Mona Abdelsalam

A Recurrent Neural Network Approach for Predicting Glucose Concentration in Type-1 Diabetic Patients

Estimation of future glucose concentration is important for diabetes management. To develop a model predictive control (MPC) system that measures the glucose concentration and automatically inject the amount of insulin needed to keep the glucose level within its normal range, the accuracy of the predicted glucose level and the longer prediction time are major factors affecting the performance of the control system. The predicted glucose values can be used for early hypoglycemic/hyperglycemic alarms for adjustment of insulin injections or insulin infusion rates of manual or automated pumps. Recent developments in continuous glucose monitoring (CGM) devices open new opportunities for glycemia management of diabetic patients. In this article a new technique, which uses a recurrent neural network (RNN) and data obtained from CGM device, is proposed to predict the future values of the glucose concentration for prediction horizons (PH) of 15, 30, 45, 60 minutes. The results of the proposed technique is evaluated and compared relative to that obtained from a feed forward neural network prediction model (NNM). Our results indicate that, the RNN is better in prediction than the NNM for the relatively long prediction horizons.

Empowerment-Based Diabetes Self-Management Education to Maintain Glycemic Targets During Ramadan Fasting in People With Diabetes Who Are on Conventional Insulin: A Feasibility Study

Objective. This study aims to assess the feasibility of promoting safe Ramadan fasting through diabetes self-management education (DSME) and to determine the effect of such education on hypoglycemic episodes. Design and methods. This prospective study included subjects attending Ramadan reinforcement sessions for participants in the Educational Program for People with Diabetes (EPPWD) at the Ain-Shams University Diabetes Center in Cairo, Egypt. The DSME sessions started 2–3 weeks before Ramadan and included one experimental fasting day during the first week and one during the second week. Participants’ A1C and serum fructosamine levels were measured before and after Ramadan, and they completed weekly self-monitoring of blood glucose (SMBG) logs. Results. Among 21 participants who were intending to fast for Ramadan, 14 completed the program. Their mean A1C was 6.7 ± 1.6%, and SMBG results showed a statistically nonsignificant difference in mean blood glucose levels before and after Ramadan (123.84 ± 39.96 and 123.84 ± 25.92 mg/dL, respectively; P >0.05). Serum fructosamine after Ramadan declined by 10% from pre-Ramadan levels. The mean number of hypoglycemic events before Ramadan was 3 ± 1.04, which declined to 1.4 ± 0.5 during Ramadan. Differences between group 1 (those without hypoglycemia, n = 8) and group 2 (those with hypoglycemia, n = 6) were nonsignificant for all variables, including A1C. Conclusion. Ramadan fasting is feasible for people with diabetes who are on a multiple daily injection insulin regimen and participate in the EPPWD. The number of hypoglycemic events per month declined with the attainment of DSME.

Blood glucose regulation using a neural network predictor with a fuzzy logic controller

Current insulin therapy for patients with type 1 diabetes often results in high variability in blood glucose concentration and may cause hyper-and hypoglycemic episodes. Closing the glucose control loop with a fully automated control system improves the quality of life for insulin-dependent patients. This paper presents a nonlinear model predictive control technique for glucose regulation in type 1 diabetic patients. The proposed technique uses a neural network as a nonlinear model for prediction of future glucose values and a fuzzy logic controller FLC to determine the insulin dose required to regulate the blood glucose level, especially after unmeasured meals. In the proposed technique, to avoid errors of meal estimation, the patient is not required to enter any data such as the meal time and size which was, in previous systems, necessary to determine the insulin bolus. The use of neural networks in predicting future glucose levels helps the proposed control strategy to handle delays associated with insulin absorption and time-lag between subcutaneous glucose readings and the plasma glucose level. The FLC uses the predicted glucose values to determine the required insulin bolus. A feed forward neural network FFNN and a recurrent neural network RNN are tested and evaluated as nonlinear glucose prediction models. Simulation results for three meal challenges are demonstrated. our results indicate that, the use of a neural network as a predictor along with a FL controller can decrease the postprandial glucose concentration, avoids hyper glycemia, and dynamically responds to glycemic challenges. The simulation results also indicate that, the use of a RNN in glucose prediction gives better results than the use of a FFNN. The RNN provides much better prediction performance than the FFNN especially at longer prediction horizons.

Effect of Metformin Therapy on Serum Fetuin Levels in Insulin Resistant Type 1 Diabetics

INTRODUCTION Insulin resistance may develop with Type 1 diabetes. Insulin resistance is currently recognized by the estimated glucose disposal rate. Serum fetuin has been accused as a risk factor for metabolic syndrome. AIM To determine the relationship between the serum fetuin and insulin resistance in Type 1 diabetes subjects and the effect of short-term Metformin therapy on this relationship. METHODS 40 T1DM male ≥ 18 years of age were screened for insulin resistance (defined using estimated glucose disposal rate). 20 subjects (Group I) were insulin resistant with a mean estimated glucose disposal rate of (7.15±0.37 mg/kg/min) while 20 subjects (Group II) were non-insulin resistant with a mean estimated glucose disposal rate of (9.08±0.42 mg/kg/min). Fasting blood sugar, 2 hours-post prandial blood sugar, HbA1c%, C-peptide, lipid profile, highly sensitive-C reactive protein, and serum fetuin were assessed. Group I were given 1gm Metformin twice daily for 3 months as an add-on to their insulin regimen. All anthropometric and laboratory parameters were reassessed at the end of the 3 months. RESULTS Group I had a higher age, BMI and waist/hip ratio, FBS, PPBS, HbA1c%, TC, LDL-C, TG, Hs-CRP and serum fetuin (ρ ≤ 0.001), and a lower C-peptide (ρ=0.001). Fetuin showed a positive correlation with age, FBS, HbA1c%, and Hs-CRP. After Metformin therapy, FBS, PPB and HbA1c%, Hs- CRP and fetuin decreased (ρ≤0.001) while eGDR and insulin dose in units/kg increased (ρ <0.001). Correlation after Metformin therapy within Group I showed that eGDR was inversely related with FBS and PPBS and fetuin showed a positive correlation with Hs-CRP. CONCLUSION Serum fetuin was elevated in insulin resistant T1DM, yet this was not associated with eGDR. Levels of fetuin-A and HsCRP decreased after Metformin therapy.

Bone mineral density in type 2 diabetes patients with Charcot arthropathy

INTRODUCTION Charcot arthropathy is one of the disabling diabetes complications. There are enigmatic areas concerning its underlying pathophysiology and risk predictors. Osteoporosis and local osteopenia has been postulated to have a role in Charcot arthropathy development, but it is still controversial AIM: To Compare bone mineral density among type 2 diabetics with and without Charcot arthropathy. METHODOLOGY Two groups with type 2 diabetes participated in this study. Group I [30] patients with charcot arhropathy while Group II [30] patients without charcot arthropathy. All patients underwent full history taking and full clinical examination with special emphasis on the feet. Laboratory investigations were done that included fasting blood sugar, post prandial blood sugar, glycosylated hemoglobin, serum calcium, serum phosphorus, and alkaline phosphatase. All patients underwent MRI for both feet and and dual energy X-ray absorptiometry scan of the lumbar spine and femur. The demographic data, clinical data, the presence or absence of comorbidities and bone mineral density were compared for both groups RESULT: Bone mineral density was significantly lower in Group I than Group II with median lumber T score (-0.15, 1.99 p <0.001), median Femur T score (0.050, 2.400, p <0.001). Group I showed higher propensity for hypertension, neuropathy, micro-albuminuria with peripheral arterial disease (23.33 %) compared to Group II (p <0.001). Multiple logistic regression analysis revealed that female gender and low femur bone mineral density can be risk predictors of the condition. CONCLUSION Bone mineral density is lower in patients with Charcot arthropathy with female gender and Femur T score as risk predictors. Peripheral arterial disease shows greater incidence in Charcot patients than was previously reported.

Relationship of planter pressure and glycemic control in type 2 diabetic patients with and without neuropathy

INTRODUCTION Foot disease is a common complication of type 2 diabetes that can have tragic consequences. Abnormal plantar pressures are considered to play a major role in the pathologies of neuropathic ulcers in the diabetic foot. AIM To examine Relationship of Planter Pressure and Glycemic Control in Type 2 Diabetic Patients with and without Neuropathy. MATERIALS AND METHODS The study was conducted on 50 type 2 diabetic patients and 30 healthy volunteers. BMI calculation, disease duration, Hemoglobin A1c and presence of neuropathy (by history, foot examination and DN4 questionnaire) were recorded. Plantar pressure was recorded for all patients using the Mat-scan (Tekscan, Inc.vers. 6.34 Boston USA) in static conditions (standing) and dynamic conditions (taking a step on the Mat-scan). Plantar pressures (kPa) were determined at the five metatarsal areas, mid foot area, medial and lateral heel areas and medial three toes. RESULTS Static and dynamic plantar pressures in both right and left feet were significantly higher in diabetic with neuropathy group than in control group in measured areas (P<0.05). Static and dynamic pressures in right and left feet were significantly higher in diabetic with neuropathy group than in diabetic without neuropathy group in measured areas (P<0.05). On comparison between controls and diabetic without neuropathy group there was a significant difference in plantar pressures especially in metatarsal areas (P<0.05). No significant correlations were present between the studied variables age, disease duration, BMI and HbA1c and plantar pressures in all studied areas. CONCLUSION Persons with diabetic neuropathy have elevated peak plantar pressure (PPP) compared to patients without neuropathy and control group. HbA1c% as a surrogate for glycemic control had no direct impact on peak planter pressure, yet it indirectly impacts neuropathy evolution through out disease duration eventually leading to the drastic planter pressure and gait biomechanics changes.