PDF
Citation
Print
INTRODUCTION
Hypoglycemia is common in patients receiving insulin therapy or insulin secretagogues and remains the major limiting factor in the glycemic management of diabetes [1]. The consequences of hypoglycemia include confusion, coma, impaired memory, loss of awareness, activation of stress and inflammatory responses, cardiovascular effects, and even death [2].
Regulatory hormonal responses are activated when blood glucose falls below 70 mg/dL (3.9 mmol/L) to prevent hypoglycemia [3]. A prior study reported that patients with type 2 diabetes mellitus (T2DM) had impaired cortisol and growth hormone (GH) responses to severe hypoglycemia, with GH impairment associated with advanced age, shorter diabetes duration, and higher body mass index (BMI) [4]. Lundqvist et al. [5] performed hyperinsulinemic euglycemic-hypoglycemic clamps in overweight/obese and lean individuals without diabetes. They found that during hypoglycemia, baseline adrenocorticotropic hormone (ACTH) did not differ, but ACTH rose to higher levels in overweight/obese individuals. Cortisol and GH responses showed different patterns between the groups, although the differences were not statistically significant [5]. These findings suggest that obesity may influence pituitary-adrenal responses to hypoglycemia.
The UK Hypoglycemia Study Group reported that in both type 1 diabetes mellitus (T1DM) and T2DM, the risk of hypoglycemia increases with longer insulin therapy and progressive pancreatic β-cell failure. Reductions in glucagon and epinephrine responses to hypoglycemia were correlated with treatment duration [6,7]. In T1DM patients, glucagon, cortisol, GH, and epinephrine concentrations were comparable at baseline but were less responsive during hypoglycemia clamps [8]. A recent study found that the cortisol response to a glucagon stimulation test was reduced in patients with prediabetes, whereas the GH response was preserved [9].
Thus, counter-regulatory hormonal failure may differ across time and between diabetes subtypes, potentially leading to variation in symptoms, recovery capacity, and the prognosis of hypoglycemia. Furthermore, interactions among hormones—particularly within the hypothalamic-pituitary-target gland axis—may further shape these responses. It remains unclear whether counter-regulatory changes in pituitary and adrenal hormones during hypoglycemia influence other pituitary-dependent hormones. In this study, we performed hyperinsulinemic euglycemic-hypoglycemic clamps in patients with T2DM, patients with T1DM, and controls without diabetes. To minimize the influence of diabetes duration and prior hypoglycemic therapy, we enrolled only newly diagnosed, drug-naive patients. We measured hormones of the pituitary-adrenal, pituitary-thyroid, pituitary-gonadal, and GH–insulin-like growth factor 1 (IGF-1) axes during the clamps to investigate differential pituitary-target gland responses to hypoglycemia in different types of diabetes.
METHODS
Subject
This study was conducted at the Nanjing First Hospital between January 2023 and December 2023. The inclusion criteria were as follows:
1. Normal participants: no previous history of diabetes or other illness; fasting plasma glucose <6.1 mmol/L and 2-hour plasma glucose <7.8 mmol/L after a 75-g oral glucose tolerance test [10]; BMI <28 kg/m2.
2. Patients with newly diagnosed T1DM or T2DM. T1DM and T2DM were diagnosed by an endocrinologist according to the Standards of Medical Care for Type 2 Diabetes in China [10] and the World Health Organization report for the classification of diabetes [11]. Patients who were insulin-dependent, had fasting C-peptide (CP) <0.60 ng/mL, and tested positive for glutamic acid decarboxylase antibodies (GAD-Ab) were diagnosed with T1DM [12]. Patients with T2DM presented varying degrees of insulin resistance and insulin hyposecretion.
3. Glycosylated hemoglobin (HbA1c) >75 mmol/mol (9.0%).
4. All participants were male, aged 18 to 60 years, and had not been treated with hypoglycemic drugs. Patients with diabetes were all newly diagnosed at the endocrinology outpatient clinic.
Exclusion criteria included: (1) insulin allergy; (2) any history of hypoglycemic agent use or current use of such agents; (3) abnormal liver or kidney function on blood or urine testing; (4) systemic corticosteroid use within 3 months; (5) recent infections or acute medical events; and (6) other major conditions or associated diseases, as determined by the researcher, including severe cardiopulmonary, endocrine, neurological, or pancreatic diseases, tumors, or a history of mental illness.
The study was reviewed and approved by the Institutional Ethics Committee of Nanjing First Hospital, Nanjing Medical University (KY20220124-07). Signed informed consent was obtained from all participants. The trial was registered at clinicaltrials. gov (NCT05290207). The study was conducted in accordance with the Helsinki Declaration, and all methods followed relevant guidelines and regulations.
Study design
After admission, baseline parameters including height, weight, age, and medical history were collected. BMI was calculated as weight/height² (kg/m²). Blood samples were obtained after overnight fasting for lipid profile, HbA1c, and fasting and 120-minute postprandial CP. Patients with diabetes were then treated with an insulin pump until blood glucose normalized (>80% of values within target range: pre-meal 3.9 to 7.0 mmol/L, postprandial 3.9 to 10.0 mmol/L). Capillary blood glucose was tested at least seven times daily, and insulin doses were adjusted by an endocrinologist. In our hospital, glucose values typically returned to the normal range within 3–5 days, and never exceeded 1 week. Hyperinsulinemic euglycemic-hypoglycemic clamps were performed the day after normalization. Normal participants also underwent overnight fasting for blood collection, followed by the clamps.
During hyperinsulinemic-euglycemic-hypoglycemic clamps, blood glucose was first maintained at 5 mmol/L for 30 minutes. CP, glucagon, GH, IGF-1, ACTH, cortisol, thyroid function, and sex hormones were measured. Insulin infusion was then accelerated, and additional measurements were obtained when blood glucose fell <3.0 mmol/L or when hypoglycemic symptoms occurred. Blood glucose was subsequently maintained in the hypoglycemic range (2.7 to 3.0 mmol/L [13], or less than or equal to the blood glucose level with hypoglycemic symptoms) for 30 minutes, with hormones measured every 10 minutes three times. Insulin infusion was then stopped, and glucose solution was administered until blood glucose normalized.
Laboratory measurements
Total cholesterol, triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol were measured using an autoanalyzer (Modular E170, Roche, Mannheim, Germany). HbA1c was determined by high-performance liquid chromatography (Bio-Rad Laboratories, Hercules, CA, USA). CP, follicle-stimulating hormone, luteinizing hormone, estradiol, prolactin, testosterone, and thyroid-stimulating hormone (TSH) were all measured using chemiluminescent immunoassay (Architect System, Abbott Diagnostics, Abbott Park, IL, USA). Additional assays included GAD-Ab (reference <10 IU/mL; YHLO, Shenzhen, China), ACTH (Siemens, Manchester, UK), IGF-1 (Siemens), cortisol (Autobio, Zhengzhou, China), and GH (Roche, Indianapolis, IN, USA), also measured by chemiluminescent immunoassay.
Statistical analysis
This was an exploratory study; no formal sample size calculation was performed. Statistical analyses were conducted using SPSS version 22.0 (IBM Corp., Armonk, NY, USA). Variables were tested for normality and expressed as mean±standard error of the mean or median (interquartile range). Changes in hormones during clamps were analyzed using repeated-measures analysis of variance (ANOVA) to test for group differences. When significance was detected (P<0.05), post hoc comparisons were performed using the least significant difference test at each time point. Hormonal area under the curve (AUC) during hypoglycemia was calculated from 0 to 30 minutes after the onset of hypoglycemia. Baseline hormone and clinical characteristics were compared among groups using ANOVA or the Kruskal-Wallis H test. All comparisons were two-sided, with statistical significance set at P<0.05.
RESULTS
Clinical characteristics
The study included six normal controls, 21 patients with T2DM, and eight patients with T1DM (Fig. 1). The clinical characteristics of different groups are shown in Table 1. Age did not differ significantly among the groups. BMI in the T1DM group was significantly lower than in both the T2DM group and controls (both P<0.001). Patients with T2DM exhibited higher TG levels than the other groups (all P<0.05), whereas HDL-C was significantly higher in T1DM patients than in T2DM patients (P=0.003). HbA1c was elevated in both diabetes groups relative to controls (both P<0.001), but there was no significant difference between T1DM and T2DM. Fasting and postprandial CP levels were markedly lower in T1DM patients than in T2DM patients (both P<0.001), while fasting CP levels in T2DM patients were comparable to that in controls.
To evaluate the effect of BMI on hormone levels, patients with T2DM were further divided into obese (n=10, BMI >28 kg/m²) and non-obese (n=11, BMI ≤28 kg/m²) subgroups. As shown in Supplemental Table S1, TG was significantly higher in the obese subgroup than in the non-obese subgroup (P=0.046), whereas no significant differences were observed in age, HbA1c, or CP levels (all P>0.05).
Blood glucose levels and hypoglycemic symptoms
Blood glucose levels during euglycemic and hypoglycemic clamps were comparable across T2DM, T1DM, and control groups (P=0.602 and P=0.114 by repeated-measures ANOVA) (Fig. 2A). The glucose infusion rate during euglycemia was lower in T2DM patients than in controls (P=0.045), but no significant difference was observed during hypoglycemia (P=1.000) (Fig. 2B). Blood glucose levels and glucose infusion rates were similar between obese and non-obese T2DM subgroups during both euglycemia and hypoglycemia (all P>0.05). Hypoglycemic symptoms were reported in 76.2% of patients with T2DM, 37.5% of those with T1DM, and 33.3% of controls (P=0.056). Symptom frequency did not differ significantly between obese (80.0%) and non-obese (72.7%) T2DM subgroups (P=1.000).
Hormone levels during clamps
Hormone levels were first assessed after blood glucose was maintained at 5 mmol/L for 30 minutes. Under euglycemic conditions, CP levels were lower in both diabetes groups compared with controls, with the lowest values in the T1DM group (P<0.001). GH levels were elevated in both diabetes groups relative to controls. Patients with T2DM showed higher estradiol and lower testosterone levels than both T1DM patients and controls (Table 2).
During hypoglycemic clamps, CP, TSH, estradiol, and testosterone decreased, while prolactin, ACTH, cortisol and GH increased significantly (P<0.05 for all) (Table 3) according to repeated-measures ANOVA. Moreover, these three groups had different CP, ACTH, cortisol, GH, estradiol, and testosterone responses to hypoglycemia (Table 3, Fig. 3).
In non-diabetes controls and the T2DM group, CP decreased within 10 minutes of hypoglycemia and remained suppressed. In contrast, CP levels in the T1DM group were unchanged and consistently lower than in other groups (Table 3, Fig. 3A).
GH increased earlier in T1DM patients (10 minutes post-hypoglycemia) than in T2DM patients and controls (30 minutes) (Table 3). Consequently, GH levels in the T1DM group were significantly higher than in the normal group from euglycemia to 20 minutes after hypoglycemia (P<0.05 for all). Moreover, the AUC of GH during hypoglycemia was much higher in the T1DM group than in the other two groups (Fig. 3B).
ACTH and cortisol responses were most pronounced in the T2DM group (Table 3, Fig. 3C, D), with cortisol levels remaining elevated throughout hypoglycemia compared to controls (P<0.05 for all) (Table 3).
In the non-diabetes group, the testosterone level was significantly lower at 0 minutes of hypoglycemia (P<0.001) and remained at a lower level during hypoglycemia (Table 3). However, the changes in testosterone levels were not significant in either diabetes group (P>0.05 for all). Consequently, testosterone levels were higher in T1DM patients than in controls from hypoglycemia onset (Table 3). The T2DM group exhibited a lower testosterone AUC than the T1DM group (Fig. 3E).
Obesity-related differences in T2DM
To further assess obesity-related effects, hormonal responses were compared between obese and non-obese T2DM patients. Obese T2DM patients exhibited higher CP and estradiol levels during both euglycemic and hypoglycemic clamps (Supplemental Tables S2, S3), and ACTH responses were attenuated in obese T2DM compared with non-obese T2DM during hypoglycemia (P=0.05) (Fig. 4, Supplemental Table S3).
DISCUSSION
Our study elucidated distinct hormonal counter-regulatory responses to hypoglycemia in individuals with T1DM, T2DM, and controls without diabetes, underscoring critical pathophysiological differences between diabetes subtypes as well as the influence of obesity in T2DM. Patients with T1DM demonstrated an accelerated and exaggerated GH response to hypoglycemia compared with the other groups. By contrast, patients with T2DM exhibited rapid and sustained increases in ACTH and cortisol, accompanied by elevated estradiol and suppressed testosterone levels.
GH secretion is triggered when blood glucose levels fall to 3.7–3.2 mmol/L, stimulating gluconeogenesis, ketone body synthesis, and lipolysis in the liver and adipose tissue [14]. In our study, patients with T1DM displayed elevated basal GH and amplified GH increments during hypoglycemia. This phenomenon may partly explain the pronounced post-hypoglycemic hyperglycemia often observed in T1DM [15]. Such fluctuations may exacerbate endothelial injury and heighten cardiovascular risk [16]. Previous studies have documented increased basal and GH-releasing hormone-stimulated GH secretion, along with reduced circulating IGF-1 levels in T1DM [17,18]. However, these investigations did not specifically assess GH dynamics during hypoglycemia. Battezzati et al. [8] reported impaired GH responses to hypoglycemia in T1DM, in contrast to our findings. Notably, their study included patients with an average diabetes duration of 14 years, whereas our cohort comprised newly diagnosed individuals. Raisingani et al. [17] proposed that GH hypersecretion in T1DM may be driven by low IGF-1 levels. We also observed a slight decline in basal IGF-1 among patients with T1DM, although IGF-1 levels did not fluctuate during hypoglycemia. Since insulin therapy can increase IGF-1 and suppress GH [19], one might attribute elevated GH to insulin deficiency. However, in our study, patients received insulin pump therapy to maintain euglycemia before clamp testing, suggesting that the elevated basal GH in T1DM cannot be explained solely by insulin deficiency or hyperglycemia. Further research is warranted to clarify the mechanisms underlying GH–IGF-1 dysregulation during hypoglycemia in T1DM.
Patients with T2DM exhibited blunted GH responses to hypoglycemia, consistent with previous evidence that 73.2% of T2DM patients demonstrate impaired GH secretion during severe hypoglycemia [4]. Although basal GH levels were higher in T2DM than in controls in the present study, attenuated secretion during hypoglycemia resulted in comparable GH AUC values between groups. Conversely, T2DM patients showed pronounced activation of the ACTH–cortisol axis. This finding is consistent with prior reports of exaggerated cortisol responses during hypoglycemic clamps in overweight or obese individuals without diabetes [5]. The researchers further suggested that such exaggerated cortisol responses might contribute to T2DM development. The study also found that overweight/obese individuals experienced more frequent hypoglycemic symptoms during clamps, with symptom severity positively correlated with cortisol response and independent of BMI [5]. We confirmed this phenomenon in our T2DM cohort. Additionally, prolactin increased progressively during hypoglycemia, especially in T2DM. May et al. [20] observed prolactin responses to hypoglycemia in healthy subjects, but not in patients with pituitary tumors or idiopathic galactorrhea. The prolactin response to insulin-induced hypoglycemia is therefore considered a sensitive marker of hypothalamic–pituitary function [21]. Collectively, our findings indicate that patients with T2DM exhibit a more pronounced hypothalamic–pituitary–adrenal response to hypoglycemia. Moreover, T2DM patients reported more frequent hypoglycemic symptoms than T1DM patients or controls, despite having similar HbA1c levels and glycemic control before clamps. In early T2DM, hypoglycemia is common due to delayed biphasic insulin secretion. The aberrant autonomic and ACTH–cortisol responses we observed may further drive food intake and weight gain, thereby accelerating T2DM progression.
Testosterone deficiency has been linked to hyperglycemia in men [22,23], but the role of testosterone in hypoglycemic responses remains poorly understood. Only one case report, by Rabijewski et al. [24], described hypoglycemia following testosterone replacement in a patient with Klinefelter syndrome. Our study provides novel evidence that testosterone exhibits dynamic sensitivity to acute changes in glycemia. In participants without diabetes, testosterone declined rapidly during hypoglycemia, even before glucose fell below 3.0 mmol/L, consistent with reports that oral glucose or mixed meals reduce testosterone in eugonadal men [25]. Previous studies suggest that testosterone decreases insulin resistance in men [21] and enhances glucose uptake via activation of the glucose transporter type 4 (GLUT4) pathway in skeletal muscle [22], which could predispose to hypoglycemia. Thus, the decline in testosterone observed during hypoglycemia may represent a feedback regulation and protective mechanism. The mechanisms underlying this phenomenon warrant further investigation. Importantly, we found this testosterone response was absent in both T1DM and T2DM patients. Earlier studies showed reduced testosterone levels in men with T2DM compared to those with T1DM [22] and latent autoimmune diabetes in adults [23]. However, to our knowledge, testosterone responses to hypoglycemia have not been previously reported.
The present study showed that estradiol levels increased in the T2DM group, especially in obese patients. The result is consistent with a previous study [24]. In contrast, estradiol decreased during hypoglycemia in the non-diabetes group following a transient elevation. Compared to testosterone, estradiol changes occurred later and were of shorter duration. Cheng et al. [25] reported that estradiol impairs hypothalamic responses to hypoglycemia in both women and rats. However, other studies have suggested that estradiol stimulates hypoglycemia-associated hyperglucagonemia and hypercorticosteronemia in female rats, thereby protecting the brain from hypoglycemia through modulation of hindbrain metabolic sensors and signaling of neuro-energetic instability [26,27]. Our previous study also found that older men with T2DM had an increased risk of nocturnal hypoglycemia during intensive insulin pump therapy, whereas this was not observed in women [28]. The inconsistent findings regarding the role of estrogen in hypoglycemia may reflect variability in estrogen levels across diabetes subtypes as well as the complex fluctuations in estradiol observed during hypoglycemia in the present study.
In contrast to findings in individuals without diabetes [5], our study demonstrated a tendency toward reduced ACTH and GH responses to hypoglycemia in obese patients with T2DM, although the differences did not reach statistical significance. This suggests that the heightened ACTH–cortisol response observed in T2DM may be attenuated by higher body weight. Conversely, whether the elevated GH levels in T1DM are linked to lower body mass remains unclear. Importantly, recruitment of newly diagnosed obese T1DM patients was difficult because of their rarity, representing a limitation of this study. A major strength of our work was the inclusion of newly diagnosed patients who were treated exclusively with insulin pumps to achieve normoglycemia, thereby minimizing confounding from hypoglycemic medications or long disease duration on pituitary–adrenal responses. However, the strict inclusion criteria restricted our sample size, particularly in the control and T1DM groups, limiting both statistical power and generalizability. Another limitation is that only male participants were included, so the findings cannot be extrapolated to women. Future studies with larger, more diverse cohorts, including both sexes and patients at different stages of disease progression, are needed to assess the dynamic changes of the pituitary–adrenal axis over time in diabetes.
In conclusion, during hypoglycemic clamp testing, GH responses were enhanced in T1DM, whereas ACTH–cortisol responses were amplified in T2DM. These distinct hormonal patterns may partially explain the different glycemic profiles observed in T1DM and T2DM and highlight the need for individualized strategies for glycemic management. Furthermore, the role of sex hormones in counter-regulation during hypoglycemia warrants further investigation.
XML Download