Decision Support Systems and Closed Loop
2019; Mary Ann Liebert, Inc.; Volume: 21; Issue: S1 Linguagem: Inglês
10.1089/dia.2019.2504
ISSN1557-8593
AutoresRevital Nimri, Alexander Ochs, Jordan E. Pinsker, Moshe Phillip, Eyal Dassau,
Tópico(s)Hyperglycemia and glycemic control in critically ill and hospitalized patients
ResumoDiabetes Technology & TherapeuticsVol. 21, No. S1 Original ArticlesFree AccessDecision Support Systems and Closed LoopRevital Nimri, Alexander R. Ochs, Jordan E. Pinsker, Moshe Phillip, and Eyal DassauRevital NimriJesse Z and Sara Lea Shafer Institute for Endocrinology and Diabetes, Schneider Children's Medical Center of Israel, Petah Tikva, IsraelSearch for more papers by this author, Alexander R. OchsSansum Diabetes Research Institute, Santa Barbara, CASearch for more papers by this author, Jordan E. PinskerSansum Diabetes Research Institute, Santa Barbara, CASearch for more papers by this author, Moshe PhillipJesse Z and Sara Lea Shafer Institute for Endocrinology and Diabetes, Schneider Children's Medical Center of Israel, Petah Tikva, IsraelSackler Faculty of Medicine, Tel Aviv University, Tel Aviv, IsraelSearch for more papers by this author, and Eyal DassauSansum Diabetes Research Institute, Santa Barbara, CAHarvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MAJoslin Diabetes Center, Boston, MASearch for more papers by this authorPublished Online:20 Feb 2019https://doi.org/10.1089/dia.2019.2504AboutSectionsPDF/EPUB Permissions & CitationsPermissionsDownload CitationsTrack CitationsAdd to favorites Back To Publication ShareShare onFacebookTwitterLinked InRedditEmail IntroductionThis year was notable for great advances in the implementation of closed-loop systems for clinical use. The Medtronic Minimed 670G hybrid closed-loop (HCL) system is now in regular use in clinical practice in the United States. The Tandem predictive low-glucose suspend (PLGS) system, branded as Basal-IQ, uses the Tandem X2 insulin pump and the Dexcom G6 continuous glucose monitor (CGM). This system was approved by the U.S. Food and Drug Administration and then released to consumers in August 2018 after showing a 31% reduction in percent time below 70 mg/dL compared with sensor-augmented pump (SAP) therapy in a randomized crossover outpatient trial (1). Clinical trial results of using the Omnipod personalized model predictive control algorithm, planned for eventual use in the Insulet OmniPod® Horizon automated glucose control system, a single-hormone HCL system, have also been published (2,3). Multiple other commercial and academic long-term trials of closed-loop systems are under way, including four U.S. National Institutes of Health–awarded multicenter studies with sites from Europe and the United States (FlorenceM system/Cambridge, UK; inControl/Virginia, United States; iLet system/Boston, United States; and Flair/Medtronic & MD-Logic system) designed to be the potential last steps for regulatory approval of these systems. All aim to demonstrate clinical effectiveness to facilitate regulatory approval and future reimbursement of these devices (4). Closed-loop systems have now been tested in various settings, for hospitalized patients as well as throughout pregnancy to include delivery and the postpartum period (5,6).Additionally, this year's article adds a section on decision support systems. These tools can support treatment decisions for a wide range of patients, whether they are treated with injections or pump and monitor glucose with glucometer or CGM or any other modality, as some patients will not be able or want to use a closed-loop system. The systems include smart pens with automated titration and dosing adjustment recommendations, analysis tools built into web-based or app software that can make dosing recommendations for patients on multiple daily injections or using insulin pumps, or mobile software with glucose prediction capabilities. By providing automated tools to optimize clinical outcomes, decision support systems have the potential to improve clinical outcomes and may increase access to care as they become accessible to more patients, enhancing utilization of health-care resources by integrating e-health and telemonitoring programs. It should not be surprising that some closed-loop systems have integrated automated decision support systems, adjusting baseline basal rates based on feedback from the use of closed-loop control or by adding automated input from additional sensors such as heart rate and activity monitoring into their algorithms, and are listed as such below.Key Articles Reviewed for the ArticleAdjusting insulin doses in patients with type 1 diabetes that use insulin pump and continuous glucose monitoring: variations among countries and physiciansNimri R, Dassau E, Segall T, Muller I, Bratina N, Kordonouri O, Bello R, Biester T, Dovc K, Tenenbaum A, Brener A, Šimunović M, Sakka SD, Nevo Shenker M, Passone CG, Rutigliano I, Tinti D, Bonura C, Caiulo S, Ruszala A, Piccini B, Giri D, Stein R, Rabbone I, Bruzzi P, Omladič JŠ, Steele C, Beccuti G, Yackobovitch-Gavan M, Battelino T, Danne T, Atlas E, Phillip MDiabetes Obes Metab 2018;20: 2458–2466Continuous glucose monitoring and insulin informed advisory system with automated titration and dosing of insulin reduces glucose variability in type 1 diabetes mellitusBreton MD, Patek SD, Lv D, Schertz E, Robic J, Pinnata J, Kollar L, Barnett C, Wakeman C, Oliveri M, Fabris C, Chernavvsky D, Kovatchev BP, Anderson SMDiabetes Technol Ther 2018;20: 531–540Decision support in diabetes care: the challenge of supporting patients in their daily living using a mobile glucose predictorPérez-Gandía C, García-Sáez G, Subías D, Rodríguez-Herrero A, Gómez EJ, Rigla M, Hernando MEJ Diabetes Sci Technol 2018;12: 243–250Twelve-week 24/7 ambulatory artificial pancreas with weekly adaptation of insulin delivery settings: effect on hemoglobin A1c and hypoglycemiaDassau E, Pinsker JE, Kudva YC, Brown SA, Gondhalekar R, Dalla Man C, Patek S, Schiavon M, Dadlani V, Dasanayake I, Church MM, Carter RE, Bevier WC, Huyett LM, Hughes J, Anderson S, Lv D, Schertz E, Emory E, McCrady-Spitzer SK, Jean T, Bradley PK, Hinshaw L, Laguna Sanz AJ, Basu A, Kovatchev B, Cobelli C, Doyle III, FJDiabetes Care 2017;40: 1719–1726Randomized outpatient trial of single- and dual-hormone closed-loop systems that adapt to exercise using wearable sensorsCastle JR, El Youssef J, Wilson LM, Reddy R, Resalat N, Branigan D, Ramsey K, Leitschuh J, Rajhbeharrysingh U, Senf B, Sugerman SM, Gabo V, Jacobs PGDiabetes Care 2018;41: 1471–1477Closed-loop insulin delivery for glycemic control in noncritical careBally L, Thabit H, Hartnell S, Andereggen E, Ruan Y, Wilinska ME, Evans ML, Wertli MM, Coll AP, Stettler C, Hovorka RN Engl J Med 2018;379: 547–556Overnight glucose control with dual- and single-hormone artificial pancreas in type 1 diabetes with hypoglycemia unawareness: a randomized controlled trialAbitbol A, Rabasa-Lhoret R, Messier V, Legault L, Smaoui M, Cohen N, Haidar ADiabetes Technol Ther 2018;20: 189–196Closed-loop control during intense prolonged outdoor exercise in adolescents with type 1 diabetes: the artificial pancreas ski studyBreton MD, Cherñavvsky DR, Forlenza GP, DeBoer MD, Robic J, Wadwa RP, Messer LH, Kovatchev BP, Maahs DMDiabetes Care 2017;40: 1644–1650Overnight closed-loop control improves glycemic control in a multicenter study of adults with type 1 diabetesBrown SA, Breton MD, Anderson SM, Kollar L, Keith-Hynes P, Levy CJ, Lam DW, Levister C, Baysal N, Kudva YC, Basu A, Dadlani V, Hinshaw L, McCrady-Spitzer S, Bruttomesso D, Visentin R, Galasso S, Del Favero S, Leal Y, Boscari F, Avogaro A, Cobelli C, Kovatchev BPJ Clin Endocrinol Metab 2017;102: 3674–3682Closed-loop glucose control in young people with type 1 diabetes during and after unannounced physical activity: a randomised controlled crossover trialDovc K, Macedoni M, Bratina N, Lepej D, Nimri R, Atlas E, Muller I, Kordonouri O, Biester T, Danne T, Phillip M, Battelino TDiabetologia 2017;60: 2157–2167Fully closed-loop multiple model probabilistic predictive controller artificial pancreas performance in adolescents and adults in a supervised hotel settingForlenza GP, Cameron FM, Ly TT, Lam D, Howsmon DP, Baysal N, Kulina G, Messer L, Clinton P, Levister C, Patek SD, Levy CJ, Wadwa RP, Maahs DM, Bequette BW, Buckingham BADiabetes Technol Ther 2018;20: 335–343Impact of macronutrient content of meals on postprandial glucose control in the context of closed-loop insulin delivery: a randomized cross-over studyGingras V, Bonato L, Messier V, Roy-Fleming A, Smaoui MR, Ladouceur M, Rabasa-Lhoret RDiabetes Obes Metab 2018;20: 2695–2699Evaluation of an artificial pancreas with enhanced model predictive control and a glucose prediction trust index with unannounced exercisePinsker JE, Laguna Sanz AJ, Lee JB, Church MM, Andre C, Lindsey LE, Doyle III, FJ, Dassau EDiabetes Technol Ther 2018;20: 455–464Ketone production in children with type 1 diabetes, ages 4–14 years, with and without nocturnal insulin pump suspensionWadwa RP, Chase HP, Raghinaru D, Buckingham BA, Hramiak I, Maahs DM, Messer L, Ly T, Aye T, Clinton P, Kollman C, Beck RW, Lum J; for the In Home Closed Loop Study GroupPediatr Diabetes 2017;18: 422–427Predictive hyperglycemia and hypoglycemia minimization: in-home double-blind randomized controlled evaluation in children and young adolescentsForlenza GP, Raghinaru D, Cameron F, Bequette BW, Chase HP, Wadwa RP, Maahs DM, Jost E, Ly TT, Wilson DM, Norlander L, Ekhlaspour L, Min H, Clinton P, Njeru N, Lum JW, Kollman C, Beck RW, Buckingham BA; for the In-Home Closed-Loop (IHCL) Study GroupPediatr Diabetes 2018;19: 420–428Clinical Decision Support SystemsAdjusting insulin doses in patients with type 1 diabetes that use insulin pump and continuous glucose monitoring: variations among countries and physiciansNimri R1, Dassau E2, Segall T3, Muller I3, Bratina N4, Kordonouri O5, Bello R1, Biester T5, Dovc K4, Tenenbaum A1,6, Brener A1, Šimunović M7, Sakka SD8, Nevo Shenker M1, Passone CG9, Rutigliano I10, Tinti D11, Bonura C12, Caiulo S12, Ruszala A13, Piccini B14, Giri D15, Stein R16, Rabbone I11, Bruzzi P17, Omladič JŠ4, Steele C18, Beccuti G19, Yackobovitch-Gavan M1, Battelino T4,20, Danne T5, Atlas E3, Phillip M1,61Jesse Z and Sara Lea Shafer Institute for Endocrinology and Diabetes, National Center for Childhood Diabetes, Schneider Children's Medical Center of Israel, Petah Tikva, Israel2Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA3DreaMed-Diabetes Ltd, Petah Tikva, Israel4Department of Pediatric Endocrinology, Diabetes and Metabolism, University Medical Centre – University Children's Hospital, Ljubljana, Slovenia5Diabetes Centre for Children and Adolescents, Auf der Bult, Kinder- und Jugendkrankenhaus, Hannover, Germany6Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel7Department of Pediatrics, University Hospital Centre Split, Split, Croatia8Department of Endocrinology and Diabetes, Evelina London Children's Hospital, London, UK9Instituto da Criança – HCFMUSP, University of Sao Paulo, Sao Paulo, Brazil10Pediatrics IRCCS Casa Sollievo della Sofferenza, San Giovanni Rotondo, Italy11Centre of Pediatric Diabetes, Department of Pediatrics, University of Turin, Turin, Italy12San Raffaele Hospital, Vita-Salute San Raffaele University, Milan, Italy13Department of Pediatric and Adolescent Endocrinology, Institute of Pediatrics, Jagiellonian University Medical College, Krakow, Poland14Diabetology Unit, Meyer Children's Hospital, Florence, Italy15Department of Paediatric Endocrinology and Diabetes, Bristol Royal Hospital for Children, Bristol, UK16Paediatric Endocrinology and Diabetes Unit, Dana-Dwek Children's Hospital, Sourasky Medical Centre, Tel Aviv, Israel17Departments of Medical and Surgical Sciences of Mothers, Children and Adults, Azienda Ospedaliero-Univeristaria of Modena Policlinico, Paediatric Unit, Modena, Italy18Paediatric Endocrinology and Diabetes, Leeds Children's Hospital, Leeds, UK19Division of Endocrinology, Diabetes and Metabolism, Department of Medical Sciences, University of Turin, Turin, Italy20Faculty of Medicine, University of Ljubljana, Ljubljana, SloveniaDiabetes Obes Metab 2018;20: 2458–2466BackgroundThere can be considerable variability among physicians when making recommendations for adjusting insulin pump settings based on clinical analysis of CGM, insulin pump, and glucometer data for patients with type 1 diabetes. Automated insulin adjustments based on computerized analysis of these data offer the potential of greater access to care and reduced variability in recommended dose changes.MethodsTwenty-six physicians from 16 centers across Europe, Israel, and South America were asked to adjust insulin dosing based on device download data from 15 patients (mean age 16.2±4.3 years, 6 female, mean glycated hemoglobin [HbA1c] 8.3±0.9%) gathered over a 3-week period. These recommendation adjustments were compared for relative changes in the basal rates, carbohydrate/insulin ratio (CR), and correction factor (CF) among physicians and among centers, as well as between doctors and DreaMed Advisor Pro, an automated algorithm. Results were calculated from the percentage of comparison points for which all methods agreed on the trend of insulin dose adjustments (same trend) and those for which there was partial agreement (increase/decrease vs no change) and full disagreement (opposite trends).ResultsFull agreement between physicians on the trend of insulin adjustments occurred at a rate of 41±9% for basal, 45±11% for CR, and 45.5±13% for CF plans. Complete disagreement percentages were 12±7%, 9.5±7%, and 10±8%, respectively. Similar results were found when comparing the physicians and Advisor Pro. The magnitude of the Advisor insulin dose changes was never greater than those proposed by physicians.ConclusionsThe automated advice of the DreaMed Advisor Pro did not differ significantly from the advice given by the physicians in the direction or magnitude of the insulin dosing.CommentsAutomated analysis of diabetes device data (insulin pump, CGM, and glucometer) holds the promise of vastly increasing access to quality medical care for patients with diabetes. In this trial, the DreaMed Advisor Pro analyzed the same dataset as the physicians and came up with similar dose adjustment recommendations. Recently, the DreaMed Advisor Pro was approved by the U.S. Food and Drug Administration for integration with the popular Glooko software package, helping health-care providers make recommendations for changes in insulin pump therapy (7). This is the first step in moving toward increasing levels of automated dose advising, with the ultimate goal that patients can have the advisor software periodically analyze their device data, rather than waiting for clinician input to make adjustments, thus improving clinical care.Continuous glucose monitoring and insulin informed advisory system with automated titration and dosing of insulin reduces glucose variability in type 1 diabetes mellitusBreton MD, Patek SD, Lv D, Schertz E, Robic J, Pinnata J, Kollar L, Barnett C, Wakeman C, Oliveri M, Fabris C, Chernavvsky D, Kovatchev BP, Anderson SMCenter for Diabetes Technology, University of Virginia, Charlottesville, VADiabetes Technol Ther 2018;20: 531–540BackgroundGlucose variability (GV) is a key limiting factor in successful diabetes management. Use of CGM and connected insulin delivery devices offers potential for expert systems to analyze this data and may further improve glucose outcomes and GV beyond use of these systems alone.MethodsA total of 24 patients with type 1 diabetes (T1D) (15 women, 37±11 years of age, HbA1c 7.2% ±1%, total daily insulin 46.7±22.3 U) using either an insulin pump or multiple daily injections with carbohydrate counting completed the study carried out over two randomized crossover 48-h visits at the University of Virginia. Each patient wore a Dexcom G4 CGM and used either usual care or the UVA decision support system (DSS). DSS comprised a combination of automated insulin titration, bolus calculation, and carbohydrate treatment advice. Patients were exposed to meals of various size and contents in addition to two 45-min bouts of exercise. GV and glucose control outcomes were measured using CGM.ResultsThe use of DSS significantly reduced GV (coefficient of variation 0.36±0.08 vs 0.33±0.06; P=0.045) and maintained glycemic control (average CGM 155.2±27.1 mg/dL vs 155.2±23.2 mg/dL) by reducing exposure to hypoglycemia ( 250 mg/dL) overnight (5.3%±9.5% vs 1.9% ±4.6%) and at mealtime (11.3%±14.8% vs 5.8% ±9.1%).ConclusionsThe device-informed advisory system was shown to be safe and feasible in a cohort of 24 subjects with T1D. Use of the system may reduce GV and improve protection against hypoglycemic events.CommentThe decision support systems used in this study consisted of two real-time advisors (CGM-Informed Bolus Advisor and Exercise Advisor) and a retrospective insulin titration tool. The system was able to significantly reduce GV, likely through the reduction of exposure to hypoglycemia, without increasing average glycemia or exposure to hyperglycemia. Remarkably, it did this in a very short time (in a 48-hour crossover study). Thus, it appears that automated insulin titration, coupled with dosing and hypoglycemia real-time advice based on CGM, is safe, feasible, and may positively impact glucose control in T1D subjects using continuous subcutaneous insulin infusion or multiple daily injections.Decision support in diabetes care: the challenge of supporting patients in their daily living using a mobile glucose predictorPérez-Gandía C1,2, García-Sáez G1,2, Subías D3, Rodríguez-Herrero A1,2, Gómez EJ1,2, Rigla M3, Hernando ME1,21CIBER-BBN: Networking Research Centre for Bioengineering, Biomaterials and Nanomedicine, Madrid, Spain2Centro de Tecnología Biomédica, ETSI de Telecomunicación, Universidad Politécnica de Madrid, Spain3Endocrinology and Nutrition Department, Parc Tauli Sabadell University Hospital, Sabadell, SpainJ Diabetes Sci Technol 2018;12: 243–250BackgroundIndividuals with T1D actively manage their condition and need to have the knowledge to make decisions fitting their day-to-day insulin requirements. Artificial intelligence applications can aid decision making for patients and allow them to adjust more quickly than scheduled face-to-face visits. This work presents a DSS based on glucose prediction that assists patients via a mobile phone platform.MethodsThe system's impact on therapeutic corrective actions was evaluated through a randomized crossover pilot study focusing on between-meal periods. Twelve T1D subjects using insulin pumps participated in two phases. During the experimental phase participants used the DSS to alter initial corrective decisions in the presence of hypoglycemic or hyperglycemic events. During the control phase patients were directed to make corrective decisions without the DSS glucose prediction. A telemedicine interface allowed participants to record glucose monitoring data and decisions while endocrinologists also supervised data from the hospital. The study period was defined as a postprediction (PP) time window.ResultsWhen provided with the glucose prediction, patients modified their initial decision 20% of the time. No statistically significant differences were found in the PP Kovatchev's risk index change (ΔRI calculated for 1 h before the start and 1 h before the end of the postprandial time window): −1.23±11.85 in experimental phase vs −0.56±6.06 in control phase. In a usability questionnaire after the study, participants expressed positive opinions about the DSS, assigning it an average score higher than 7 (out of 9 total).ConclusionsThe DSS had a relevant impact in the participants' decision making while aiding T1D management and showed a high confidence of patients in the use of glucose prediction.CommentsAlthough not achieving clinically significant improvements in glycemic results, this study highlights the importance of patient acceptance in clinical DSS recommendations. Based on the system recommendations, participants modified their initial decision 20% of the time. They also had a positive opinion about the DSS, with a high average score in a usability questionnaire despite having only limited use of the system. These usability issues and learning to have confidence in the system will be of paramount importance as more patients gain access to DSSs for home use.Closed-Loop Systems With Automated Decision Support SystemsTwelve-week 24/7 ambulatory artificial pancreas with weekly adaptation of insulin delivery settings: effect on hemoglobin A1c and hypoglycemiaDassau E1,2, Pinsker JE2, Kudva YC3, Brown SA4, Gondhalekar R1,2, Dalla Man C5, Patek S4, Schiavon M5, Dadlani V3, Dasanayake I2,6, Church MM2, Carter RE7, Bevier WC2, Huyett LM2,6, Hughes J4, Anderson S4, Lv D4, Schertz E4, Emory E4, McCrady-Spitzer SK3, Jean T2, Bradley PK2, Hinshaw L3, Laguna Sanz AJ1,2, Basu A3, Kovatchev B4, Cobelli C5, Doyle III, FJ1,21Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA2William Sansum Diabetes Center, Santa Barbara, CA3Endocrine Research Unit, Mayo Clinic, Rochester, MN4Center for Diabetes Technology, University of Virginia, Charlottesville, VA5Department of Information Engineering, University of Padova, Padua, Italy6Department of Chemical Engineering, University of California Santa Barbara, Santa Barbara, CA7Department of Health Sciences Research, Mayo Clinic, Rochester, MNDiabetes Care 2017;40: 1719–1726BackgroundBecause insulin analogs take time to reach peak serum concentration, adjustment of basal rate (either reduction/suspension or increase in insulin delivery) is not always enough to compensate for inaccuracies in meal bolus dosing, exercise, illnesses, stress, or other activities that change insulin sensitivity. The authors hypothesized that having patients on near optimal basal rates helps the artificial pancreas (AP) algorithms to be as effective as possible, as they work within a given constraint when reducing/suspending or giving extra insulin. To that end, they developed and tested a novel adaptive AP in an outpatient, single-arm, uncontrolled multicenter clinical trial lasting 12 weeks.MethodsThirty adults with T1D completed a 1-week sensor-augmented pump (SAP) run-in period. After the AP was started, basal insulin delivery settings used by the AP for initialization were adapted weekly, and carbohydrate ratios were adapted every 4 weeks by an algorithm running on a cloud-based server, with automatic data upload from devices. Adaptation recommendations were reviewed by expert study clinicians and participants before being implemented. The primary endpoint was change in HbA1c.ResultsTwenty-nine patients completed the trial. Mean HbA1c was 7.0±0.8% at the start of AP use and decreased to 6.7±0.6% after 12 weeks (−0.3 [95% CI −0.5 to −0.2]; P<0.001). Compared with the SAP run-in, CGM time spent in the hypoglycemic range improved during the day from 5.0% to 1.9% (−3.1 [95% CI −4.1 to −2.1]; P<0.001) and overnight from 4.1% to 1.1% (−3.1 [95% CI −4.2 to −1.9]; P<0.001). Approximately 10% of adaptation recommendations were manually overridden by study physicians.ConclusionsUse of the novel adaptive AP yielded significant reductions in HbA1c and hypoglycemia.CommentsIn this study, a cloud-based system seamlessly integrated data from AP use into an automated adaptation framework for basal rates and carbohydrate ratios throughout the 12-week period, potentially obviating the need for clinician involvement prior to or during use of AP to optimize open-loop settings. Percent time with glucose <70 mg/dL significantly decreased during the day and overnight, while at the same time HbA1c decreased (−0.3% [95% CI −0.5 to −0.2]; P<0.001). Although single arm and uncontrolled, this study showed automated adaptations of insulin pump settings can be performed safely and effectively to improve AP performance. Further studies are needed to best delineate how often these optimizations should occur and if they can be done in a completely automated manner without physician review in the future.Randomized outpatient trial of single- and dual-hormone closed-loop systems that adapt to exercise using wearable sensorsCastle JR1, El Youssef J1,2, Wilson LM1, Reddy R2, Resalat N2, Branigan D1, Ramsey K3, Leitschuh J2, Rajhbeharrysingh U1, Senf B1, Sugerman SM1, Gabo V1, Jacobs PG21Harold Schnitzer Diabetes Health Center, Division of Endocrinology, Oregon Health and Science University, Portland, OR2Department of Biomedical Engineering, Oregon Health and Science University, Portland, OR3Oregon Clinical and Translational Research Institute Biostatistics and Design Program, Oregon Health and Science University, Portland, ORDiabetes Care 2018;41: 1471–1477This manuscript is also discussed in the article on Advances in Exercise, Physical Activity, and Diabetes Mellitus, page S-112.BackgroundExercise-related hypoglycemia remains a challenge, even with the use of AP systems. This study aimed to determine whether a dual-hormone closed-loop system with wearable sensors to detect exercise and adjust dosing to reduce exercise-related hypoglycemia would outperform other forms of closed-loop and open-loop therapy.MethodsThe study consisted of four arms—dual-hormone, single-hormone, PLGS, and continuation of current care—which participants completed in randomized order over 4 outpatient days. Each arm included three moderate-intensity aerobic exercise sessions. Physical activity and heart rate were captured with the ZephyrLife BioPatch and were incorporated into the AP algorithms during the study period. Primary outcomes were percent time in hypoglycemia (<70 mg/dL) and percent time in target blood glucose range (70–180 mg/dL) assessed across the entire study and from the start of the in-clinic exercise until the next meal.ResultsTwenty adults with T1D completed all four arms. Mean time (SD) in hypoglycemia was lowest with dual-hormone during the exercise period: 3.4% (4.5) vs 8.3% (12.6) with single-hormone (P=0.009) vs 7.6% (8.0) with PLGS (P<0.001) vs 4.3% (6.8) with current care, allowing preexercise insulin adjustments (P=0.49). Across the entire study, time in hypoglycemia was the lowest with dual-hormone treatment as well: 1.3% (1.0) vs 2.8% (1.7) for single-hormone treatment (P < 0.001) vs 2.0% (1.5) for PLGS (P=0.04) vs 3.1% (3.2) for current care (P=0.007). Time in range during the entire study was the highest with single-hormone vs dual-hormone treatment: 74.3% (8.0) vs 72.0% (10.8) (P=0.44).ConclusionsIn physically active adults with T1D, the addition of glucagon to a closed-loop system with automated exercise detection resulted in less time spent in hypoglycemia.CommentsThis study shows the potential of adding additional signals such as heart rate and activity monitoring into AP, comparing all modern modalities of insulin delivery systems, to include dual-hormone AP, single-hormone AP, PLGS, and SAP with dosing adjustments allowed prior to exercise. In a prior study without automated exercise detection, the authors showed that both dual-hormone AP as well as SAP with dosing adjustment allowed in advance of exercise could reduce the incidence of hypoglycemia equally as well (8). In this study, with the addition of automated exercise detection, the dual-hormone AP performed best, minimizing hypoglycemia both during exercise and during the entire 4-day period, highlighting how well AP can function with additional signals added as input into the AP algorithm. Future studies with more portable exercise monitors (such as wristwatches) that are conformable for patient wear will allow for long-term use of these devices in the outpatient setting.Closed-Loop SystemsClosed-loop insulin delivery for glycemic control in noncritical careBally L1,2,3, Thabit H3,6,7, Hartnell S5, Andereggen E1, Ruan Y3, Wilinska ME3,4, Evans ML3,5, Wertli MM2, Coll AP3,5, Stettler C1, Hovorka R3,4Departments of 1Diabetes, Endocrinology, Clinical Nutrition, and Metabolism and 2General Internal Medicine, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland3Wellcome Trust–MRC Institute of Metabolic Science and the 4Department of Pediatrics University of Cambridge, Cambridge, UK5Wolfson Diabetes and Endocrine Clinic, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK6Manchester University Hospitals NHS Foundation, Manchester Academic Health Science Centre, and the 7Division of Diabetes, Endocrinology, and Gastroenterology, Faculty of Biology, Medicine, and Health, University of Manchester, Manchester, UKN Engl J Med 2018;379: 547–556BackgroundFor patients with diabetes, hospitalization can lead to poor glycemic control, adversely affecting outcomes. There already exists a significant evidence base for AP improving glycemic control in both the inpatient and outpatient settings in patients with T1D. This study investigated whether a closed-loop system could also improve glycemic control in hospitalized patients with type 2 diabetes (T2D) that were receiving noncritical care.MethodsThis randomized, open-label trial was conducted on general wards in two tertiary hospitals in the United Kingdom and Switzerland. A total of 136 adults with T2D who required subcutaneous insulin therapy were assigned to either the closed-loop insulin delivery group (70 patients) or conventional subcutaneous insulin therapy group (66 patients), according to local clinical practice. The primary endpoint was the percentage of time that the sensor glucose measurement was within the target range of 100–180 mg/dL (5.6–10.0 mmol/L) for up to 15 days or until hospital discharge.ResultsMean (±SD) percent time in the target range was 65.8±16.8% in the closed-loop group and 41.5±16.9% in the control group, a difference of 24.3±2.9 percentage points ([95% CI 18.6–30.0]; P<0.001). Values above the target range were found in
Referência(s)