RT‐CGM in conjunction with CSII vs MDI in optimizing glycaemic control in T1DM: Systemic review and meta‐analysis

Abstract Introduction To determine the impact of real‐time continuous glucose monitoring (RT‐CGM) in conjunction with ‘Open loop’‐ continuous subcutaneous insulin infusion (CSII) as compared to conventional multiple daily injections (MDI) in type 1 diabetes. Methods We explored the COCHRANE database, MEDLINE, WEB OF SCIENCE, GOOGLE SCHOLARS, PUBMED, EMBASE, and cited literature in articles retrieved (2010–2021) for all randomized controlled trials and real‐world trials of more than 6 months duration in patients with type 1 diabetes that compared RT‐CGM+CSII vs RT‐ CGM+MDI. A total of 1645 publications have been identified; however, only 3 trials fulfilled our inclusion criteria with a total number of 150 patients (72 patients using RT‐CGM+CSII and 78 patients on RT‐CGM+MDI). A Systematic Review and Meta‐analysis were carried out. Results No statistically significant reduction in HbA1c was found on comparing RT‐CGM+CSII vs RT‐ CGM + MDI, with p‐value = .75. Likewise, impact on TIR, weight and insulin usage was found to be statistically insignificant with p‐value of 0.15, 0.75 and 0.20 respectively. There was an overall homogeneity between the 3 trials in respect to all previous variables with I 2 being 0%. Conclusions Real‐time continuous glucose monitors in conjunction with MDI open‐loop CSII had a similar impact on HbA1c, weight, insulin usage and TIR. In addition, RT‐CGM when combined with CSII was associated with higher costs and reduced quality of life, hence RT‐ CGM+MDI can be considered as a cheaper, safer yet equivalent substitute. Review Registration This study was registered in PROSPERO (International prospective register of systematic reviews). Registration Name: RT‐CGM in conjunction with CSII vs MDI in optimizing glycaemic control in T1DM: a systematic review. Registration No: CRD42021255333. Accessible at: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021255333. Amendments: Few amendments to the above‐mentioned registration were made: (1) Title (Meta‐analysis was added). (2) Prof. Gleeson was added as an author. (3) Real‐world trials were included. (4) Outcomes required in studies as per our inclusion criteria amended to include at least 1 outcome. (5) Bias risk was assessed by the CASP tool.


| INTRODUC TI ON
In 2019, the global estimate of people with diabetes was 463 million.
This number is projected to reach 578 million by 2030 and 700 million by 2045. 1 Of these, approximately 10% have type 1 diabetes. 1 Historically, the first commercially available self-monitoring blood glucose (SMBG) device became available early in the 1980s, and later in 1999, the FDA approved the first CGM. 2 Nonetheless, it was not until 12 years later that the Endocrine Society had released its first CGM guidelines in the year 2011 and recommended its use due to data showing its effectiveness in adults with type 1 diabetes, who can utilize these devices on a nearly daily basis. 3 Recently, the Advanced Technologies & Treatments for Diabetes (ATTD) consensus established recommendations for the relevant aspects of CGM data utilization and reporting among the various diabetes populations and endorsed a standardized template (Ambulatory Glucose Profile) for data presentation and visualization to help unify the information in question. 4 Furthermore, lately, Continuous

Subcutaneous Insulin Infusions (CSII) and Sensor-Augmented
Insulin-Pump (SAP) with a threshold-suspend feature have been introduced, allowing automated suspension of basal insulin delivery in response to a predicted or detected low glucose level, which has been found to exhibit eminent results in terms of glycaemic control. 5 Nevertheless, with the implementations of CGM in conjunction with intensive insulin therapy Sensor-Augmented Insulin Regimen (SAIR) using CSII or MDI, a new uncertainty was recognized; which method of SAIR has better outcomes, is more cost-effective and has less frequent hypoglycaemic episodes?
A non-randomized prospective real-life clinical trial by Rodbard et al. 6 compared glycaemic outcome when CGM in used with CSII vs MDI and concluded that both therapy groups have similar and statistically indistinguishable responses with an improvement of glycaemic variability mean by 15.02% and 11.39% respectively. Furthermore, a multicentre, open-label, randomized controlled trial involving 15 paediatric National Health Service (NHS) diabetes services in England and Wales, with participants aged between 7 months and 15 years comparing CSII to MDI, concluded that during the first year following type 1 diabetes diagnosis, no clinical benefit of CSII over MDI was identified in children and young people in the UK setting, and treatment with either regimen was suboptimal in achieving HbA1c thresholds. In addition, CSII was not cost-effective with a mean total cost being higher by £1863 (95% confidence interval £1620 to £2137) for CSII than for MDI; with the most of this difference (£1177) from the additional cost of consumables and devices (undiscounted annual cost of £600 for CSII vs £80 for MDI). 7 This was contrary to both, STAR 3 study and the Eurythmic Trials, which concluded that switching from optimized MDI to SAP therapy allowed rapid and safe A1c reduction that was observed by 3 months and persisted throughout the study period (18 months). 8,9 Nevertheless, in both trials, only the patients using CSII were on CGM, and therefore, conclusions were not specific whether glycaemic improvement was due to CGM or CSII use.
The up-to-date recommendation from the ADA in 2021 10 concluded that RT-CGM or intermittently scanned CGM (isCGM) in conjunction with MDI or CSII can lower and maintain HbA1c, and reduce hypoglycaemia in youth and adults with diabetes to replace SMBG when used properly. However, no consensus is yet available to guide the selection of the appropriate insulin administration method and it continues to be individually based, although socioeconomic status, in addition to other factors like race, ethnicity, private health insurance and education influence this decision. 11 Therefore, conducting a Systematic Review/Meta-analysis on the most recent high-quality studies will aid us in answering our queries concerning the impact of RT-CGM when combined with CSII as compared to MDI on glycaemic control, cost-effectiveness, and rates of hypoglycaemia. This will enhance our practice in the field of diabetes management, the quality of life of the diabetic population worldwide and achieve required glycaemic targets by utilizing the most recent innovations.

| AIM & OBJEC TIVE S
Our primary objective is to assess the impact of RT-CGM+open-loop CSII vs RT-CGM+MDI on HbA1c/GV in type 1 diabetes. However, our secondary objectives are to assess the impact on several other variables:

| Search strategy
A comprehensive review of the existing literature was undertaken using the Preferred Reporting Items for Systematic Reviews and Meta-analysis-PRISMA guidelines. 12 Several search engines were explored using specific keywords as shown below to obtain required RCTs/ real-world clinical trials. For each database, distinct and comprehensive search strategies were constructed using subjectheading mapping. In addition, a repeat search was made just before the final analyses and adaptations for British and American English were made for all searches.

| Study selection
Publications were eligible if they fulfilled the inclusion criteria, in addition to reporting at least one of the outcomes in question.

| Data extraction
Through the search strategy being implemented only articles that fulfilled the inclusion criteria were included in the study. Included trials were downloaded, and data were extracted from each study using customized data extraction forms created to include relevant data required for subsequent data analysis (Tables 1-3).

| Assessment of risk of bias
CASP tool (accessed 1st September 2021), was utilized to help critically appraise articles and reduce the risk of bias ( Table 4). The CASP tool contains checklists that assist researchers in critically appraising various types of evidence. All studies that were included in this review have been deemed credible and non-prejudiced through CASP tool's checklists.

| Data synthesis
Data were processed through meta-analysis and represented in a forest plot, with a primary outcome of interest being HbA1c/ GV. Furthermore, the treatment effect was estimated with a mean difference in the ultimate values of HbA1c/GV between the RT-CGM+CSII group and the RT-CGM+MDI group.
I 2 statistics were measured to assess heterogeneity across the studies. MIX 2.0 Pro (version 2.014) was used to yield the data, and 95% confidence intervals (95% CI) was calculated. The mean difference (MD) as an indicator of efficacy has been reported.

| Study selection
The literature search based on the specific keywords yielded 1541 references, with additional 104 articles identified from reference lists. However, after further assessment of articles as per inclusion criteria, only 3 studies 13-15 were found eligible. Figure 1 schematically demonstrates the search strategy used to identify trials for inclusion.

| Quality appraisal and publication bias
All 3 included studies have shown high quality through CASP tool, showing a good methodological approach. Nevertheless, as it is not possible to blind the participants or the experimenters from intervention protocol, all studies did not score in this criterion.
Moreover, the likelihood of publication bias was not feasible to assess since only a small number of studies were eligible to be included in our meta-analysis.

| Participant characteristics
The overall sample size was 150, distributed into two groups, RT-CGM+CSII and RT-CGM+MDI (78 and 72) respectively. The mean age of the participants in the included trials was 37.1 years (range 32.3-46 years). In addition, the mean in HbA1c in overall RT-CGM+CSII and RT-CGM+MDI groups was 63.3 ± 9.2 (mmol/mol) and 63.5 ± 10.2 (mmol/ mol) respectively. More emphasis on the overall characteristics of trials included in the meta-analysis is provided in Table 5.

| Synthesis of results
Our primary outcome was to assess the impact difference on HbA1c in Type 1 diabetes between patients on RT-CGM+MDI compared with those on RT-CGM+CSII through carrying out a meta-analysis on the data extracted from studies in question. Upon assessment of the results from the 3 trials included, there was no statistically significant reduction in HbA1c on comparing RT-CGM+CSII vs RT-CGM+MDI with homogeneity in outcome throughout all studies (I 2 = 0%, p-value =.75) as shown in Figure 2.
Furthermore, regarding our secondary outcomes, the impact of RT-CGM+CSII in comparison to RT-CGM+MDI on TIR, weight and insulin usage, were the only assessable outcomes in our meta-analysis due to the limited data availability in the included studies. The impact on all these variables was found to be statistically insignificant with pvalue being 0.15, 0.75 and 0.20 respectively (Fig. S1-S3). In addition, all three studies have shown homogeneity when compared to each other in respect to all previous variables with I 2 being 0% (0%-40% generally considered unimportant; 50%-90%, substantial; Fig. S4). Nevertheless, since the results within the meta-analysis have shown overall homogeneity and considering the small number of studies and the relatively low combined sample size, further covariate analysis to reduce in-group error and eliminate confounding factors will probably be affected and therefore less useful. Another secondary outcome in question was to explore whether RT-CGM+CSII and RT-CGM+MDI would increase the frequency of severe hypoglycaemic episodes or ketoacidosis episodes; however, because these unfavourable complications were uncommon in all 3 trials, it was not explored.

| Limitations
This current study adds to the existing literature of the most comprehensive accumulation of published evidence regarding Abbreviations: "+" meaning yes; "±" meaning partially present; "−" meaning no/can't tell.