Insulin is an essential hormone for animal activity and survival, and it controls the metabolic functions of the entire body. Throughout the evolution of metazoan animals and the development of their brains, a sustainable energy supply has been essential to overcoming the competition for survival under various environmental stresses. Managing energy for metabolism, preservation, and consumption inevitably involves high oxidative stress, causing tissue damage in various organs. In both mice and humans, excessive dietary intake can lead to insulin resistance in various organs, ultimately displaying metabolic syndrome and type 2 diabetes. Insulin signals require thorough regulation to maintain metabolism across diverse environments. Recent studies demonstrated that two types of forkhead-box family transcription factors, FOXOs and FOXKs, are related to the switching of insulin signals during fasting and feeding states. Insulin signaling plays a role in supporting higher activity during periods of sufficient food supply and in promoting survival during times of insufficient food supply. The insulin receptor depends on the tyrosine phosphatase feedback of insulin signaling to maintain adipocyte insulin responsiveness. α4, a regulatory subunit of protein phosphatase 2A (PP2A), has been shown to play a crucial role in modulating insulin signaling pathways by regulating the phosphorylation status of key proteins involved in these pathways. This short review summarizes the current understanding of the molecular mechanism related to the regulation of insulin signals.

With the rapid development of computer science, there is an increasing demand for the use of causal inference methods and machine learning in the research of endocrine disorders and their long-term health outcomes. However, studies on the effective and appropriate applications of these approaches in real-world data and clinical settings are still limited. This review will illustrate the use of causal inference and machine learning in epidemiological research within the field of endocrinology and metabolism. It will examine each concept of causal inference and machine learning through application examples of endocrine disorders. Subsequently, the paper will discuss the integration of machine learning within the causal inference framework, including (i) the estimation of treatment effects or the causal relationship between exposure and outcomes, and (ii) the evaluation of heterogeneity in such treatment effects (or exposure-outcome causal relationship) based on individuals’ characteristics. Accurately assessing causal relationships and their heterogeneity across different individuals is crucial not only for determining effective interventions, but also for the appropriate allocation of medical resources and reducing healthcare disparities. By illustrating some application examples in endocrinology, this review aims to enhance readers’ understanding and application of causal inference and machine learning in future epidemiological studies focusing on endocrine disorders.

The effectiveness of a hybrid closed-loop (HCL) system in improving glycemic control is unclear in Japanese individuals. Therefore, we assessed the effect impact of the MiniMed 770G HCL system on glycemic control in this population. This prospective, single-center, 24-week observational study (registration number: UMIN000047394) enrolled 23 individuals with type 1 diabetes mellitus using the Medtronic MiniMed 640G system. The primary endpoint was the improvement in time in the range of 70–180 mg/dL after transitioning to the MiniMed 770G HCL system. We observed an increase in time in range (from 64.1 [55.8–69.5] to 70.9 [67.1–74.4] %, interquartile range 25–75%, p < 0.001) and a decrease in glycated hemoglobin level (from 7.4 [7.0–7.9] to 7.1 [6.8–7.4] %, p = 0.003). There was a significant reduction in time above the range (181–250 mg/dL: 25.8 [20.9–28.6] to 19.5 [17.1–22.1] %, p < 0.001; >251 mg/dL: 8.7 [4.0–13.0] to 4.7 [3.6–9.1] %, p < 0.001). Time below the range remained unchanged (54–69 mg/dL: 1.8 [0.4–2.4] to 2.1 [0.4–3.9] %, p = 0.24; <54 mg/dL: 0.2 [0.0–1.0] to 0.5 [0.1–1.3] %, p = 0.14). In a subgroup of 12 patients with a high HCL implementation rate, the basal insulin infusion decreased immediately after mealtime insulin administration and increased after approximately 120 minutes. The ratings from questionnaires assessing treatment burden, satisfaction, and quality of life remained unchanged. The MiniMed 770G HCL system improved glycemic control and optimized insulin delivery, particularly in patients with high implementation rates.

We previously reported that a high HbA1c level 3 months before vitrectomy for vitreous hemorrhage or a large preoperative decrease in the HbA1c level over 3 months tended to increase the risk of rebleeding in diabetic retinopathy patients evaluated between 2010 and 2014. Here, we aimed to confirm these results with an extended study period and an increased number of operated eyes. This study included 121 diabetic patients who were admitted to Osaka University Hospital between 2010 and 2019 and who underwent vitrectomy for vitreous hemorrhage. Binomial logistic regression analysis was performed with the presence of postoperative bleeding as the outcome. The present study showed that the duration of the operation was associated with rebleeding (odds ratio = 1.02, p = 0.0016). A high HbA1c level just before vitrectomy tended to be associated with the bleeding (odds ratio = 1.27, p = 0.05), while preoperative HbA1c changes were not associated with rebleeding. The results of this study suggest that a high preoperative HbA1c level just before vitrectomy, not a decrease in HbA1c levels, in addition to the duration of the operation may increase the risk of postoperative bleeding after vitrectomy in diabetic retinopathy patients.

There have been no reports comparing neonatal external genitalia of 5α-reductase deficiency (5αRD) with those of other 46,XY differences of sex differentiation (DSD). This study enrolled 31 Japanese cases of 46,XY DSD whose external genitalia was examined during the neonatal period; four were diagnosed as 5αRD and 15 were defined as non-5αRD by genetic analysis of SRD5A2 or urinary steroid metabolites. We compared the following characteristics between 5αRD and non-5αRD groups, adjusting the severity of undermasculinization of the external genitalia: stretched penile length (SPL), glans width, location of the external urethral opening, and proportion of undescended testis. The external genitalia of all the 5αRD cases were Quigley classification grade 2 or 3. We compared the phenotypes between the four 5αRD cases and 11 non-5αRD cases with grade 2 or 3. The median (range) of SPL in the 5αRD group (14 mm [11-16]) was significantly lower than that in the non-5αRD group (22 mm [15-29]) (p = 0.003). An SPL cut-off value of <15 mm yielded a sensitivity of 50% (95% confidence interval [CI]; 7–93%) and specificity of 100% (95% CI, 72–100%) for discriminating between the groups. The median glans width, location of the external urethral opening, and proportion of undescended testis were not significantly different between the groups. The SPL of 5αRD in Quigley classification grade 2 or 3 was significantly shorter than that of other 46,XY DSDs with the equivalent grade.

This study aimed to evaluate the associations of fasting plasma glucose (FPG) and glycosylated hemoglobin (HbA1c) levels at <24 weeks of gestation with hypertensive disorders of pregnancy (HDP) and compare the strengths of the associations of HDP with FPG and HbA1c levels. Totally, 1,178 participants were included in this prospective cohort study. HDP, FPG, HbA1c, and potential confounding factors were included in multiple logistic regression models. The number of HDP cases was 136 (11.5%). When FPG and HbA1c were included in the model separately, quartile 4 (Q4) of FPG (87–125 mg/dL) and HbA1c (5.2–6.3% [33–45 mmol/mol]) levels had higher odds of HDP than quartile 1. The odds ratios (ORs) were 1.334 (95% confidence interval [CI]: 1.002–1.775) for Q4 of FPG and 1.405 (95% CI: 1.051–1.878) for Q4 of HbA1c. When the participants were divided into two categories based on the cut-off value with the maximum Youden Index of FPG or HbA1c, the ORs for high FPG (≥84 mg/dL) or high HbA1c (≥5.2% [33 mmol/mol]) were 1.223 (95% CI: 1.000–1.496) and 1.392 (95% CI: 1.122–1.728), respectively. When both FPG and HbA1c were included in the model simultaneously, the statistical significance of Q4 of FPG disappeared, whereas that of HbA1c remained. In two-category models, the same results were obtained. High FPG and HbA1c levels at <24 weeks of gestation were risk factors for HDP in pregnant Japanese women. In addition, high HbA1c levels were more strongly associated with HDP than high FPG levels.

Cushing’s syndrome, a clinical condition characterized by hypercortisolemia, exhibits distinct clinical signs and is associated with cyclic cortisol secretion in some patients. The clinical presentation of cyclic Cushing’s syndrome can be ambiguous and its diagnosis is often challenging. We experienced a 72-year-old woman with cyclic ACTH-dependent Cushing’s syndrome caused by a pulmonary carcinoid tumor. Diagnosis was challenging because of the extended trough periods, and the responsible lesion was initially unidentified. A subsequent follow-up computed tomography revealed a pulmonary lesion, and ectopic ACTH secretion from this lesion was confirmed by pulmonary artery sampling. Despite the short peak secretion period of ACTH (approximately one week), immunostaining of the surgically removed tumor confirmed ACTH positivity. Interestingly, stored plasma chromogranin A levels were elevated during both peak and trough periods. The experience in evaluating this patient prompted us to investigate the potential use of plasma chromogranin A as a diagnostic marker of ACTH-dependent Cushing’s syndrome. A retrospective study was conducted to determine the efficacy of plasma chromogranin A in three patients with ectopic ACTH syndrome (EAS), including the present case, and six patients with Cushing’s disease (CD) who visited our hospital between 2018 and 2021. Notably, plasma chromogranin A levels were higher in patients with EAS than in those with CD. Additionally, a chromogranin A level in the present case during the trough phase was lower than that in the peak phase, and was similar to those in CD patients. The measurement of plasma chromogranin A levels could aid in differentiating EAS from CD.

A 38-year-old Japanese woman with a history of abnormal thyroid function of non-autoimmune origin, pituitary endocrine tumor, and untreated diabetes mellitus was referred to our outpatient clinic when she became pregnant with twins. Physical findings consistent with Cushing’s syndrome (CS) were absent at the time of presentation. Although baseline plasma adrenocorticotropic hormone, serum cortisol, and 24-hour urinary free cortisol excretion levels were above the upper limits of normal non-pregnant reference ranges, we could not exclude a physiological increase associated with pregnancy. No medical or surgical intervention for hypercortisolism was performed during pregnancy. Spontaneous vaginal delivery resulted in the normal delivery of live twins. A diagnosis of Cushing’s disease (CD) was established when papery skin developed postpartum. Transsphenoidal surgery was performed and the hypercortisolism partially resolved post-operatively. The patient’s abnormal thyroid function also resolved. Pregnancy in women with endogenous CS is rare, with less than 300 cases reported. Most reported cases of CS during pregnancy are of adrenal origin. Only two cases of twin pregnancies with CD have been reported. Therefore, we reported the third case of CD in a twin pregnancy and reviewed the diagnostic and therapeutic challenges associated with CD during pregnancy.