Introduction

Diabetes mellitus is one of the leading causes of chronickidney disease (CKD), with approximately 30 to 40% ofpeople living with diabetes developing CKD [1–3]. In the adult population, diabetes mellitus remains a primary cause of end-stage kidney disease (ESKD), a condition character-ized by the irreversible failure of kidney function, necessi-tating dialysis or transplantation for survival [1, 3].

    Glycaemic management in patients with ESKD is chal-lenging because of defects in glucose and insulin metab-olism due to uremia; decreased gluconeogenesis and impaired insulin clearance by the kidney, resulting also in greater susceptibility to iatrogenic hypoglycaemia; impaired counterregulatory hormone responses (i.e. cortisol, growth hormone); increased erythrocyte glucose uptake during hemodialysis; and the underlying mild inflammatory state, which may predispose patients to hyperglycaemia [4, 5].

    People with diabetes mellitus and ESKD have a higher risk of developing severe hypoglycaemic events and hyper-glycaemic crises requiring emergency room visits and/or hospitalisation than individuals at high risk non-ESKD [4]. Hypoglycaemic episodes are associated with a higher risk of recurrent hypoglycaemia and mortality after initiation of dialysis and also impose a significant burden of care [6].

    Automated insulin delivery (AID) systems have revo-lutionized diabetes care, particularly in people with type 1 diabetes (T1D) [3]. Meta-analyses undertaken from trial data show that AID systems outperform non-automated sys-tems with improvements in time spent in the target glucose range of approximately 8–12% points, reduced time spent in hyperglycaemia, reduced mean glucose and either a reduc tion or no increase in time in hypoglycaemia [3, 7]. Exten-sive evidence supports the benefits of current AID systems in improving glycaemic management, decreasing hypo-glycaemia risk and fear of hypoglycaemia, and improving Quality of Life (QOL) for people with diabetes mellitus [3, 7–14].

    A recent Clinical Practice Consensus document recom-mends the use of AID in patients with chronic renal failure, even if the evidence on the use of these devices in people with diabetes mellitus and ESKD is still limited [15]. Two small studies conducted in adults with type 2 diabetes (T2D) receiving haemodialysis showed significant improvements in Time in range (TIR) without increasing the risk of hypo glycaemia in patients in treatment with AID compared with conventional insulin therapy [16, 17]. These small studies demonstrate the glycaemic benefit of AID in adults with T2D undergoing haemodialysis, but further investigations are needed in this population, as well as in T1D population receiving haemodialysis and in individuals receiving peri-toneal dialysis.

    The primary objectives of this longitudinal retrospective pilot real-world study were to describe the glycaemic out-comes and the QOL observed in a cohort of patients with diabetes mellitus undergoing dialysis using AID. The sec-ondary outcome was to describe the AID settings used in subjects undergoing dialysis.

Materials and methods

Study design and setting

This is a retrospective multicentric real-world analysis of anonymized data of T1D or T2D insulin-treated individuals undergoing dialysis, who began AID therapy to optimize glycaemic control. The pilot real-world study was con-ducted at the Department of Medical and Surgical Sci -ences of the University “Magna Graecia” of Catanzaro, the ASP of Catanzaro (Italy), and the Annunziata hospital in Cosenza (Italy). AID therapy was proposed to T1D or T2D  insulin -treated subjects undergoing hemodialysis or peri-toneal dialysis, without a history of dementia or cognitive impairment. Other therapy exclusion criteria were non-self-sufficient or elderly people (>75 years) without a dedicated caregiver, subjects with only basal insulin (<10 units/day), recent hospitalization for acute events, presence of acute infections or malignant diseases. All subjects in the analysis were treated with the MiniMed™ 780G system for at least 3 months.

    All the individuals received education on diabetes man-agement and on the use of the AID system, accurately assessed and periodically reassessed by diabetologists.

    This retrospective analysis was approved by the Regional Territorial Ethics Committee (Comitato Etico Territoriale Regione Calabria, approval code: 213/2025) and was con-ducted in accordance with the 1964 Declaration of Helsinki and its subsequent amendments.

Onboarding and follow-up protocol

Subjects with sub-controlled diabetes undergoing hemodi-alysis or peritoneal dialysis were identified by nephrologists and referred to diabetologists for consultation. After being deemed suitable for AID therapy, they were first educated on using CGM only (Medtronic Guardian™ 4 sensor) at the dialysis clinics. After two weeks, they were re-evaluated and, if deemed ready, started using the insulin pump. For 14 days, they used the manual mode, followed by activa-tion of the SmartGuard™ feature. In all users, the following system settings were used: glucose target=100 mg/dL and AIT=2.30 h, Patients were then followed up at the dialysis clinics at 3, 6, and 12 months by diabetologists.

    The same protocol had been followed for subjects who underwent peritoneal dialysis, but the diabetological ambu-latories were used instead of the dialysis ones.

Laboratory determinations

HbA1c was measured with high-performance liquid chro-matography using an NGSP-certified automated analyzer (Adams HA-8160 HbA1c analyzer, Menarini, Italy).

Glucose metrics

Glucose data were extracted from Medtronic CareLink™ Clinic, a dedicated web-cloud platform (https://carelink.mi nimed.eu/app/login).

    Glucose metrics, such as mean sensor glucose (SG), TIR (70–180 mg/dL), TBR (<70 mg/dL), TAR (>180 mg/dL), TAR180 (180–250 mg/dL), TAR250 (>250 mg/dL), and GMI, were collected at baseline, 3, 6, and 12 months after initiating the SmartGuard™ feature. The baseline was con-sidered a 14-day run-in period during which the subjects used only the CGM.

Quality of Life

At the last follow-up, the validated Italian version of the WHOQoL-Bref questionnaire was administered to assess the perceived QOL in this fragile population. This question-naire was chosen because it compares the current therapies with the previous one and measures changes across time in the impact of disease and impairment on daily activities.

Statistical analysis

Descriptive statistics were used to summarize results. These include mean and SD for continuous variables and counts and percentages for categorical variables. Summary statis-tics were reported with a maximum of 2 decimal places, as appropriate. Mean changes for each pairwise period comparison were estimated using linear mixed models, to account for the within-patient correlation. Estimates along with their 95% Confidence Intervals (CIs) are provided. SAS software, version 9.4, (SAS Institute Inc., Cary, NC, USA) was used to perform all statistical analyses.

Results

Twenty adult subjects were started on CGM and among these, fourteen started AID therapy and were included in this analysis. Six refused an insulin pump or were not suit-able for AID and continued to use CGM only.

    Out of the 14 people, 8 reached the 12-month follow-up. The remaining 6 subjects were enrolled at a later date, and the timing of their enrolment in the study did not allow them to reach the 12-month follow-up in the data analyses.

    The mean age was 60.6 years, with females comprising 21.4% (n=3) of the cohort. The average duration of diabetes was 21.9 years. Most of the subjects were affected by T2D (85.7%). The average HbA1c was 7.6% and the average BMI was 29.7 kg/m2 . Most of them had at least one com-plication: in particular, 71.4% of the subjects had a history of cardiovascular disease; 50% had retinopathy: and 21.4% had diabetic foot. An estimated 28.6% of the subjects in the study had peripheral artery disease and 21.4% had a history of major amputation. Of these, 71.4% had hypertension and 57.1% had dyslipidemia. Most of the subjects had a second-ary school diploma (63.7%) and most of them had a care-giver (71.4%). Thirteen subjects underwent hemodialysis, and one underwent peritoneal dialysis. The average duration of dialysis was 4.3 years, and the majority followed a three-times-week dialysis rhythm (92.3%). All baseline charac-teristics are summarized in Table 1. During the follow-up, no hospitalization, acute diabetes-related complications or adverse events were reported among the patients in study.

Glycemic outcomes

Fourteen subjects with diabetes mellitus undergoing dialysis using the AID system exhibited improvement in glycemic outcomes over time. Interestingly, we observed clinically meaningful improvements in glycemic parameters after just 3 months of use of the AID system. Notably, patients exhibited a clinically relevant increase of TIR from 63% at

baseline to 72.4% at 3 months (mean change 9.4% [95%CI − 1.1, 19.8]), and a reduction of TAR from 36.5% to 27.4% (mean change−9.3% [95% CI − 19.7, 1.2]), TAR180 from 29.8% to 23.4% (mean change−6.5% [95% CI −14.2, 1.2]), TAR 250 from 6.7% to 4% (mean change−2.7% [95% CI − 7.6, 2.2]), and TBR from 0.3% to 0.2% (mean change−0.1% [95% CI − 0.4, 0.1]) (Fig. 1; Tables 2 and 3).

    Furthermore, after 3 months of AID treatment, patients showed a decrease of GMI from 7.4% at baseline to 7.1% (mean change−0.3% [95% CI − 0.6, 0.0]), sensor glucose (mean change−13.2 mg/dL [95% CI − 6.8, 5.0]) and CV

(mean change−0.2% [95% CI − 4.1, 3.8]) (Fig. 1; Tables 2 and 3).

    In subjects who reached the 12-month follow-up, we observed a continued improvement in glycemic outcomes with a clinically meaningful increase of TIR from 63% at baseline to 69% (mean change 6.7% [95% CI − 5.5, 18.9]), and reduction of TBR from 0.3% to 0.0% (mean change−0.3% [95% CI − 0.05, 0.0]), TAR from 36.5% to 30.8% (mean change−6.4% [95% CI − 18.7, 5.8]), TAR180 from 29.8% to 26.9% (mean change−3.1% [95% CI − 11.9, 5.8]), and TAR 250 from 6.7% to 3.9% (mean change−3.2% [95% CI − 8.9, 2.5]) (Fig. 1; Tables 2 and 3).

    Moreover, we observed a decrease of GMI from 7.4% at baseline to 7.1% (mean change−0.2% [95% CI − 0.6, 0.0]), sensor glucose (mean change−10 mg/dL [95% CI − 26, 6]) and CV (mean change−1.3% [95% CI − 5.7, 3.2]) (Fig. 1; Tables 2 and 3).

    At the 12-month follow-up, seven out of eight sub-jects achieved all three International Glycemic Targets for this fragile population (TIR>50%, TBR<1% and TAR250<10%) (18).

Glycemic outcome during dialysis vs. non-dialysis days

We also compared glycaemic parameters during the dialysis days with the non-dialysis days in all the AID system users, who underwent hemodialysis and had at least one follow-up, considering the last 14 days of sensor data extracted from CareLink™ Clinic.

    Out of 11 subjects that underwent hemodialysis and who had at least one follow-up visit, 10 of them (90.9%) under-went dialysis for 6 days out of the 14-days period consid-ered. Only one of them (9.1%) underwent dialysis only for 4 out of the 14 days considered. For the sake of the analysis, glycemic outcomes were stratified first by dialysis vs. non-dialysis day, and then averaged at the patient level.

    We found that TIR was higher (72.4% vs. 63.2%) and TAR was lower (23.4% vs. 29.8%) during the dialysis days compared with non-dialysis days, whereas no differences were observed in TBR (Fig. 2).

Setting of AID system in subjects undergoing dialysis

All the subjects with diabetes mellitus undergoing dialy-sis using the AID system considered in this analysis used a glycemic target of 100 mg/dl, an active insulin time (AIT) before and during dialysis of 2.30 h, while no Temporary Target was used during dialysis days.

Quality of Life evaluation

At the last follow-up, all subjects completed the WHOQOL-BREF questionnaire, regardless of the month of follow-up. After starting to use the AID system, 91.7% of the users indicated a good, or very good QOL, whereas only 8.3% had a good QOL and no one had a very good QOL with the previous treatment (Fig. 3a). Moreover, 58.3% of the users were satisfied or very satisfied with their health status versus only 25% before using AID system (Fig. 3a). Additionally, there was an increase in physical, psychological and envi-ronment scores in all subjects after use of the AID system compared with the previous treatment (Fig. 3b).

Discussion

This longitudinal retrospective multicentric pilot real-world study provides evidence that in patients with diabetes under-going dialysis the use of the AID system is associated with an improvement of glycaemic outcomes and overall QOL. Notably, after just 3 months of use of AID, we observed clinically meaningful improvements in CGM metrics with a decrease of GMI of 0.3%, and an increase of the Time in range of 9.4%. Moreover, we found a clinically relevant reduction of TAR of 9.3% and TBR of 1%. This improve-ment in glycaemic parameters was maintained over time. Particularly, almost all subjects who reached the 12-month follow-up achieved all three CGM metrics targets recom mended for this fragile population [18]. Specifically, we observed a clinically meaninful reduction of TBR from 0.3% to 0.0% over time with no hypoglycaemic episodes detected by CGM or reported by the users.

    These findings hold clinical significance as hypoglycae-mia presents a substantial obstacle to the optimization of insulin treatment. People with diabetes mellitus undergoing dialysis have high risk of developing severe hypoglycae-mic events because of impaired kidney gluconeogenesis, decreased metabolism and clearance of insulin, coexistence of comorbidities (e.g., diabetic gastroparesis and malnu-trition), intradialytic glucose shifts into erythrocytes dur-ing hemodialysis treatment, use of lower dialysate glucose concentrations over time, and limited access to food dur-ing in-center hemodialysis, and frequently exhibit impaired awareness of hypoglycemia [4–7, 19]. Hypoglycemia is even associated with a higher risk of recurrent hypogly cemia and mortality after initiation of dialysis and also imposes a significant burden of care [6].

    More generally, glycemic management in patients with dialysis is challenging and individuals have a risk of hyper-glycemic crises requiring hospitalization due to increased insulin resistance, impaired insulin secretion, post-dialysis rebound hyperglycemia and a counter-regulatory hormone

response in response to the hypoglycemia during the hemo-dialysis session, and exposure to high dialysate glucose concentrations in those receiving peritoneal dialysis [4, 19]. Hypoglycemic and hyperglycemic episodes that frequently occur during and after dialysis are often unrecognised.

    In our study, we had also compared glycaemic param-eters during the dialysis days with non-dialysis days. We observed that in all the AID system users who underwent hemodialysis TIR was higher (72.4% vs. 63.2%) and TAR was lower (23.4% vs. 29.8%) during the dialysis days com-pared with non-dialysis days. Interestingly, no time spent in hypoglycemia was observed during dialysis days. A plausi-ble explanation for the enhanced glycemic control observed could be attributed to the influence of dialysate glucose con-centration on glycemic values, leading to an adaptation of the pump’s automatic insulin delivery algorithm. Additionally, dietary intake during dialysis differs from non-dialysis days. Furthermore, the insulin delivery automation of the AHCL system could mitigate post-dialysis hyperglycemia. These findings underscore the significance of an AID system that continuously adapts to glucose fluctuations during and after dialysis treatment, as well as varying insulin requirements from day to day. Moreover, this analysis provides evidence of the safety and the efficacy of the use of ambitious settings (i.e., glycemic target=100 mg/dL and AIT=2.30 h), also in vulnerable patients, such as those undergoing dialysis.

    Our data confirm and expand the results of two previ-ous studies demonstrating significant improvements in TIR without increasing the risk of hypoglycaemia in adults with T2D in haemodialysis treated with the AID system [16, 17]. Additionally, this study presents evidence in a setting of patients for whom there is limited data on the utilization of AID systems, including, albeit in small numbers, those on dialysis with T1D and one patient undergoing peritoneal dialysis. Furthermore, our findings align with prior evidence demonstrating the attainment of glycemic control targets with the AID system in the general diabetic population [20].

    It should be also noted that, at 12 months, in AID users undergoing dialysis, no adverse events and hospitalizations were reported.

    It is also truly surprising to note the findings of the enhancement of QOL after the use of the AID system. Nota-bly, the WHOQol-Bref questionnaire submitted to all sub-jects undergoing dialysis revealed that 91.7% of the AID users indicated a good, or very good QOL, whereas only 8.3% had a good QOL and no one had a very good QOL with the previous treatment. Moreover, 58.3% of the users were satisfied or very satisfied with their health status versus only 25% before using the AID system. Additionally, there was an increase in physical, psychological and environment scores in all subjects after use of the AID system compared with previous treatment. These findings hold greater signifi-cance given the substantial patient burden associated with dialysis, particularly considering the presence of numerous comorbidities [6, 7]. The use of the AID system improved QOL independently of improvements in glycaemic control, as it enabled PWD to reduce distress related to hypo and hyperglycaemia, as well as to decrease the number of injec-tions and fingersticks. Indeed, most of the patients had at least one complication; in particular, 71.4% of the subjects had a history of cardiovascular disease, 21.4% had diabetic foot and history of major amputation. It becomes even more important in people with many comorbidities reach glyce- mic target without hypoglycemia and adverse events and simplify the management of diabetes especially in dialysis

    Furthermore, the clinical pathway adopted, which con-sists of a close collaboration between the department of nephrology and diabetologists, was crucial to reduce patients’ burden and improve their acceptance of the new AID therapy; as they were treated only in the dialysis clinic during the whole process of enrollment and follow-up. Our approach could form a model to be followed also in other healthcare systems for treating more patients, improving the management of diabetes mellitus and developing a new standard of care for these people.

    This study has some strengths that merit consideration. Our group includes people with T2D and, although fewer in number, also people with T1D on haemodialysis and one patient on peritoneal dialysis, all for whom further investi-gations on the use of the AID system are needed. Moreover, our analysis includes glycaemic evaluation, also during dial-ysis sessions. Furthermore, all the analysis and the WHO-Qol-Bref questionnaires were collected by examiners who were blinded to the clinical data of the study participants.

    The present study also has some limitations. First, the retrospective design and the small sample size do not permit any causal inferences. Additionally, the analysis includes only Caucasian individuals, thus limiting the generaliz-ability of the present results to other ethnicities. However, our model of close collaboration between the department of nephrology and diabetologists could be followed also in other healthcare systems. Moreover, the WHOQol-Bref questionnaires, collected only at the last follow-up, may have caused some bias. Further randomized controlled tri-als and large well-designed real-world studies are needed to confirm these findings.

    In conclusion, this analysis has shown how the use of the AID system is safe and can help to improve glycaemic outcomes, with no hypoglycaemia observed in patients with diabetes undergoing hemodialysis. Additionally, the use of the AID system is associated with a consistent improvemen

of QOL scores in these patients. It is imperative that more attention be paid to the organizational model for the treat-ment of this fragile population.

Author contributions Study concept and design: E.S., G.C. and R.M. Acquisition of data: E.S., G.C., P.S., F.B., I.G., V.M., G.F., F.I., R.A., G.M., R.N., M.C., G.Z., M.A., R.M. and F.A. Analysis of data, Writ-ing and editing the manuscript: E.S., G.C. and R.M. Critical revision of the manuscript: P.S., F.B., I.G., V.M., G.F., F.I., R.A., G.M., R.N., M.C., G.Z., M.A., F.A. All authors have read and approved the final

Funding Open access funding provided by Università degli studi "Magna Graecia" di Catanzaro within the CRUI-CARE Agreement. This research received no external funding.

Declarations

Conflict of interest The authors declare that they have no known com- peting financial interests or personal relationships that could have ap- peared to influence the work reported in this paper.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.o rg/licenses/by/4.0/.

 This article is excerpted from the 《Acta Diabetologica》 by Wound World.