Prediction of cardiovascular risk factors from retinal fundus photographs: Validation of a deep learning algorithm in a prospective non-interventional study in Kenya.

AIM: Hypertension and diabetes mellitus (DM) are major causes of morbidity and mortality, with growing burdens in low-income countries where they are underdiagnosed and undertreated. Advances in machine learning may provide opportunities to enhance diagnostics in settings with limited medical infrastructure. MATERIALS AND METHODS: A non-interventional study was conducted to develop and validate a machine learning algorithm to estimate cardiovascular clinical and laboratory parameters. At two sites in Kenya, digital retinal fundus photographs were collected alongside blood pressure (BP), laboratory measures and medical history. The performance of machine learning models, originally trained using data from the UK Biobank, were evaluated for their ability to estimate BP, glycated haemoglobin, estimated glomerular filtration rate and diagnoses from fundus images. RESULTS: In total, 301 participants were enrolled. Compared with the UK Biobank population used for algorithm development, participants from Kenya were younger and would probably report Black/African ethnicity, with a higher body mass index and prevalence of DM and hypertension. The mean absolute error was comparable or slightly greater for systolic BP, diastolic BP, glycated haemoglobin and estimated glomerular filtration rate. The model trained to identify DM had an area under the receiver operating curve of 0.762 (0.818 in the UK Biobank) and the hypertension model had an area under the receiver operating curve of 0.765 (0.738 in the UK Biobank). CONCLUSIONS: In a Kenyan population, machine learning models estimated cardiovascular parameters with comparable or slightly lower accuracy than in the population where they were trained, suggesting model recalibration may be appropriate. This study represents an incremental step toward leveraging machine learning to make early cardiovascular screening more accessible, particularly in resource-limited settings.

What are the factors associated with long-term glycaemic control in patients with type 2 diabetes and elevated glycated haemoglobin (≥7.0%) at initiation of second-line therapy? Results from the DISCOVER study.

AIMS: Glycaemic control is a cornerstone of type 2 diabetes (T2D) management. We assessed factors associated with good long-term glycaemic control in patients with glycated haemoglobin (HbA1c) ≥7.0% at initiation of second-line glucose-lowering therapy, using data from DISCOVER, a global, prospective, 3-year observational study of patients with T2D. MATERIALS AND METHODS: This analysis included patients with HbA1c ≥7.0% at baseline (initiation of second-line therapy). Multivariable regression models assessed factors associated with having HbA1c <7.0% at 3 years in two distinct groups: patients with (a) HbA1c ≥7.0% and <9.0%, and (b) HbA1c ≥9.0% at baseline. RESULTS: In total, 7575 patients with baseline HbA1c ≥7.0% were included (2233 with baseline HbA1c ≥9.0%). At 6 months, 43.7% and 24.2% of patients had an HbA1c level <7.0% in groups a and b, respectively; the corresponding proportions at 3 years were 45.8% and 29.3%. Having HbA1c <7.0% at 6 months (vs. ≥7.0%) was the strongest predictor of having HbA1c <7.0% at 3 years in both group a and group b [odds ratio (95% confidence interval): 2.01 (1.77-2.27) and 2.68 (2.10-3.41), respectively]. Longer T2D duration was associated with a decreased likelihood of having HbA1c <7.0% at 3 years. CONCLUSIONS: In patients with poor glycaemic control at initiation of second-line therapy, early attainment of HbA1c <7.0% appears predictive of long-term glycaemic control, suggesting that timely modification of treatment strategies in patients with elevated HbA1c after 6 months is important to minimize therapeutic inertia.

Health-related quality of life in patients with type 2 diabetes initiating a second-line glucose-lowering therapy: The DISCOVER study.

AIM: To investigate factors associated with health-related quality of life (HRQoL) in patients with type 2 diabetes mellitus (T2D) at initiation of second-line glucose-lowering therapy. METHODS: DISCOVER is a 3-year, prospective observational study of patients with T2D initiating second-line glucose-lowering therapy, conducted in 38 countries. HRQoL at baseline was assessed using the physical and mental component summary (PCS; MCS) scores of the 36-Item Short Form Health Survey version 2 (SF-36v2) in 31 countries (n = 8309) and the Hypoglycaemia Fear Survey-II (HFS-II) in 23 countries (n = 6516). Factors associated with differences in HRQoL were assessed using multivariable hierarchical regression models. RESULTS: Mean PCS and MCS scores were 48.0 (standard deviation [SD]: 7.8) and 45.5 (SD: 10.4), respectively. Factors associated with significantly lower SF-36v2 scores included being female, having a history of macrovascular complications and first-line treatment with oral combinations (vs metformin monotherapy). Mean HFS-II behaviour and worry scores were 8.2 (SD: 9.9) and 7.3 (SD: 11.8), respectively. Increased fear of hypoglycaemia was significantly associated with lower SF-36v2 scores. CONCLUSIONS: Several patient-, disease- and treatment-related characteristics correlated with HRQoL, indicating that a multifactorial approach is needed to maintain HRQoL in patients with T2D.

Associations between second-line glucose-lowering combination therapies with metformin and HbA1c, body weight, quality of life, hypoglycaemic events and glucose-lowering treatment intensification: The DISCOVER study.

AIM: To explore the effects of second-line combination therapies with metformin on body weight, HbA1c and health-related quality of life, as well as the risks of hypoglycaemia and further treatment intensification in the DISCOVER study, a 3-year, prospective, global observational study of patients with type 2 diabetes initiating second-line glucose-lowering therapy. MATERIALS AND METHODS: Adjusted changes from baseline in weight, HbA1c and 36-item Short Form Health Survey version 2 (SF-36v2) summary scores at 6, 12, 24 and 36 months were assessed using linear mixed models. Risk of hypoglycaemia and further intensification were assessed using interval censored analyses. RESULTS: At baseline, 7613 patients received metformin in combination with a sulphonylurea (SU; 40.9%), a dipeptidyl peptidase-4 (DPP-4) inhibitor (48.3%), a sodium-glucose co-transporter-2 (SGLT-2) inhibitor (8.3%) or a glucagon-like peptide-1 (GLP-1) receptor agonist (2.4%). After 36 months, all combinations showed similar reductions in HbA1c (0.8%-1.0%), however, metformin plus a DPP-4 inhibitor, an SGLT-2 inhibitor or a GLP-1 receptor agonist was associated with greater weight loss (1.9, 2.9 and 5.0 kg, respectively) than metformin plus an SU (1.3 kg, P < .0001). Proportions of further treatment intensification were similar across combinations (19.9%-26.2%). Patients prescribed metformin plus an SU more often reported one or more hypoglycaemic events (11.9%) than other combinations (3.9%-6.4%, P < .0001). SF-36v2 summary scores were typically lowest among patients prescribed metformin and an SU. CONCLUSIONS: Combinations of metformin with an SU were associated with the lowest weight reduction, highest risk of hypoglycaemia and lower SF-36v2 scores.

Pakistan: the new target of terrorism. Are Karachi's emergency medical response systems adequately prepared?

OBJECTIVE: To assess the efficacy and preparedness of the pre-hospital and hospital emergency medical systems and post graduate trainees in the city to deal with a massive terrorist strike. METHODS: A cross-sectional survey of postgraduate trainees was conducted at Jinnah Postgraduate Medical Center and Civil Hospital Karachi from 21st July 2007 to 24th July 2007, to evaluate the preparedness and self identified deficiencies of doctors involved in massive trauma casualty management. To assess the pre-hospital care in Karachi, structured telephonic interviews were conducted of administrators of two private run charity based ambulance services. RESULTS: Out of the 90 respondents questioned regarding a self assessment of their training, only 3 (3.3%) of them were confident about their management of bomb blast victims. Eighty-seven (96.6%) of the respondents felt they required some further training (44.4%) or comprehensive training (52.2%). No simulated drills or courses had been conducted for disaster management in the emergency department of the surveyed Hospitals. Most of the ambulance drivers had no paramedic training. Ambulances are equipped with a stretcher and an oxygen cylinder only. No resuscitation measures are available in the ambulances. CONCLUSION: With an increasing number of terrorist attacks in the country, massive influx of casualties in a relatively short time span has become a regular feature of the Pakistani hospital system. Lack of adequate training at pre-hospital and in hospital levels may translate into increasing morbidities and mortalities. It is imperative that training of junior doctors and paramedical staff be conducted regularly and an effective regional communication base established for efficient interdepartmental coordination.