A machine learning approach for predicting treatment response of hyponatremia.

Hyponatremia leads to severe central nervous system disorders and requires immediate treatment in some cases. However, a rapid increase in serum sodium (s-Na) concentration could cause osmotic demyelination syndrome. To achieve a safety hyponatremia treatment, we develop a prediction model of s-Na concentration using a machine learning. Among the 341 and 47 patients admitted to two tertiary hospitals for hyponatremia treatment (s-Na <130 mEq/L), those who were admitted to the general unit with urine sodium <20 mEq/L or treated with desmopressin were excluded. Ultimately, 74 and 15 patients (342 and 146 6-hourly datasets) were included in the learning and validation data, respectively. We trained the prediction model using three regression algorithms for shallow machine learning to predict s-Na every 6 h during treatment with the data of patients with hyponatremia (median s-Na: 112.5 mEq/L; range: 110.0-116.8 mEq/L) from one hospital. The model was validated externally using the data of patients with hyponatremia (median s-Na: 117.0 mEq/L; range: 112.9-120.0 mEq/L) from another hospital. Using 5-7 predictors (water intake, sodium intake, potassium intake, urine volume, s-Na concentration, serum potassium concentration, serum chloride concentration), the support vector regression model showed the best performance overall (root mean square error = 0.05396; R2 = 0.92), followed by the linear regression and regression tree models. The predicted s-Na levels, using explainable machine learning algorithms and clinically accessible parameters, correlated well with the actual levels. Thus, our model could be applied to the treatment of hyponatremia in clinical practice.

Effects of Digitization of Self-Monitoring of Blood Glucose Records Using a Mobile App and the Cloud System on Outpatient Management of Diabetes: Single-Armed Prospective Study.

BACKGROUND: In recent years, technologies promoting the digitization of self-monitoring of blood glucose (SMBG) records including app-cloud cooperation systems have emerged. Studies combining these technological interventions with support from remote health care professionals have reported improvements in glycemic control. OBJECTIVE: To assess the use of an app-cloud cooperation system linked with SMBG devices in clinical settings, we evaluated its effects on outpatient management of diabetes without remote health care professional support. METHODS: In this multicenter, open-label, and single-armed prospective study, 48 patients with diabetes (including type 1 and type 2) at 3 hospitals in Japan treated with insulin or glucagon-like peptide 1 receptor agonists and performing SMBG used the app-cloud cooperation system for 24 weeks. The SMBG data were automatically uploaded to the cloud via the app. The patients could check their data, and their attending physicians reviewed the data through the cloud prior to the patients' regular visits. The primary outcome was changes in glycated hemoglobin (HbA1c) levels. RESULTS: Although HbA1c levels did not significantly change in all patients, the frequency of daily SMBG following applying the system was significantly increased before induction at 12 (0.60 per day, 95% CI 0.19-1.00; P=.002) and 24 weeks (0.43 per day, 95% CI 0.02-0.84; P=.04). In the subset of 21 patients whose antidiabetic medication had not been adjusted during the intervention period, a decrease in HbA1c level was observed at 12 weeks (P=.02); however, this significant change disappeared at 24 weeks (P=.49). The Diabetes Treatment Satisfaction Questionnaire total score and "Q4: convenience" and "Q5: flexibility" scores significantly improved after using the system (all P<.05), and 72% (33/46) patients and 76% (35/46) physicians reported that the app-cloud cooperation system helped them adjust insulin doses. CONCLUSIONS: The digitization of SMBG records and sharing of the data by patients and attending physicians during face-to-face visits improved self-management in patients with diabetes. TRIAL REGISTRATION: Japan Registry of Clinical Trials (jRCT) jRCTs042190057; https://jrct.niph.go.jp/en-latest-detail/jRCTs042190057.

Improved glycemic control after the use of flash glucose monitoring accompanied by improved treatment satisfaction in patients with non-insulin-treated type 2 diabetes: A post-hoc analysis of a randomized controlled trial.

AIMS: In our previously reported randomized controlled trial in patients with noninsulin-treated type 2 diabetes, the use of flash glucose monitoring (FGM) improved glycated hemoglobin (HbA1c), and the improvement was sustained after the cessation of glucose monitoring. In this post-hoc analysis, we examined data from our trial to identify the factors that influenced FGM efficacy. METHODS: We analyzed data for 48 of 49 participants of the FGM group who completed the trial to clarify the changes in various parameters and factors related to HbA1c improvement with the use of FGM. RESULTS: Analyses of the FGM data during the 12-week FGM provision period showed that the weekly mean blood glucose levels considerably decreased as early as at 1 week compared with the baseline values, and this decline continued for 12 weeks. An enhancement in the Diabetes Treatment Satisfaction Questionnaire regarding "willingness to continue the current treatment" score was significantly associated with the improvement in HbA1c at 12 (p = 0.009) and 24 weeks (p = 0.012). CONCLUSIONS: Glycemic control was improved soon after FGM initiation, accompanied by improved satisfaction with continuation of the current treatment in patients with noninsulin-treated type 2 diabetes.

Resting energy expenditure depends on energy intake during weight loss in people with obesity: a retrospective cohort study.

Objective: Resting energy expenditure (REE) decreases if there is reduced energy intake and body weight (BW). The decrease in REE could make it difficult for patients with obesity to maintain decreased BW. This study aimed to investigate the correlation among changes in REE, energy intake, and BW during the weight loss process in patients with obesity. Materials and methods: We conducted a retrospective cohort study of patients hospitalized for the treatment of obesity in Japan. Patients received fully controlled diet during hospitalization and performed exercises if able. REE was measured once a week using a hand-held indirect calorimetry. Energy intake was determined by actual dietary intake. Results: Of 44 inpatients with obesity, 17 were included in the analysis. Their BW decreased significantly after 1 week (-4.7 ± 2.0 kg, P < 0.001) and 2 weeks (-5.7 ± 2.2 kg, P < 0.001). The change in REE after 1 and 2 weeks was positively correlated with the energy intake/energy expenditure ratio (r = 0.66, P = 0.004 at 1 week, r = 0.71, P = 0.002 at 2 weeks). Using a regression equation (y = 0.5257x - 43.579), if the energy intake/energy expenditure ratio within the second week was 82.9%, the REE after 2 weeks was similar to the baseline level. There was no significant correlation between the change in REE and BW. Conclusion: Our data suggest that changes in REE depend on energy intake/energy expenditure ratio and that the decrease in REE can be minimized by matching energy intake to energy expenditure, even during the weight loss process.

Resting energy expenditure depends on energy intake during weight loss in people with obesity: a retrospective cohort study

Abstract Objective: Resting energy expenditure (REE) decreases if there is reduced energy intake and body weight (BW). The decrease in REE could make it difficult for patients with obesity to maintain decreased BW. This study aimed to investigate the correlation among changes in REE, energy intake, and BW during the weight loss process in patients with obesity. Materials and methods: We conducted a retrospective cohort study of patients hospitalized for the treatment of obesity in Japan. Patients received fully controlled diet during hospitalization and performed exercises if able. REE was measured once a week using a hand-held indirect calorimetry. Energy intake was determined by actual dietary intake. Results: Of 44 inpatients with obesity, 17 were included in the analysis. Their BW decreased significantly after 1 week (−4.7 ± 2.0 kg, P < 0.001) and 2 weeks (−5.7 ± 2.2 kg, P < 0.001). The change in REE after 1 and 2 weeks was positively correlated with the energy intake/energy expenditure ratio (r = 0.66, P = 0.004 at 1 week, r = 0.71, P = 0.002 at 2 weeks). Using a regression equation (y = 0.5257x - 43.579), if the energy intake/energy expenditure ratio within the second week was 82.9%, the REE after 2 weeks was similar to the baseline level. There was no significant correlation between the change in REE and BW. Conclusions: Our data suggest that changes in REE depend on energy intake/energy expenditure ratio and that the decrease in REE can be minimized by matching energy intake to energy expenditure, even during the weight loss process.

Adult-onset autoimmune diabetes identified by glutamic acid decarboxylase autoantibodies: a retrospective cohort study.

AIMS/HYPOTHESIS: Patients with GAD antibodies (GADAb) showing clinical features of type 2 diabetes typically exhibit progression to an insulin-dependent state in several months or years. This condition is diagnosed as slowly progressive insulin-dependent (type 1) diabetes mellitus (SPIDDM) or latent autoimmune diabetes in adults, a subtype of adult-onset autoimmune diabetes. However, some patients diagnosed with adult-onset autoimmune diabetes do not progress to an insulin-dependent state. We conducted a retrospective cohort study to identify patients with non-insulin-dependent diabetes among those diagnosed with adult-onset autoimmune diabetes using measurable indicators in routine clinical practice. METHODS: We surveyed data from the electronic medical records of all patients with GADAb from eight medical centres in Japan for selecting and analysing patients who matched the diagnostic criteria of SPIDDM. RESULTS: Overall, 345 patients were analysed; of these, 162 initiated insulin therapy (insulin therapy group), whereas 183 did not (non-insulin therapy group) during the follow-up period (median 3.0 years). Patients in the non-insulin therapy group were more likely to be male and presented a later diabetes onset, shorter duration of diabetes, higher BMI, higher blood pressure levels, lower HbA1c levels, lower GADAb levels and lesser antidiabetic agent use than those in the insulin therapy group when GADAb was first identified as positive. A Cox proportional hazards model showed that BMI, HbA1c levels and GADAb levels were independent factors for progression to insulin therapy. Kaplan-Meier analyses revealed that 86.0% of the patients with diabetes having GADAb who presented all three factors (BMI ≥ 22 kg/m2, HbA1c < 75 mmol/mol [9.0%] and GADAb <10.0 U/ml) did not require insulin therapy for 4 years. CONCLUSIONS/INTERPRETATION: Higher BMI (≥22 kg/m2), lower HbA1c (<75 mmol/mol [9.0%]) and lower GADAb levels (<10.0 U/ml) can predict a non-insulin-dependent state for at least several years in Japanese patients with diabetes having GADAb.