Clinical care and other categories posters: Real-world study

Effect of bariatric surgery on diagnosed chronic kidney disease and cardiovascular events in microalbuminuria patients with insulin-treated type 2 diabetes: A retrospective cohort study from a large UK primary care database

M Alkharaiji^1,2, UC Anyanwagu¹, R Donnelly¹, I Idris¹

¹Division of Medical Sciences & Graduate Entry Medicine, School of Medicine, University of Nottingham, Royal Derby Hospital Centre, Derby, UK, ²Faculty of Public Health, College of Health, The Saudi Electronic University, Riyadh, Saudi Arabia

Aim To compare the effect of bariatric surgery on renal and cardiovascular (CV) outcomes among obese patients with insulin-treated type 2 diabetes with and without microalbuminuria (i.e. uACR >3.0 mg/mmol).

Methods A retrospective cohort study was conducted among 11,125 patients with type 2 diabetes from The Health Improvement Network (THIN) database. Propensity score matching (up to 1:6 ratio) was used to identify patients who underwent bariatric surgery (N=131) with a non-bariatric cohort (N=579). Follow-up was undertaken for 10 years (6,487 person-years) to compare differences in risk of CV events and renal outcomes.

Results The mean age was 52(SD:13) years; (60% female); weight 116(SD:25)kg, BMI 41(SD:9)kg/m², estimated glomerular filtration rate (eGFR) 70.4 ± 20 ml/min/1.73 m², and median albumin-creatinine ratio (uACR) 2.0 mg/mmol (IQR: 0.9–5.2 mg/mmol). Bariatric surgery was associated with a 54% reduction in developing Chronic Kidney Disease (CKD) compared to the non-bariatric cohort (aHR: 0.46; 95%CI: 0.24–0.85, p=0.02). Among patients with microalbuminuria at baseline, bariatric surgery was protective against CKD (aHR: 0.42, 95%CI: 0.18–0.99, p=0.050). eGFR was significantly increased from baseline favouring the bariatric group (mean difference: 4.1 ml/min/1.73 m²; p < 0.05). However, little or no significant change was observed with non-fatal CV events (aHR: 0.36, 95%CI: 0.11–1.13, p=0.079). uACR and total protein levels had little or no association with the intervention.

Conclusion Bariatric surgery may protect patients with diabetes with/out microalbuminuria against the risk of CKD; is associated with improvements in overall renal outcomes (e.g. eGFR); and with a modest protective effect on non-fatal CVD risk.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

Oral glucose tolerance test results are altered by the size of the last meal before the test

WP Nakanga^1,3, RC Andrews¹, A Jones¹, BM Shields¹, P Hughes³, A Karabarinde³, AT Hattersley¹, MJ Nyirenda^2,3

¹National Institute for Health Research (NIHR), Exeter Clinical Research Facility, University of Exeter, Exeter, UK, ²London School of Hygiene and Tropical Medicine (LSHTM), London, UK, ³Medical Research Council, Uganda Virus and Research Institute and LSHTM Uganda Research Unit, Entebbe, Uganda

Background The oral glucose tolerance test (OGTT) is widely used in epidemiological studies in sub-Saharan Africa to determine prevalence of diabetes and glucose intolerance. In many households in sub-Saharan Africa, obtaining sufficient food remains a challenge. It is not clear how restrictions in food availability affects OGTT results.

Aims To determine the impact of the size of evening meal the day before OGTT.

Methods 40 adult participants (mean age 43.7years, mean BMI 23.3/kg/m², 21 males) with no history of diabetes were recruited from rural Uganda for a randomised cross-over study. The day prior to the OGTT participants were admitted and given a normal midday meal (1,065 kcal, 139 g carbohydrate) and then randomly either a normal (1,065 kcal, 139 g carbohydrate) or a small (58 kcal, 14 g carbohydrate) evening meal. OGTT was performed the next morning with 30-min glucose sampling. The protocol was repeated after one week with the alternative evening meal.

Results Glucose tolerance was worse after the small evening meal (mean area under plasma glucose-time curve 925 mmol min/l(small) v 841 mmol min/l(normal) p < 0.001) with a higher 2h OGTT value (7.2 vs 6.3 mmol/l p=0.003) and the diagnosis of impaired glucose tolerance (>7.8 mmol/l) doubled (10 vs 5 p=0.18).

Conclusion This randomised cross over study demonstrates that abnormal glucose tolerance results may not reflect sustained hyperglycaemia, but rather response to the glucose load as a result of the previous short-term food restriction. This indicates that glucose tolerance results might not reflect the actual burden of diabetes in some African populations.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

Using repeated measurements to predict cardiovascular risk in patients with type 2 diabetes

K Gokhale^1,2, P Tiňo², T Taverner¹, C Sainsbury¹, K Nirantharakumar¹

¹Institute of Applied Health Research, University of Birmingham, Birmingham, UK, ²School of Computer Science, University of Birmingham, Birmingham, UK

Aims Cardiovascular disease (CVD) risk assessment models perform sub-optimally in patients with type 2 diabetes, therefore resulting in inappropriate statin treatment allocation. We aimed to improve the performance of CVD risk prediction by incorporating time dependent measurements.

Methods We extracted incident and prevalent patients with type 2 diabetes aged 25–85 years, not on statin treatment and had no CVD diagnosis at baseline from ‘The Health Improvement Network’ database. Outcomes were first diagnosis of Ischaemic heart disease or stroke/transient ischaemic attack. We included risk factors ranging from socio-demographic characteristics, medical history and diagnoses, physical measurements and laboratory test results and diabetes treatment as prognostic markers for our analysis. We split the data into training and testing sets. We employed Cox's hazards model for time-dependent covariates to estimate the hazard rates for each risk factor and calculated a 10-year risk score using the training data. We measured performance and then internally validated our model using the test dataset.

Results Study included 198,833 (108,245 males with 12,156 outcomes. 90,588 females with 8,258 outcomes) patients with type 2 diabetes. The incidence of CVD was 0.02 in males and 0.015 in females (per 100,000 person-years). The models were well calibrated and the time-dependent AUROC (95% confidence-interval) was 0.75 (0.74–0.76) for females and 0.73 (0.72–0.74) for males.

Conclusion Our CVD risk prediction model for type 2 diabetes patients using repeatedly measured data, performed better than the ones currently used in clinical practice. Future research should explore survival versions of explainable deep learning models to further improve CVD risk prediction in patients with type 2 diabetes.

Projecting total potential future life years lost to diabetes: Outcomes from analysis of the latest National Diabetes Audit (NDA) and Office of National Statistics (ONS) data

M Lunt², M Stedman¹, G Rayman³, R Gadsby⁴, A Heald^5,6

¹Health Research, Res Consortium, Andover, UK, ²The School of Medicine and Manchester Academic Health Sciences Centre, Manchester University, Manchester, UK, ³The Ipswich Diabetes Centre and Research Unit, Ipswich Hospital NHS Trust, Ipswich, UK, ⁴Warwick Medical School, University of Warwick, Warwick, UK, ⁵The School of Medicine and Manchester Academic Health Sciences Centre, Manchester University, Manchester, UK, ⁶Department of Endocrinology and Diabetes, Salford Royal Hospital, Salford, UK

Refer to Oral number A52.

Burden of influenza-related admissions to secondary care in patients with diabetes

I Gibbons, C Davidson, J Clark-Wright

Patient Access, Sanofi Pasteur, Reading, UK

Refer to Oral number A33.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

Predictors of genital mycotic infections with SGLT2 inhibitors

AP Mcgovern, JM Dennis, BM Shields, AT Hattersley, AG Jones

¹University of Exeter Medical School, University of Exeter, Exeter, UK

Aims Genital mycotic infection is one of the most common adverse effects of SGLT2-inhibitors. However, there has been little exploration of which people with type 2 diabetes are at the highest risk. We assessed infection risk factors and absolute risk across key risk groups.

Methods We explored predictors of genital mycotic infection using Cox regression in 21,004 people with type 2 diabetes initiated on SGLT2-inhibitors in UK primary care data (CPRD), compared with 55,471 people initiating DPP4-inhibitors. We report absolute genital infection rates in females and males with and without a history of prior genital infection.

Results Genital mycotic infection was more common in those treated with SGLT2-inhibitors (8.1% within one year) than DPP4-inhibitors (1.8%). Key predictors of infection with SGLT2-inhibitors were; female gender (hazard ratio 3.64; 95%CI 3.23–4.11; p < 0.001), and history of genital infection; within the last year (HR 4.38; 3.73–5.13; p < 0.001), 1–5 years (HR 3.04; 2.64–3.51; p < 0.001), and >5 years (HR 1.79–1.55, 2.07; p < 0.001). Baseline HbA1c was not associated with infection risk with SGLT2-inhibitors in contrast to DPP4-inhibitors where risk increased with baseline HbA1c. One-year absolute risk of genital infection with SGLT2-inhibitors was highest for those with a history of prior infection (females 23.7%, males 12.1%), compared to those without (females 10.8%, males 2.7%).

Conclusions Genital mycotic infections with SGLT2-inhibitors are common in primary care and female gender and history of infection can be used to stratify risk. High baseline HbA1c does not increase infection risk with SGLT2-inhibitors.

The benefit of GP-based practice pharmacists in tackling clinical inertia in people with type 2 diabetes

G Foreshaw¹, L Hauser², A Mucha², B Thompson³, G Syan⁴, S Rao⁴, P English⁵, WD Strain⁶

¹Coleridge Medical Practice, NHS, Ottery St Mary, UK, ²Livewell Southwest, NHS, Devon, UK, ³Boots Pharmacy, Queens Surgery, Devon, UK, ⁴Soar Beyond, insight 2 implementation network, Hertfordshire, UK, ⁵Diabetes, Endocrinology, Acute and General Internal Medicine, Derriford Hospital, Plymouth, UK, ⁶Diabetes and Vascular Research Centre, University of Exeter Medical School, Exeter, UK

Aims The potential role of the practice pharmacist to improve glycaemic control for people with type 2 diabetes is postulated in the five-year general practice forward view and the NHS long-term plan. We aimed to test this in a real-world quality-improvement project.

Methods Nine general practice pharmacists in Devon, supported remotely by expert pharmacists and two consultants, identified 4 key target groups of people with type 2 diabetes:

• Monotherapy with HbA1c >53 mmol/mol

• Dual therapy with HbA1c >58 mmol/mol

• HbA1c <48 on insulin and/or sulfonylurea therapy at risk of hypoglycaemia

• People >75 years of age with a HbA1c <58 mmol/mol on insulin or sulfonylureas

Over 12 months these individuals were invited to attend pharmacist-led clinics.

Results At baseline 498, 222, 97 and 88 patients were identified in groups 1–4 respectively from a cumulative practice population of 73,450. Four pharmacists completed 12-months in this unfunded project. 72% of potential patents were reached (range 66–88%) in these practices. In groups 1 and 2, HbA1c fell by 4 mmol/mol and 6 mmol/mol respectively with 59.6% of individuals achieving their target HbA1c compared to 22.0% in the same practices who had not attended a clinic. In groups 3 and 4, there were only 3 hypoglycaemia-related admissions after intervention compared to 11 in the same period prior to intervention.

Conclusion Approximately half of pharmacists completed 12-months due to competing pressures, however in those practices, almost 60% of people with type 2 diabetes achieved target HbA1c and hospitalisations due to hypoglycaemia were substantially reduced. In order achieve this benefit in clinical practice, adequate support and resource allocation is required.

Establishing the relationship between diabetes and pancreatic cancer in the UK biobank health database cohort: A prospective study

AWE Cheang¹, CD Byrne^1,2, ZZR Hamady^1,3

¹Human Development and Health Academic Unit, Faculty of Medicine, University of Southampton, Southampton, UK, ²Endocrinology and Metabolic Medicine, University of Southampton NHS FT, Southampton, UK, ³Upper Gastrointestinal Surgery, University of Southampton NHS FT, Southampton, UK

Refer to Oral number A34.

Patterns of change in insulin requirement and impact on glycaemic control after commencing continuous subcutaneous insulin infusion (CSII) in type 1 diabetes

YRN Sredharan¹, S Philip², PS Yap², P Mcdonald², L Browne², L Cameron², L Mccauley², B Macdonald², P Abraham²

¹University of Aberdeen, Aberdeen, UK, ²Diabetes and Endocrinology, NHS Grampian, Aberdeen, UK

Background This study aimed to compare patterns of changes in insulin requirement following a year of CSII.

Methods Case-note and device download review to compare change in insulin requirements before and 12 months after commencing CSII using a Medtronic pump following a structured start process between 2013 and 2016 at our centre.

Results Fifty patients with adequate follow-up information were included. Mean age was 48(±13) years, with duration of diabetes of 27(±12) years and 30(60%) were female. The mean HbA1c was 71.1(±10.7)mmol/mol which dropped by 8.5mmol/mol (p=<0.01) after CSII, with greatest reductions in those with HbA1c above 75mmol/mol [14.8(±12.2)mmol/mol]. Forty-two (84%) had lower HbA1c following CSII, and 40% achieved HbA1c below 58mmol/mol. 32(64%) gained weight with mean increase of 1.7kg (p=0.03).

The mean insulin requirement on CSII was 0.58 (±0.17)units/kg. Mean total daily dose (TDD) of insulin dropped from 53.1(±21.0) units to 47.3(±6.6) units following CSII. TDD decreased by 0.1units/kg (14.7%, p=<0.01), mean basal insulin decreased 0.03units/kg (10%, p=0.45) and bolus dose dropped by 0.07units/kg (18.4%, p=<0.01).

Thirty-four (83%) out of 41 patients reported higher insulin sensitivity factor (1.5–2.5) after CSII (p=<0.01). Thirty-eight (80.9%) increased their carbohydrate to insulin ratio [8.7(3.1)–10.9(3.7)] (p=<0.01). Proportion of basal insulin increased from 43(±0.13)% to 48 ± (0.09)% (p=<0.01).

Conclusion There was an overall improvement in glycaemic control following CSII. This was associated with a fall in insulin doses following CSII, especially for bolus insulin resulting in a relatively higher proportion of basal. The study provides real world data on change in insulin requirements following CSII.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.