Original Title: Efectividad, seguridad y coste-efectividad de los sistemas de monitorización continua de glucosa en tiempo real, sistemas de asa abierta y sistemas híbridos de asa cerrada (páncreas artificial) para pacientes con diabetes

Authors' objectives: • To evaluate of the effectiveness and safety of RT-CGM systems, open loop systems (SAPsp) and hybrid closed loop systems (PA), currently marketed in Spain, for people with any type of diabetes being treated with insulin. • To evaluate the cost-effectiveness of RT-CGM systems, open loop systems (SAPsp) and hybrid closed loop systems (PA), currently marketed in Spain, for people with any type of diabetes on insulin treatment. • To identify the ethical, legal, organizational, social and environmental aspects related to RT-CGM systems, open loop systems (SAPsp) and hybrid closed loop systems (AP). • To identify research needs and standard outcome measures from the perspective of patients, families/caregivers, healthcare professionals and researchers on RT-CGM systems, open-loop systems (SAPsp) and hybrid closed-loop systems (AP).

Authors' results and conclusions: RESULTS: EFFECTIVENESS/SAFETY: [ RT-CGM/SAPsp ] Twenty-five articles (26 references) were included. In adults with DM1, three RCTs, one non-RCT, and two controlled OS obtained results significantly favorable to CGM in all or almost all glycemic variables (HbA1c and CGM variables) compared to SMBG or FM. A controlled OS did not obtain differences in HbA1c or serious acute events, and another did not perform statistical contrasts, showing small differences. There were also significant benefits in fear of hypoglycemia (two studies) and satisfaction with treatment (two studies), but not in perception of hypoglycemia (two studies). Only one controlled OS that exclusively included minors was included. A significantly favorable result was obtained for the intervention against FM in HbA1c. In the rest of the glycemic variables, no statistical contrasts were performed, although the differences are clinically important in favor of the intervention. In the case of T2DM, two RCTs and one controlled OS were included. One obtained significantly favorable differences between CGM and SMBG in all glycemic variables except variability. The other one only evaluated HbA1c, obtaining a significant benefit. The OS did not obtain significant differences in HbA1c or serious acute events. A study with women with GDM did not obtain significant differences in any glycemic variable, nor in birth or neonatal complications. An RCT with adults at risk of hypoglycemia obtained a significant reduction in the incidence rate of severe hypoglycemia and the time in hypoglycemia and fear of hypoglycemia, and a small but significant reduction in TIR, compared to insulin pump plus SMBG. Another with adolescents and young adults obtained significant reductions in time in hypoglycemia with SAPsp compared to SAP. These two studies were already included in the previous report. A retrospective controlled OS with 5 years of follow-up observed a difference favorable to SAPsp versus SAP in HbA1c, but did not reach significance (-0.40%, p = 0.06). The same occurred with severe hypoglycemia events (4.2% vs. 10.3%). A study with pregnant women with DM1 did not obtain significant differences between SAPsp and SAP in any glycemic or self-reported variable. Eight uncontrolled OSs evaluated the safety of the Eversense® implantable device, with no serious adverse events observed. One RCT compared this device versus FTM/SMBG. In adults with DM1, there were no significant differences in any glycemic or self-reported variables. In adults with a tendency toward biochemical hypoglycemia in the previous two weeks, there was a significant reduction in favor of the implantable device in time <54 mg/dl. A small crossover RCT found no relevant differences compared to the Dexcom® G5 device. [ ARTIFICIAL PANCREAS (PA) ] Twenty-six studies were included, all with patients with DM1 except for a small uncontrolled OS with patients with cystic fibrosis: 15 RCTs, 7 controlled OSs, and 4 uncontrolled OSs (two of them with two overlapping subsamples of patients with DM2). The MA carried out with RCTs and OSs compared to any type of control group showed results significantly favorable to PA in HbA1c (MD = -0.31%, 95% CI: -0.57, -0.21), TIR (MD = 11.28%, 95% CI: 8.38, 14.17), T>180 mg/dl (-8.56%, 95% CI: -11.59, -5.54), T<70 mg/dl (MD = -1.05%, 95% CI: -1.57, -0.54), T<54 mg/dl (MD = -0.36%, 95% CI: -0.57, -0.15), standard deviation (MD = -7.16, 95% CI: -9.53, -4.79) and coefficient of variation (MD = -3.09, 95% CI: -4.30, -1.88) of glucose levels, and diabetes-related quality of life (SMD = -0.23, 95% CI: -0.39, -0.07). There were no significant differences in fear of hypoglycemia or satisfaction with treatment, or in safety variables (severe hypoglycemia, ketoacidosis, other serious acute events). The results described show a high statistical heterogeneity in all cases, which was partly explained by the type of control group. The mentioned differences are less intense in the comparisons against SAP and SAPsp, but the benefits on HbA1c, TIR, and T>180 mg/dl are relevant compared to all comparators, both in adults and in minors. In the latter, the results on the reduction of time in hypoglycemia are more uncertain. No studies on GDM were included, and in the case of uncontrolled OS with DM2 and cystic fibrosis patients, significant pre-post changes were obtained, although the scarcity and uncontrolled nature of this evidence does not allow recommendations to be supported. COST-EFFECTIVENESS: Twelve economic evaluations were included where the devices were evaluated in subjects with DM1 through cost-utility analysis. The methodological quality was considered moderate, although most studies suffer from a lack of transparency, for example, by not reporting the cost of the devices. No studies carried out in people with DM2 or carried out in Spain were identified. Five studies evaluated the cost-effectiveness of CGM. In all of them, Dexcom G6 was evaluated in comparison with SMBG in adults with DM1 using the IQVIA Core Diabetes Model, for the entire life of the patient and including only direct healthcare costs. A limitation common to all studies is using the DIAMOND RCT as a source of effectiveness in which a previous version of the device, the Dexcom® G4, was evaluated. Two studies also compared Dexcom with FM. All studies were funded by the manufacturer Dexcom. In all studies, the ICERs obtained were below the cost-effectiveness thresholds specified for each of the countries and, once transformed into 2023 euros, below the threshold of €25,000/QALY. The studies further found that the probability of Dexcom being cost-effective for their respective thresholds is above 89%. The results are especially sensitive to changes in the effect on HbA1c, the utilities associated with fear of hypoglycemia and fingertip prick, the frequency of SMBG, and the cost of CGM. Seven studies evaluated the cost-effectiveness of AP. All studies consisted of Markov models evaluating the MiniMed™® 670G or 780G (from Medtronic). The main limitation of the studies is that the majority took the evidence on effectiveness from observational studies. Four studies approached the analysis from a social perspective while the other three only included direct health costs. In 6 studies, MiniMed™® was evaluated in adults and for the entire patient’s life in comparison with various alternatives: insulin pump plus SMBG, SAPsp, FM and MDI, FM with MDI or insulin pump, MDI with or without CGM. In another study, it was evaluated in a 12-year-old population in comparison with insulin pump or MDI (with SMBG or CGM). Five studies were funded by the manufacturer Medtronic. In all studies, the ICERs obtained were below the cited cost-effectiveness thresholds or even dominance was obtained in the base case in a study that compared the MiniMed™ 780G with the SAPsp. The probability that AP is cost-effective for these thresholds is above 84%, with the exception of a study carried out in the United Kingdom where the probability that MiniMed™® 670 G compared to insulin pump is cost-effective is 45.1% when the threshold is 20,000 GBP/QALY and 99.8% when the threshold is 30,000 GBP/QALY. When converting the original ICERs to 2023 euros, in two studies that compared the MiniMed™® 780G with FM the ratios exceed the cost-effectiveness threshold of €25,000/QALY. With respect to the deterministic sensitivity analysis, utilities, the effect on hypoglycemic event rates, baseline HbA1c, the effect on HbA1c, the time horizon, and the cost of treatments were identified as determinants. In the two independent industry studies, ICERs were obtained that were sensitive to variations in key parameters such as the cost of BP or the effect of treatment on severe hypoglycemia in adults with DM1 and to variations in non-severe hypoglycemia in children and adolescents with DM1, for example. ASPECTS RELATED TO THE PATIENTS’ PERSPECTIVE, ETHICAL, LEGAL, SOCIAL AND ORGANIZATIONAL: Nine studies were included in the qualitative systematic review on closed-loop systems. The results have been grouped into two thematic blocks: acceptability of the systems and equity. The acceptability of the devices analyzed is generalized for all the technologies and populations analyzed. There is a wide variability in preferences depending on personal abilities and attitudes towards the management of the disease, so that the selection of specific devices could take into account the values and preferences of the patients. For its implementation, it is necessary to take into account the existence of a period of adaptation to the new devices in which patients may require additional support, apart from the training they currently receive. Additionally, healthcare professionals, especially those not specialized in diabetes, may require additional training or strategy to ensure continuity of care. The main reason for the lack of equity in access to these devices, according to the literature, is economic. No qualitative studies were found that addressed CGM systems for the subgroup of people with diabetes other than type 1 and 2. RESEARCH NEEDS: Six studies were identified that propose a prioritization of key outcome measures for different subpopulations (DM1, DM2, GDM). CONCLUSIONS: • No study has been identified that compared the devices evaluated against FTM with alarms (FreeStyle Libre® 2), the device used by the majority of patients in Spain, so no conclusions can be drawn about their relative effectiveness. RT-CGM IN T1DM: • Compared with SMBG and no-alarm FTM, the evidence identified for the effectiveness of RT-CGM in adults with T1D included patients with poor glycemic control, except for one study with patients who had a recent episode of severe hypoglycemia. All studies evaluated the Dexcom® G6 device (along with the G5 in two of them), generally showing statistically significant and clinically relevant benefits on HbA1c and other variables related to glycemic control, as well as on the fear of hypoglycemia and satisfaction with treatment. These results support the conclusions of the previous report on the favorable risk/benefit balance in this population. • Regarding cost-effectiveness in patients with DM1, the included studies also focus on patients with poor glycemic control and the Dexcom® G6 device, obtaining ICER below the cost-effectiveness thresholds specified for each of the countries where were carried out against SMBG (5 studies), FTM without alarms (3 studies) and FTM with alarms (one study). The studies also found that the probability that Dexcom® is cost-effective for their respective thresholds is above 89%. These results (contrary to those obtained by the SESCS with its own economic model in the previous report) are especially sensitive to changes in key parameters such as the effect on HbA1c, the utilities associated with the fear of hypoglycemia and the fingertip prick. in the case of SMBG, the frequency of this and the cost of RT-CGM. • No studies were identified that evaluated the new sensors marketed by Medtronic, although it is reasonable to assume their non-inferiority compared to the versions evaluated in the previous report, in which they were included in the favorable recommendation for patients with DM1 with poor glycemic control and/or risk of severe hypoglycemia. • The Eversense® implantable RT-CGM device shows a good safety profile and effectiveness results that are not significantly different from alarm-free FTM in patients with DM1, but the evidence is still too scarce to support a recommendation. No evidence was identified regarding the rest of the devices on the market. • In the case of children with DM1, only one observational study was included with favorable results of RT-CGM versus FTM without alarms, so the evidence remains of low quality. RT-CGM IN T2DM: • In the case of T2DM, although two RCTs with clinically relevant results compared to SMBG were included, the generalization of their results is subject to uncertainty given the populations studied (patients only with basal insulin or without insulin treatment). In the coming years, more evidence will be available in this population, based on the ongoing RCTs identified. • A single study provided by the industry evaluated Dexcom® One compared to SMBG in adults with T2DM in Spain. According to this study, not yet published, this technology is cost-effective with ICERs much lower than the threshold of €25,000/QALY. RT-CGM IN OTHER POPULATIONS: • The evidence of the effectiveness of RT-CGM in GDM is limited to one RCT without significant results, and no study has been identified with other insulin-dependent patients (not DM1, DM2 or GDM). SAPSP: • The randomized evidence on SAPsp in patients with DM1 has not changed compared to that identified in the previous report (2020), referring to the Minimed™ 640G device. An RCT with adults at risk of hypoglycemia obtained a significant reduction in the incidence rate of severe hypoglycemia, hypoglycemia time and fear of hypoglycemia compared to SMBG+ISCI, results that supported a favorable recommendation in this population from a clinical perspective (although with contrary recommendation based on the cost-effectiveness model carried out). A new retrospective OS identified, in adults with DM1, with a mean follow-up of 5 years, observed an improvement in HbA1c compared to FTM without alarms that was only significant at 90% confidence, with no differences in acute and long-term complications. No studies were identified with version 740G that met the inclusion criteria. • No EEs were identified for this device. In the previous report carried out by the SESCS, a specific EE was carried out for Spain in which SAPsp was compared with SMBG+ISCI in a population with DM1 and risk of severe hypoglycemia. It was concluded that the SAPsp was not cost-effective as very high ratios were obtained in terms of euros per QALY. ARTIFICIAL PANCREAS: • In the case of AP, the evidence focuses almost exclusively on DM1. In general, these devices appear superior to their alternatives in the reduction of HbA1c, time in hyper- and hypoglycemia and glycemic variability, and the increase in time in therapeutic range, both in adults and in minors, although in the latter the effects on biochemical hypoglycemia are more uncertain. • No evidence has been identified on the cost-effectiveness of AP in DM1 compared to FTM with alarms. Compared to FTM without alarms and SMBG, the evidence obtained is of moderate quality; Most studies suffer from a lack of transparency, for example by not reporting the cost of devices or offering limited sensitivity analyses. The only study carried out for Spain obtained an ICER slightly above the cost-effectiveness threshold of €25,000/QALY. • The evidence on the effectiveness and cost-effectiveness of PA in DM2, GDM and other insulin-dependent populations other than DM1 is very scarce or non-existent.

Authors' recommendations: • The conditional recommendation in favor of RT-CGM established in the previous report for the treatment of adult patients with DM1 at risk of severe hypoglycemia is maintained. Given that only one small effectiveness study was identified in this population that used SMBG as a comparator (with favorable results) and there were no cost-effectiveness studies, it is not possible to establish a strong recommendation. • The conditional recommendation in favor of RT-CGM established in the previous report for the treatment of adult patients with DM1 with poor glycemic control is maintained. A strong recommendation is not established because, although the evidence of effectiveness against SMBG and FTM without alarms has been reinforced, it has not been compared against FTM with alarms, and there is still uncertainty about its cost-effectiveness in Spain. • The recommendation established in the previous report on the SAPsp (MiniMed™ 640G, extensible to the 740G version) in the treatment of adult patients with DM1 at risk of severe hypoglycemia is maintained: favorable from a clinical perspective but unfavorable from a cost-effectiveness one. The new 740G version can be considered equivalent from an effectiveness/safety perspective, since it uses the same automatic insulin suspension technology, but there are no data on its cost-effectiveness. • It is not possible to establish a recommendation for or against RT-CGM for minors with DM1. Given the foreseeable absence of RCTs in the near future, decisions about this population must be based on observational evidence and clinical considerations about the needs and characteristics of each age subgroup within minors (young children, pubescent children, adolescents). • It is not possible to establish a recommendation for or against RT-CGM for patients with DM2, GDM and other insulin-dependent populations. In the coming years, more evidence will likely be available in T2DM and GDM, given the number of ongoing RCTs identified. • It is not possible to make a recommendation for or against the Eversense® implantable device until more evidence is available. ARTIFICIAL PANCREAS: • A conditional recommendation is established in favor of PA in the treatment of patients with DM1 (adults and minors) with poor glycemic control and/or risk of severe hypoglycemia. Annex 13 shows the assessment carried out for each of the dimensions of the EtD framework of the GRADE system. The favorable direction of the recommendation is based on a positive risk/benefit balance and the good general acceptability of these devices by patients and professionals, as well as the improvement of equity in the treatment of diabetes and its easy implementation. The existing uncertainty about its cost-effectiveness in Spain does not allow for a strong recommendation. • It is not possible to establish a recommendation for or against PA for patients with DM2, GDM and other insulin-dependent populations.

Authors' methods: A systematic review (SR) was carried out in the MEDLINE, EMBASE and WOS databases from 2015 to June 2023. The following were included: 1) Effectiveness: Randomized controlled trials (RCT) and non-randomized (nRCT), controlled and uncontrolled longitudinal observational studies (OS), and economic evaluations (EE). The studies were selected by two researchers independently. Risk of bias assessment was also carried out independently by two reviewers. The RCTs were evaluated using the Cochrane Collaboration’s RoB-II tool, the nRCTs and OSs with the Joanna Briggs Institute (JBI) checklist, and the EEs following the criteria of Drummond et al. (2005). A meta-analysis of effectiveness and safety results was carried out if the number of studies and their methodological homogeneity permitted it. On the other hand, another bibliographic search was carried out in the same databases to evaluate the ethical, legal, organizational, social, patient, and environmental aspects related to the technology, as well as a consultation with the Spanish Cystic Fibrosis Federation. The methodological limitations of the qualitative studies were evaluated with the CASPe tool. A thematic analysis was carried out taking into account criteria of relevance, coherence and triangulation of the results. To analyze the environmental impacts of the technology, companies were asked for information on environmental impact statements and other issues. Finally, research needs were assessed through a search of the James Lind Alliance (JLA), COMET and ICHOM databases.

Project Status: Completed

Year Published: 2023

URL for published report: https://sescs.es/informe-sistemas-de-monitorizacion-continua-de-glucosa-en-tiempo-real/

English language abstract: An English language summary is available

Publication Type: Full HTA

Country: Spain

Organisation Name: Canary Health Service

Contact Address: Dirección del Servicio. Servicio Canario de la Salud, Camino Candelaria 44, 1ª planta, 38109 El Rosario, Santa Cruz de Tenerife

Contact Name: sescs@sescs.es

Contact Email: sescs@sescs.es