Diabetes Management and Outcomes among Patients with Type 2 Diabetes Attending a Renal Service

Background Patients with comorbid type 2 diabetes mellitus (T2DM) and renal disease, particularly those treated with insulin, often require complex pharmacological treatment and management of other diabetes complications. Aims To assess the achievement of metabolic targets and compare the current management of renal service attenders with insulin- and noninsulin-treated T2DM. Methods Single-centre retrospective cross-sectional study involving medical record review of patients with T2DM aged ≥18 years who visited a metropolitan renal outpatient clinic in 2017. Univariable analysis and multivariable logistic regression were used to identify factors associated with insulin treatment. Results Among 268 patients (45.5% insulin-treated), mean HbA1c was higher in insulin-treated vs. noninsulin-treated patients (8.0 ± 1.8% (64 mmol/mol) vs. 6.8 ± 1.2% (51 mmol/mol), p < 0.001). Significantly fewer insulin-treated patients had HbA1c ≤ 7.0% (53 mmol/mol; 31.8% vs. 69.3%, p < 0.001). More insulin-treated patients had ischaemic heart disease (46.7% vs. 33.6%, p = 0.028), diabetic foot disease (15.6% vs. 4.8%, p = 0.003), retinopathy (40.2% vs. 11.0%, p < 0.001), and emergency attendance for severe hypoglycaemia (3.8% vs. 0% p = 0.042). Insulin treatment was more associated with chronic kidney disease stages 4-5 (adjusted odds ratio (aOR) 2.41, 95% CI 1.07-5.43), retinopathy (aOR 3.10, 95% CI 1.04-9.27), and podiatry review (aOR 5.06, 95% CI 1.20-21.38). Only 38 (14.2%) individuals were seen by a colocated public multidisciplinary diabetes service in 2017. Conclusions Renal clinic attenders with T2DM, particularly if insulin-treated, remained at increased risk of diabetes-related complications, including severe hypoglycaemia, with limited input from the colocated hospital diabetes team. Approaches to increase coordination of diabetes care among renal patients should be investigated.


Introduction
A major complication of diabetes mellitus is end-stage renal disease (ESRD). Diabetes is the leading cause of ESRD in Australia, accounting for 38% of new cases in 2017 [1]. Furthermore, 50% of those with severe chronic kidney disease (CKD) from any cause requiring renal service attention have diabetes [1]. Morbidity from diabetes-related complications, including diabetic foot disease and retinopathy, is particularly high when CKD is present among people with type 2 diabetes mellitus (T2DM) [2,3]. There is also an eightfold increased risk of cardiovascular and all-cause mortality, and as CKD progresses, the risk of acute events including severe hypoglycaemia increases [2]. Renal impairment in T2DM is independently associated with the reduced achievement of metabolic targets and complicates treatment choices due to interference with glucose metabolism, impaired renal clearance, and the metabolism of medications [4,5]. Clearly, those with T2DM and renal disease have an increased complexity of care derived from greater morbidity and mortality, difficulty in achieving metabolic targets, and adjusting pharmacological treatments.
Standards for evidence-based care for patients with T2DM have been defined in the Royal Australian College of General Practitioners (RACGP) guidelines, the National Institute for Health and Care Excellence (NICE) guideline NG28 from the United Kingdom (UK), and the American Diabetes Association (ADA) Standards of Medical Care in Diabetes [6][7][8]. The Kidney Disease Improving Global Outcomes (KDIGO) Diabetes in CKD guidelines also have specific recommendations for individuals with comorbid diabetes and CKD [2]. The guideline recommendations for those patients with CKD include tight glycaemic control (HbA1c ≤ 7:0%), blood pressure (BP) (≤130/80 mmHg), and lipid control (total cholesterol < 4:0 mmol/L; LDL <2.0 mmol/L), risk factor reduction, and renin-angiotensinaldosterone system (RAAS) blockade. Achievement of metabolic targets, as defined by RACGP guidelines, reduces the risk or slows the progression of CKD in T2DM [6,9,10].
Insulin therapy is an important therapeutic option for individuals with advanced CKD, although such treatment has shown mixed efficacy in achieving glycaemic control in T2DM [2,[11][12][13]. Insulin therapy is associated with weight gain and a greater risk of hypoglycaemia, both of which are associated with increased cardiovascular and other risks [2,11,12].
Diabetes mellitus is a significant problem in South Western Sydney (SWS) with a prevalence of 6.7%, compared with a national prevalence of 5.1% [14]. As a result, T2DM and its complications are a major strain on local hospital services. In view of the complexities in the care of those with insulintreated T2DM, the aims of this study were to compare the local achievement of metabolic targets, diabetes management, and associated risks between renal patients with insulin-and noninsulin-treated T2DM.

Materials and Methods
This was a single-centre retrospective cross-sectional study among patients with diabetes mellitus attending an outpatient renal specialist service in SWS. The renal service is on the same hospital site as a major public multidisciplinary diabetes outpatient service. The diabetes service is staffed by endocrinologists, diabetes educators, podiatrists, and dieticians. Patients of the renal service have their diabetes mellitus managed by a mix of general practitioners (GPs) and private and public diabetes specialist care at the discretion of the GP and the person with diabetes.
The study was approved by the Quality Improvement Committee (reference: CT01_2019) of the South Western Sydney Local Health District Health Research Ethics Committee.

Data Collection.
Deidentified quantitative data of all attending patients with diabetes mellitus aged ≥18 years between January 1st and December 31st, 2017 were collected following the review of electronic and paper medical records from both the local renal and diabetes services. Data collected included patient demographics, medications, clinical encounters, comorbidities, anthropometric measures, and metabolic parameters. Ethnicity was not ascertained.
The last result in 2017 by the date of collection was recorded for laboratory examinations. Where available, all estimated glomerular filtration rate (eGFR) and urinary albumin-to-creatinine ratio (UACR) data were recorded for the previous three years. The stage of CKD was determined based on the last recorded eGFR. Metabolic targets were defined in accordance with the RACGP guidelines (Supplementary Table 1), and body mass index (BMI) was defined by the World Health Organisation classification [6,15]. Medications were recorded according to the last documented regimen. Attendance at a public diabetes service was recorded if it occurred in 2017 and included, but was not limited to the local hospital diabetes service. Attendance at public allied health services including podiatry, dietetics, and diabetes educators was available from electronic records. Private clinical encounters were recorded when available in the documentation.
2.2. Statistics. As this was a pilot quality improvement activity including all clinic attenders in the index year, no power calculations were undertaken. Analyses were undertaken using SPSS, version 25 (SPSS Inc., Chicago, IL, USA). Categorical variables were described using numbers and percentages, and Pearson's chi-squared test or Fischer's exact test, where appropriate, was used to compare groups. Continuous variables were described using mean ± standard deviation or median and range. Differences between groups were compared using the independent t-test. When skewed, the Mann-Whitney test or geometric mean was used. Stepwise binomial logistic regression including the variables with a significant difference in the univariable analysis was conducted to compare insulin-and noninsulin-treated patients. Nagelkerke's R square was computed for the proportion of variance accounted for in the model. Statistical significance at p < 0:05 was considered significant, and all statistical tests were two-tailed.

Results
The rate of decline of eGFR per year was faster in the insulin-treated group (-5 mL/min/1.  Table 3 shows that metformin and sulphonylureas were used significantly less among insulin-treated patients (30.3% vs. 45.9%, p = 0:009; 15.6% vs. 42.5%, p < 0:001, respectively). There were no significant differences in the use of antihypertensive agents based on insulin treatment status. Renin-angiotensin-aldosterone system (RAAS) blockers were prescribed to 70.1% of the cohort. There was a high use of statins, particularly among those with insulin-treated T2DM (86.1% vs. 69.9%, p = 0:002).

Discussion
We have shown in this first Australian study, and one of few international studies, that compared with noninsulin-treated renal patients with T2DM, those treated with insulin had worse hyperglycaemia, more hypoglycaemia, and experienced a greater burden of diabetes-related complications including foot problems, hospitalisation, and a faster decline in renal function. There were low rates of documented diabetes risk monitoring, and few patients were documented to be receiving care from the local multidisciplinary diabetes service, while those that did, received more modern pharmacotherapy. The study was undertaken to understand whether renal services might benefit from working more closely with colocated diabetes specialist services. Our findings suggest that such coworking is likely to be beneficial.
Insulin treatment is a major risk factor for hypoglycaemia in those with CKD, and less stringent HbA1c targets may be appropriate in individuals who experience

5
Journal of Diabetes Research hypoglycaemia in view of its association with increased mortality [6,25]. Additionally, higher individualised HbA1c targets may be appropriate in individuals who are older, with a longer duration of diabetes and severe established vascular complications, as is reflected in the insulin-treated cohort [5]. Clinical inertia including delay or failure to intensify insulin treatment may also play a role in the suboptimal glycaemic control found in this study [26].
That diabetes care can be improved is suggested by those attending the public diabetes service being more likely to be treated with modern diabetes treatment including DPP4i and GLP-1 RA, which are beneficial in reducing the risk of hypoglycaemia. Nevertheless, there was a relatively low uptake of sodium-glucose cotransporter 2 inhibitors (SGLT2i) which are now considered first-line therapy in combination with metformin for individuals with eGFR ≥ 30 mL/min/1.73 m 2 due to their reduced risk of renal and cardiovascular morbidity and mortality [2,8,10,27]. These recommendations, however, date more recently than this study, and thus we would expect a future comparative study to reflect the changes in antihyperglycaemic therapy recommendations.   In contrast to the relatively low achievement of glycaemic targets, our cohort of patients was more likely to achieve the BP target than those in earlier studies, albeit with room for improvement. Hypertension is a significant contributor to the development of renal disease and its progression [19,23]. Sustained blood pressure control is the single most effective intervention to reduce the risk and progression of CKD and simultaneously reduce cardiovascular risk [28]. Individuals with comorbid T2DM and CKD usually achieve worse blood pressure control than those without CKD, notwithstanding the greater use of antihypertensive agents [4]. Cohort studies of patients with T2DM and CKD have shown that 14.3-20.8% of patients achieved BP ≤ 130/80 mmHg, 20.8-41.0% achieved ≤140/80 mmHg, and 38.8-60.7% achieved BP ≤ 140/90 mmHg [17,18,20,21,23]. There were insufficient data to identify the reasons behind this better BP control in our study. Studies of the epidemiology of hypertension in the background population and any wider secular trends such as mortality could be helpful. This discordance between the achievement of blood pressure and glycaemic targets in T2DM has recently been shown elsewhere, with the suggestion that glycaemic (vs. blood pressure) management may warrant greater inclusion of diabetes specialist service intervention [29].
The RACGP, KDIGO, and ADA recommend patients with T2DM and CKD use RAAS blockers as they have been shown to slow the progression of renal disease [2,6,8]. The uptake of RAAS blockade (70.1%) in this study was similar to that in other cohorts of patients with CKD (49.0-79.7%), except for one French cohort with T2DM and CKD which showed uptake as high as 91.5% [17-20, 22, 23, 30]. Higher uptake in this French cohort may be related to the exclusion of patients with stage 5 CKD [23]. Participation in integrated multidisciplinary care has shown improvements in RAAS blocker uptake [18,31].
Renal specialist services can significantly slow the progression of CKD, improve survival, and control blood pressure while increasing the uptake of RAAS blockers and statins among patients with T2DM [32,33]. However, achievement of metabolic targets (set to reduce morbidity) and management of other complications including acute foot events, retinopathy, and hypoglycaemia would likely have benefitted from coordination with diabetes specialist services. Further, there were low rates of documented diabetes risk monitoring, although information regarding private health encounters was not available beyond the correspondence that had been received by the renal physician. Renal   [34]. This raises the question of whether there is a need for better coordination between renal and diabetes services. One way to reduce risks and enhance coordination between renal and diabetes care may be to establish a joint renal-diabetes clinic, involving multidisciplinary endocrinology and nephrology specialist care, which have demonstrated improvements in achieving metabolic targets and slowing the progression of renal disease [18,20,31]. A qualitative evaluation of one such service in Australia has shown improved integration of care and perception of improved health and management of health [35]. Evaluation of metabolic and complication outcomes is awaited. Developing a similar model of care locally, particularly targeting those receiving insulin treatment, is likely to improve the way diabetes care is being delivered. Other models of care to consider include self-management programmes for those with comorbid diabetes and renal disease. These have demonstrated improved glycaemic control, maintenance of renal function, and improvements in quality of life [36]. Case conferencing with specialists and upskilling GPs may also play a role, considering that large proportions of those with comorbid T2DM and CKD are seen in primary care [37].
The strengths of the study are being the first in Australia to explore the achievement of metabolic targets and assess the management of patients with T2DM and CKD based on insulin treatment status. Further strengths are the inclusion of a complete clinic cohort with a modest catchment population size, in alignment with other similar studies [21,22,24,38]. Additionally, the data collected includes a comprehensive register of measurable diabetes care-related parameters and reflects real-world clinical practice.
Limitations of this study include that it is a single-centre study, relaying data specific to an Australian healthcare system which may restrict its relevance to other centres. Due to the retrospective cross-sectional study design, there were missing data, and causality was not able to be determined. Private health encounters may not have all been recorded, underestimating the number of individuals engaged in diabetes risk monitoring. As a result of the comanagement of diabetes, it is difficult to assess who and what was contributed by the varying healthcare professionals. Additionally, the data were not adjusted for time spent under the renal service, and this may affect the management of their diabetes and CKD. The potential inaccuracy of HbA1c measurement in CKD should also be noted. This occurs secondary to rapid cell turnover and may result in falsely lower HbA1c values [39]. Nevertheless, HbA1c remains preferable to other measurements including fructosamine and glycated albumin [39].
It would be useful to conduct another study in this cohort in the future to determine the effects of changing clinical practices, including the routine use of SGLT2 inhib-itors as first-line therapy. It may be useful to assess the influence of individual nephrologists on the diabetes care and management of patients by comparing their initial nephrologist consultation to a point in time in the future. Additionally, the efficacy of joint diabetes-renal clinics in an Australian healthcare setting needs to be evaluated.

Conclusion
Renal clinic attendees with insulin-treated T2DM had a higher burden of diabetes-related complications including diabetic foot disease, retinopathy, hypoglycaemia, and hospitalisation. In contrast to good blood pressure control, glycaemic management, monitoring of diabetes risk, and use of more modern pharmacotherapy were limited. A joint renal diabetes clinic may be one way to reduce T2DM-related morbidity and mortality. Studies are needed to assess the efficacy of such joint clinics in the Australian healthcare system.

Data Availability
Access to data is restricted. The data and distribution of this audit have not been approved under the terms of the ethics approval.

Conflicts of Interest
The authors declare that they have no conflicts of interest.