Assessing the predictive value of the suppressed 1 mg overnight dexamethasone suppression test in success of bariatric surgery

Abstract

Bariatric surgery has been proven to be a successful intervention for managing obesity. There are numerous studies in the literature aiming to predict the factors influencing the success of bariatric surgery. Our study aims to determine whether preoperative 1 mg overnight dexamethasone suppression test (1 mg-DST) serum cortisol levels can serve as predictors of the effectiveness of bariatric surgery in severe obese patients without Cushing syndrome. A total of 98 patients who underwent bariatric surgery were included in the study. The preoperative 1 mg-DST levels, insulin levels, thyroid function tests, and lipid profiles of the patients were recorded. The patients’ preoperative, postoperative 3rd, and 6th month weights were recorded and the percent total weight loss (%TWL) is calculated. Patients were categorized into 2 groups based on their TWL at 6 months. The 1 mg-DST results were significantly lower in the high-TWL-6 group (0.93 ± 0.37 μg/dL) compared to the low-TWL-6 group (1.09 ± 0.36 μg/dL, P = .040). Similarly, Homeostatic Model Assessment for Insulin Resistance values were lower in the high-TWL-6 group (5.63 ± 2.21) compared to the low-TWL-6 group (6.63 ± 2.55, P = .047). The optimal cutoff value found for 1 mg-DST level was 0.97 µg/dL, providing 50% sensitivity and 70% specificity. This study is the first to examine the predictive role of suppressed 1 mg-DST levels on postoperative weight loss in nondiabetic patients. The most prominent result of this study was that we observed a negative correlation between 1 mg-DST levels and %TWL.

1. Introduction

Over the past few decades, there has been a significant increase in the prevalence of obesity in both Western and developing countries. It is estimated that by the year 2035, the projected obesity prevalence among children and adults worldwide will be 24%.^[1] Obesity not only affects the quality of life of individuals but also contributes to increased morbidity and mortality rates. In 2019, it is estimated that high body mass index (BMI) levels beyond the optimal range led to around 5 million deaths due to noncommunicable diseases such as cardiovascular diseases, diabetes mellitus, cancers, neurological disorders, chronic respiratory diseases, and digestive disorders.^[2] The treatment of obesity generally consists of lifestyle modifications, medical therapies, and bariatric surgery.^[3] According to the latest guideline revision by the American Society for Metabolic and Bariatric Surgery and the International Federation for the Surgery of Obesity and Metabolic Disorders bariatric surgery is recommended for individuals with a BMI ≥ 35 kg/m², irrespective of the severity or presence of obesity-related complications.^[4] Prior to bariatric surgery, assessing patients for hormonal and metabolic abnormalities that could contribute to obesity is crucial. According to the Endocrine work-up in obesity guideline from the European Society of Endocrinology, preoperative screening for hypercortisolism is recommended for patients undergoing bariatric surgery.^[5] The 1 mg overnight dexamethasone suppression test (1 mg-DST) is recommended as a screening test for hypercortisolism prior to bariatric surgery.^[5] This test is considered adequately sensitive for excluding hypercortisolism, with a post-dexamethasone level threshold of ≤1.8 µg/dL.^[6] However, it is not appropriate to rigidly classify the cortisol value obtained from the 1 mg-DST. In some studies, patients exhibiting cortisol levels that fail to suppress after the 1 mg-DST have been divided into 2 subgroups. These subgroups include individuals with 1 mg-DST levels ranging between 1.9–4.9 µg/dL, and those with levels measuring 5 µg/dL or above.^[7,8] As seen, despite the establishment of certain cutoffs based on 1 mg-DST results, this distinction has not yet been fully clarified. Our study aims to examine whether 1 mg-DST results below 1.8 µg/dL during the preoperative period can serve as predictors of the effectiveness of bariatric surgery.

2. Methods

Our study was approved by the Ethics Committee of Sutcu Imam University Faculty of Medicine dated December 14, 2022 and numbered 08. All clinical investigations were conducted in accordance with the principles of the Declaration of Helsinki.

All participants presented with a BMI of ≥40 kg/m², accompanied by obesity-related health issues, and had not achieved significant weight reduction despite undergoing medical therapy for over 6 months. Patients with Cushing syndrome, other hormonal abnormalities, those undergoing steroid therapy, as well as patients with diabetes mellitus or using hypoglycemic agents, were excluded from the study. In addition to these, patients with chronic liver disease, chronic kidney disease, pituitary disease, a history of malignancy, pregnant, and lactating women, as well as those with a history of medication altering dexamethasone metabolism, were not included in the study. The preoperative insulin levels, thyroid function tests, and lipid profiles of the patients were recorded. The 1 mg-DST results routinely performed before bariatric surgery in our clinic were also recorded. Serum cortisol was measured at 08.00 to 09.00 am after the administration of 1 mg dexamethasone at 23.00 pm on the previous day. The patients’ preoperative, postoperative 3rd, and 6th month weights were recorded. The percent total weight loss (%TWL) was calculated as 100 × (operative weight − follow-up weight)/operative weight.

Recent studies suggest that a successful gastric bypass surgery should result in a 6-month % Total Weight Loss (%TWL-6) of ≥25.^[9,10] In this context, patients were categorized into 2 groups based on their total weight loss at 6 months: those with a low-TWL-6 < 25 and those with a high-TWL-6 ≥ 25.

2.1. Statistical analysis

Data analysis was performed using R software (Version 4.0.2, R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were reported as mean ± standard deviation, and categorical variables were reported as frequency (percentage). Independent tests appropriate for continuous variables and the chi-squared test for categorical variables were used. The normality of numerical data was evaluated using both graphical and statistical methods. The predictive power of cortisol levels for %TWL-6 ≥ 25 was analyzed using receiver operating characteristic curves, and the area under the curve (AUC) with 95% confidence intervals was calculated. The optimal cutoff point was determined, and the sensitivity and specificity rates were calculated. Factors influencing a %TWL-6 of ≥25 were identified through both univariate and multivariate analyses. To ensure the validity of the multivariate model, tests for collinearity among variables were conducted. Variables that passed the collinearity tests were included in the analysis, ensuring that the model remained robust and accurate.

3. Results

3.1. Patient demographics and clinical outcomes following bariatric surgery

This retrospective observational study encompasses 98 patients who underwent bariatric surgery. The distribution of participants by sex was not statistically significant between groups: 60% (24 individuals) in the low-TWL-6 group and 69% (40 individuals) in the high-TWL-6 group, with an overall female representation of 65.3% (64 individuals; P = .484). The mean age was 33.70 ± 8.92 years in the low-TWL-6 group and 32.14 ± 9.18 years in the high-TWL-6 group, with no significant age difference observed (P = .463). BMI averages were 44.46 ± 3.80 kg/m² in the low-TWL-6 group and 46.55 ± 5.63 kg/m² in the high-TWL-6 group, approaching statistical significance (P = .074).

The 1 mg-DST results were significantly lower in the high-TWL-6 group (0.93 ± 0.37 μg/dL) compared to the low-TWL-6 group (1.09 ± 0.36 μg/dL, P = .040) (Figure 1). Similarly, Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) values were lower in the High-TWL-6 group (5.63 ± 2.21) compared to the low-TWL-6 group (6.63 ± 2.55, P = .047). No significant differences were detected in thyroid function tests, lipid profiles, or other laboratory parameters between the groups. The details of these findings are summarized in Table 1.

Figure 1.

Cortisol levels after 1 mg dexamethasone suppression test by percent total weight loss in 6-month (%TWL-6) groups.

Table 1.

Demographic and laboratory characteristics of the TWL-6 groups.

Variables	Total n = 98	Low-TWL-6 n = 40	High-TWL-6 n = 58	P value
Sex (female)	%65.3 (64)	%60.0 (24)	%69.0 (40)	.484
Age (years)	32.78 ± 9.07	33.70 ± 8.92	32.14 ± 9.18	.463
BMI (kg/m²)	45.70 ± 5.05	44.46 ± 3.80	46.55 ± 5.63	.074
%TWL-3	16.17 ± 4.25	13.16 ± 3.35	18.24 ± 3.51	.000
%TWL-6	28.09 ± 6.68	22.11 ± 4.28	32.21 ± 4.58	–
1 mg-DST (μg/dL)	1.00 ± 0.37	1.09 ± 0.36	0.93 ± 0.37	.040
Insulin (mIU/L)	24.69 ± 9.85	24.72 ± 9.67	24.67 ± 10.05	.980
HOMA-IR	6.04 ± 2.39	6.63 ± 2.55	5.63 ± 2.21	.047
TSH (mIU/L)	2.50 ± 0.69	2.64 ± 0.66	2.40 ± 0.70	.092
FT4 (ng/dL)	1.16 ± 0.19	1.16 ± 0.13	1.16 ± 0.22	.169
FT3 (pg/mL)	3.32 ± 0.56	3.35 ± 0.55	3.29 ± 0.56	.639
LDL (mg/dL)	117.29 ± 37.12	115.72 ± 36.65	118.36 ± 37.72	.831
HDL (mg/dL)	47.99 ± 15.33	49.23 ± 17.97	47.14 ± 13.30	.756
TG (mg/dL)	130.83 ± 58.35	127.30 ± 57.90	133.26 ± 59.04	.633

%TWL-3 = 3-month % total weight loss, %TWL-6 = 6-month % total weight loss, 1 mg-DST = 1 milligram dexamethasone suppression test, BMI = body mass index, FT3 = free triiodothyronine, FT4 = free thyroxine, HDL = high-density lipoprotein, high-TWL-6 = 6-month % total weight loss of ≥25, LDL = low-density lipoprotein, low-TWL-6 = 6-month % total weight loss of <25, TG = triglycerides, TSH = thyroid-stimulating hormone.

3.2. Factors correlating with percentage of total weight loss

The correlation analysis between demographic and laboratory parameters and the %TWL at 3 (TWL-3) and 6 months (TWL-6) is detailed in Table 2. No significant correlations were observed between %TWL and the following parameters at either time point: age, BMI, free triiodothyronine (FT3), high-density lipoprotein, low-density lipoprotein, triglycerides, and thyroid-stimulating hormone (TSH); all exhibited P-values above .05. A negative correlation was detected between free thyroxine (FT4) levels and %TWL at 6 months (r = ‐0.182, P = .074). Although this correlation was not statistically significant, it approached significance, suggesting a trend where increased FT4 levels are associated with reduced weight loss. A similar nonsignificant negative correlation was observed between insulin levels and %TWL at 3 months (R = 0.128, P = .208).

Table 2.

Correlation analysis of demographic and laboratory parameters with %TWL values at 3 and 6 months.

Variable	Correlation TWL-3	Correlation TWL-6
Age (years)	R = 0.011, P = .916	R = 0.003, P = .977
BMI (kg/m²)	R = 0.064, P = .533	R = 0.071, P = .488
TSH (mIU/L)	r = -0.093, P = .362	r = -0.137, P = .180
FT3 (pg/mL)	r = -0.001, P = .992	R = 0.030, P = .769
FT4 (ng/dL)	r = -0.193, P = .057	r = -0.182, P = .074
HDL (mg/dL)	r = -0.058, P = .571	r = -0.060, P = .559
HOMA-IR	r = -0.124, P = .225	r = -0.222, P = .028
LDL (mg/dL)	r = -0.071, P = .486	R = 0.031, P = .763
TG (mg/dL)	R = 0.007, P = .942	r = -0.022, P = .827
1 mg-DST (μg/dL)	r = -0.198, P = .051	r = -0.286, P = .004
Insulin (mIU/L)	R = 0.128, P = .208	r = -0.045, P = .662

%TWL-3 = 3-month % total weight loss, %TWL-6 = 6-month % total weight loss, 1 mg-DST = 1 milligram dexamethasone suppression test, BMI = body mass index, FT3 = free triiodothyronine, FT4 = free thyroxine, HDL = high-density lipoprotein, LDL = low-density lipoprotein, TG = triglycerides, TSH = thyroid-stimulating hormone.

Significant findings include a negative correlation between the HOMA-IR and %TWL at 6 months (r = ‐0.222, P = .028), indicating that higher insulin resistance is associated with less weight loss. Additionally, 1 mg-DST results were significantly negatively correlated with %TWL at 6 months (r = ‐0.286, P = .004), suggesting that higher cortisol levels are associated with reduced weight loss.

3.3. Preoperative predictors of weight loss efficacy 6 months post-bariatric surgery

Receiver operating characteristic curve analysis was conducted to determine the cutoff value of 1 mg-DST level for predicting the group where %TWL is ≥25, aiming to forecast a suitable event (Figure 2). The optimal cutoff value found for 1 mg-DST level was 0.97 µg/dL, providing 50% sensitivity and 70% specificity. The calculated AUC value was 0.610 with a 95% CI: [0.495, 0.719]. This AUC value indicates that cortisol levels are somewhat useful in predicting cases where %TWL is ≥25; however, it also suggests that the predictive power of the model is not excellent. To enhance predictive performance, additional biomarkers may need to be included. While the AUC value demonstrates a certain level of utility in predicting cases where 1 mg-DST levels are indicative of %TWL being ≥25, it also signifies that the model’s predictive power is not optimal.

Figure 2.

Cortisol levels after 1 mg dexamethasone suppression test and prediction of 6-month % total weight loss of ≥25 (%TWL ≥ 25): receiver operating characteristic (ROC) curve analysis.

In this study, univariate and multivariate logistic regression analyses were conducted sequentially to identify independent risk factors for individuals with a TWL of 25% or more over a 6-month period. The analyses included variables such as sex (male), age, BMI, cortisol, insulin, TSH, FT4, FT3, low-density lipoprotein, high-density lipoprotein, and triglycerides. The univariate logistic regression results particularly highlighted significant relationships with the group having a TWL ≥ 25%. For BMI, an odds ratio (OR) of 1.096 (95% CI 1.001–1.201, P = .048) was calculated, and for 1 mg-DST, an OR of 0.311 (95% CI 0.098–0.983, P = .047) was observed. These results suggest a potential relationship between these variables and the group with TWL ≥ 25%. Variables with a P-value <.2 in the univariate analysis were subsequently included in the multivariate logistic regression model. In this model, the 1 mg-DST variable maintained its significance, demonstrating an independent effect with an OR of 0.292 (95% CI 0.087–0.973, P = .045), which was statistically significant. While BMI did not reach statistical significance in the model, it showed a noteworthy effect with an OR of 1.093 (95% CI 0.991–1.206, P = .074). Additional findings are presented in Table 3.

Table 3.

Univariate and multivariate analyses of independent predictive variables for 6-month % total weight loss of ≥25 (high-TWL-6).

Factors	Univariate analysis		Multivariate analysis
	OR (95% CI)	P	OR (95% CI)	P
Sex (female)	0.675 (0.291, 1.567)	.360
Age (years)	0.981 (0.938, 1.026)	.401
BMI (kg/m²)	1.096 (1.001, 1.201)	.048*	1.093 (0.991, 1.206)	.074
1 mg-DST (μg/dL)	0.311 (0.098, 0.983)	.047*	0.292 (0.087, 0.973)	.045
Insulin (mIU/L)	0.999 (0.959, 1.041)	.979
HOMA-IR	0.834 (0.697, 0.997)	.046*	0.861 (0.710, 1.043)	.127
TSH (mIU/L)	0.597 (0.324, 1.100)	.098*	0.726 (0.384, 1.371)	.323
FT4 (ng/dL)	0.802 (0.096, 6.667)	.838
FT3 (pg/mL)	0.837 (0.402, 1.743)	.635
LDL (mg/dL)	1.002 (0.991, 1.013)	.728
HDL (mg/dL)	0.991 (0.965, 1.018)	.507
TG (mg/dL)	1.002 (0.995, 1.009)	.618

1 mg-DST = 1 milligram dexamethasone suppression test, BMI = body mass index, CI = confidence interval, FT3 = free triiodothyronine, FT4 = free thyroxine, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OR = odds ratio, TG = triglycerides, TSH = thyroid-stimulating hormone.

*The P-values of variables with P < 0.2.

4. Discussion

Obesity, which is increasingly recognized as a global health concern, is commonly attributed to an imbalance between energy intake and expenditure, influenced by a myriad of environmental and genetic factors. However, recent findings indicate that cortisol may serve as a pivotal factor in this epidemic.^[11] Recent studies has extensively examined different facets of this subject. The research conducted by Wester et al revealed elevated hair cortisol levels in obese individuals compared to normal-weight individuals.^[12] In another cross-sectional study involving 2527 participants, it was demonstrated that there is a positive relationship between basal cortisol levels and body weight, BMI, and abdominal obesity.^[13] As the topic gained popularity, studies on the relationship between bariatric surgery and cortisol have also commenced. The study conducted by Bando et al suggests that basal cortisol levels examined before bariatric surgery may be useful in predicting postoperative weight loss in male patients. However, the same correlation was not observed in female patients.^[14] The authors hypothesized that the absence of such a relationship in women may be attributed to the influence of estrogen on cortisol-binding globulin. In another study conducted by Ozturk et al, it was shown that preoperative basal cortisol levels were not useful in predicting postoperative weight loss.^[15] Between these 2 studies, there were differences in both the patient population and the study design. The study conducted by Ozturk et al primarily enrolled patients with diabetes mellitus. In the Bando study, the success of bariatric surgery was evaluated using %TWL, while in the Ozturk study, percent excess weight loss was utilized. Another finding observed in the study by Ozturk et al was that patients with more suppressed 1 mg-DST tended to experience statistically higher weight loss. Despite excluding patients with diabetes mellitus from our study due to its potential influence on the hypothalamic-pituitary-adrenal axis, the observed result consistent with the findings of our study. To our knowledge, there are no other studies investigating the predictive value of a 1 mg-DST for the success of bariatric surgery in nondiabetic patients. In this study, we observed that the correlation of 1 mg-DST results was statistically close to significant in TWL-3, while it was significant in TWL-6. In some studies, the effect of suppressed 1 mg-DST levels on cortisol-related diseases has been evaluated. In a study conducted by Castro et al it was shown that the prevalence of cardiovascular disease increased in patients with 1 mg-DST results above 0.9 µg/dL.^[16] In another study on depression, which is also one of the cortisol-related diseases, it was shown that it is more frequently observed in patients with 1 mg-DST levels above 0.9 µg/dL.^[17] In our study, similar to these studies, we showed that the optimal cutoff value for the 1 mg-DST level to predict the success of bariatric surgery was 0.97 µg/dL. However, we did not encounter a result that explains why relatively high cortisol levels at 1 mg-DST are associated with an increased risk of cortisol-related diseases. Increased activity of 11β-hydroxysteroid dehydrogenase type 1 in adipose tissue has been associated with central obesity and metabolic complications, as it amplifies local cortisol concentrations without significantly altering circulating levels.^[18] It is possible that the activity of these hormone synthesis pathways in the adipose tissue itself affects the effectiveness of surgery.

As far as we can determine, the impact of preoperative BMI on postoperative weight loss is controversial in the literature; some studies find an association, while others report no relationship.^[10,19,20] Patients’ follow-up durations, variances in surgical procedures performed, and the study design could be contributing factors to the differences observed in these studies. In our study, no significant correlation was found between preoperative BMI and TWL-6. In univariate analysis, BMI appeared to be a potential risk factor for patients who experienced more than 25% weight loss at 6 months, whereas in multivariate analysis, BMI was not identified as an independent risk factor, though it approached significance. This suggests that BMI may interact with other biochemical and demographic factors, diminishing its standalone effect. To comprehensively understand the impact of preoperative BMI on postoperative weight loss, more extensive and long-term studies are required.

In the literature, there are varying results in studies evaluating the effect of gender on the success of bariatric surgery. In the study conducted by Nielsen et al, female gender was shown to result in greater weight loss post-surgery, whereas in the study by Matteo et al, it was demonstrated that gender had no effect on the success of the surgery.^[21,22] In our study, we also found that gender did not have any significant impact on postoperative weight loss. The impact of age on postoperative weight loss also yields varying results in studies. While some studies suggest that older age decreases the likelihood of surgical success,^[23,24] others demonstrate that age does not influence postoperative weight loss.^[25,26] In our study as well, we observed that age did not have any impact on weight loss. We attribute these differences in the literature to differences in study design, variations in surgical procedures, and the sample sizes of the included patients.

Obesity is associated with several endocrine abnormalities, including thyroid dysfunction. We also investigate the effect of preoperative TSH, FT3, and FT4 levels in euthyroid patients on weight loss after bariatric surgery. We observed that none of these 3 parameters were statistically significant based on the weight loss at 6 months. The literature also shows that our results are consistent with other studies.^[27] Since all patients in our study were euthyroid, we do not expect thyroid hormones to have an effect on the weight loss of patients after bariatric surgery. Additionally, in our study, we observed that there was no relationship between preoperative lipid profiles of patients and postoperative weight loss.

One of the parameters we examined in our study was the relationship between postoperative weight loss and preoperative HOMA-IR, which is an indicator of insulin resistance. We observed a negative correlation between the HOMA-IR and %TWL at 6 months, indicating that higher insulin resistance is associated with less weight loss. In the study conducted by Casa-Tapia et al, an association was found between a higher preoperative HOMA-IR and increased weight loss 1 year after surgery.^[28] The difference between the 2 studies could be attributed to the different patient groups included in the studies. Our study was conducted with nondiabetic patients, while the other study included patients with diabetes mellitus. Additionally, the average BMI of the patients in our study was higher compared to the other study. Other studies in the literature also indicate conflicting results regarding the association between nonsurgical weight loss methods and insulin resistance.^[29,30] It is evident that further comprehensive studies are required on this topic.

This study has some limitations. It was retrospective in nature, restricted to a single center, and had a relatively small sample size. Also we did not evaluate certain clinical parameters that might affect postoperative weight loss, such as mental state, alcohol consumption, diet adherence, and physical activity. Furthermore, the follow-up duration for the patients included in the study was relatively short, approximately 6 months.

5. Conclusions

There are numerous studies in the literature aiming to predict the factors influencing the success of bariatric surgery. This study is the first to examine the predictive role of suppressed 1 mg-DST levels on postoperative weight loss in nondiabetic patients. The most interesting result of this study was that we observed a negative correlation between 1 mg-DST levels and %TWL. This result suggests that the interpretation of 1 mg-DST should consider it as a continuous variable rather than a categorical (yes/no) one. Due to limitations in our study, there is a need for more extensive, long-term, prospective studies to further elucidate this issue.

Author contributions

Conceptualization: Hüseyin Avni Findikli, Bahtiyar Muhammedoğlu.

Data curation: Vehbi Şirikçi, Bahtiyar Muhammedoğlu.

Formal analysis: Hüseyin Avni Findikli.

Investigation: Vehbi Şirikçi.

Methodology: Cem Onur Kiraç.

Project administration: Cem Onur Kiraç, Hüseyin Avni Findikli.

Resources: Vehbi Şirikçi.

Writing – review & editing: Vehbi Şirikçi.

Writing – original draft: Cem Onur Kiraç.