Clinical care: Addressing risk

Serum hyaluronic acid improves the performance of the Fibrosis-4 index in the identification of advanced liver disease in type 2 diabetes: The Edinburgh Type 2 Diabetes Study

SM Grecian¹, S McLachlan¹, JA Fallowfield², IN Guha³, JR Morling³, RM Williamson⁴, RM Reynolds⁵, NN Zammitt⁶, JF Price¹, MWJ Strachan⁴, PC Hayes², S Glancy⁷, BM Frier⁵

¹Usher Institute of Population Health Sciences and Informatics, University of Edinburgh, Edinburgh, UK, ²Centre for Inflammation Research, Queen's Medical Research Institute, University of Edinburgh, Edinburgh, UK, ³NIHR Nottingham Biomedical Research Centre, Nottingham University Hospitals NHS Trust and University of Nottingham, Nottingham, UK, ⁴Metabolic Unit, Western General Hospital, Edinburgh, UK, ⁵Centre for Cardiovascular Science, Queen's Medical Research Institute, University of Edinburgh, Edinburgh, UK, ⁶Department of Diabetes and Endocrinology, Royal Infirmary of Edinburgh, Edinburgh, UK, ⁷Department of Radiology, Western General Hospital, Edinburgh, UK

Background Type 2 diabetes is associated with increased cirrhosis and hepatocellular carcinoma (HCC), mostly secondary to non-alcoholic fatty liver disease (NAFLD). There is increasing interest in risk stratification to identify who may benefit from hepatologist review. Current risk stratification tools are suboptimal and perform worse in people with diabetes than in those without.

Aim To identify whether the addition of complementary biomarker(s) improves performance of current risk stratification tools to identify cirrhosis/HCC in type 2 diabetes.

Methods The Edinburgh Type 2 Diabetes Study investigated men and women with type 2 diabetes (n = 1,066, age 60–75 at baseline). Biomarkers were measured at baseline and one year. Cases of cirrhosis/HCC were identified over 11 years of follow-up. Model association with cirrhosis/HCC was assessed using logistic regression.

Results Of existing scores examined, the Fibrosis-4 (FIB-4) index and the AST:platelet ratio index (APRI) performed best in our cohort. Of additional baseline biomarkers assessed, combining hyaluronic acid and gamma-glutamyltransferase with FIB-4 or APRI improved model performance by AIC most. Using cut-points, the only model which retained a false negative rate of ≤25% was FIB-4 (cut-point ≥1.3) with hyaluronic acid (cut-point ≥50g/l). This model in isolation, or as the fibrosis marker within the European screening algorithm, reduced those falsely identified as ‘high-risk’ for incident disease by ~50% while the false negative rate increased from 14 to 19%.

Conclusions Hyaluronic acid and FIB-4 combined fibrosis assessment reduced those inappropriately identified as ‘high-risk’ for cirrhosis/HCC in a community population of people with type 2 diabetes. These findings require validation.

Acknowledgement Edinburgh Type 2 Diabetes Study

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

Non-linear fast glomerular filtration rate decline is associated with Afro-Caribbean ethnicity and HbA1c variability in patients with type 2 diabetes who develop end-stage renal disease

SI Stoilov, N Fountoulakis, A Panagiotou, S Thomas, J Karalliedde

Diabetes and Endocrinology Department, Guy's and St Thomas’ NHS Foundation Trust, London, UK

Aims To describe features and patterns of estimated glomerular filtration rate (eGFR) trajectories over time in an ethnically diverse population of patients with type 2 diabetes and diabetic nephropathy (DN). 

Methods A total of 398 patients with type 2 diabetes and DN with baseline eGFR > 30ml min-1 attending a secondary care diabetes renal clinic were evaluated. The median follow-up was 7 years. Of the cohort, 71 (18%) reached end-stage renal disease (ESRD). eGFR trajectories were determined using the generalised additive model (GAM) function in R. Multinomial regression analysis was used to determine predictors of trajectories. 

Results In the ESRD cohort (61% male, 47% Afro-Caribbean), baseline mean ± standard deviation: age was 58.4 ± 9.4 years and duration of diabetes 13.1 ± 7.7 years. The majority of patients (83.1%) received uninterrupted renin-angiotensin system (RAS) inhibition therapy and 91.4% had a confirmed diagnosis of diabetic retinopathy. Non-linear eGFR trajectories were identified in 59% of the cohort and 4 distinct patterns were observed [‘linear’ (n = 29), ‘accelerator’ (n = 16), ‘decelerator’ (n = 9), ‘stepwise’ (n = 17)]. A rapid decline (accelerator) pattern was associated with Afro-Caribbean ethnicity (OR = 10.6, 95% CI 2.14–52.7, p < 0.001). HbA1c variability (coefficient of variation of HbA1c over follow-up) was higher for patients exhibiting a stepwise eGFR trajectory (OR = 1.15, 95% CI 1.03–1.3, p = 0.02).

Conclusions In our cohort, eGFR trajectories exhibit significant non-linearity. Afro-Caribbean ethnicity and higher glycaemic variability are independent predictors of an accelerating and a stepwise pattern of eGFR loss respectively. Further studies are needed to examine patterns of eGFR loss in type 2 diabetes and identify associated patient features and risk factors to enable more precise risk stratification.

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

Aims Increased all-cause mortality due to suboptimal condition management is an acknowledged consequence of diabetes. Data from NDA Complications and Mortality 2015–2016 (NDA-CM) provides opportunity to quantify the potential future lost life years (LLY) based on historic treatment levels in people with diabetes in England.

Methods Relative mortality rates for type 1 diabetes/type 2 diabetes by age group and gender from the NDA-CM were combined with ONS overall population mortality data to compare overall life expectancy of the current diabetic populations with equivalent non-diabetic populations and determine potential LLYs for each group.

Results The study covered 6,917 general practices supporting 53.8 million patients of which 223,000 were on the type 1 diabetes register and 2.71 million on type 2 diabetes register, with 102,000 recorded deaths. An ‘average’ type 1 diabetes person (aged 42.8 years) had life expectancy of 32.6 years compared to 40.2 years if they were non-diabetes: future LLY of 7.6 years equals 1.7 million total type 1 diabetes LLY.

An ‘average’ type 2 diabetes person (aged 65.4 years) has life expectancy of 18.6 years compared to 20.3 years if they were non-diabetes: equals 4.2 million type 2 diabetes LLY. Women's LLY/person were 20% greater for type 1 diabetes and 50% for type 2 diabetes compared to men.

NDA reports 70% of type 1 diabetes people and 33% of type 2 diabetes had HbA1c > 58mmol/mol. Simple pro-rata calculation shows that for both type 1 diabetes/type 2 diabetes, one year with suboptimal glycaemic control could bring their mortality forward by 100 days.

Conclusion These 6.4 million LLY come mainly from the smaller group of ‘harder to treat’ individuals. These findings might encourage clinicians to find more suitable therapies and for at-risk patients to engage with those therapies.

Can retinal vascular measurements predict global vascular complications in patients with type 2 diabetes? A systematic review and meta-analysis

AS Nar¹, Y Huang¹, G George¹, S Gan¹, ATN Nair¹, AL Rajandrakumar¹, G Pradeepa³, E Trucco², ASF Doney¹

¹Division of Population Health and Genomics, University of Dundee, Dundee, UK, ²School of Science and Engineering, University of Dundee, Dundee, UK, ³Madras Diabetes Research Foundation, Chennai, India

Aim To investigate the relationship between retinal vascular measurements (RMs) from diabetes retinal screening photographs and microvascular and macrovascular complications of type 2 diabetes.

Methods PubMed and Web of Science were used to search for relevant studies published until 31 December 2018. RMs considered for this study were vessel width, tortuosity, fractal dimensions and branching geometry. Studies which report risks in terms of OR, HR, RR along with 95% CI, and used RMs as independent and continuous variables were selected for meta-analysis. A random effects model was used.

Results The only studies found that satisfied the criteria for our meta-analysis were for Diabetic Retinopathy (DR) (four studies) and stroke (eight studies). For DR, three studies (n = 5,793) explored association with vessel width in 216 (3.7%) patients with DR. A significant relationship between arteriolar vessel width and risk of DR was found [OR: 1.36 (95% CI: 1.12–1.57)]. Two studies (n = 1,092) explored the association between branching geometry and DR in 355 (32.50%) patients with DR and showed a significant relationship between venular branching geometry and risk of DR [OR: 1.15 (95% CI: 1.0–1.32)].

Seven studies (n = 26,081) explored the association between retinal vessel width in 1,385 (18.83%) with stroke and showed significant association between arteriolar and venular vessel width and risk of stroke, OR: 1.22 (95% CI: 1.03–1.46) and OR: 1.13 (95% CI: 1.00–1.29) respectively.

Conclusion Our study indicates the RMs may be significant predictors of vascular complications in type 2 diabetes and that there remain significant gaps in knowledge in this area.

Acknowledgement INSPIRED