Risk of diabetic retinopathy and diabetic macular oedema with sodium-glucose cotransporter 2 inhibitors and glucagon-like peptide 1 receptor agonists in type 2 diabetes: a real-world data study from a global federated database.

AIMS/HYPOTHESIS: A protective role of sodium-glucose cotransporter 2 inhibitors (SGLT2is) and glucagon-like peptide 1 receptor agonists (GLP1-ra) in the development of diabetic retinopathy and diabetic macular oedema has been described in some recent studies, which may extend beyond glycaemic control. We aimed to review the clinical impact of SGLT2i and GLP1-ra therapy on the risk of diabetic retinopathy and diabetic macular oedema in individuals with type 2 diabetes taking insulin. METHODS: This is a retrospective cohort analysis of approximately two million people with type 2 diabetes receiving insulin across 97 healthcare organisations using a global federated health research network (TriNetX, Cambridge, USA). Two intervention cohorts (SGLT2i + insulin, n=176,409; GLP1-ra + insulin, n=207,034) were compared against a control cohort (insulin with no SGLT2i/GLP1-ra, n=1,922,312). Kaplan-Meier survival analysis was performed and estimated HRs were reported for each outcome. Propensity score was used to 1:1 match for age, sex, ischaemic heart disease, hypertension, microvascular complications, chronic kidney disease, HbA1c, BMI and use of pioglitazone, lipid modifying agents, antilipemic agents, ACE inhibitors, angiotensin II inhibitors and metformin. A sub-analysis comparing the two intervention cohorts was also performed. RESULTS: SGLT2i with insulin was associated with a reduced HR (95% CI) for diabetic macular oedema compared with the control cohort (0.835; 0.780, 0.893), while GLP1-ra with insulin demonstrated a lack of signal with no statistical significance to the HR (1.013; 0.960, 1.069). SGLT2i with insulin was not associated with a clinically significant increase in the risk of developing diabetic retinopathy (1.076; 1.027, 1.127), while GLP1-ra with insulin increased diabetic retinopathy risk (1.308; 1.261, 1.357). Compared with SGLT2i with insulin, GLP1-ra with insulin was associated with higher risk of diabetic retinopathy (1.205; 1.153, 1.259) and diabetic macular oedema (1.130; 1.056, 1.208). CONCLUSIONS/INTERPRETATION: Our study suggests that the combination of SGLT2i and insulin is associated with lower risk of developing diabetic macular oedema. However, the use of GLP1-ra was associated with an increased risk of diabetic retinopathy in individuals with type 2 diabetes also taking insulin. A comparative analysis showed favourable outcomes with SGLT2i and insulin in the development of diabetic macular oedema and diabetic retinopathy. RCTs using dedicated  retinal imaging are required to determine the causal relationship with these therapies.

Obesity pharmacotherapy in older adults: a narrative review of evidence.

The prevalence of obesity in older adults (people aged >60 years) is increasing in line with the demographic shift in global populations. Despite knowledge of obesity-related complications in younger adults (increased risk of type 2 diabetes, liver and cardiovascular disease and malignancy), these considerations may be outweighed, in older adults, by concerns regarding weight-loss induced reduction in skeletal muscle and bone mass, and the awareness of the 'obesity paradox'. Obesity in the elderly contributes to various obesity-related complications from cardiometabolic disease and cancer, to functional decline, worsening cognition, and quality of life, that will have already suffered an age-related decline. Lifestyle interventions remain the cornerstone of obesity management in older adults, with emphasis on resistance training for muscle strength and bone mineral density preservation. However, in older adults with obesity refractory to lifestyle strategies, pharmacotherapy, using anti-obesity medicines (AOMs), can be a useful adjunct. Recent evidence suggests that intentional weight loss in older adults with overweight and obesity is effective and safe, hence a diminishing reluctance to use AOMs in this more vulnerable population. Despite nine AOMs being currently approved for the treatment of obesity, limited clinical trial evidence in older adults predominantly focuses on incretin therapy with glucagon-like peptide-1 receptor agonists (liraglutide, semaglutide, and tirzepatide). AOMs enhance weight loss and reduce cardiometabolic events, while maintaining muscle mass. Future randomised controlled trials should specifically evaluate the effectiveness of novel AOMs for long-term weight management in older adults with obesity, carefully considering the impact on body composition and functional ability, as well as health economics.

Brain alterations in regions associated with end-organ diabetic microvascular disease in diabetes mellitus: A UK Biobank study.

BACKGROUND: Diabetes mellitus (DM) is associated with structural grey matter alterations in the brain, including changes in the somatosensory and pain processing regions seen in association with diabetic peripheral neuropathy. In this case-controlled biobank study, we aimed to ascertain differences in grey and white matter anatomy in people with DM compared with non-diabetic controls (NDC). METHODS: This study utilises the UK Biobank prospective, population-based, multicentre study of UK residents. Participants with diabetes and age/gender-matched controls without diabetes were selected in a three-to-one ratio. We excluded people with underlying neurological/neurodegenerative disease. Whole brain, cortical, and subcortical volumes (188 regions) were compared between participants with diabetes against NDC corrected for age, sex, and intracranial volume using univariate regression models, with adjustment for multiple comparisons. Diffusion tensor imaging analysis of fractional anisotropy (FA) was performed along the length of 50 white matter tracts. RESULTS: We included 2404 eligible participants who underwent brain magnetic resonance imaging (NDC, n = 1803 and DM, n = 601). Participants with DM had a mean (±standard deviation) diagnostic duration of 18 ± 11 years, with adequate glycaemic control (HbA1C 52 ± 13 mmol/mol), low prevalence of microvascular complications (diabetic retinopathy prevalence, 5.8%), comparable cognitive function to controls but greater self-reported pain. Univariate volumetric analyses revealed significant reductions in grey matter volume (whole brain, total, and subcortical grey matter), with mean percentage differences ranging from 2.2% to 7% in people with DM relative to NDC (all p < 0.0002). The subcortical (bilateral cerebellar cortex, brainstem, thalamus, central corpus callosum, putamen, and pallidum) and cortical regions linked to sensorimotor (bilateral superior frontal, middle frontal, precentral, and postcentral gyri) and visual functions (bilateral middle and superior occipital gyri), all had lower grey matter volumes in people with DM relative to NDC. People with DM had significantly reduced FA along the length of the thalamocortical radiations, thalamostriatal projections, and commissural fibres of the corpus callosum (all; p < 0·001). INTERPRETATION: This analysis suggests that anatomic differences in brain regions are present in a cohort with adequately controlled glycaemia without prevalent microvascular disease when compared with volunteers without diabetes. We hypothesise that these differences may predate overt end-organ damage and complications such as diabetic neuropathy and retinopathy. Central nervous system alterations/neuroplasticity may occur early in the natural history of microvascular complications; therefore, brain imaging should be considered in future mechanistic and interventional studies of DM.

The synergistic impact of type 2 diabetes and MASLD on cardiovascular, liver, diabetes-related and cancer outcomes.

BACKGROUND AND AIMS: We examined the impact of a co-diagnosis of metabolic dysfunction-associated steatotic liver disease (MASLD) and type 2 diabetes (T2D) on patient outcomes. METHODS: Using TriNetX, a global federated research network (n = 114 million), we undertook two retrospective cohort studies, using time-to-event analysis. Analysis 1 compared MASLD with T2D to MASLD alone; analysis 2 compared T2D with MASLD to T2D alone. Propensity score matching using greedy nearest neighbour (calliper .1) balanced the cohorts (1:1) for significant covariates. Primary outcomes were cardiovascular, liver, diabetes-related, and cancer events over 5 years. RESULTS: Analysis 1 (n = 95 275): a co-diagnosis of T2D significantly increased the risk of ischaemic heart disease (IHD) (HR 1.39; CI: 1.34, 1.44), ischaemic stroke (HR 1.45; CI: 1.35, 1.56), heart failure (HR 1.42; CI: 1.36, 1.49), atrial fibrillation (HR 1.09; CI: 1.03, 1.16), hepatocellular carcinoma (HR 1.96; CI: 1.69, 2.27), pancreatic cancer (HR 1.25; CI: 1.06, 1.48) and liver-related complications over 5 years from MASLD diagnosis. Analysis 2 (n = 15 208): a co-diagnosis of MASLD significantly increased risk of all-cause mortality (HR 1.11; CI: 1.02, 1.22), IHD (HR 1.181; CI: 1.08, 1.29), hepatocellular (HR 50.31; CI: 6.94, 364.72), pancreatic (HR 1.78; CI: 1.12, 2.84), breast (HR 1.43; CI: 1.09, 1.88) and renal cancer (HR 2.01; CI: 1.24, 3.26), and diabetic neuropathy (HR 1.17; CI: 1.09, 1.27) over 5 years from metformin initiation. CONCLUSIONS: T2D significantly potentiates the risk of cardiovascular, malignancy and liver-related outcomes in people with MASLD. The effect of MASLD on people with T2D, although less dramatic, still potentiated risk of death, IHD, malignancy and peripheral neuropathy.

Denosumab, for osteoporosis, reduces the incidence of type 2 diabetes, risk of foot ulceration and all-cause mortality in adults, compared with bisphosphonates: An analysis of real-world, cohort data, with a systematic review and meta-analysis.

AIM: To evaluate the impact of denosumab on (i) the incidence of type 2 diabetes (T2D), and (ii) long-term health outcomes (microvascular [neuropathy, retinopathy, nephropathy] and macrovascular [cardiovascular disease, cerebrovascular accident] complications, and all-cause mortality) in patients with T2D, before (iii) combining results with prior studies using meta-analysis. METHODS: A retrospective analysis of data in a large global federated database (TriNetX; Cambridge, MA) was conducted from 331 375 patients, without baseline T2D or cancer, prescribed either denosumab (treatment, n = 45 854) or bisphosphonates (control, n = 285 521), across 83 healthcare organizations. Propensity score matching (1:1) of confounders was undertaken that resulted in 45 851 in each cohort. Secondary analysis further evaluated the impact of denosumab on long-term health outcomes in patients with T2D. Additionally, we systematically searched prior literature that assessed the association between denosumab and T2D. Estimates were pooled using random-effects meta-analysis. Risk of bias and evidence quality were assessed using Cochrane-endorsed tools. RESULTS: Denosumab (vs. bisphosphonates) was associated with a lower risk of incident T2D over 5 years (hazard ratio 0.83 [95% confidence interval {CI} 0.78-0.88]). Secondary analysis showed significant risk reduction in all-cause mortality (0.79 [0.72-0.87]) and foot ulceration (0.67 [0.53-0.86]). Also, pooled results from four studies (three observational, one randomized controlled trial) following meta-analysis showed a reduced relative risk (RR [95% CI]) for incident T2D in patients prescribed denosumab (0.83 [0.79-0.87]) (I2 = 10.76%). CONCLUSIONS: This is the largest cohort study to show that denosumab treatment is associated with a reduced RR of incident T2D, as well as an associated reduced RR of all-cause mortality and microvascular complications, findings that may influence guideline development in the treatment of osteoporosis, particularly in patients who are at a high risk of T2D.

Effect of combination pioglitazone with sodium-glucose cotransporter-2 inhibitors or glucagon-like peptide-1 receptor agonists on outcomes in type 2 diabetes: A systematic review, meta-analysis, and real-world study from an international federated database.

AIMS: To evaluate the efficacy and cardiovascular outcomes of combination pioglitazone with either a glucagon-like peptide-1 receptor agonist (GLP-1RA) or a sodium-glucose cotransporter-2 (SGLT2) inhibitor in individuals with type 2 diabetes (T2D) by conducting a systematic review, meta-analysis, and analysis of a large international real-world database. METHODS: We searched MEDLINE, SCOPUS and Web of Science to identify relevant articles for inclusion (PROSPERO [CRD: 42023483126]). Nineteen studies assessing pioglitazone + SGLT2 inhibitors or GLP-1RAs versus controls were identified, 16 of which were randomized controlled trials. Risk of bias was assessed using Cochrane-endorsed tools and quality of evidence was assessed using GRADE. We additionally performed a retrospective cohort study of all individuals aged 18 years or over with T2D, using the TriNetX platform. We included propensity-score-matched individuals who were treated for at least 1 year with pioglitazone and a GLP-1RA or pioglitazone and an SGLT2 inhibitor, compared against GLP-1RA and SGLT2 inhibitor monotherapy. Outcomes were all-cause mortality, heart failure, chronic kidney disease and composite stroke and transient ischaemic attack. RESULTS: The average follow-up in the included studies ranged from 24 to 52 weeks. Combination of pioglitazone with a GLP-1RA reduced glycated haemoglobin (HbA1c) and weight greater than in controls: mean differences -1% (95% confidence interval [CI] -1.27, -0.74) and -1.19 kg (95% CI -1.80, -0.58), respectively. There was no statistically significant difference in systolic blood pressure (SBP) or mortality between groups: mean difference - 1.56 mmHg (95% CI -4.48, 1.35; p = 0.30) and relative risk (RR) 0.29 (95% CI 0.07-1.15; p = 0.08), respectively. Combination of pioglitazone with SGLT2 inhibitors reduced HbA1c, weight and SBP to a greater extent than control treatment: mean differences -0.48% (95% CI -0.67, -0.28), -2.3 kg (95% CI -2.72, -1.88) and -2.4 mmHg (95% CI -4.1, -0.7; p = 0.01), respectively. There was no statistically significant difference in mortality between groups (RR 1.81, 95% CI 0.30-10.97; p = 0.52). The included trials demonstrated a reduction in risk of heart failure with combination treatment. Similarly, from the real-world database (n = 25 230 identified), pioglitazone and SGLT2 inhibitor combination therapy was associated with reduced risk of heart failure compared to monotherapy alone (hazard ratio 0.50, 95% CI 0.38-0.65; p < 0.001). CONCLUSION: Both our systematic review/meta-analysis and the real-world dataset show that combination of pioglitazone with either GLP-1RAs or SGLT2 inhibitors is associated with increased weight loss and reduced risk of heart failure compared with monotherapy.

Artificial intelligence-enhanced electrocardiogram analysis for identifying cardiac autonomic neuropathy in patients with diabetes.

AIM: To develop and employ machine learning (ML) algorithms to analyse electrocardiograms (ECGs) for the diagnosis of cardiac autonomic neuropathy (CAN). MATERIALS AND METHODS: We used motif and discord extraction techniques, alongside long short-term memory networks, to analyse 12-lead, 10-s ECG tracings to detect CAN in patients with diabetes. The performance of these methods with the support vector machine classification model was evaluated using 10-fold cross validation with the following metrics: accuracy, precision, recall, F1 score, and area under the receiver-operating characteristic curve (AUC). RESULTS: Among 205 patients (mean age 54 ± 17 years, 54% female), 100 were diagnosed with CAN, including 38 with definite or severe CAN (dsCAN) and 62 with early CAN (eCAN). The best model performance for dsCAN classification was achieved using both motifs and discords, with an accuracy of 0.92, an F1 score of 0.92, a recall at 0.94, a precision of 0.91, and an excellent AUC of 0.93 (95% confidence interval [CI] 0.91-0.94). For the detection of any stage of CAN, the approach combining motifs and discords yielded the best results, with an accuracy of 0.65, F1 score of 0.68, a recall of 0.75, a precision of 0.68, and an AUC of 0.68 (95% CI 0.54-0.81). CONCLUSION: Our study highlights the potential of using ML techniques, particularly motifs and discords, to effectively detect dsCAN in patients with diabetes. This approach could be applied in large-scale screening of CAN, particularly to identify definite/severe CAN where cardiovascular risk factor modification may be initiated.

Insomnia and short sleep duration, but not chronotype, is associated with chronic widespread pain: Mendelian randomization study.

OBJECTIVES: Sleep disturbance has been associated with chronic widespread pain (CWP), but their causal relationship remains unclear. We aimed to examine the causal relationship and direction between CWP and sleep traits, namely insomnia, sleep duration and chronotype, using Mendelian Randomization. METHOD: We used genetic association data from ~0.5 million individuals and up to 1.8 million controls from the UK Biobank (UKB). All traits were defined predominantly by self-report. Short sleep duration was defined as average ≤6 hours per 24 hours. Chronotype refers to the inclination to sleep at certain times where some wake and go to bed early ('morning' person), and others wake and go to sleep later ('evening' person). To permit use of the largest available genetic association data, we used the Causal Analysis Using Summary Effect estimates (CAUSE) method, which allows for sample overlap. RESULTS: Insomnia (OR 1.009, 95% credible interval 1.005, 1.014; p = 0.018 that the causal model is a better fit than non-causal model) and short sleep duration (OR 1.060, 95%CrI 1.038, 1.083; p = 0.040) were causally associated with increased risk of CWP, with limited evidence for reverse causation. There was no evidence in support of long sleep duration or chronotype being associated with CWP. CONCLUSIONS: This study suggest that insomnia and short sleep duration (≤6 hours) are associated with an increased risk of CWP. Improving short sleep duration and insomnia, rather than chronotype, may be effective in reducing the risk of CWP, although these results should be replicated in epidemiological and interventional studies.

SGLT2i and GLP-1 RA therapy in type 1 diabetes and reno-vascular outcomes: a real-world study.

AIMS/HYPOTHESIS: Insulin is the primary treatment for type 1 diabetes. However, alternative glucose-lowering therapies are used adjunctively, but importantly are off-label in type 1 diabetes. Little work has previously been undertaken to evaluate safety with long-term efficacy and cardio-renal benefits of such therapies. We sought to investigate the real-world impact of sodium-glucose cotransporter 2 inhibitor (SGLT2i) and glucagon-like peptide-1 receptor agonist (GLP-1 RA) therapy in individuals with type 1 diabetes in relation to effect on blood glucose levels, adverse events and cardio-renal outcomes. METHODS: We performed a retrospective cohort study of all patients aged 18 or over with type 1 diabetes on the TriNetX platform, a global collaborative network providing access to real-time, anonymised medical records. We included patients who had been treated with an SGLT2i or GLP-1 RA for at least 6 months and analysed the efficacy, safety and cardio-renal outcomes 5 years after initiation of therapy. RESULTS: We identified 196,691 individuals with type 1 diabetes, 13% of whom were treated with adjunctive glucose-lowering therapy in addition to insulin. Included in the core analysis were 1822 patients treated with a GLP-1 RA and 992 individuals treated with an SGLT2i. Both agents provided clinically meaningful reductions in HbA1c (-2.6 mmol/mol [-0.2%] with SGLT2i and -5.4 mmol/mol [-0.5%] with GLP-1 RA). The SGLT2i treated cohort showed preservation of eGFR over a 5-year period compared with the GLP-1 RA treated cohort (+3.5 ml/min per 1.73 m2 vs -7.2 ml/min per 1.73 m2, respectively), including patients with established chronic kidney disease (CKD). The SGLT2i treated cohort experienced higher rates of diabetic ketoacidosis (DKA) (RR 2.08 [95% CI 1.05, 4.12] p=0.0309) and urinary tract infection/pyelonephritis (RR 2.27 [95% CI 1.12, 4.55] p=0.019) compared with the GLP-1 RA treated cohort. However, the SGLT2i treated cohort were less likely to develop heart failure (RR 0.44 [95% CI 0.23, 0.83] p=0.0092), CKD (RR 0.49 [95% CI 0.28, 0.86] p=0.0118) and be hospitalised for any cause (RR 0.59 [95% CI 0.46, 0.76] p≤0.0001) when compared with the GLP-1 RA treated cohort. CONCLUSIONS/INTERPRETATION: Both SGLT2is and GLP-1 RAs have potential benefits as adjunctive agents in type 1 diabetes. SGLT2is provide cardio-renal benefits, despite an increase in the risk of DKA and urinary tract infection compared with GLP-1 RA therapy. Long-term evaluation of the efficacy and safety of these adjunctive therapies is required to guide their use in individuals with type 1 diabetes.

Metformin for the prevention of diabetes among people with HIV and either impaired fasting glucose or impaired glucose tolerance (prediabetes) in Tanzania: a Phase II randomised placebo-controlled trial.

AIMS/HYPOTHESIS: In sub-Saharan Africa (SSA), 5% of adults are living with type 2 diabetes and this is rising sharply, with a greater increase among people with HIV. Evidence on the efficacy of prevention strategies in this cohort is scarce. We conducted a Phase II double-blind placebo-controlled trial that aimed to determine the impact of metformin on blood glucose levels among people with prediabetes (defined as impaired fasting glucose [IFG] and/or impaired glucose tolerance [IGT]) and HIV in SSA. METHODS: Adults (≥18 years old) who were stable in HIV care and found to have prediabetes (IFG and/or IGT) and who were attending hospitals in Dar es Salaam, Tanzania, were randomised to receive sustained-release metformin, 2000 mg daily, or matching placebo between 4 November 2019 and 21 July 2020. Randomisation used permuted blocks. Allocation was concealed in the trial database and made visible only to the Chief Pharmacist after consent was taken. All participants, research and clinical staff remained blinded to the allocation. Participants were provided with information on diet and lifestyle and had access to various health information following the start of the coronavirus disease 2019 (COVID-19) pandemic. Participants were followed up for 12 months. The primary outcome measure was capillary blood glucose measured 2 h following a 75 g glucose load. Analyses were by intention-to-treat. RESULTS: In total, 364 participants (182 in each arm) were randomised to the metformin or placebo group. At enrolment, in the metformin and placebo arms, mean fasting glucose was 6.37 mmol/l (95% CI 6.23, 6.50) and 6.26 mmol/l (95% CI 6.15, 6.36), respectively, and mean 2 h glucose levels following a 75 g oral glucose load were 8.39 mmol/l (95% CI 8.22, 8.56) and 8.24 mmol/l (95% CI 8.07, 8.41), respectively. At the final assessment at 12 months, 145/182 (79.7%) individuals randomised to metformin compared with 158/182 (86.8%) randomised to placebo indicated that they had taken >95% of their medicines in the previous 28 days (p=0.068). At this visit, in the metformin and placebo arms, mean fasting glucose levels were 6.17 mmol/l (95% CI 6.03, 6.30) and 6.30 mmol/l (95% CI 6.18, 6.42), respectively, and mean 2 h glucose levels following a 75 g oral glucose load were 7.88 mmol/l (95% CI 7.65, 8.12) and 7.71 mmol/l (95% CI 7.49, 7.94), respectively. Using a linear mixed model controlling for respective baseline values, the mean difference between the metformin and placebo group (metformin-placebo) was -0.08 mmol/l (95% CI -0.37, 0.20) for fasting glucose and 0.20 mmol/l (95% CI -0.17, 0.58) for glucose levels 2 h post a 75 g glucose load. Weight was significantly lower in the metformin arm than in the placebo arm: using the linear mixed model adjusting for baseline values, the mean difference in weight was -1.47 kg (95% CI -2.58, -0.35). In total, 16/182 (8.8%) individuals had a serious adverse event (Grade 3 or Grade 4 in the Division of Acquired Immunodeficiency Syndrome [DAIDS] adverse event grading table) or died in the metformin arm compared with 18/182 (9.9%) in the placebo arm; these events were either unrelated to or unlikely to be related to the study drugs. CONCLUSIONS/INTERPRETATION: Blood glucose decreased over time in both the metformin and placebo arms during the trial but did not differ significantly between the arms at 12 months of follow up. Metformin therapy was found to be safe for use in individuals with HIV and prediabetes. A larger trial with longer follow up is needed to establish if metformin can be safely used for the prevention of diabetes in people who have HIV. TRIAL REGISTRATION: The trial is registered on the International Standard Randomised Controlled Trial Number (ISRCTN) registry ( www.isrctn.com/ ), registration number: ISCRTN76157257. FUNDING: This research was funded by the National Institute for Health Research using UK aid from the UK Government to support global health research.

Comparison of amitriptyline supplemented with pregabalin, pregabalin supplemented with amitriptyline, and duloxetine supplemented with pregabalin for the treatment of diabetic peripheral neuropathic pain (OPTION-DM): a multicentre, double-blind, randomised crossover trial.

BACKGROUND: Diabetic peripheral neuropathic pain (DPNP) is common and often distressing. Most guidelines recommend amitriptyline, duloxetine, pregabalin, or gabapentin as initial analgesic treatment for DPNP, but there is little comparative evidence on which one is best or whether they should be combined. We aimed to assess the efficacy and tolerability of different combinations of first-line drugs for treatment of DPNP. METHODS: OPTION-DM was a multicentre, randomised, double-blind, crossover trial in patients with DPNP with mean daily pain numerical rating scale (NRS) of 4 or higher (scale is 0-10) from 13 UK centres. Participants were randomly assigned (1:1:1:1:1:1), with a predetermined randomisation schedule stratified by site using permuted blocks of size six or 12, to receive one of six ordered sequences of the three treatment pathways: amitriptyline supplemented with pregabalin (A-P), pregabalin supplemented with amitriptyline (P-A), and duloxetine supplemented with pregabalin (D-P), each pathway lasting 16 weeks. Monotherapy was given for 6 weeks and was supplemented with the combination medication if there was suboptimal pain relief (NRS >3), reflecting current clinical practice. Both treatments were titrated towards maximum tolerated dose (75 mg per day for amitriptyline, 120 mg per day for duloxetine, and 600 mg per day for pregabalin). The primary outcome was the difference in 7-day average daily pain during the final week of each pathway. This trial is registered with ISRCTN, ISRCTN17545443. FINDINGS: Between Nov 14, 2017, and July 29, 2019, 252 patients were screened, 140 patients were randomly assigned, and 130 started a treatment pathway (with 84 completing at least two pathways) and were analysed for the primary outcome. The 7-day average NRS scores at week 16 decreased from a mean 6·6 (SD 1·5) at baseline to 3·3 (1·8) at week 16 in all three pathways. The mean difference was -0·1 (98·3% CI -0·5 to 0·3) for D-P versus A-P, -0·1 (-0·5 to 0·3) for P-A versus A-P, and 0·0 (-0·4 to 0·4) for P-A versus D-P, and thus not significant. Mean NRS reduction in patients on combination therapy was greater than in those who remained on monotherapy (1·0 [SD 1·3] vs 0·2 [1·5]). Adverse events were predictable for the monotherapies: we observed a significant increase in dizziness in the P-A pathway, nausea in the D-P pathway, and dry mouth in the A-P pathway. INTERPRETATION: To our knowledge, this was the largest and longest ever, head-to-head, crossover neuropathic pain trial. We showed that all three treatment pathways and monotherapies had similar analgesic efficacy. Combination treatment was well tolerated and led to improved pain relief in patients with suboptimal pain control with a monotherapy. FUNDING: National Institute for Health Research (NIHR) Health Technology Assessment programme.

Childhood, adolescent, and adulthood adiposity are associated with risk of PCOS: a Mendelian randomization study with meta-analysis.

STUDY QUESTION: What is the influence of body composition during childhood, adolescence, and adulthood, as well as metabolic parameters, on incident polycystic ovary syndrome (PCOS)? SUMMARY ANSWER: Excess body fat, even during childhood/adolescence, and metabolic parameters, suggestive of hyperinsulinaemia/insulin resistance, significantly impact the risk of PCOS in a linear fashion. WHAT IS KNOWN ALREADY: Observational and Mendelian randomization (MR) data have demonstrated an association between adulthood overweight/obesity and development of PCOS. However, the contribution of body composition in childhood/adolescence to incident PCOS is unclear, as is the influence of childhood overweight/obesity. STUDY DESIGN, SIZE, DURATION: We conducted a systematic review and meta-analysis and integrated our results with a previously published systematic review. Two blinded investigators screened abstracts published between November 2010 and May 2021. Furthermore, we incorporated summary statistics from genome-wide association study (GWAS) data in subjects of European ancestry. Adult overweight was defined as BMI ≥ 25 kg/m2 and obesity as BMI ≥ 30 kg/m2; in Asian subjects, overweight was defined as BMI ≥ 23 kg/m2 and obesity as BMI ≥ 25 kg/m2. PARTICIPANTS/MATERIALS, SETTING, METHODS: We utilized meta-analysis and MR together to allow synthesis of genetic and observational data. For the systematic review, the search revealed 71 studies, of which 63 were included in meta-analysis by calculating odds ratios (ORs) using the random-effects model. Furthermore, we conducted a two-sample MR study of GWAS data to determine the impact of childhood and adult body size (defined categorically by BMI and childhood body size proportions), abnormal body composition and metabolic parameters (higher fasting serum insulin or lower sex hormone-binding globulin (SHBG) concentration) on the odds of incident PCOS via the inverse-variance weighted method. MAIN RESULTS AND THE ROLE OF CHANCE: Significant associations were shown between body composition and PCOS incidence. From the systematic review/meta-analysis, women with overweight (OR 3.80, 2.87-5.03), obesity (OR 4.99, 3.74-6.67), and central obesity (OR 2.93, 2.08-4.12) had increased odds of PCOS. For adolescents with overweight and/or obesity, the PCOS odds were greater than for adults. From MR, for every standard deviation increase in BMI (4.8 kg/m2), the odds of PCOS increased by 2.76 (2.27-3.35). Childhood body size had an independent effect on PCOS odds after adjusting for adult body size (OR: 2.56, 1.57-4.20). Genetically determined body fat percentage (OR 3.05, 2.24-4.15), whole body fat mass (OR 2.53, 2.04-3.14), fasting serum insulin (OR 6.98, 2.02-24.13), and SHBG concentration (OR 0.74, 0.64-0.87) were all significantly associated with PCOS in a linear relation. LIMITATIONS, REASONS FOR CAUTION: The meta-analysis included studies which were cross-sectional and retrospective, limiting our ability to determine causality. MR was limited by interrogating subjects only of European ancestry and including cases classified by either self-diagnosis or diagnostic criteria. WIDER IMPLICATIONS OF THE FINDINGS: Our study demonstrates for the first time a critical role of the impact of excess childhood/adolescent adiposity on the pathophysiology of adult PCOS. Our results, driven by genetically determined childhood/adolescent body composition, higher BMI, hyperinsulinaemia, and lower SHBG, clearly favour obesity driving the metabolic, but not reproductive, PCOS phenotype. Overall, effective weight maintenance, even from the early years, is likely to reduce the risk of this reproductive endocrine disorder. STUDY FUNDING/COMPETING INTEREST(S): S.S.Z. was funded by a National Institute for Health and Care Research (NIHR) Academic Clinical Lectureship. U.A. is chair of the NIHR Steering Committee Trial-CASSANDRA-DN. No other authors declare any sources of funding or relevant conflicts of interest. The authors declare that the research was conducted in the absence of any commercial or financial relations that could be construed as a potential conflict of interest. TRIAL REGISTRATION NUMBER: N/A.

Multimodal testing reveals subclinical neurovascular dysfunction in prediabetes, challenging the diagnostic threshold of diabetes.

AIM: To explore if novel non-invasive diagnostic technologies identify early small nerve fibre and retinal neurovascular pathology in prediabetes. METHODS: Participants with normoglycaemia, prediabetes or type 2 diabetes underwent an exploratory cross-sectional analysis with optical coherence tomography angiography (OCT-A), handheld electroretinography (ERG), corneal confocal microscopy (CCM) and evaluation of electrochemical skin conductance (ESC). RESULTS: Seventy-five participants with normoglycaemia (n = 20), prediabetes (n = 29) and type 2 diabetes (n = 26) were studied. Compared with normoglycaemia, mean peak ERG amplitudes of retinal responses at low (16-Td·s: 4.05 µV, 95% confidence interval [95% CI] 0.96-7.13) and high (32-Td·s: 5·20 µV, 95% CI 1.54-8.86) retinal illuminance were lower in prediabetes, as were OCT-A parafoveal vessel densities in superficial (0.051 pixels/mm2 , 95% CI 0.005-0.095) and deep (0.048 pixels/mm2 , 95% CI 0.003-0.093) retinal layers. There were no differences in CCM or ESC measurements between these two groups. Correlations between HbA1c and peak ERG amplitude at 32-Td·s (r = -0.256, p = 0.028), implicit time at 32-Td·s (r = 0.422, p < 0.001) and 16-Td·s (r = 0.327, p = 0.005), OCT parafoveal vessel density in the superficial (r = -0.238, p = 0.049) and deep (r = -0.3, p = 0.017) retinal layers, corneal nerve fibre length (CNFL) (r = -0.293, p = 0.017), and ESC-hands (r = -0.244, p = 0.035) were observed. HOMA-IR was a predictor of CNFD (ß = -0.94, 95% CI -1.66 to -0.21, p = 0.012) and CNBD (ß = -5.02, 95% CI -10.01 to -0.05, p = 0.048). CONCLUSIONS: The glucose threshold for the diagnosis of diabetes is based on emergent retinopathy on fundus examination. We show that both abnormal retinal neurovascular structure (OCT-A) and function (ERG) may precede retinopathy in prediabetes, which require confirmation in larger, adequately powered studies.

All-cause mortality and cardiovascular outcomes with sodium-glucose Co-transporter 2 inhibitors, glucagon-like peptide-1 receptor agonists and with combination therapy in people with type 2 diabetes.

AIM: To assess the relationship of sodium-glucose cotransporter-2 inhibitors (SGLT2i), glucagon-like peptide-1 receptor analogues (GLP-1RA) and their combination (SGLT2i + GLP-1RA) with 5-year risk of all-cause mortality, hospitalization and cardiovascular/macrovascular disease in people with type 2 diabetes. MATERIALS AND METHODS: Retrospective cohort analysis of 2.2 million people with type 2 diabetes receiving insulin across 85 health care organizations using a global federated health research network. Three intervention cohorts (SGLT2i, GLP-1RA and SGLT2i + GLP-1RA) were compared against a control cohort (no SGLT2i/GLP-1RA). Propensity score matching for age, ischaemic heart disease, sex, hypertension, chronic kidney disease, heart failure and glycated haemoglobin was used to balance cohorts 1:1 (SGLT2i, n = 143 600; GLP-1RA, n = 186 841; SGLT-2i + GLP-1RA, n = 108 504). A sub-analysis comparing combination and monotherapy cohorts was also performed. RESULTS: The intervention cohorts showed a reduced hazard ratio (HR, 95% confidence interval) over 5 years compared with the control cohort for all-cause mortality (SGLT2i 0.49, 0.48-0.50; GLP-1RA 0.47, 0.46-0.48; combination 0.25, 0.24-0.26), hospitalization (0.73, 0.72-0.74; 0.69, 0.68-0.69; 0.60, 0.59-0.61) and acute myocardial infarct (0.75, 0.72-0.78; 0.70, 0.68-0.73; 0.63, 0.60-0.66), respectively. All other outcomes showed a significant risk reduction in favour of the intervention cohorts. The sub-analysis showed a significant risk reduction in all-cause mortality for combination therapy versus SGLT2i (0.53, 0.50-0.55) and GLP-1RA (0.56, 0.54-0.59). CONCLUSIONS: SGLT2i, GLP-1RAs or combination therapy confers mortality and cardiovascular protection in people with type 2 diabetes over 5 years. Combination therapy was associated with the greatest risk reduction in all-cause mortality versus a propensity matched control cohort. In addition, combination therapy offers a reduction in 5-year all-cause mortality when compared directly against either monotherapy.

Artificial intelligence utilising corneal confocal microscopy for the diagnosis of peripheral neuropathy in diabetes mellitus and prediabetes.

AIMS/HYPOTHESIS: We aimed to develop an artificial intelligence (AI)-based deep learning algorithm (DLA) applying attribution methods without image segmentation to corneal confocal microscopy images and to accurately classify peripheral neuropathy (or lack of). METHODS: The AI-based DLA utilised convolutional neural networks with data augmentation to increase the algorithm's generalisability. The algorithm was trained using a high-end graphics processor for 300 epochs on 329 corneal nerve images and tested on 40 images (1 image/participant). Participants consisted of healthy volunteer (HV) participants (n = 90) and participants with type 1 diabetes (n = 88), type 2 diabetes (n = 141) and prediabetes (n = 50) (defined as impaired fasting glucose, impaired glucose tolerance or a combination of both), and were classified into HV, those without neuropathy (PN-) (n = 149) and those with neuropathy (PN+) (n = 130). For the AI-based DLA, a modified residual neural network called ResNet-50 was developed and used to extract features from images and perform classification. The algorithm was tested on 40 participants (15 HV, 13 PN-, 12 PN+). Attribution methods gradient-weighted class activation mapping (Grad-CAM), Guided Grad-CAM and occlusion sensitivity displayed the areas within the image that had the greatest impact on the decision of the algorithm. RESULTS: The results were as follows: HV: recall of 1.0 (95% CI 1.0, 1.0), precision of 0.83 (95% CI 0.65, 1.0), F1-score of 0.91 (95% CI 0.79, 1.0); PN-: recall of 0.85 (95% CI 0.62, 1.0), precision of 0.92 (95% CI 0.73, 1.0), F1-score of 0.88 (95% CI 0.71, 1.0); PN+: recall of 0.83 (95% CI 0.58, 1.0), precision of 1.0 (95% CI 1.0, 1.0), F1-score of 0.91 (95% CI 0.74, 1.0). The features displayed by the attribution methods demonstrated more corneal nerves in HV, a reduction in corneal nerves for PN- and an absence of corneal nerves for PN+ images. CONCLUSIONS/INTERPRETATION: We demonstrate promising results in the rapid classification of peripheral neuropathy using a single corneal image. A large-scale multicentre validation study is required to assess the utility of AI-based DLA in screening and diagnostic programmes for diabetic neuropathy.

ASSOCIATIONS BETWEEN DYSGLYCEMIA, RETINAL NEURODEGENERATION, AND MICROALBUMINURIA IN PREDIABETES AND TYPE 2 DIABETES.

PURPOSE: To explore the association between retinal neurodegeneration and metabolic parameters in progressive dysglycemia. METHOD: A cross-sectional study was performed on 68 participants: normal glucose tolerance (n = 23), prediabetes (n = 25), and Type 2 diabetes without diabetic retinopathy (n = 20). Anthropometric assessment and laboratory sampling for HbA1c, fasting glucose, insulin, c-peptide, lipid profile, renal function, and albumin-to-creatinine ratio were conducted. Central and pericentral macular thicknesses on spectral domain optical coherence tomography were compared with systemic parameters. RESULTS: Baseline demographic characteristics were similar across all groups. Cuzick's trend test revealed progressive full-thickness macular thinning with increasing dysglycemia across all three groups (P = 0.015). The urinary albumin-to-creatinine ratio was significantly correlated with full-thickness superior (R = -0.435; P = 0.0002), inferior (R = -0.409; P = 0.0005), temporal (R = -0.429; P = 0.003), and nasal (R = -0.493; P < 0.0001) pericentral macular thinning, after post hoc Bonferroni adjustment. There was no association between macular thinning and waist circumference, body mass index, blood pressure, lipid profile, or insulin resistance. CONCLUSION: Progressive dysglycemia is associated with macular thinning before the onset of visible retinopathy and occurs alongside microalbuminuria. Retinal neurodegenerative changes may help identify those most at risk from dysglycemic end-organ damage.

Lipids and peripheral neuropathy.

PURPOSE OF REVIEW: Hyperlipidaemia is associated with the development of neuropathy. Indeed, a mechanistic link between altered lipid metabolism and peripheral nerve dysfunction has been demonstrated in a number of experimental and clinical studies. Furthermore, post hoc analyses of clinical trials of cholesterol and triglyceride-lowering pharmacotherapy have shown reduced rates of progression of diabetic neuropathy. Given, there are currently no FDA approved disease-modifying therapies for diabetic neuropathy, modulation of lipids may represent a key therapeutic target for the treatment of diabetic nerve damage. This review summarizes the current evidence base on the role of hyperlipidaemia and lipid lowering therapy on the development and progression of peripheral neuropathy. RECENT FINDINGS: A body of literature supports a detrimental effect of dyslipidaemia on nerve fibres resulting in somatic and autonomic neuropathy. The case for an important modulating role of hypertriglyceridemia is stronger than for low-density lipoprotein cholesterol (LDL-C) in relation to peripheral neuropathy. This is reflected in the outcomes of clinical trials with the different therapeutic agents targeting hyperlipidaemia reporting beneficial or neutral effects with statins and fibrates. The potential concern with the association between proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor therapy and cognitive decline raised the possibility that extreme LDL-C lowering may result in neurodegeneration. However, studies in murine models and data from small observational studies indicate an association between increased circulating PCSK9 levels and small nerve fibre damage with a protective effect of PCSK9i therapy against small fibre neuropathy. Additionally, weight loss with bariatric surgery leads to an improvement in peripheral neuropathy and regeneration of small nerve fibres measured with corneal confocal microscopy in people with obesity with or without type 2 diabetes. These improvements correlate inversely with changes in triglyceride levels. SUMMARY: Hyperlipidaemia, particularly hypertriglyceridemia, is associated with the development and progression of neuropathy. Lipid modifying agents may represent a potential therapeutic option for peripheral neuropathy. Post hoc analyses indicate that lipid-lowering therapies may halt the progression of neuropathy or even lead to regeneration of nerve fibres. Well designed randomized controlled trials are needed to establish if intensive targeted lipid lowering therapy as a part of holistic metabolic control leads to nerve fibre regeneration and improvement in neuropathy symptoms.

The prevalence of cardiac autonomic neuropathy in prediabetes: a systematic review.

AIMS/HYPOTHESIS: Cardiac autonomic neuropathy (CAN) is independently associated with silent myocardial ischaemia, major cardiovascular events, myocardial dysfunction and cardiovascular mortality. Several studies have highlighted the increased prevalence of CAN in prediabetes (impaired glucose tolerance and/or impaired fasting glucose). Considering the exponential rise of prediabetes, we aimed to determine the prevalence of CAN through a systematic literature review. METHODS: This systematic review was registered with PROSPERO (CRD42019125447). An electronic literature search was performed using MEDLINE, EMBASE, PubMed, Web of Science, Scopus and Cochrane databases. Published full text, English language articles that provide CAN prevalence data of studies in individuals with prediabetes and aged over 18 years were included. Prevalence data for normal glucose tolerance and diabetes were also extracted from the selected articles, if present. All articles were screened by two independent reviewers using a priori criteria. Methodological quality and risk of bias were evaluated using a critical appraisal tool. RESULTS: Database searches found 4500 articles; subsequently, 199 full text articles were screened, 11 of which fulfilled the inclusion criteria (4431 total participants, 1730 people with prediabetes, 1999 people with normal glucose tolerance [NGT] and 702 people with predominantly type 2 diabetes). Six of the selected studies reported definite CAN prevalence data (9-39%). Only a single large population-based study by Ziegler et al (KORA S4 study, 1332 participants) determined definite CAN based on two or more positive autonomic function tests (AFTs), with a mean prevalence of 9% in all prediabetes groups (isolated impaired glucose tolerance 5.9%; isolated impaired fasting glucose 8.1%; impaired fasting glucose plus impaired glucose tolerance 11.4%), which was higher than NGT (4.5%). This study is most likely to provide a reliable population-specific estimate of CAN in prediabetes. There was a higher than expected prevalence of CAN in prediabetes (9-38%) when compared with normal glucose tolerance (0-18%) within the same studies (n = 8). There was a wide prevalence of possible CAN based on one positive AFT (n = 5). There was heterogeneity between the studies with variations in the definition of CAN, methodology and characteristics of the populations, which likely contributed to the diversity of prevalence estimates. The overall risk of bias was low. CONCLUSIONS/INTERPRETATION: There is a higher than expected prevalence of CAN in prediabetes. Early detection of CAN in prediabetes through population screening needs careful consideration in view of the excess morbidity and mortality risk associated with this condition. Graphical abstract.

Vitamin D deficiency is associated with painful diabetic neuropathy.

BACKGROUND: The aetiology of painful diabetic neuropathy is unclear. We have evaluated vitamin D levels in diabetic patients with and without painful neuropathy. METHODS: Forty-three patients with type 1 diabetes and painless (DPN) (n = 20) or painful (PDN) (n = 23) neuropathy and 14 non-diabetic healthy control subjects (C) underwent assessment of neurologic deficits, quantitative sensory testing (QST), electrophysiology, skin biopsy, corneal confocal microscopy (CCM) and measurement of serum 25(OH)D. RESULTS: There were no significant differences for age, BMI, HbA1c , lipids, neurological deficits, QST, electrophysiology, intra-epidermal nerve fibre density (IENFD) and corneal nerve morphology between patients with DPN and PDN. Both positive (hyperalgesia and allodynia) and negative symptoms (paraesthesia and numbness) of diabetic neuropathy were greater in PDN compared with DPN (P = .009 and P = .02, respectively). Serum 25(OH)D levels were significantly lower in PDN (24.0 ± 14.1 ng/mL) compared with DPN (34.6 ± 15.0 ng/mL, P = .01) and controls (34.1 ± 8.6 ng/mL, P = .03). The odds ratio in favour of painful diabetic neuropathy was 9.8 [P = .003 (95% CI, 2.2-76.4)] for vitamin D deficiency (<20 ng/mL) and 4.4 [P = .03 (95% CI, 1.1-19.8)] for vitamin D insufficiency (<30 ng/mL). CONCLUSIONS: This study suggests that vitamin D deficiency and insufficiency are associated with painful diabetic neuropathy.

Greater small nerve fibre damage in the skin and cornea of type 1 diabetic patients with painful compared to painless diabetic neuropathy.

BACKGROUND AND AIM: Damage to small nociceptive fibres may contribute to painful diabetic neuropathy. We aimed to compare large and small nerve fibre measurements together with skin biopsy and corneal confocal microscopy in patients with type 1 diabetes and painful or painless diabetic neuropathy. METHODS: We have assessed the McGill pain questionnaire, neuropathy disability score, vibration perception threshold, warm and cold sensation thresholds, electrophysiology, corneal confocal microscopy and skin biopsy in participants with type 1 diabetes and painful (n = 41) or painless (n = 50) diabetic neuropathy and control subjects (n = 50). RESULTS: The duration of diabetes, body mass index, glycated haemoglobin (HbA1c), blood pressure and lipid profile did not differ between subjects with painful and painless neuropathy. Neuropathy disability score and vibration perception threshold were higher and sural nerve conduction velocity was lower, but sural nerve amplitude, peroneal nerve amplitude and conduction velocity and cold and warm sensation thresholds did not differ between patients with painful compared to painless diabetic neuropathy. However, intraepidermal nerve fibre density, corneal nerve fibre density, corneal nerve branch density and corneal nerve fibre length were significantly lower in subjects with painful compared to painless diabetic neuropathy. CONCLUSIONS: There is evidence of more severe neuropathy, particularly small fibre damage in the skin and cornea, of patients with painful compared to painless diabetic neuropathy.