Diabetes epidemiology and management最新文献

Patients with type 2 diabetes mellitus (T2DM) are exposed to a high risk of atherosclerotic cardiovascular disease, heart failure and chronic kidney disease. The incidence of these complications increases markedly with the duration of diabetes and aging. Sodium-glucose cotransporter 2 inhibitors (SGLT2is) showed a remarkable reduction in hospitalization for heart failure and progression of kidney disease in large prospective placebo-controlled trials. Post hoc analyses of these trials demonstrated that cardiorenal protection occurred independently of age. The present comprehensive review analyzes the effects of SGLT2is on cardiovascular and renal outcomes among older patients with T2DM in cohort studies and real-life conditions. SGLT2is were associated with a significant reduction in hospitalization for heart failure (alone or combined with mortality) and in a composite renal outcome, including end-stage renal disease when compared to other oral glucose-lowering drugs, dipeptidyl peptidase-4 inhibitors and glucagon-like peptide-1 receptor agonists in patients aged ≥ 65 years and even ≥ 75 years. Several observational studies worldwide compared cardiorenal outcomes in people aged ≥ 65 years versus < 65 years and showed a similar relative benefit of SGLT2is in older versus younger patients with T2DM. These favourable results were obtained while the safety profile of SGLT2is in older patients was acceptable and almost comparable with that reported in younger patients. In conclusion, observational studies in real-life conditions confirm previous results reported in placebo-controlled trials and a positive benefit/risk balance in elderly patients with T2DM at risk of heart failure and chronic kidney disease.

Introduction

Big data sources represent an opportunity for diabetes research. One example is the French national health data system (SNDS), gathering information on medical claims of out-of-hospital health care and hospitalizations for the entire French population (66 million). Currently, a validated algorithm based on antidiabetic drug reimbursement is able to identify people with pharmacologically-treated diabetes in the SNDS. But it cannot distinguish type 1 from type 2 diabetes. Differentiating type 1 and type 2 diabetes is crucial in diabetes surveillance, because they carry differences in their prevention, populations at risk, disease natural history, pathophysiology, management and risk of complications.

This article investigates the development of a type 1/type 2 diabetes classification algorithm using artificial intelligence and its application to estimate the prevalence of type 1 and type 2 diabetes in France.

Methods

The final data set comprised all diabetes cases from the CONSTANCES cohort (n = 951). A supervised machine learning method based on eight steps was used: final data set selection, target definition (type 1), coding features, final data set splitting into training and testing data sets, feature selection and training and validation and selection of algorithms. The selected algorithm was applied to SNDS data to estimate the type 1 and type 2 diabetes prevalence among adults 18–70 years of age.

Results

Among the 3481 SNDS features, 14 were selected to train the different algorithms. The final algorithm was a linear discriminant analysis model based on the number of reimbursements for fast-acting insulin, long-acting insulin and biguanides over the previous year (specificity 97% and sensitivity 100%). In 2016, after adjusting for algorithm performance, type 1 and type 2 diabetes prevalence in France was estimated to be 0.3% and 4.4%, respectively.

Conclusion

Our type 1/type 2 classification algorithm was found to perform well and to be applicable to any prescription or medical claims database from other countries. Artificial intelligence opens new possibilities for research and diabetes prevention.

Aims

The study aimed to identify weight-based insulin requirements for dexamethasone-induced hyperglycemia in COVID-19 infection stratified by hemoglobin A1c (HbA1c).

Methods

This retrospective study assessed hospitalized patients ≥ 18 years admitted with COVID-19 and receiving ≥ 1 dose of dexamethasone 6 mG. Daily blood glucose (BG) and insulin doses were collected and organized by HbA1c.

Results

Among 45 patients with available HbA1c, 100% [HbA1c ≥ 7%] and 72% [HbA1c < 7%] developed hyperglycemia (BG ≥180 mG/dL). Median daily insulin (Interquartile Range) (units/kG/day) was 0.03 (0, 0.32) [HbA1c 6–6.9%], 0.1 (0.06, 0.36) [HbA1c 7–7.9%], 0.66 (0.39, 0.69) [HbA1c 8–8.9%], and 0.72 (0.63, 0.78) [HbA1c ≥ 9%]. On day 10 of dexamethasone, when majority of patients were at goal BG, patients required 0.07 (0.01, 0.31) [HbA1c 6–6.9%], 0.59 (0.11, 0.75) [HbA1c 7–7.9%], 1.15 (0.95, 1.35) [HbA1c 8–8.9%], and 1.14 units/kG/day [HbA1c ≥ 9%]. Of 24 patients completing 10 days of dexamethasone, 25% experienced hypoglycemia (BG < 70 mG/dL) upon discontinuation.

Conclusion

Patients with higher HbA1c experienced greater dexamethasone-induced hyperglycemia and required higher insulin doses. Inpatient insulin dosing algorithms should take into consideration baseline HbA1c to avoid delays in achieving normoglycemia.

Aims

To assess the prognosis and risk factors for diabetic ketoacidosis (DKA) in Tampere University Hospital (Tays) in a retrospective case-control study.

Methods

All 282 patients (age ≥15 years) treated for DKA in Tays during the period 2014–2020 were included. A total of 846 controls adjusted for age, gender, diabetes type and municipality, and without any DKA during follow-up were collected from the Finnish National Diabetes Registry. HbA1c, mental and behavioural disorders, and mortality obtained from the Finnish National Diabetes Registry were compared between patients with and without DKA.

Results

Patients’ median age was 36 years. Ten percent of the patients with DKA died during the median follow-up time of three years. Mortality rate was sixfold higher in patients with DKA than among the controls (OR 6.28; 95% CI 3.17–12.42). Patients with DKA had higher rates of substance abuse (OR 4.68; 95% CI 3.23–6.78) and depression (OR 2.24; 95% CI 1.58–3.18), and higher median HbA1c levels (84 vs. 61 mmol/mol, p < 0.001). Nineteen percent of the DKA patients (n = 53) had recurrent DKA.

Conclusions

DKA is a strong indicator for premature death. Poor glycaemic control, depression and substance abuse are risk factors for DKA.

Background

Gold and Clarke questionnaire are originally developed to assess impaired awareness of hypoglycaemia (IAH) in type 1 diabetes. Present study examined the similarities and differences between the two questionnaires when administered to insulin-treated type 2 diabetes patients.

Methods

A total of 153 insulin-treated type 2 diabetes patients with mean age of 61.0±9.4 years and mean HbA1c of 8.4±1.5% completed questionnaire in diabetes outpatient clinics of tertiary-care hospital. Factor analysis was conducted to examine the psychometric properties of Clarke questionnaire. Spearman's correlation was used to examine convergent validity of Clarke questionnaire with Gold method.

Results

Bifactorial structure for Clarke questionnaire was identified, namely Awareness of Hypoglycaemia (Factor 1) and Experience of Hypoglycaemia (Factor 2). Clarke Factor 1 correlated strongly with Gold scores (r_s=0.77, p<0.001), and yielded 22.9% prevalence of IAH using cut-off score of ≥2.5, which is comparable to Gold method of 19.6%.

Conclusions

Gold single-item questionnaire assesses hypoglycaemia awareness only while Clarke questionnaire assesses both hypoglycaemia awareness and severe hypoglycaemia events. There is a high degree of convergence between Gold and Clarke in hypoglycaemia awareness assessment among insulin-treated type 2 diabetes. Hence, these two questionnaires are similar but not interchangeable due to bifactorial nature of Clarke questionnaire.

Background

Many works are ongoing with the aim of obtaining a more convenient way than the parenteral injection for administering insulin.

Purpose

To review the biopharmaceutics and clinical outcomes of the various emerging dosage forms of insulin so as to identify the promising formulations.

Method

A systematic literature search with analysis was carried out to obtain information on the biopharmaceutics and clinical outcomes of the emerging dosage forms.

Results

Intraperitoneal insulin was found to be characterized by direct drug delivery through the portal vein to the liver having bioavailability of 60%, but its clinical application is limited by the high risk of infection. The bioavailability of transdermal insulin has been enhanced using electrical, mechanical and physical techniques; and such formulations could achieve up to 39.5% blood glucose reduction. Oral insulin, known to be the most convenient, has its bioavailability limited to 1% by enzymatic degradation and poor absorption. Its challenges however, have been addressed by various interventions to achieve different levels of bioavailability up to 73.1%. Buccal insulin has shown potentials in managing postprandial hyperglycaemia without posing hypoglycaemic risk but its clinical applicability has not been established; whereas the long transit time, lower levels of peptidases and incorporation of permeation-enhancers have been shown to be responsible for the good treatment outcome of colon-targeted insulin. Rectal insulin with bioavailability of 11% has been shown to be considerably safe but not cost-effective while the ocular insulin is limited by poor absorption. Nasal tolerance and high rate of treatment failures were shown to be limiting intranasal insulin while the pulmonary insulin is being limited by peripheral drug retention and insulin resistance.

Conclusion

The biopharmaceutical profiles and clinical outcomes of transdermal, oral and colon-targeted insulin are superior to those of the other dosage forms. Further research works could be done towards the full development of these promising formulations.

Background

Malaria and haemolysis have been linked to a preponderance of altered glycaemic indices. This study set out to estimate the association between asymptomatic malaria and the Fasting Plasma Glucose (FPG), glycated haemoglobin (HBA1c) and Oral Glucose Tolerance (OGT) tests.

Methods

A cross-sectional survey was conducted at a general hospital in eastern Uganda. Eligible participants were patients aged 30–75 years, seeking care at the outpatient department, of unknown diabetes status. Participants were tested for FPG, OGT and HBA1c tests. Multiple linear regression and ROC curve analysis were conducted for the three tests.

Results

A total of 504 participants were enrolled on the study, of whom 78.4% (395) were female. After adjusting for age, sex, and BMI, individuals with asymptomatic malaria had lower average HBA1c [-5 mmol/mol (95% CI, -7 -2) and OGT tests levels [-1.75 mmol/l (-2.6, -0.8)]. The optimal cut-off points for diabetes among individuals with asymptomatic malaria were lower for the HBA1c test [6.5% (47 mmol/mol) versus 6.6% (49 mmol/mol), respectively] but higher for the FPG test (6.6 mmol/l versus 6.2 mmol/l, respectively).

Conclusions

These findings may have implications for diabetes screening in malaria-endemic settings.

Aims

The objective of the study was to compare screening performances of HbA1c, fasting plasma glucose (FPG), and two-hour plasma glucose (2hPG) in heart failure (HF) patients.

Methods

We included 237 HF patients aged >20 years without history of diabetes, using National Health and Nutrition Examination Survey data (2005–2016). American Diabetes Association diabetes screening criteria were used: (1) HbA1c ≥6.5%, (2) FPG ≥126 mg/dL, and (3) 2hPG ≥200 mg/dL. Sensitivity, specificity, predictive values, and Receiver Operating Characteristic (ROC) curves for HbA1c and FPG were examined against reference methods.

Results

N = 50 patients (20.5%) met at least 1 of 3 clinical criteria for diabetes. 2hPG alone identified 70.5% of patients, whereas HbA1c alone identified only 27.0% of patients. Sensitivity and specificity using a HbA1c cutoff at ≥6.5% were 24.4% and 97.6%, respectively. The Youden's J statistic for HbA1c was maximized at 6.1%. The area under the ROC curve of HbA1c against 2hPG was significantly lower compared to FPG (0.79, 95% CI 0.70-0.88; 0.89, 95% CI 0.84-0.94, respectively; p = 0.04).

Conclusions

Blood glucose criteria are more sensitive than HbA1c when screening HF patients for diabetes. Future studies should test performance of a HbA1c cutoff at 6.1% when FPG or 2hPG cannot be completed.

Background

Patients’ acceptance of artificial intelligence (AI) based health-related technologies depend strongly on their perception and trust of AI. This research field has not been studied extensively, especially among people living with diabetes. A large proportion of them frequently use health technologies in their everyday lives to manage their condition, which may make them more prepared to adopt AI-based solutions. Our study aimed to investigate the perception of AI-based solutions in healthcare, and characteristics associated with positive attitudes towards AI among people with and without diabetes.

Methods

An online survey was sent to 12,755 participants in the Health in Central Denmark cohort, including 10 questions and six scenarios related to current technology use, data sharing, and AI. The question on benefits and risks of AI, and the responses to the scenarios were used as outcomes. Multinomial logistic regression was used to examine which characteristics were associated with seeing the benefit of AI over the risks, including diabetes status, age, sex, education, health literacy, the use of wearable devices, and views on data sharing. A similar analysis was conducted on the acceptance of AI-based solutions in healthcare-related scenarios.

Findings

8,420 participants responded to the survey. Most participants (88%) had previously heard about AI. 46% of participants agreed with the statement that the benefits of AI outweigh the risks, while only 2% agreed with the opposite statement, and 30% were unsure. We did not find evidence for a differential opinion by diabetes status. Having diabetes was associated with less openness to replace healthcare professionals by AI-based technologies, although most people were still open to AI if controlled by humans.

Interpretation

Despite the generally positive perception of AI and its benefits to healthcare, human interaction seemed to play an important role in defining positive attitudes to AI across different healthcare scenarios, especially among people with diabetes. This highlights the pressing need for a patient-centered development process of AI-based solutions in the future.