Intranasal Oxytocin for Obesity

Abstract

BACKGROUND

Accumulating preclinical and preliminary translational evidence shows that the hypothalamic peptide oxytocin reduces food intake, increases energy expenditure, and promotes weight loss. It is currently unknown whether oxytocin administration is effective in treating human obesity.

METHODS

In this randomized, double-blind, placebo-controlled trial, we randomly assigned adults with obesity 1:1 (stratified by sex and obesity class) to receive intranasal oxytocin (24 IU) or placebo four times daily for 8 weeks. The primary end point was change in body weight (kg) from baseline to week 8. Key secondary end points included change in body composition (total fat mass [g], abdominal visceral adipose tissue [cm²], and liver fat fraction [proportion; range, 0 to 1; higher values indicate a higher proportion of fat]), and resting energy expenditure (kcal/day; adjusted for lean mass) from baseline to week 8 and caloric intake (kcal) at an experimental test meal from baseline to week 6.

RESULTS

Sixty-one participants (54% women; mean age ± standard deviation, 33.6±6.2years; body-mass index [the weight in kilograms divided by the square of the height in meters], 36.9±4.9) were randomly assigned. There was no difference in body weight change from baseline to week 8 between oxytocin and placebo groups (0.20 vs. 0.26 kg; P=0.934). Oxytocin (vs. placebo) was not associated with beneficial effects on body composition or resting energy expenditure from baseline to week 8 (total fat: difference [95% confidence interval], 196.0 g [−1036 to 1428]; visceral fat: 3.1 cm² [−11.0 to 17.2]; liver fat: −0.01 [−0.03 to 0.01]; resting energy expenditure: −64.0 kcal/day [−129.3 to 1.4]). Oxytocin compared with placebo was associated with reduced caloric intake at the test meal (−31.4 vs. 120.6 kcal; difference [95% confidence interval], −152.0 kcal [−302.3 to −1.7]). There were no serious adverse events. Incidence and severity of adverse events did not differ between groups.

CONCLUSIONS

In this randomized, placebo-controlled trial in adults with obesity, intranasal oxytocin administered four times daily for 8weeks did not reduce body weight. (Funded by the National Institute of Diabetes and Digestive and Kidney Diseases and others; ClinicalTrials.gov number, NCT03043053.)

Introduction

Obesity is a highly prevalent chronic disease associated with cardiometabolic complications and increased mortality.^1,2 Although glucagon-like peptide-1 receptor agonists effectively reduce weight, their use is limited by variable individual responses and side effects,³ highlighting the need to expand antiobesity drug options. Oxytocin, an anorexigenic neurohormone that drives weight loss in animals,⁴ is considered a promising potential therapeutic for obesity.⁵

Oxytocin is produced in the hypothalamus and released in brain regions involved in appetite and body weight control, including the hypothalamic melanocortin system, nucleus of the solitary tract, and ventral tegmental area, and to the peripheral circulation where oxytocin acts on target metabolic organs.^4,6,7 Oxytocin-deficient mice and humans with Prader–Willi syndrome, characterized by reduced number and activity of oxytocinergic neurons, develop obesity and metabolic disruptions, including decreased insulin sensitivity and dyslipidemia.^8–10 Extensive evidence in diet-induced obese animals indicates that acute oxytocin administration reduces food intake,^4,5 and chronic oxytocin administration (vs. placebo) leads to body weight reduction accompanied by increased energy expenditure, brown adipose tissue thermogenesis, reduced white adipose tissue volume,¹¹ and metabolic improvements (e.g., decreased insulin resistance and triglyceride levels).¹² In humans, single-dose intranasal oxytocin modulates functional magnetic resonance imaging (MRI) activation and connectivity of brain regions regulating eating behavior^13,14 and reduces food intake.^15–17 A pilot randomized clinical trial (RCT) in adults with obesity (n=9 receiving oxytocin vs. n=11 receiving placebo for 8weeks) showed a 3.2±1.9 reduction in body-mass index (BMI, the weight in kilograms divided by the square of the height in meters).¹⁸ In this 8-week randomized, double-blind, placebo-controlled trial, we evaluated the efficacy, safety, and mechanisms of intranasal oxytocin administered four times daily for treating obesity.

Methods

TRIAL DESIGN AND PARTICIPANTS

This trial¹⁹ was approved by the Mass General Brigham Institutional Review Board and conducted at Massachusetts General Hospital under a U.S. Food and Drug Administration Investigational New Drug application in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines and overseen by a Data Safety and Monitoring Board. The trial was preregistered at ClinicalTrials.gov (NCT03043053).

Inclusion criteria were 18 to 45 years of age, BMI greater than or equal to 30, and willingness to maintain current diet and physical activity level for the trial duration to enable testing of underlying mechanisms associated with oxytocin-induced weight loss. Exclusion criteria included a history of cardiovascular disease, bariatric surgery, or eating disorder; current psychiatric disorder; oral contraceptive use; medications altering metabolism, blood glucose, or appetite (stable dose of metformin for ≥3 months was allowed); and fasting glucose greater than or equal to 126 mg/dl or glycated hemoglobin (HbA1c) greater than or equal to 7% (a complete list is in Supplementary Methods in the Supplementary Appendix). Participants gave written informed consent before procedures. Participants were randomly assigned by an unblinded research pharmacist 1:1 to receive intranasal oxytocin (24 IU) or placebo stratified by self-reported sex assigned at birth and obesity class (I: BMI≥30 to <35; II: BMI≥35 to <40; III: BMI≥40). Data on sex, age, race, and ethnic group of individuals with obesity were obtained from the World Health Organization and U.S. National Center for Health Statistics to confirm the representativeness of the trial sample and generalizability of the findings. Additional information is available in the protocol provided with the full text of this article at evidence.nejm.org.

TRIAL END POINTS

The primary end point was change in body weight (kilograms) from baseline to week 8 (end of treatment) following administration of oxytocin versus placebo. Key secondary end points were changes from baseline to end of treatment in body composition (total fat mass [g], abdominal visceral adipose tissue [cm²], and liver fat fraction [proportion; range, 0 to 1; higher values indicate higher proportion of fat]), resting energy expenditure (kcal/day; adjusted for lean mass), and caloric intake (kcal) at an experimental test meal¹⁵ (administered at baseline and week 6). Additional secondary end points included changes from baseline to the end of treatment in total lean mass (g), self-reported daily caloric intake (kcal), metabolic profile (glucose [mg/dl], insulin [uIU/ml], the Homeostatic Model Assessment for Insulin Resistance, lipid panel [mg/dl], and high-sensitivity C-reactive protein [mg/l]), depressive and anxiety symptoms, eating disorder psychopathology (baseline to week 4), hedonic drive to eat (caloric intake [kcal] at a Cookie Taste Test,²⁰ baseline to week 4), and quality of life. Details of the specific scales used to assess these parameters are described below. Of note, on-treatment assessment time points before the end of treatment were chosen for selected outcomes to allow for studying underlying mediating mechanisms of oxytocin-induced improvements of primary and key secondary end points. We also analyzed fasting plasma oxytocin levels (pg/ml) to assess the impact of intranasal oxytocin administration on endogenous oxytocin levels over the course of the trial. Safety assessments comprised changes from baseline to week 8 in vital signs (systolic and diastolic blood pressure [mmHg] and pulse [beats per minute]) and electrocardiogram corrected QT (QTc) interval (ms). We systematically assessed safety throughout the trial (including an off-treatment follow-up visit 6 weeks after trial drug discontinuation and additional safety visits in case of early termination), including assessment of medical history, adverse events, psychiatric symptoms, physical examination, laboratory tests for electrolytes and renal and liver function, and urine pregnancy test (in women) at all visits.

TRIAL PROCEDURES

Following a screening visit to determine eligibility, participants completed baseline assessments of all end point measures; on-treatment assessments at weeks 2, 4, 6, and 8; and off-treatment follow-up at week 14 to assess sustained effects and safety of oxytocin.¹⁹

Trial Drug and Adherence

Intranasal oxytocin (Syntocinon nasal spray) and placebo were purchased from Victoria Apotheke (Zürich/Switzerland). Participants were instructed to administer three sprays per nostril (24 IU) four times daily (20 to 30 minutes before meals and bedtime) for 8 weeks. Adherence was defined as greater than or equal to 80% self-reported trial drug administration on the basis of diaries.

Anthropometric Measurements, Body Composition, and Resting Energy Expenditure

At all visits, trained research staff obtained fasting anthropometric measurements using calibrated instruments. At baseline and week 8, the following assessments were performed: fasting abdominal visceral adipose tissue using 3T MRI, liver fat fraction using breath-hold single-voxel hydrogen 1 magnetic resonance spectroscopy, resting energy expenditure using indirect calorimetry, and postmeal total fat and lean mass by dual-energy x-ray absorptiometry.

Caloric Intake

Caloric intake was assessed using a validated experimental test meal (Breakfast Test Meal¹⁵) at baseline and week 6 after a 10-hour fast. Trial drug administration was withheld before this test meal to capture chronic oxytocin effects. Participants self-reported daily food intake using a food record for 4 days at baseline, all on-treatment visits, and the off-treatment follow-up.

Metabolic Profile

From fasting blood samples, the following were assessed: HbA1c at baseline (to characterize the trial cohort); glucose at baseline and weeks 2, 4, 8, and 14; and insulin, lipid profile, and high-sensitivity C-reactive protein at baseline and weeks 4, 8, and 14.

Psychological, Behavioral, and Quality-of-Life Assessments

Depressive and anxiety symptoms were assessed at baseline and weeks 4, 8, and 14 using the Beck Depression Inventory II (total score range, 0 to 63; higher scores indicate more severe depressive symptomatology)²¹ and the State-Trait Anxiety Inventory²² Trait Scale (trait score range, 20 to 80; higher scores indicate higher trait anxiety), respectively. Eating disorder psychopathology was measured with the Eating Disorder Examination Questionnaire (score range, zero to six; higher scores indicate more pronounced psychopathology)²³ at baseline and weeks 4 and 14. Participants reported their total amount of exercise (hours per week) at all trial visits. The Cookie Taste Test,²⁰ where participants rated the taste of cookies and calories consumed were calculated as a measure of hedonic drive to eat beyond satiety (higher calorie intake [kcal] indicates more hedonic eating), was used at baseline and week 4 following a standardized snack to eliminate homeostatic drive to eat. Quality of life was assessed with the 36-Item Short-Form Health Survey (SF-36; Physical and Mental Component Summary scores are norm-based and transformed to a scale on which the 2009 general population of the United States has a mean score of 50 and a standard deviation [SD] of 10; higher scores indicate better quality of life)²⁴ at baseline and weeks 4, 8, and 14.

Plasma Oxytocin

From fasting blood samples, plasma oxytocin levels were analyzed at baseline and weeks 4, 6, 8, and 14.

Treatment Allocation Guess

At week 14, participants were asked to guess to which treatment group they had been assigned. Response options were “oxytocin,” “placebo,” and “not sure.”

BIOSTATISTICAL CONSIDERATIONS

Power Analysis

The primary end point was change in body weight (kg) from baseline to week 8 for the oxytocin group versus the placebo group. An a priori power analysis using two-sample t-tests at a two-tailed significance level of α=0.05 and an SD of approximately 4kg (reported by a pilot study of 8-week intranasal oxytocin administration in adults with obesity¹⁷) yielded that with greater than or equal to 50 participants completing the treatment and end-of-treatment visit, we would have 80% power to detect a between-group difference in weight loss of greater than or equal to 3.2 kg, which is 50% of what the pilot study¹⁷ observed and clinically meaningful.

Data Analysis

Statistical analyses were performed using Stata Statistical Software (version 17.0; StataCorp). All end points were analyzed with linear mixed effects models, including the factors Time (as a categorical variable), Group, and the interaction term Time times Group controlled for sex and obesity class (I, II, and III) using z statistics with an unstructured covariance matrix. The change from baseline to end of treatment (week 8) — or the latest on-treatment time point available — was the key end point. For resting energy expenditure, lean mass was added as a covariate of no interest. Treatment efficacy was determined using a modified intention-to-treat approach including all randomly assigned participants who completed at least one visit while on the trial drug. For the primary end point body weight, a linear regression analysis with the week 8 body weight measurement as outcome and group, the baseline body weight measurement, sex, and obesity class as the predictors was also performed to corroborate the linear mixed effects model analysis. Additionally, a sensitivity analysis and an analysis of sustained treatment effects at the 14-week off-treatment follow-up were performed. Accuracy of treatment allocation guess was compared between oxytocin and placebo groups using Fisher’s exact test. The analyses of secondary outcomes did not include a provision for correction for multiplicity. Therefore, the results are reported as point estimates and unadjusted confidence intervals (CIs) and should not be used to infer treatment effects.

Results

PARTICIPANTS

Between October 2017 and March 2022, 61 participants were randomly assigned to an 8-week course of intranasal oxytocin (n=31) or placebo (n=30) and received at least one dose of trial drug (safety cohort). The mean age was 33.6 years of age (SD, 6.2 years of age), 54% were female. The mean BMI was 36.9 (SD, 4.9). Eleven participants (18.0%) had an HbA1c level in the prediabetes range (≥5.7 to <6.5%), and one (1.6%) had an HbA1c level of 6.7%, consistent with diabetes (≥6.5%). One participant took metformin, seven received antihypertensive therapy, and four took cholesterol-modifying medications. Age, sex, ethnicity, clinical characteristics, and fasting plasma oxytocin levels were similar between treatment groups (Table 1; expanded data are presented in Table S1); the oxytocin group had a greater proportion of Black/African American and Asian participants than the placebo group. Overall, the diversity of the trial cohort was representative of the studied population, namely, adults with obesity (Table S2).

Table 1.

Demographic and Clinical Characteristics of the Participants at Baseline.^*

Characteristic	Oxytocin (n=31)	Placebo (n=30)
Age — yr	34.0 ± 6.1	33.2 ± 6.4
Sex — no. (%)
Male	14 (45.2)	14 (46.7)
Female	17 (54.8)	16 (53.3)
Race — no. (%)†
Asian	4 (12.9)	2 (6.7)
Black/African American	10 (32.3)	3 (10.0)
White	15 (48.4)	23 (76.7)
Other/more than one	2 (6.5)	1 (3.3)
Asked but unknown	0 (0.0)	1 (3.3)
Hispanic/Latino ethnic group — no. (%)†	6 (19.4)	7 (23.3)
Body weight — kg	108.9 ± 18.5	109.7 ± 20.1
Body-mass index‡	36.9 ± 4.0	37.0 ± 5.7
Waist circumference — cm	122.6 ± 8.4	123.4 ± 10.1
Waist-to-hip ratio	0.94 ± 0.06	0.94 ± 0.09
Total fat mass — g§	44,740 ± 10,587	45,258 ± 11,629
Total lean mass — g§	63,809 ± 13,773	64,219 ± 13,088
Abdominal visceral adipose tissue — cm2¶	144.7 ± 80.6	162.8 ± 92.7
Liver fat fraction||	0.09 ± 0.09	0.11 ± 0.12
Resting energy expenditure — kcal/day**	1,539 ± 273.7	1,582 ± 277.5
Glucose — mg/dl††	93.7±11.7	85.9 ± 15.2
Insulin — μIU/ml‡‡	11.4 ± 10.0	11.5 ± 6.6
HOMA-IR§§	2.5 ± 1.7	2.4 ± 1.5
Glycated hemoglobin — %	5.5 ± 0.42	5.4 ± 0.35
Lipid profile — mg/dl
Total cholesterol¶¶	177.0 ± 28.9	177.8 ± 31.9
HDL cholesterol	45.7 ± 7.6	44.3 ± 10.1
LDL cholesterol||||	111.0 ± 24.9	113.0 ± 27.3
Triglycerides	104.6 ± 70.8	101.5 ± 44.0
Liver enzymes — U/l
ALT***	25.1 ± 15.8	31.3 ± 27.8
AST†††	21.2 ± 7.3	25.0 ± 16.3
High-sensitivity C-reactive protein — mg/l‡‡‡	4.1 ± 2.4	3.1 ± 2.0
Plasma oxytocin — pg/ml§§§	361.8 ± 233.5	389.2 ± 268.5
Medications — no. (%)
Metformin	1 (3.2)	0 (0.00)
Antihypertensive	4 (12.9)	3 (10.0)
Cholesterol modifying	1 (3.2)	3 (10.0)

^*Plus-minus values are means ±SD. Shown are the demographic and clinical characteristics for the safety cohort, which was defined as all randomly assigned participants who received at least one dose of trial drug. Baseline was defined as the most recent nonmissing measurement before the first dose of trial drug. HDL denotes high-density lipoprotein; and SD, standard deviation.

^†Race and ethnic group were reported by the participant.

^‡The body-mass index is the weight in kilograms divided by the square of the height in meters.

^§Dual-energy x-ray absorptiometry (Hologic Horizon A; Hologic Inc.) was used for analysis.

^¶The 3T magnetic resonance imaging (Siemens Trio; Siemens Medical Systems) was performed using VITRAK software (Merge; eFilm).

^||Breath-hold single-voxel hydrogen 1 magnetic resonance spectroscopy using LCModel software (version 6.3–0K/S; Provencher) was used for analysis. Liver fat fraction ranges from zero to one, with higher values representing a higher proportion of fat, and it is reported for 28 participants in the oxytocin group and 29 participants in the placebo group.

^**Indirect calorimetry (VMAX Encore 29 metabolic cart; Viasys Healthcare/Carefusion) was used for analysis.

^††Glucose levels are reported for 22 participants in the oxytocin group and 22 participants in the placebo group.

^‡‡Insulin levels are reported for 25 participants in the oxytocin group and 24 participants in the placebo group.

^§§Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) is calculated as glucose times insulin divided by 405, and it is reported for 18 participants in the oxytocin group and 17 participants in the placebo group.

^¶¶Total cholesterol is reported for 30 participants in the oxytocin group and 29 participants in the placebo group.

^||||Low-density lipoprotein (LDL) cholesterol is reported for 30 participants in the oxytocin group and 29 participants in the placebo group.

^***Alanine transaminase (ALT) levels are reported for 30 participants in the oxytocin group and 30 participants in the placebo group.

^†††Aspartate transaminase (AST) levels are reported for 27 participants in the oxytocin group and 29 participants in the placebo group.

^‡‡‡High-sensitivity C-reactive protein levels are reported for 25 participants in the oxytocin group and 26 participants in the placebo group.

^§§§Plasma oxytocin levels are reported for 31 participants in the oxytocin group and 29 participants in the placebo group.

Sixty participants (oxytocin group: n=31, placebo group: n=29) completed at least one visit while on the trial drug (modified intention-to-treat cohort). Fifty-one participants (oxytocin group: n=29, placebo group: n=22) completed the 8-week treatment period and the end-of-treatment assessment. Fifty participants (oxytocin group: n=28, placebo group: n=22) completed the 14-week off-treatment follow-up visit (Fig. 1). Trial drug–related dis-continuations (n=4) were because of reports from participants regarding an odor associated with the trial drug. Four participants, who were on the trial drug at the time, were instructed to discontinue the trial drug to allow for an investigation. Following the investigation, revealed no concerns, the trial resumed.

Figure 1. CONSORT Diagram.

For the modified intention-to-treat cohort, adherence on the basis of the trial drug diary was 93.1% (95% CI, 90.2 to 96.0) in the oxytocin group and 92.0% (95% CI, 88.2 to 95.8) in the placebo group. Two participants in the oxytocin group and four participants in the placebo group reported less than 80% adherence. Change in amount of exercise from baseline to week 8 did not differ between groups (oxytocin group: 0.64 hours/week, placebo group: −0.17 hours/week; difference, 0.81 hours/week; 95% CI, −0.67 to 2.3), consistent with instructions to abstain from lifestyle changes.

BODY WEIGHT AND KEY SECONDARY END POINTS

Eight weeks of treatment with intranasal oxytocin (24 IU four times daily) did not impact body weight. Body weight change from baseline to week 8 was 0.20 kg in the oxytocin group and 0.26 kg with placebo (difference, −0.06 kg; 95% CI, −1.4 to 1.3; P=0.934) (Fig. 2). Oxytocin (vs. placebo) was not associated with beneficial effects on body composition or resting energy expenditure (Fig. 3A to 3D and Table 2; absolute values of all primary and secondary end points as well as key safety assessments by group at baseline and the end of treatment are in Table S3). The oxytocin group was associated with reduced caloric intake at the Breakfast Test Meal from baseline to week 6 compared with the placebo group, with participants in the oxytocin group consuming an average of 1146 kcal at baseline and 1114 kcal at week 6 and participants in the placebo group consuming an average of 1076 kcal at baseline and 1197 kcal at week 6 (−31.4 vs. 120.6 kcal; difference, −152.0 kcal; 95% CI, −302.3 to −1.7) (Fig. 3E).

Figure 2. Effects of Intranasal Oxytocin (24IU Four Times Daily) versus Placebo on the Primary End Point Body Weight.

Displayed are the individual changes in body weight (kilograms) from baseline to week 8 among all participants of the modified intention-to-treat cohort with available data at both time points (oxytocin group: n=29, placebo group: n=22). Note that change in body weight had a value of zero for one participant of the control group and that for this participant, the assigned space does not show a bar. There was no difference in body weight change from baseline to week 8 between oxytocin and placebo groups (0.20 vs. 0.26 kg; difference, −0.06 kg; 95% confidence interval, −1.4 to 1.3; P=0.934).

Figure 3. Effects of Intranasal Oxytocin (24 IU Four Times Daily) versus Placebo on the Key Secondary End Points.

Displayed are the individual changes in total fat mass (Panel A), abdominal visceral adipose tissue (Panel B), liver fat fraction (Panel C), and resting energy expenditure (Panel D) from baseline to week 8 and caloric intake at an experimental test meal (Panel E) from baseline to week 6 among all participants of the modified intention-to-treat cohort with available data at both time points (total fat mass and resting energy expenditure: oxytocin group: n=29, placebo group: n=22; abdominal visceral adipose tissue: oxytocin group: n=28, placebo group: n=19; liver fat fraction: oxytocin group: n=25, placebo group: n=19; caloric intake at an experimental test meal: oxytocin group: n=29, placebo group: n=24). Note that change in caloric intake at an experimental test meal had a value of zero for one participant of the control group and that for this participant, the assigned space does not show a bar. Oxytocin (vs. placebo) was not associated with beneficial effects on body composition or resting energy expenditure from baseline to week 8 (total fat: difference, 196.0 g; 95% confidence interval [CI], −1036 to 1428; abdominal visceral adipose tissue: difference, 3.1 cm²; 95% CI, −11.0 to 17.2; liver fat fraction: difference, −0.01; 95% CI, −0.03 to 0.01; resting energy expenditure: difference, −64.0 kcal/day; 95% CI, −129.3 to 1.4). Oxytocin compared with placebo was associated with reduced caloric intake at the experimental test meal from baseline to week 6 (−31.4 vs. 120.6 kcal; difference, −152.0 kcal; 95% CI, −302.3 to −1.7). The analyses of the secondary outcomes did not include a provision for correction for multiplicity. Therefore, the results should not be used to infer treatment effects.

Table 2.

Change from Baseline to End of Treatment (Week 8) in Primary, Key Secondary, and Additional Secondary End Points as well as Key Safety Assessments.^*

End Point	Oxytocin (n=31)	Placebo (n=29)	Difference
Primary end point: body weight — kg	0.20	0.26	−0.06 (−1.4 to 1.3), P=0.934
Key secondary end points
Body composition
Total fat mass — g†	248.5	52.5	196.0 (−1036 to 1428)
Abdominal visceral adipose tissue — cm2‡	2.2	−0.90	3.1 (−11.0 to 17.2)
Liver fat fraction§	−0.01	0.0004	−0.01 (−0.03 to 0.01)
Resting energy expenditure — kcal/day¶	−52.0	12.0	−64.0 (−129.3 to 1.4)
Caloric intake at Breakfast Test Meal — kcal||	−31.4	120.6	−152.0 (−302.3 to -1.7)
Additional secondary end points
Body composition: total lean mass — g†	−677.1	−469.3	−207.8 (−1242 to 826.8)
Self-reported daily caloric intake — kcal**	−149.3	91.5	−240.8 (−582.8 to 101.3)
Metabolic profile
Glucose — mg/dl	−2.6	2.1	−4.7 (−13.2 to 3.7)
Insulin — μIU/ml††	−0.33	0.41	−0.75 (−3.7 to 2.2)
HOMA-IR‡‡	−0.27	0.36	−0.63 (−1.5 to 0.25)
Total cholesterol — mg/dl	−6.4	−2.8	−3.6 (−14.2 to 7.0)
HDL cholesterol — mg/dl	−0.36	3.2	−3.5 (−6.8 to -0.32)
LDL cholesterol — mg/dl	−5.9	−6.3	0.39 (−8.1 to 8.8)
Triglycerides — mg/dl	−4.8	6.1	−10.9 (−32.7 to 10.9)
High-sensitivity C-reactive protein — mg/l§§	0.17	0.03	0.14 (−0.66 to 0.93)
Psychological, behavioral, and quality-of-life assessments
BDI-II Total Score¶¶	−1.1	0.17	−1.3 (−3.8 to 1.2)
STAI Trait Score||||	−0.83	−0.15	−0.68 (−3.3 to 2.0)
EDE-Q Restraint Score***	−0.60	−0.51	−0.09 (−0.74 to 0.56)
EDE-Q Eating Concern Score***	−0.19	−0.37	0.18 (−0.31 to 0.68)
EDE-Q Shape Concern Score***	−0.32	−0.37	0.05 (−0.40 to 0.50)
EDE-Q Weight Concern Score***	−0.20	−0.27	0.07 (−0.37 to 0.50)
EDE-Q Global Score***	−0.33	−0.40	0.07 (−0.31 to 0.46)
Caloric intake at Cookie Taste Test — kcal†††	19.4	7.5	11.8 (−54.7 to 78.3)
SF-36 Physical Component Summary‡‡‡	−0.02	0.75	−0.77 (−3.5 to 1.9)
SF-36 Mental Component Summary‡‡‡	1.7	−1.5	3.2 (0.25 to 6.1)
Plasma oxytocin — pg/ml	33.6	24.1	9.5 (−53.4 to 72.5)
Key safety assessments
Vital signs
Blood pressure — mm Hg
Systolic	4.5	2.9	1.6 (−4.5 to 7.6)
Diastolic	−2.1	−3.2	1.1 (−3.5 to 5.6)
Pulse — beats/minute	−0.42	−2.5	2.1 (−3.3 to 7.5)
EKG QTc interval — ms§§§	−2.2	−3.1	0.94 (−10.0 to 11.9)

^*Valúes are estimated mean changes for oxytocin and placebo groups and differences in change scores (95% confidence intervals) between the treatment groups from linear mixed effects models with the factors Time (baseline and all available on-treatment time points for each respective measure), Group (oxytocin/placebo), and the interaction term Time times Group controlled for sex and obesity class (I, II, and III). Shown are the changes in primary, key secondary, and additional secondary end points as well as key safety assessments for the sample included in the modified intention-to-treat analysis, which includes all randomly assigned participants who completed at least one visit while on the trial drug. Baseline was defined as the most recent nonmissing measurement before the first dose of the trial drug. The analyses of the secondary outcomes did not include a provision for correction for multiplicity. Therefore, the results should not be used to infer treatment effects. HDL denotes high-density lipoprotein; and LDL, low-density lipoprotein.

^†Dual-energy x-ray absorptiometry (Hologic Horizon A; Hologic Inc.) was used for analysis.

^‡The 3T magnetic resonance imaging (Siemens Trio; Siemens Medical Systems) was performed using VITRAK software (Merge; eFilm).

^§Breath-hold single-voxel hydrogen 1 magnetic resonance spectroscopy using LCModel software (version 6.3–0K/S; Provencher) was used for analysis. Liver fat fraction ranges from zero to one, with higher values representing a higher proportion of fat.

^¶Indirect calorimetry (VMAX Encore 29 metabolic cart; Viasys Healthcare/Carefusion) was used for analysis and statistically adjusted for lean mass.

^||Energy intake at the Breakfast Test Meal was assessed as change from baseline to week 6.

^**Daily caloric intake was on the basis of the Four-Day Food Record. At baseline, the Four-Day Food Record was obtained twice, and data were averaged across all available 8 days.

^††Insulin levels are reported on the basis of the linear mixed effects model with 31 participants in the oxytocin group and 28 participants in the placebo group.

^‡‡Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) is calculated as glucose times insulin divided by 405 and reported on the basis of the linear mixed effects model with 29 participants in the oxytocin group and 26 participants in the placebo group.

^§§High-sensitivity C-reactive protein levels are reported on the basis of the linear mixed effects model with 27 participants in the oxytocin group and 27 participants in the placebo group.

^¶¶The Beck Depression Inventory II (BDI-II) Total Score ranges from 0 to 63, with higher scores indicating more severe depressive symptomatology. Scores are reported on the basis of the linear mixed effects model with 28 participants in the oxytocin group and 24 participants in the placebo group.

^||||The State-Trait Anxiety Inventory (STAI) Trait Score ranges from 20 to 80, with higher scores indicating higher trait anxiety.

^***Scores on the Eating Disorder Examination Questionnaire (EDE-Q) range from zero to six, with higher scores indicating more pronounced psychopathology, and they were assessed as change from baseline to week 4.

^†††Caloric intake at the Cookie Taste Test was obtained as an assessment of hedonic drive to eat, with a higher caloric intake indicating more hedonic eating as change from baseline to week 4, and it is reported on the basis of the linear mixed effects model with 30 participants in the oxytocin group and 29 participants in the placebo group.

^‡‡‡Scores on the 36-Item Short-Form Health Survey (SF-36) are norm based and transformed to a scale on which the 2009 general population of the United States has a mean score of 50 and a standard deviation of 10. Higher scores indicate better quality of life.

^§§§From electrocardiogram (EKG) data, the corrected QT (QTc) interval was calculated using the Fridericia formula.

ADDITIONAL SECONDARY END POINTS AND KEY SAFETY ASSESSMENTS

A slight reduction in high-density lipoprotein (HDL) cholesterol from baseline to week 8 was observed in the oxytocin group, whereas an increase in HDL cholesterol was observed in the placebo group (−0.36 vs. 3.2 mg/dl; difference, −3.5 mg/dl; 95% CI, −6.8 to −0.32). The oxytocin group was associated with a greater change in mental health–related quality of life from baseline to week 8 versus placebo (SF-36 Mental Component Summary: 1.7 vs. −1.5; difference, 3.2; 95% CI, 0.25 to 6.1). There were no other group differences in secondary or safety outcomes from baseline to week 8 or latest available on-treatment time point, including no group differences in changes in plasma oxytocin levels from baseline to weeks 4, 6, and 8 (Table 2; a complete report of the effects of oxytocin vs. placebo from baseline to all on-treatment assessment time points within the modified intention-to-treat cohort together with reports of main effects of Time and Group is in Table S4). Sensitivity and off-treatment follow-up analyses are included in Supplementary Results and Tables S5 and S6.

ADVERSE EVENTS

There were no serious adverse events. Ten moderate adverse events were documented in seven participants, one of which occurred in the oxytocin group (headache), whereas the other nine were in the placebo group (Supplementary Results). All other adverse events were mild. Other than nasal discomfort reported by six participants in the oxytocin group (19.4%) versus two participants in the placebo group (6.7%), the incidence of adverse events was similar across treatment groups (Table 3; an expanded report of adverse events is in Table S7).

Table 3.

Adverse Events.^*

System Organ Class	Oxytocin (n=31)	Placebo (n=30)
Cardiac disorders	0 (0.0%)	1 (3.3%)
Ear and labyrinth disorders	0 (0.0%)	1 (3.3%)
Gastrointestinal disorders	5 (16.1%)	8 (26.7%)
General disorders and administration site conditions	1 (3.2%)	1 (3.3%)
Infections and infestations	2 (6.5%)	3 (10.0%)
Injury, poisoning, and procedural complications	3 (9.7%)	4 (13.3%)
Investigations	6 (19.4%)	7 (23.3%)
Metabolism and nutrition disorders	1 (3.2%)	3 (10.0%)
Musculoskeletal and connective tissue disorders	3 (9.7%)	2 (6.7%)
Nervous system disorders	7 (22.6%)	8 (26.7%)
Psychiatric disorders	2 (6.5%)	3 (10.0%)
Renal and urinary disorders	2 (6.5%)	0 (0.0%)
Reproductive system and breast disorders	2 (6.5%)	2 (6.7%)
Respiratory, thoracic, and mediastinal disorders	11 (35.5%)	6 (20.0%)
Skin and subcutaneous tissue disorders	5 (16.1%)	5 (16.7%)
Vascular disorders	2 (6.5%)	2 (6.7%)

^*Values are numbers of participants (percentages). Adverse events are reported for the safety cohort, which was defined as all randomly assigned participants who received at least one dose of trial drug.

TREATMENT ALLOCATION GUESS

Of the assessed 28 participants in the oxytocin group, 3 correctly guessed allocation to oxytocin, 19 incorrectly guessed allocation to placebo, and 6 indicated that they were not sure. Of the assessed 21 participants in the placebo group, 4 incorrectly guessed allocation to oxytocin, 11 correctly guessed allocation to placebo, and 6 answered that they were not sure.

Discussion

In this RCT, 8-week intranasal oxytocin in adults with obesity did not impact body weight. Compared with placebo, oxytocin was associated with reduced caloric intake at an experimental test meal but not with beneficial changes in body composition, resting energy expenditure, self-reported food intake, or metabolic parameters. Adverse events were generally mild.

We selected the dosing regimen on the basis of previous investigations in humans showing that 24 IU intranasal oxytocin modulated neural activity of brain circuitry governing eating behavior^13,14,16 and reduced experimental food intake in fasted and fed states^15,17,25 as well as a small RCT of 24 adults with obesity resulting in 8.9-kg weight loss using the same regimen as the one used in this trial.¹⁸ We did not replicate the findings of the prior pilot study, which was limited by substantial differences in baseline characteristics, namely, older age and lower BMI in the placebo group, and inclusion of only nine individuals who received oxytocin. Similar to our trial, a crossover RCT of 16 to 24 IU intranasal oxytocin three times daily for 8 weeks in 10 individuals with tumor-induced hypothalamic obesity showed no effect on body weight.²⁶ Although Zhang et al.¹⁸ reported a decrease in total and low-density lipoprotein cholesterol levels, we did not detect changes in lipid profile with oxytocin other than a small decrease in HDL cholesterol levels, which is unlikely to be clinically meaningful. Unlike Espinoza et al.,²⁷ who showed an oxytocin-induced increase in lean mass without affecting body weight in 21 older patients with sarcopenic obesity, we did not see effects of oxytocin on body composition. This could be because of differences in the trial population, with a preferential effect of oxytocin on lean mass in individuals who have sarcopenia and/or are older given lower oxytocin levels with advanced age.²⁸

There are several possible explanations for the observed lack of oxytocin effects. The weight loss effects of oxytocin consistently demonstrated in animal models may not translate to humans. Alternatively, different doses of oxytocin may be necessary to achieve weight loss.²⁹ In rodents, continuous intracerebroventricular infusion of oxytocin results in downregulation of oxytocin receptors³⁰; however, both continuous^11,12 oxytocin delivery and intermittent³¹ oxytocin delivery promote weight loss. Our dosing regimen provided intermittent pulses of oxytocin with a half-life of several minutes in the periphery³² and 20minutes in the central nervous system,³³ and it might be expected to avoid receptor downregulation. Although even single-dose daily oxytocin results in weight loss in animal studies,³⁴ a longer-acting oxytocin receptor agonist might be more effective in humans. We did not observe an effect of the treatment regimen on fasting plasma oxytocin levels and thus, found no evidence for repetitive exogenous oxytocin administration counteracting metabolic effects by suppressing endogenous oxytocin secretion.

Although the acute anorexigenic effects of oxytocin are well established,⁴ there is evidence that effects may wane with time in animal models. However, in our trial, oxytocin was associated with a reduction in caloric intake at an experimental test meal after 6 weeks of treatment (vs. an increase in caloric intake with placebo). We withheld trial drug administration before the test meal to examine chronic effects of oxytocin administration and might have observed greater suppression of caloric intake if participants had administered oxytocin before the meal. In contrast, compared with placebo, oxytocin was not associated with a difference in self-reported daily food intake, which may be limited by misreporting errors with interindividual variability.³⁵ Participants were instructed to maintain their current diet, which could have limited the opportunity to detect changes in daily food consumption. Finally, oxytocin has been shown to promote cognitive control over behavioral impulses in adults with obesity,³⁶ and it is possible that oxytocin is most effective at reducing caloric intake when behavioral self-control is required as in the experimental test meal when participants were given double portions of their chosen meal, requiring them to suppress habit-driven and/or hedonic eating beyond satiety. Of note, in this trial, oxytocin was not associated with changes in hedonic drive to eat as measured by the Cookie Taste Test. However, this assessment is limited to offering a specific food and might not be sensitive enough to capture changes in overall hedonic drive. Finally, the data showed a numerical decrease in resting energy expenditure in the oxytocin group relative to the placebo group (although the 95% CI around the estimate was wide and included zero) that might warrant further investigation.

Consistent with other studies,^18,27 our safety data were reassuring, with no serious adverse events and no differences in adverse events between participants receiving oxytocin and placebo other than more reports of nasal irritation with oxytocin. Adherence was excellent. A majority of participants were unable to correctly indicate their treatment allocation. In this relatively small sample, oxytocin did not cause hyponatremia, vital sign changes, or a systematically prolonged QTc interval. We observed greater change in mental health–related quality of life in the oxytocin group compared with placebo, which may indicate subjectively perceived benefits when averaged across mental health domains, although a positive finding in one of multiple secondary outcomes should be considered only hypothesis-generating.

Strengths of our trial include its randomized, controlled design, which was well powered to investigate chronic effects of intranasal oxytocin on body weight, eating behavior, and metabolism in humans with obesity using the same regimen reported to reduce weight in a previous small study.¹⁸ Our methodology was rigorous. We recruited a diverse trial cohort. Our trial included comprehensive evaluation to provide safety data. We did not include recommendations for healthy nutrition and physical activity and asked participants to maintain their lifestyle during this brief intervention to isolate the mechanistic effects of oxytocin administration on eating behavior and metabolism. Antiobesity pharmacotherapy has been evaluated as an adjuvant therapy to lifestyle modification, with a more potent effect when exercise is actively combined with medications compared with medications alone.³⁷ Oxytocin treatment could potentially augment adherence to behavioral interventions because it increases self-control in individuals with obesity.³⁶ Only one participant had type 2 diabetes, a condition that may be more responsive to oxytocin treatment because of associated loss of oxytocinergic neurons in the hypothalamic paraventricular nucleus³⁸ and lower endogenous oxytocin levels.³⁹ The trial duration was relatively short, and it is possible that a longer course of oxytocin treatment is needed for the effects on caloric intake seen at the experimental test meal to translate to weight loss.

A recent study suggests that administering oxytocin in combination with supraphysiologic doses of magnesium enhances oxytocin effects by potentiating oxytocin–oxytocin receptor binding efficacy.⁴⁰ Modified synthetic oxytocin may improve ligand-receptor specificity, in vivo stability, and potency.⁴¹ Given that our trial did not demonstrate weight loss with intranasal oxytocin, future investigations could use different dosing regimens, optimized formulations, and/or selective longer-acting agonists to determine whether oxytocin-induced weight loss seen in animal studies translates to humans.

Supported by the National Institute of Diabetes and Digestive and Kidney Diseases (grant numbers R01DK1009932 and P30DK040561), the National Institute of Mental Health (grant number K24MH120568), and the National Center for Advancing Translational Sciences (grant number UL1TR001102).