Sex differences in glucose-regulating neurohormonal pathways—potential impact for type 2 diabetes development Authors: Martin H Lundqvist, Maria J Pereira, Urban Wiklund, Susanne Hetty, Jan W Eriksson Published: 2026 ABSTRACT AbstractAimMen have a higher risk to develop type 2 diabetes (T2D) than women. The aim of this study was to investigate if there are sex differences in counter-regulatory hormone dynamics that might affect insulin resistance and hence propensity for T2D. Further, we addressed the roles of sex hormones, menopausal status and metabolic phenotypes for such sex differences.Methods22 men (7 T2D, BMI 28.8 kg/m2) and 28 women (16 postmenopausal, 9 T2D, BMI 31.5 kg/m2) underwent euglycemic-hyperinsulinemic-hypoglycemic and hyperglycemic clamps with assessments of counter-regulatory hormones and heart rate variability (HRV). Sex hormones (eg, testosterone, estradiol) were analyzed from fasting plasma samples.ResultsMen vs women had higher responses of glucagon, ACTH, cortisol and GH but attenuated HRV responses during hypoglycemia and higher ACTH, more sympathetic dominance and lower GH levels during normo- and hyperglycemia. Post- vs premenopausal women had lower GH levels and less favorable HRV profiles but did not otherwise differ. Within-sex correlations between neurohormonal responses and sex hormones were generally weak. In multilinear regressions, M-value and waist-to-hip ratio were the strongest predictors of cortisol, ACTH and GH responses, whereas sex hormones did not explain within- or between-sex variations in neurohormonal responses.ConclusionIn healthy individuals as well as those with obesity or T2D, glucose-dependent counter-regulatory hormonal responses are elevated in men vs both pre- and postmenopausal women. This may contribute to insulin resistance and hence an increased risk of T2D development in men. Central adiposity, but not sex hormone profiles, may partly explain the observed sex differences.Keywords: counter-regulatory, hypoglycemia, hyperglycemia, sex, sex hormones Although there are many known risk factors for type 2 diabetes (T2D), including obesity, physical inactivity and male sex [1], there are considerable knowledge gaps regarding what mechanisms are involved in the disease development. Besides traditionally proposed mechanisms, including ectopic fat accumulation and inflammatory pathways [2], the role of the brain in the regulation of glucose and development of T2D has been receiving growing attention [3, 4]. Hypoglycemia elicits a brain-coordinated response, characterized by an increase of so-called counter-regulatory hormones—glucagon, catecholamines, GH and cortisol—and sympathetic nerve activity. These hormones and neural systems share the capacity to increase systemic glucose levels, essentially by counter-acting insulin, and the net effect of this counter-regulatory response is therefore to normalize glucose levels and alleviate hypoglycemia. However, it has also been proposed that maladaptation of these systems may be involved in the development of T2D. In support of this hypothesis, our research group have previously demonstrated that the glucose-dependent responses of glucagon, ACTH, cortisol and catecholamines are elevated in T2D and at early stages of T2D development—obesity, insulin resistance and prediabetes [5-8]. T2D is more common and presents at an earlier age in men compared to women [9]. This has been attributed to the beneficial actions of estrogen, that declines rapidly in postmenopausal women in parallel with unfavorable changes in body composition and an increasing risk of cardiometabolic diseases such as T2D [9]. It has been previously shown that counter-regulatory responses to hypoglycemia are elevated in men compared to premenopausal and metabolically healthy women [10-15]. However, there is considerably weaker evidence suggesting that these sex differences are attenuated after the menopause [16, 17] and a potential mechanistic role of sex hormones has not been explored in humans, to our knowledge. In this study, we have therefore aimed to further dissect the relative impact of sex, menopausal status, circulatory sex hormones and metabolic characteristics on glucose-dependent activity of neuroendocrine systems. We hypothesized that, (1) glucose-dependent counter-regulatory activity is higher in men compared to both pre- and postmenopausal women across various stages of metabolic health—from normal glucose metabolism to T2D—and that, (2) this is explained by differences in circulatory levels of sex hormones. To this end, we present data from multimodal glucose clamp experiments performed in a cohort of men and women, representing considerable diversity in terms of demography, metabolic and reproductive health. Materials and methods Participants This cross-sectional study took place at the Uppsala University Hospital and the Department of Medical Sciences at Uppsala University between March 2018 and May 2024. In this study, individuals with and without T2D have undergone 1 stepwise hyperinsulinemic-euglycemic-hypoglycemic clamp (henceforth denoted hypoglycemic clamp) and 1 stepwise hyperglycemic clamp on 2 occasions. The current cohort has been expanded in relation to previous cohorts described in other publications [5, 7, 8], where recruitment methods and other details are provided. Inclusion and exclusion criteria for the main study are summarized in Supplement S1 [18]. Premenopausal women underwent a pregnancy test before inclusion. From a total of 58 individuals eligible for the current analysis, women on hormonal replacement therapy or oral contraceptives (n = 4) as well as women with polycystic ovarian syndrome (n = 4) were excluded, resulting in a study populations of 50 individuals (22 men, 28 women). Women were classified as premenopausal (n = 12), perimenopausal (n = 1) or postmenopausal (n = 15) based on medical history, date of last menstruation period and hormonal levels according to an algorithm developed for the WISE study [19]. In the statistical analyses of group differences, peri- and postmenopausal women were aggregated and henceforth, the term postmenopausal refers to both peri- and postmenopausal women. One postmenopausal woman did not complete the hypoglycemic clamp for personal reasons and 1 premenopausal woman did not complete the hyperglycemic clamp due to difficulty in obtaining venous access. Thus, 49/50 individuals completed the hypoglycemic clamp and 49/50 individuals completed the hyperglycemic clamp. Experimental procedures The clamp experiments and other study procedures have been described in full detail previously [5, 7, 8] and will only be summarized here. After obtaining baseline fasting blood samples, anthropometrics and hemodynamics, the clamps were initiated at approx. 09:00 Am. In the hypoglycemic clamp, the target glucose levels, indicated in Fig. 1A, were achieved by a fixed intravenous infusion of insulin (Actrapid®), potassium (8 mmol/hour) and a variable infusion of glucose (200 mg/mL). In the hyperglycemic clamp, glucose (200 mg/mL) was infused to achieve the target glucose levels in Fig. 1B. Glucose levels were measured every 5 minutes during both clamps with a Contour Glucose Meter (Bayer Healthcare, Leverkusen, Germany). At regular intervals during both clamps (Fig. 2), blood samples from an arterialized vein and blood pressure were obtained. ECG was recorded for 5 minutes before clamp initiation and continuously during the clamps to assess heart rate variability (HRV). In the hypoglycemic clamp, subjects were asked to assess hypoglycemic symptoms, using the Edinburgh Hypoglycemia Symptom Scale [20]. Glucose (A, B), insulin (C, D) and glucose infusion rate (GIR, E, F) in the hypoglycemic clamp (left panels) and hyperglycemic clamp (right panels). Data shown as geometric mean and standard deviation (A-D) or median and interquartile range (E, F) for men (n = 22, squares, solid lines) and women (n = 27, triangles, dashed lines). Abbreviations: Iso, isoglycemic phase; Rec, recovery phase; FFM, fat free mass. Counter-regulatory hormones during hypoglycemic (left panels) and hyperglycemic (right panels) clamps for men (n = 22, squares, solid lines) and women (n = 27, triangles, dashed lines). Data displayed as geometric mean and standard deviation. *P < .05, **P < .01 for AUC, †P < .05, ††P < .01 for ΔAUC in men vs women. Abbreviations: Iso, isoglycemic phase, Rec, recovery phase. Glucagon samples missing as described in Supplement S2. Biochemistry Biochemical analyses regarding metabolism and counter-regulatory hormones were performed and processed as described previously [5, 7, 8]. For this study, frozen EDTA-plasma was thawed to analyze FSH, LH, testosterone, estradiol, progesterone SHBG, IGF-1, prolactin, androstenedione and dehydroepiandrosterone sulfate (DHEA-S) at Uppsala Clinical Chemistry laboratory using the Siemens Immulite 2000XPI (androstenedione), Diasorin LiasonXL (IGF-1), Roche Cobas8000 e602 (DHEA-S) and Roche Cobas Pro e601 (all else) instruments. Progesterone was analyzed for premenopausal women only to validate or determine the menstrual phase. Fasting samples from before the hypoglycemic clamp were analyzed. Free testosterone was calculated according to Vermeulen et al [21]. Using the same principle, free estradiol was calculated based on mean association constants with albumin and SHBG from previous studies [22]. Values under the limit of quantification (LOQ) were set to LOQ/2. Heart rate variability ECG recordings were processed as described previously to assess the high frequency (HF) and low frequency (LF) components of the HRV [5, 8]. High frequency (.15-0.5 Hz) mainly reflects the parasympathetic part of cardiac autonomic modulation, while LF (0.04-0.15) reflects a combination of sympathetic and parasympathetic activity. The ratio LF:HF represents the balance between sympathetic and parasympathetic activity. Statistical analysis According to power calculations, a sample size of 32 individuals (14 men, 18 women) would be needed to detect a significant 25% difference between men and women in glucagon and cortisol AUC, with 90% power at an alpha level of .05 assuming a CV of 0.2 as stipulated previously [7]. P-values <.05 were considered significant. Group differences and associations are significant unless stated otherwise. No adjustment for multiple comparisons was undertaken. Group differences (men vs women; pre- vs postmenopausal women) were analyzed with Mann-Whitney U tests for continuous variables and Fisher's exact test for categorical variables. For repeated measures, trapezoidal area under the curve (AUC), ΔAUC (subtracting area under baseline), mean, Δmean or peak, and Δpeak was calculated for the hypoglycemic (80-185 minutes, Fig. 1A) and hyperglycemic (30-165 minutes, Fig. 1B) phases of each clamp. The impact of fasting levels of sex hormones and other clinical variables on neurohormonal responses were investigated by Spearman correlations, stratified by sex. Group comparisons were performed in SPSS for Mac (v.30, IBM corp., Armonk, NY). Correlation and regression analyses were performed using the Hmisc (v.5.2.4, Harrel, 2025) and lme4 package (v. 1.1.34, Bates, 2015), respectively, in R (v. 4.2.2, R Core Team, 2021) to investigate the relative impact of sex, menopausal status and sex hormones compared to other clinical variables on neurohormonal responses. Potential biases behind observed group differences were explored in additional models, adjusted for T2D status, BMI, glycemic control, medication and tobacco use. Variable selection methods and other method details are provided in Supplement S2 [18]. To further visualize sex differences in glucose-dependent neurohormonal responses, data from the hypoglycemic and hyperglycemic clamp were pooled and neurohormonal assessments were plotted against mean glucose levels from the preceding 10 minutes. Linear mixed models were performed using the lme4 package in R with glucose and group (men, women) and their interaction as fixed effects and id as random effects. Details are provided in Supplement S2 [18]. Figures were constructed in GraphPad Prism (v.9.1.0, GraphPad Software, San Diego, Ca) and R using the ggplot2 package (v. 3.5.1, Wickham, 2016). Ethics All study procedures were performed in accordance with the Declaration of Helsinki. The study was approved by the local Research Ethics Committee of Uppsala (DNR 2017/550). Amendments were approved by the Swedish Ethical Review Authority (DNR 2019-04166, 2021-05714-02). Subjects were informed verbally and in writing about the study and signed an Informed Consent Form. Abnormal laboratory results were handled according to clinical practice. Results Participant characteristics Clinical characteristics and fasting hormone levels are presented in Tables 1 and 2, respectively, and stratified by T2D status in Supplement S3 and S4 [18], respectively. Men exhibited higher waist-to-hip ratio (WHR), blood pressure and less favorable cholesterol profiles compared to women, who had higher body fat % and a trend toward higher BMI. T2D prevalence was similar between sexes (7 men, 12 women), with comparable disease duration (median 21 vs 17 months), although there was a trend toward higher metformin use in men than in women (6/7 vs 9/12), with the remaining participants being drug-naïve due to recent diagnosis. Further, men with T2D had significantly inferior glycemic control (Median HbA1c 58 vs 47 mmol/mol). Regardless of T2D status, there was a trend toward higher use of antihypertensives and statins, as well as a higher prevalence of smoking, in men. At fasting, before the clamps, glucagon and ACTH levels were higher, while GH levels were lower in men. Table 1Clinical characteristicsMen (n = 22)All women (n = 28)Postmenopausal (n = 16)Premenopausal (n = 12)T2D, n (%)7 (31.8%)9 (32.1%)8 (50%)1 (8.3)†Age, years52 (43;60)51 (39;56)55 (51;60)37 (29;47)†††BMI, kg/m228.8 (26.5;35.9)31.5 (24.8;37.2)32.6 (27.3;36.1)27.8 (22.8;41.8)WHR, ratio0.99 (0.95;1.05)0.90 (0.85;0.96)**0.92 (0.86;0.97)0.90 (0.78;0.96)Body fat, %28.4 (22.1;33.4)41 (30.6;46.3)***42.2 (38.3;47.6)39.2 (25.9;45.5)Fasting plasma glucose, mmol/L6.1 (5.7;7.6)6.0 (5.4;6.7)6.5 (5.5;7.3)5.7 (5.2;6.0)†IFCC HbA1c, mmol/mol37 (34;51)37 (33;45)42 (36;48)35 (31;36)††Insulin, μIU/mL12.1 (7.5;17.8)10.6 (6.0;19.7)12.3 (6.7;21.0)8.0 (5.7;15.0)C-peptide, nmol/L0.9 (0.6;1.3)1.0 (0.6;1.4)1.1 (0.8;1.5)0.7 (0.5;1.4)HOMA-IR3.4 (2.2;5.7)2.9 (1.5;5.1)3.8 (1.8;6.2)2.1 (1.3;4.1) a M-value, mg kgFFM−1 minute−16.5 (4.1;9.7)9.0 (3.7;11.7)6.6 (3.4;11.5)9.7 (5;11.8) b Total cholesterol, mmol/L4.2 (3.9;5)4.2 (3.6;4.9)4.4 (3.9;5.2)3.7 (3.2;4.8)HDL cholesterol, mmol/L1.1 (0.9;1.2)1.4 (1.1;1.6)1.4 (1.1;1.8)1.3 (1;1.4)LDL cholesterol mmol/L2.7 (2.3;3.6)2.5 (1.9;2.9)2.5 (1.9;2.9)2.4 (1.8;2.9)LDL/HDL, ratio2.9 (2.2;3.6)1.9 (1.3;2.5)1.8 (1.2;2.6)1.9 (1.4;2.5)Triglycerides, mmol/L1.3 (0.8;1.7)0.9 (0.6;1.3)1.1 (0.7;1.5)0.7 (0.5;0.9)Smoking, n (%)3 (13.6%)1 (3.6%)1 (6.3%)0 (0%)Smokeless tobacco, n (%)1 (4.5%)2 (7.1%)1 (6.3%)1 (8.3%)Antihypertensives, n (%)6 (27.3%)4 (14.3%)3 (18.8%)0 (0%)Statins, n (%)6 (27.3%)3 (10.7%)4 (25.0%)0 (0%)Metformin, n (%)6 (27.3%)6 (21.4%)5 (31.3%)1 (8.3%)Obtained at the first visit or mean of both visits (fasting glucose, insulin, HOMA-IR). Data presented as median (interquartile range) or n (%).Abbreviations: FFM, fat free mass; HDL, high density lipoprotein; HOMA-IR, homeostatic model for assessment of insulin resistance; LDL, low density lipoprotein; T2D, type 2 diabetes; WHR, waist-to-hip ratio. * P < .05. ** P < .01. *** P < 0.001 vs men. † P < .05. †† P < .01. ††† P < 0.001 vs postmenopausal women. a Measurements missing for 1 postmenopausal woman. b Measurement missing for 4/3/1 (men/post-/premenopausal women). Clinical characteristics Obtained at the first visit or mean of both visits (fasting glucose, insulin, HOMA-IR). Data presented as median (interquartile range) or n (%). Abbreviations: FFM, fat free mass; HDL, high density lipoprotein; HOMA-IR, homeostatic model for assessment of insulin resistance; LDL, low density lipoprotein; T2D, type 2 diabetes; WHR, waist-to-hip ratio. * P < .05. ** P < .01. *** P < 0.001 vs men. † P < .05. †† P < .01. ††† P < 0.001 vs postmenopausal women. a Measurements missing for 1 postmenopausal woman. b Measurement missing for 4/3/1 (men/post-/premenopausal women). Table 2Fasting hormone levels and HRV indices at restMen (n = 22)Women (n = 28)Postmenopausal (n = 16)Premenopausal (n = 12)Total testosterone, nmol/L14.8 (9.7;19.0)0.6 (0.2;1.0)***0.5 (0.1;0.9)0.7 (0.3;1.1)Free testosterone, nmol/L0.32 (0.24;0.39)0.01 (0.00;0.02)***0.01 (0.00;0.02)0.01 (0.00;0.02)Total estradiol, pmol/L116 (91;144)96 (54;384)60 (27;93)483 (231;905)†††Free estradiol, pmol/L3.4 (2.6;4.4)2.8 (1.3;12.8)1.5 (0.7;2.4)17.5 (6.5;35.4)†††Testosterone:estradiol, ratio137.7 (90.3;159.4)3.1 (0.9;10.8)***7.7 (1.4;16.4)1.8 (0.5;3.2)††SHBG, nmol/L31 (22;40)35 (22;64)30 (20;38)49 (34;70)FSH, IU/mL4.7 (3.1;6.2)39.5 (4.5;63.0)57.0 (42.5;71.3)4.1 (2.5;6.3)LH, IU/mL3.8 (3.4;5.0)18.7 (6.5;36.0)32.5 (20.4;42.0)6.2 (3.9;7.9)Progesterone, nmol/LNA0.6 (0.3;5.8)NA0.6 (0.3;5.8)Prolactin, mIU/L154 (121;190)217 (171;305)**181 (144;257)292 (205;414)††Androstenedione, nmol/L2.0 (0.5;3.1)1.6 (0.5;3.1)1.3 (0.5;2.4)2.0 (1.2;5.0)DHEA-S, μmol/L3.9 (2.6;5.4)3.5 (1.7;5.3)3.3 (1.0;5.2)4.4 (3.2;5.9) a Glucagon, pmol/L12.4 (10.9;15.9)8.2 (5.3;11.2)**8.2 (5.5;10.8)8.3 (4.8;13.0)Cortisol, nmol/L248 (212;298)224 (197;309)246 (197;322)219 (171;302)ACTH, pmol/L4.1 (2.9;7.2)2.5 (2.0;3.9)**3.0 (2.1;4.2)2.3 (1.7;2.8)GH, μg/L0.07 (0.03;0.25)0.55 (0.11;2.90)***0.33 (0.10;1.23)2.34 (0.33;4.29)IGF-1, μg/L139 (124;171)139 (115;180)125 (102;145)180 (138;206)†† b Adrenaline, pg/mL22.0 (16.0;31.0)13.0 (5.8;26.4)13.7 (6.4;31.4)10.4 (3.7;20.0) b Noradrenaline, pg/mL293.8 (183.5;335.8)208.6 (165.7;305.9)212.2 (164.3;319.5)182.1 (158.9;238.5) b,c Heart rate, bpm65.1 (55.4;68.2)66.5 (60.2;73.5)68.3 (59.9;75.9)65.2 (60.1;69.3) c HF, ms2 log2.38 (1.78;2.93)2.67 (2.22;2.92)2.42 (1.97;2.77)2.80 (2.59;3.15)† c LF:HF, ratio0.34 (0.04;0.62)0.09 (−0.01;0.26)0.10 (−0.01;0.38)0.08 (0.00;0.21)SBP, mmHg137 (129;152)125 (113;141)*132 (119;145)116 (109;127)†DBP, mmHg88 (80;92)82 (73;90)*85 (76;90)75 (69;88)Counter-regulatory hormones and HRV indices are mean from both visits, all other hormones obtained at the hypoglycemic clamp visit. Data presented as median (interquartile range).Abbreviations: DBP, diastolic blood pressure; DHEA-S, dehydroepiandrosterone sulfate; HF, power of the high frequency component of HRV; LF:HF, ratio between power of the low and high frequency components of HRV; SBP, systolic blood pressure. * P < .05. ** P < .01. *** P < .001 vs men. † P < .05. †† P< .01. ††† P < .001 vs postmenopausal women. a Measurements missing for 3/0/0 (men/post-/premenopausal women). b Measurements missing for 13/6/6. c Measurements missing for 5/3/2. Fasting hormone levels and HRV indices at rest Counter-regulatory hormones and HRV indices are mean from both visits, all other hormones obtained at the hypoglycemic clamp visit. Data presented as median (interquartile range). Abbreviations: DBP, diastolic blood pressure; DHEA-S, dehydroepiandrosterone sulfate; HF, power of the high frequency component of HRV; LF:HF, ratio between power of the low and high frequency components of HRV; SBP, systolic blood pressure. * P < .05. ** P < .01. *** P < .001 vs men. † P < .05. †† P< .01. ††† P < .001 vs postmenopausal women. a Measurements missing for 3/0/0 (men/post-/premenopausal women). b Measurements missing for 13/6/6. c Measurements missing for 5/3/2. Postmenopausal women (n = 16, median 7 years since last menstruation) had higher blood pressure, HbA1c, and T2D prevalence than premenopausal women (n = 12). Details on hormonal contraceptive use and menstrual cycle phase are provided in Supplement S5 [18]. Fasting sex hormone levels differed as expected by sex and menopausal status. Primary outcomes: counter-regulatory responses to glycemic challenges Sex differences in hormonal responses At similar achieved glucose and insulin levels in both clamps (Fig. 1), men had numerically higher responses of all counter-regulatory hormones (significant for glucagonAUC, glucagonΔAUC, cortisolAUC, ACTHAUC, GHΔAUC) during the hypoglycemic clamp, as well as higher ACTHAUC and lower GHAUC in the hyperglycemic clamp (Fig. 2, Supplement S6 [18]). When analyzing pooled data from both clamps in relation to clamp glucose levels, glucagon, ACTH and noradrenaline were all higher in men throughout the glycemic range (Fig. 3, model details in Supplement S7 [18]). At normoglycemia, GH was lower in men while cortisol and adrenaline did not differ; all these hormones rose more markedly during hypoglycemia. Hormonal responses in relation to clamp glucose levels for men vs women. Data from both clamps have been pooled. Glucose levels from the coinciding and 2 preceding time points have been averaged. Y-axes are on log2 scale in (A), (C), (D), and (E). Data displayed as observed data (open squares = men, n = 22; open triangles = women, n = 28) and estimated marginal means (solid lines, blue = men, red = women) with 95% confidence intervals (dashed lines and filled areas, light blue = men, light red = women) from linear mixed models (summarized in Supplement S7). *P < .05, **P < .01, ***P < .001. Abbreviations: ns, not significant referring to Type III fixed effects; Sex|Glucose, interaction between sex and glucose. Glucagon and cathecolamine measurements missing as described in Supplement S2. Sex differences in autonomic and symptom responses Despite higher hormonal responses, men exhibited attenuated cardiac autonomic responses to hypoglycemia, characterized by smaller increases in heart rate and LF:HF ratio and less pronounced HF reduction (Fig. 4, Supplement S6 [18]). Men also maintained higher mean blood pressure in both clamps and higher LF:HF during hyperglycemia. Total hypoglycemic symptom scores were similar between sexes, but men reported lower peak malaise symptoms (Supplement S8 [18]). Heart rate variability responses during hypoglycemic (left panels) and hyperglycemic (right panels) clamps for men (n = 22, squares, solid lines), women (n = 27, triangles, dashed lines). Data displayed as median and interquartile range (E, F) or geometric mean and standard deviation (all else). *P < .05 for mean, †P < .05, †††P < .001 for Δmean in men vs women. Abbreviations: Iso, isoglycemic phase; Rec, recovery phase; HR, heart rate; HF, power of the high frequency component of HRV; LF:HF, ratio between power of the low and high frequency components of HRV. Measurements missing as described in Supplement S2. Effect of menopausal status In comparison with premenopausal women, postmenopausal women had lower GHAUC and HF in both clamps, while mean glucose levels and LF:HF were higher in the hyperglycemic clamp (Supplement S9-S12 [18]). Secondary analyses: mechanisms underlying group differences Independence from clinical confounders Linear regression models adjusted for T2D status, HbA1c, BMI, medications (metformin, antihypertensives, statins), and tobacco use demonstrated that sex differences in counter-regulatory responses were largely independent of these clinical factors, identified as potential confounders due to skewed distribution between groups (Supplement S13-S14 [18]). Notably, significant sex × T2D interactions emerged for glucagon responses during hypoglycemia; stratified analyses revealed that sex differences in glucagon were significant only in individuals with T2D (Supplement S15 [18]). Adjustment for clinical confounders minimally attenuated observed sex effects, and these variables provided little additional explanatory value beyond sex alone. Similarly, menopausal status effects in women overall remained robust after adjustment for these factors (Supplement S16 [18]). Role of sex hormones vs metabolic phenotype To test whether sex hormone levels drive the observed sex differences, we examined correlations and performed regression analyses including sex hormones and metabolic variables. In both sexes, fasting levels of main sex hormones did not correlate strongly with hypoglycemic neurohormonal responses (Supplement S17 [18]). Significant correlations were observed for testosterone with ACTHAUC (rho −0.47), FSH with GHΔAUC (rho −0.49) and heart rateΔMEAN (rho −0.66) in men as well as estradiol with GHAUC (rho 0.40), LH with GHAUC (rho −0.47) and HFMEAN (rho −0.43). Notably, estradiol levels displayed positive trends with hormonal responses in women, in particular GH, whereas inconsistent directions were observed for testosterone in men. By contrast, measures of insulin resistance and adiposity correlated strongly and, overall, similarly across both sexes with elevated cortisol and ACTH responses during hypoglycemia, elevated glucagon responses during hyperglycemia and reduced parasympathetic tone (lower HF with blunted hypoglycemic decreases) in both clamps (Supplement S18 [18]). Sex-specific patterns also emerged: in men only, dysmetabolic phenotypes correlated strongly with higher glucagon responses during hypoglycemia as well as higher ACTH during hyperglycemia; in women only, inverse correlations were observed between such metabolic phenotypes and GH responses during hyperglycemia. Regression analyses of pooled data from both sexes confirmed these patterns. Adjusting for sex hormone levels did not substantially attenuate sex effects or improve model fit (Supplement S13 [18]). By contrast, adjusting for WHR, HbA1c and M-value substantially reduced the magnitude of associations between sex and ACTH, GH, and heart rate responses during hypoglycemia while simultaneously improving model fit (Table 3). Nevertheless, male sex remained independently associated with higher glucagon and cortisol and blunted decrease of HF during hypoglycemia, as well as lower GH and higher LF:HF during hyperglycemia (Supplement S13-S14 [18]). Table 3Selected linear regression modelsVariablesACTHAUCaGHΔAUCHRΔMEANbSex (ref. women), std. β0.220.22−0.21HbA1c, std. β−0.06−0.27−0.25WHR, std. β0.33*0.56**0.04M-Value, std. β−0.31*0.290.38* R 2 0.442†,§0.307†,§0.357†,§Models indicating superior prediction of metabolic variables compared to sex with regard to hypoglycemic neurohormonal responses in men (n = 22) and women (n = 27).Abbreviations: HR, heart rate; WHR, waist-hip ratio. * P < .05 for standardized beta coefficient (std. β).**P < .01 for standardized beta coefficient (std. β). †Model P < .05. § P < .05 vs model including only sex by likelihood ratio test. a Log transformed prior to analysis. b Measurements missing for 5 men and 5 women. Selected linear regression models Models indicating superior prediction of metabolic variables compared to sex with regard to hypoglycemic neurohormonal responses in men (n = 22) and women (n = 27). Abbreviations: HR, heart rate; WHR, waist-hip ratio. * P < .05 for standardized beta coefficient (std. β). **P < .01 for standardized beta coefficient (std. β). †Model P < .05. § P < .05 vs model including only sex by likelihood ratio test. a Log transformed prior to analysis. b Measurements missing for 5 men and 5 women. In sex-stratified regression models, sex hormones were generally not significantly associated with neurohormonal responses after adjustment for metabolic variables. Exceptions included FSH (associated with GHΔAUC and heart rateΔMEAN in men) and LH (associated with GHAUC in women), all observed during the hypoglycemic clamp (Supplement S19-S20 [18]). Discussion In this study, we have investigated neurohormonal responses to glucose variations in men and women with a variety of metabolic phenotypes. We hypothesized that (1) previous reports of elevated counter-regulatory responses in healthy men compared to premenopausal women are also applicable in individuals with compromised metabolic health and that (2) these sex differences are explained by differences in the circulatory levels of sex hormones. In support of our first hypothesis, the current results indicate that men have higher counter-regulatory hormonal responses to hypoglycemia while autonomic nervous system responses to glucose variations are attenuated in comparison with women. These differences were observed in individuals with a wide variety of metabolic phenotypes and were largely consistent in individuals with and without T2D. However, we were unable to fully substantiate our second hypothesis; neither menopausal status nor circulating sex hormones had a substantial impact on dynamic neurohormonal responses to hypo- or hyperglycemia. By contrast, our findings point to android metabolic phenotypes as potential mechanistic factors. Thus, the current results render further support to previous evidence of sex differences in the glucose-dependent regulation of neurohormonal system and adds valuable insight regarding possible underlying mechanisms. Comparison with previous research Previous investigations of sex differences in hormonal responses to hypoglycemia—largely in line with the present results—have mainly been conducted in healthy individuals of reproductive age [10-17]. By contrast, postmenopausal women [16] and women with T2D [17] have not been found to have differential hypoglycemic responses in comparison with age-matched men of similar clinical characteristics [16, 17]. As the main novelty of the current work, the observed sex differences were consistent in individuals with and without T2D and dynamic responses to hypo- and hyperglycemia were overall similar in post- compared to premenopausal women. The greater sample size in the current work as well as differences in subject selection—for instance the high prevalence of insulin and sulfonylurea treatment in Bremner et al [17]—which may be associated with recurrent hypoglycemia and blunting of counter-regulatory responses—could explain the discrepant results. While the sympathetic dominance of the ANS in men and postmenopausal women in the normo- and hyperglycemic range are fully compatible with previous research [23-25], the currently observed attenuation of hypoglycemic cardiac autonomic responses in men have not been reported previously, to our knowledge. Our findings appear to be explained by loss of parasympathetic inhibition rather than sympathetic activation, judging by the presently reported HRV dynamics as well as the numerically augmented catecholamine response in men. In further support of this concept, other work has shown that the hypoglycemic response of pancreatic polypeptide—a putative marker of parasympathetic activity—is elevated in men compared to women [12, 13, 16]. Thus, the current results are not incompatible with previous evidence of elevated sympathetic nervous system responses in men—assessed by muscle sympathetic nerve activity—during hypoglycemia [12, 14, 16]. Methodological limitations of HRV, discussed in more detail below, should also be considered when interpreting the present findings. Addressing potential confounders With respect to the observed sex differences, we identified several sources of bias of potential importance: (1) higher BMI in women; (2) a higher prevalence of chronic cardiometabolic medication use in men (metformin, antihypertensives, statins); (3) a higher prevalence of smoking in men; (4) heterogeneity of the study population with respect to T2D status; and (5) poorer glycemic control among men with type 2 diabetes compared with women. These factors were addressed using linear regression models adjusted for sex and the corresponding confounder; for T2D, interaction terms with sex were also included. Overall, these analyses did not indicate a substantial contribution of differences in BMI, chronic medication use, or tobacco use to the observed sex differences. However, interaction analyses followed by stratified analyses showed that hypoglycemia-induced glucagon responses were higher in men than in women specifically among participants with type 2 diabetes. In line with previous findings [26], this may be related to the poorer glycemic control currently observed in men with T2D. Nevertheless, adjustment for HbA1c did not attenuate the effect of sex, suggesting that glycemic control alone did not fully explain this difference. Alternatively, these findings may indicate that hyperglucagonemia plays a more prominent mechanistic role in the development of type 2 diabetes in men than in women, consistent with previous reports [27]. Mechanistic insights and clinical implications To elucidate possible mechanisms behind the observed group differences, we performed correlation and regression analyses to investigate the impact of circulating levels of sex hormones as well as metabolic factors on glucose-dependent neurohormonal activity. To our knowledge, this has not been done before. In summary, the observed sex-differences were not mirrored by typical male and female sex hormone profiles within either sex and correlation and regression analyses suggested that metabolic factors—in particular central obesity and insulin resistance—were more strongly associated with variations in counter-regulatory responses to hypoglycemia than circulating levels of sex hormones, both within and between sexes. This is compatible with previous findings by our research group [5-8]. Based on the current findings, we hypothesize that central adiposity—more pronounced in men—may promote increased activity in counter-regulatory hormonal systems, thereby contributing to insulin resistance, hyperglycemia and, ultimately, to the increased risk of T2D associated with male sex [9]. However, the cross-sectional design precludes conclusions regarding causality, and longitudinal studies are needed to test this hypothesis. The signaling pathways linking central adiposity to increased CNS-regulated hormonal activity remain unclear, and candidate mediators such as free fatty acids and adipokines also warrant further investigation. In tandem with elevated hormonal activity, sympathetic dominance of ANS activity in men may certainly have additional unfavorable effects on glucose metabolism, given the strong association between this phenomenon and insulin resistance, components of the metabolic syndrome and T2D [28]. As for the absence of parasympathetic inhibition during hypoglycemia observed in men, the underlying mechanisms as well as the metabolic consequences are less clear. Interestingly, the male HRV pattern during hypoglycemia is similar to that associated with obesity, insulin resistance, prediabetes and T2D as reported previously [5, 8] and presently. Nevertheless, adjustment for metabolic variables did not attenuate the observed sex effect, suggesting other factors at play, such as inherent sex-differences in central autonomic regulation or peripheral vascular response to hypoglycemia. Regarding the metabolic consequences of the current finding, it should be pointed out that parasympathetic activity is a well-known stimulus for glucagon secretion [29] and remaining parasympathetic drive during hypoglycemia may therefore contribute to the concomitantly elevated glucagon response in men. While the observed sympathetic dominance in postmenopausal women may be metabolically detrimental, the present results do not support a role for counter-regulatory hormones in explaining the increased risk of T2D following menopause [28] and other factors such as weight gain, abdominal fat accumulation and impaired insulin secretion as well as signaling in target organs may be more important. We further acknowledge that limitations related to sample size, single-time point assessment of sex hormones, and the inability to capture tissue-specific hormonal actions make it difficult to exclude a role for sex hormones in regulating counter-regulatory responses. There were also residual sex differences that remained unexplained by the variables examined, suggesting the presence of additional mechanisms beyond the scope of the current study. Strengths and limitations In comparison with previous studies with a similar scope, the current sample size was heterogeneous with regard to metabolic profiles, ranging from lean, metabolically healthy individuals to obese individuals with T2D. The study also included both pre- and postmenopausal women, allowing investigation of the impact of menopausal status on glucose-dependent neurohormonal responses, which has not been done before to our knowledge. In addition, study assessments were obtained during both hypo- and hyperglycemic clamps, collectively representing most of the physiological and pathological glycemic range, which is unique. There are however several limitations—most of which have been outlined above—that need to be addressed. First, a fundamental limitation is the cross-sectional study design, which precludes causal inference regarding the observed group differences and associations. Hence, the mechanistic interpretations offered in this discussion should be considered hypothesis-generating, requiring validation in prospective and interventional studies. Second, the current sample size in this post hoc study reflect the practical and procedural constraints inherent to hyperinsulinemic clamp studies, limiting statistical power, in particular for subgroup and association analyses. Although consistent sex differences across outcomes reduce the likelihood of false negatives, null findings, especially regarding sex hormones and menopausal status, should be interpreted with caution, as modest but clinically relevant effects may have been missed. Moreover, the sample was drawn from a single center in an academic setting and consisted predominantly of individuals of European ancestry, thereby limiting the generalizability of the results to other populations. Third, multiple comparisons were made without formal adjustment of P-values, which may have led to false positive findings. This necessitates cautious interpretation of nominal P-values and further underscores the need for replication in independent cohorts. Fourth, exposure to sex hormones was based on a single measurement, which may not accurately reflect long-term hormonal exposure, particularly in premenopausal women with cyclic variation. Thus, the observed absence of associations between sex hormones and counter-regulatory responses may reflect measurement imprecision rather than a true biological null effect, which might have been mitigated by implementing repeated measurements. Fifth, beyond circulating levels, local actions of sex hormones in key target organs, such as the brain, may differ according to sex, menopausal status, and metabolic context. For example, a lower ESR1:ESR2 gene expression ratio has been observed in adipose tissue from postmenopausal compared with premenopausal women and has been associated with markers of insulin resistance and obesity in men [30, 31]. Furthermore, sex hormone–mediated processes occurring during early stages of organ development may also contribute to the observed sex differences. Sixth, HRV mainly reflects vagal activity at the sinoatrial node, does not directly measure whole-body sympathetic outflow and has poor concordance with other methods to assess ANS activity, such as MSNA [32]. Therefore, these findings should be interpreted with caution, and future studies should include direct measures of sympathetic activity. Finally, while the metabolic heterogeneity of our sample enhances generalizability and extends prior knowledge, it also introduces complexity in interpretation. Biases arising from this heterogeneity may not have been fully accounted for in the sensitivity analyses, and residual confounding may also have influenced the results. Conclusions Our current findings suggest that, in individuals with and without T2D, the secretion of glucose-raising hormones are elevated in men compared to both pre- and postmenopausal women at physiologically relevant glucose concentrations. Sex differences in circulating levels of sex hormones do not seem to account for these findings but a potential role of android metabolic features such as visceral adiposity has been highlighted. It is plausible that hypersecretion of insulin-antagonistic hormones and relative sympathetic dominance in men adversely affect glucose homeostasis, which may contribute to the higher risk of T2D observed in men, though this hypothesis requires testing in prospective studies. Further research on the interaction between sex and insulin-antagonistic factors in the development of T2D is clearly warranted. Future studies should confirm the present findings in larger cohorts and use longitudinal approaches with repeated assessments in men and women, ideally before and after interventions targeting specific neurohormonal systems, including the reproductive apparatus. Acknowledgments We are grateful for the hard work by nurses Anna Åhlander, Sofia Löfving, Carola Almström, Caroline Woxberg, Kristina Mood and Erik Sjölund in conducting the experiments. The authors used AI-assisted tools to improve language and readability. No AI tools were used for data analysis or interpretation. The authors take full responsibility for the content of the manuscript. Contributor Information Martin H Lundqvist, Email: martin.lundqvist@medsci.uu.se, Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala University Hospital, Uppsala 751 85, Sweden. Maria J Pereira, Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala University Hospital, Uppsala 751 85, Sweden. Urban Wiklund, Department of Diagnostics and Intervention, Radiation Physics, Biomedical Engineering, Umeå University, Umeå 901 85, Sweden. Susanne Hetty, Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala University Hospital, Uppsala 751 85, Sweden. Jan W Eriksson, Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala University Hospital, Uppsala 751 85, Sweden. Funding This work was funded by Diabetesfonden (DIA2021-661, DIA2024-935), Stiftelsen för strategisk forskning EXODIAB (Excellence of Diabetes Research in Sweden), Stiftelsen Familjen Ernfors Fond, P.O. Zetterlings stiftelse, Novo Nordisk Fonden (NNF20OC0063864, NNF23OC0084483), Agnes och Mac Rudbergs Stiftelse, the European Commission Horizon RIA project PAS GRAS (101080329), Vetenskapsrådet (project 2024-03344), Stiftelsen för strategisk forskning (project CMP22-0014) and the Uppsala University Hospital ALF grants. Author contributions M.H.L. assisted in designing the study, recruiting subjects, supervising experiments, managing and analyzing data and wrote the first draft of the manuscript. M.J.P. assisted in data management, planning of biochemical analyses and provided critical input on presentation of data and statistical methods. U.W. performed HRV assessments. S.H. assisted in data management and planning of biochemical analyses. J.W.E. designed and supervised the conduction of the study, gave critical input on all aspects of presentation, analysis and interpretation of data. All authors have assisted in writing the final version of the manuscript and in the interpretation of the results. Disclosures The authors have nothing to declare. Data availability Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided. References Associated Data Data Citations Data Availability Statement Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.