0.616, p < 0.001; equivalent to a difference of 0.39%, favoring CSII. I2 statistic was 20.9; τ2 = 0.016; Q = 11.378 with df = 9, indicating that heterogeneity was not significant (heterogeneity p-value = 0.251). Patients on CSII demonstrated significantly lower values (8.2±0.72 versus 7.73±0.72; p-value < 0.001 respectively). This statistical and meta-analysis favors the usage of insulin pump therapy. We concluded that patient centered approach should be used while selecting the patients for insulin pump (CSII) or MDI.
Introduction
There has been a debate that continuous subcutaneous insulin infusion (CSII) or pump therapy (first introduced in 1970s) is superior to insulin injections including multiple daily injections (MDI), intensive insulin therapy or the basal bolus therapy. Furthermore, some studies have demonstrated that insulin dosages were less in CSII with better patient satisfaction [1 ,2]. However, there are several other studies with some conflicting results and some authors have concluded that both are equally effective in term of reduction of glycated hemoglobin (HbA1c) [3–5]. The main outcome of these studies was HbA1c, which shows the control of diabetes for the past two months. HbA1c is important, as if this worsens, diabetes complications initiates and progress [6]. Under this debate, we collected randomized clinical trials (RCTs) conducted on type-1 diabetic patients comparing the HbA1c results between MDI and CSII and conducted meta-analysis.
Materials and methods
PRISMA guidelines were used for reporting of individual patient data meta-analyses [7]. We performed internet database survey (PubMed, Google Scholar) and reviewed literature. Only randomized controlled trials (RCTs) on type-1 diabetic patient were included. Observational studies, reviews, surveys, and short term studies (less than two months) were excluded. Also studies with incomplete data and those studies which did not provide complete data details (such as mean ±SD or the numbers randomized/exact number of subjects) were excluded from meta-analysis. HbA1c mean ± SD was calculated for MDI and CSII. Heterogeneity between trials was quantified by conventional Q-statistic (Cochran’s heterogeneity statistic) and Higgins I2 statistic (the degree of inconsistency in the results between studies or the percentage of variability in effect due to heterogeneity rather than sample error) with 0–40% representing negligible heterogeneity, 30–60% moderate heterogeneity, 50–90% substantial heterogeneity and 75–100% considerable heterogeneity. Additionally, tau-squared (τ2), estimates for the between-study random-effects variance was calculated as well. Standardized statistical techniques and Meta Analyst software was used to analyze the data and to conduct meta-analysis [8 – 12]. Data was also entered in SPSS to find mean HbA1c differences (t-test) for MDI and CSII. A random-effect analysis was performed on these studies to find out overall effect measure.
Results
According to inclusion criteria, ten studies were identified as RCT on type-1 diabetics, with 809 patients randomized to receive either MDI (N = 394) or CSII (N = 415). Table-1 demonstrates details and characteristics of the trials included in the meta-analysis [13-22]
Figure-1 shows a forest plot and results of aggregate meta-analysis with the effect size of all ten studies, their confidence intervals (95% CI), and the summary with overall effect measure for the mean HbA1c difference between MDI and CSII.
Forest Plot results of random effect meta-analysis model (DerSimonian-Laird random effects method) with standardized mean differences (SDM), 95% confidence intervals for percentage of glycated hemoglobin (HbA1c%) compared with insulin pump (CSII) versus MDI or basal bolus therapy(SMD = 0.441 (95% CI 0.267 to 0.616) I2 = 20.9; τ2 = 0.016; Q = 11.378 df = 9; p = 0.251)
A random-effect analysis ((DerSimonian-Laird method) performed on ten studies found that the percentage of glycated Haemoglobin (HbA1c) was lower in patients receiving continuous subcutaneous insulin infusion compared with those receiving insulin injections; standardized mean difference (SMD) was 0.441, 95% confidence interval 0.267 to 0.616, p < 0.001; equivalent to a difference of 0.39%, favoring CSII. I2 statistic was 20.9; τ2 = 0.016; Q = 11.378 with df = 9, indicating that heterogeneity was not significant (heterogeneity p-value = 0.251). When mean HbA1c values of MDI and CSII were compared, patients on CSII demonstrated significantly lower values (8.2±0.72 versus 7.73±0.72; p-value < 0.001 respectively). This statistical and meta-analysis favors the usage of insulin pump therapy.
Discussion conclusion and recommendations
Successful management of diabetes requires diabetes self-management education (DSME) and a team work. Selecting the patient for specific management strategies is an art. While selecting the patient for MDI or CSII, several factors should be considered, such as patient’s age, his glycemic profile, HbA1c, dietary pattern, other comorbid complications and a history of hypoglycemia (such as late night, early morning or daytime hypoglycemia). Patient counselling is also an essential aspect of diabetes management prior to starting or selecting specific insulin therapies or regimens, such as MDI or CSII. Although different studies in medical literature have given different conclusions, however, our meta-analysis favors the use of insulin pump in type-1 diabetics for better glycemic control. Some studies conducted in past have also concluded that insulin pump provides only satisfaction to the patients and that glycemic control was equally effective with MDI or CSII [17]. While on the other hand, some studies have reported lower risk of hypoglycemia with CSII [13]. Conversely, other authors have proved that the incidence of hypoglycemia was similar with CSII and MDI [14, 22]. Under this discussion and meta-analysis, physicians and diabtologists should use patient centered approach for managing hyperglycemia in type-1 diabetics [23, 24]. Patient’s selection for optimal therapies remains the top priority which can be achieved by DSME and counselling. Furthermore, cost effectiveness should also be considered while selecting MDI and CSII. Further studies at multicenter level are required to confirm the findings of the current study.
Data Availability
Data fro RCT
Conflict of interest
Authors declare no competing conflict of interest.
Funding
No funding was received and no organization funded this work
References
