Low-Dose ACTH Stimulation Test in Obesity: A Randomized Dose Assessment

Background The short cosyntropin test is widely used for adrenal insufficiency screening and diagnosis. Lower cosyntropin doses may have greater sensitivity vs. the standard dose in detecting adrenal dysfunction. Obesity and overweight are increasing, impacting the clinical presentation of some diseases. Currently more than 50% of the subjects diagnosed with autoimmune adrenal insufficiency have a BMI greater than 25, and hence individuals living with overweight and obesity are more frequently requiring evaluation of the adrenal cortical function. Fixed-dose cosyntropin stimulation may not be appropriate for individuals with obesity. Objective The primary objective was to compare cortisol response to a weight-adapted cosyntropin dose vs. a fixed low dose (1 µg) and a more physiologically fixed dose (10 µg). Methods Twenty individuals with obesity and 20 age-matched healthy controls underwent in a randomized sequence at least one-week apart, to the The short cosyntropin test with three different doses, 0.2 µg/kg of body weight, 1 or 10 µg fixed dose stimuli. The assessment and data analysis were blinded to the individual and the investigator. Results Cortisol response was reduced in the group with obesity with the 1 µg fixed dose stimuli at 30 minutes (median, IQR) 649.6 µg, 567.3–738.4 µg for the control group vs. 568.6 µg, 528.4–623.13 µg, p=0.04; there was a lower cortisol peak at 60′ in all the three evaluated doses, with a dose-dependent trend. A weight-adapted cosyntropin dose of 0.2 µg in obesity produces a similar response to the one observed in individuals without obesity. The 1 µg ACTH test falls short on stimulating the cortisol adrenal response in individuals with obesity.


Introduction
Although there may be still some controversies on the dosing, the short cosyntropin test is widely used for adrenal insufciency screening and diagnosis [1]. Classically, a single, 250 µg, fxed dose was considered the standard stimulus [2]; however, reduced cosyntropin doses have been compared to the insulin tolerance test (ITT) providing similar peak serum cortisol responses [3,4]. It is generally accepted that lower cosyntropin doses may have greater sensitivity vs. the 250 µg test [5,6]; a recent systematic review found similar diagnostic accuracy for low and standard doses [1]. However, the studies included in the systematic review did not take into account any of the baseline characteristics of the participants to individualize the ACTH dosing.
In the past, we proposed a more physiological dose to be used instead of 250 µg. We compared 10 μg vs. 250 μg cosyntropin to evaluate the adrenocortical response in healthy individuals and in a variety of hypothalamic-pituitary-adrenal axis disorders. Tere were no diferences in baseline or peak serum cortisol response in healthy subjects, but in some AI subjects, the results showed normal cortisol response during the standard test but abnormal cortisol response during the low-dose test. We also found better sensitivity with the low-dose vs. standard-dose cosyntropin stimulation test [7]. Graybeal and Fang studied the cortisol and ACTH responses to the insulin tolerant test, and they found that cortisol response to a 0.2 µg/kg dose mimics the response to ITT [8]. Our reasoning at the moment was to test the typical individual being investigated for adrenal insufciency, so, if planning to test a representative 50-70 kg of body weight individual, it was reasonable to use 10-14 µg cosyntropin stimuli.
Obesity and overweight prevalence are rising around the globe, and the World Health Organization estimated a combined prevalence of overweight and obesity in adults by 2016 of 52% (39% overweight, 13% with obesity) [9]. Additionally, over the past four decades, the number of children and adolescents with obesity has risen more than tenfold [10]. Predictions have been made that if this trend continues, approximately 60% of the world's adult population could be either overweight or with obesity by 2030 [11]. Tis epidemiological revolution may impact the archetypal clinical presentation of some diseases, and one of those well described examples is type 1 diabetes, in which more than 30% of the subjects afected now are either overweight or living with obesity [12]. Adrenal insufciency is not an exception and in a recent nationwide Swedish registry, Dalin et al. found more than 50% of the subjects diagnosed with autoimmune adrenal insufciency having a BMI greater than 25 (36.9% between 25 and 30, 13.8% ≥ 30) [13]. Nowadays patients may present with overweight or obesity, not just because the increase trend describe above, but also because it can represent an earlier stage of the disease. So an underweight patient its not imprescindible to suspected adrenal insufciency, since the typically wasted, underweight patient reported in the past, may represent a late stage of the disease [14].
Based on Graybeal's fndings, suggesting that a weightadjusted dose of 0.2 micrograms per kilo of body weight would elicit indistinguishable cortisol and cosyntropin response compared to the stress-induced ITT, we designed a blinded, randomized-sequence study aiming to compare the cortisol responses to 1 µg and 10 µg cosyntropin fxed doses and a weight-adjusted stimulation with 0.2 micrograms per kilo of cosyntropin in normal weight and individuals with obesity.

Participants.
Te study protocol was approved by the Institutional Research and Ethics Committee of "Dr. José Eleuterio González" University Hospital. Each volunteer gave their written informed consent before any study procedure. Twenty individuals with obesity (defned as a BMI equal or greater than 30) without comorbidities, between 18 and 60 years old, and 21 age-matched healthy controls with a BMI between 18.5 and 25 were included in the study. We excluded individuals with any endocrine disorder, use of glucocorticoids in the prior 24 months, and current use of drugs that may interfere with cortisol metabolism, transport, or the hypothalamus-pituitary-adrenal axis function. One subject on the healthy control group withdrew the consent before participating in any study procedure due to scheduling conficts and was substituted with another agematched volunteer. Individuals within the obesity group were invited to participate from the internal medicine, primary care, and endocrinology clinics of "Dr. José E. González" University Hospital once any medical condition other than obesity was ruled out.

2.2.
Methods. Prior to study start, a blinded randomization procedure was designed to receive any of the three study doses (1 µg, 10 µg, or 0.2 mcg per kilogram of body weight) of cosyntropin, ACTH 1-24 (Cortrosyn ® , Amphastar Pharmaceuticals Inc, Rancho Cucamonga, CA, USA) in a random sequence. Only the study coordinator (KCR) was aware of the testing dose allocation and was responsible for administering the blinded dose and blinding the blood samples. All cosyntropin stimulation tests were done at least seven days away from each other to avoid any carry-over efect. Te cosyntropin preparation and dilution procedures were previously described [15].
Patients were instructed to be present at the laboratory facilities after an overnight fasting. Weight and height were measured using a calibrated scale (SOEHNLE ® , Backnang, Germany) and between 0800 and 0900; while sitting in a blood drawing chair, an IV catheter (21-gauge vein needle set, BD Vacutainer ® Safety-Lok ™ Blood Collection Set, Franklin Lakes, NJ, USA) was inserted into a forearm vein and a baseline sample was taken. Te set was heparinized and subsequently, 0.2 µg/kg, 10 or 1 of cosyntropin were injected as a bolus and blood samples were collected at +5, +10, +30 and + 60 minutes, eliminating the frst 2 ml extracted to avoid dilution with the heparinized solution. Te collected samples had the serum separated and preserved at −20°C in aliquots to be processed in batches to reduce the inter-assay variability. Cortisol was measured using a commercial electrochemiluminescence kit (Modular Analytics 6000, Roche/Hitachi Diagnostics, Mannheim, Germany). Te intra-assay and inter-assay CV were 1.1% and 1.98%, respectively. Te intra-assay and inter-assay CV were 8.7% and 9.8%, respectively.

Statistical Analysis.
Te database was blinded for the analysis regarding the dose administered and their sequence and unblinded for the authors only after the statistical analysis was fnalized. Variables were tested for normality, using the Kolmogorov-Smirnov test. Median and interquartile range (IQR) are reported as central tendency and dispersion measures for continuous variables. Nominal variables are reported as proportions. Te Mann-Whitney U test for independent samples was performed to test distributions of cortisol at baseline to compare across BMI categories. Te Kruskal-Wallis test was done to compare the responses at 30 and 60 minutes to the three diferent doses of cosyntropin using as grouping variable the presence of obesity and fnally to compare the three doses within the normal BMI and groups with obesity. Friedman's two-way analysis of variance by ranks was done. For all the inferential tests, a p value ≤0.05 was considered signifcant.

Cortisol Responses.
Baseline serum cortisol was not diferent between nBMI and oBMI participants during either of the three stimulation tests (Figure 1). Figure 2(a) presents the +30′ cortisol value in response to the three diferent stimulation doses used in the study. Te oBMI group had a lower serum cortisol compared to the nBMI group, only during the 1 µg stimulation test, with non-diferent responses to the weight adjusted or the 10 µg cosyntropin fxed doses. Additionally when comparing the 30′ serum cortisol within the groups, it was lower in the 1 µg test in both oBMI and nBMI groups. +60 minutes responses are shown in Figure 2(b). All study doses at +60′ exposed diferences between oBMI and nBMI groups with lower responses in the former. Te within group comparison, aligned with the 30′ results, showed a reduced response to the 1 µg dose.

Discussion
In our study, otherwise healthy subjects with obesity had signifcantly lower serum cortisol than healthy subjects without obesity at 30 and 60 minutes after the application of a fxed dose of 1 µg of cosyntropin as compared to the fxed 10 µg cosyntropin dose. Te lower response in subjects with obesity suggests that the stimulation caused by 1 µg cosyntropin dose depends on subject's body weight. Tere was no diference at +30 minutes between the fxed 10 µg cosyntropin dose compared to the weight-adjusted dose of 0.2 µg/kg supporting the weight-related efect.
Te relationship between cosyntropin dose, adrenal response, and body weight could be in part explained by the efect of fat distribution and obesity on HPA axis [16]. Roelfsema et al. reported that there may be a reduced adrenal sensitivity to cosyntropin in premenopausal women with obesity [17]. Tis altered adrenal response has been linked to body fat distribution. Pasquali et al. studied the adrenal response to two boli of 0.2 µg/kg of BW of cosyntropin in healthy subjects with obesity reporting higher cortisol peak in subjects with abdominal obesity compared to peripheral obesity [18]. However, they did not fnd a correlation of stimulated cortisol and BMI or weight, and the study was done with a double stimulus. In our study, we did not fnd evidence of adrenal hyperresponsiveness to cosyntropin. We used BMI to defne obesity fnding a signifcant diference between groups in cortisol response at 30 and 60 minutes comparing low dose to the other studied doses without any signifcant diferences in the baseline serum cortisol and randomizing the dose order to avoid any carry-over or preconditioning. Our results contrast with previously mentioned studies suggesting an efect of weight and BMI on adrenal response.
Our results have shown that cortisol response peak with 0.2 µg/kg compared to 1 µg was higher at 30 minutes in both studied groups. Cortisol values at 60 minutes were higher with 0.2 µg/kg compared to the other doses. Tis fnding suggests that weight-adjusted cosyntropin doses stimulate adrenal gland better than other doses. Our results contrast with Dickstein previous studies fnding a maximal adrenal response with 1 µg cosyntropin dose [19]. MacCario et al. studied lower doses, fnding similar results with 1 µg dose [20]. Te anthropometric measurement adjusted to dose may infuence adrenal response to the cosyntropin stimulation test. MacCario et al. studied body surface adjusted doses, showing no diference of cortisol response between lean women and women with obesity [20]; in their study, cortisol, aldosterone, and dehydroepiandrosterone responses to high, low, and very low cosyntropin doses in women with obesity overlapped with the age-matched lean control group responses. In contrast, Gonzalbez et al. did not fnd diferences in the response to 1 µg, 250 µg, or ITT stimulation in the 30 min response vs. the peak response to the ITT [21]. We used a body weight dose approach, showing that this could be a more appropriate way to stimulate the cortisol response to cosyntropin in individuals living with obesity in whom adrenal insufciency has to be ruled out. Further studies with more frequent and earlier sampling as well as comparison to ITT response are granted.
Te strengths of our study include the design, a randomized clinical trial, the blinded analysis and the fact that we selected individuals without other comorbidities. Certainly, our study has some limitations, and we assessed the diference in response to diferent doses in healthy subjects with and without obesity. Our study fndings would need to be reproduced and confrmed in individuals with suspected AI and overweight-obesity.

Conclusions
In conclusion, our fndings indicate that 1 µg cosyntropin dose to perform adrenal function test can lead to submaximal stimulation both in subjects with and without obesity and a weight-adjusted dose using 0.2 µg/kg elicits a similar response in both the oBMI and nBMI groups. An insufcient adrenal stimulation in the context of a stimulation test might have diagnostic and therapeutic consequences which need to be confrmed in new studies. On the basis of our fndings in AI, the performance of other dynamic endocrine testing should be evaluated in individuals living with obesity.

Data Availability
Te data that support the fndings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this paper.