Clinical symptoms and cognitive functioning following adjunctive N-acetylcysteine in obsessive-compulsive disorder: a double-blind, placebo-controlled trial

Abstract

Background

Obsessive-compulsive disorder (OCD) often remains partially responsive to first-line treatments such as selective serotonin reuptake inhibitors (SSRI). N-acetylcysteine (NAC), a glutamate-modulating antioxidant, has shown promise as an adjunctive therapy, but prior studies frequently lacked standardized pharmacotherapy and cognitive outcome measures, particularly in Iranian populations.

Methods

In this 12-week, double-blind, placebo-controlled randomized clinical trial (IRCT20170123032145N7, registered on 2023-06-08), 35 adults with moderate to severe OCD were assigned to receive sertraline (200 mg/day) augmented with either NAC (2400 mg/day) or placebo. Clinical symptoms were assessed using the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) at baseline, weeks 4, 8, and 12. Cognitive functioning was evaluated before and after treatment using the Wisconsin Card Sorting Test and the D2 Test of Attention.

Results

This pilot study reveals a significant Time × Group interaction for Y-BOCS Total (p < 0.005), with the NAC group showing greater symptom reduction than placebo. The Obsession subscale demonstrates the strongest effect (p < 0.001), while Compulsion subscale improvements are comparable between groups. Clinical response rate ( ≥ 35% Y-BOCS reduction) is significantly higher in the NAC group (100%) versus placebo (92.6%, p = 0.004). Both groups improve on cognitive measures (processing speed and attention) over time, with no significant between-group differences. Adverse events are mild and comparable.

Conclusions

Adjunctive NAC may enhance early clinical response in OCD, particularly for obsessions, when added to high-dose SSRI (Sertraline) treatment. However, no cognitive advantage is observed within the 12-week period. These preliminary findings support NAC’s safety and potential utility as a short-term adjunctive treatment in OCD.

Plain Language Summary

Obsessive-compulsive disorder (OCD) causes unwanted thoughts and repetitive behaviors that interfere with daily life. Standard medications, such as sertraline, help some patients, but many continue to struggle. This study tests whether adding N-acetylcysteine (NAC), a natural dietary supplement, to standard medication could improve outcomes. We randomly assigned 35 adults with moderate to severe OCD to receive sertraline plus NAC or sertraline plus placebo for 12 weeks. People receiving NAC experienced faster relief from obsessive thoughts compared to those taking placebo, with 100% of NAC patients achieving meaningful improvement versus 92.6% in the placebo group. NAC appears safe and may accelerate symptom relief when combined with standard medication, offering patients a potential low-risk option for faster improvement in their symptoms.

Askari, Eghdami, et al. conduct a 12-week double-blind randomized trial testing adjunctive N-acetylcysteine versus placebo added to fixed-dose sertraline in adults with moderate to severe OCD. Adjunctive N-acetylcysteine yields faster and greater obsession symptom reduction than placebo, and is well tolerated, without adding cognitive benefit.

Introduction

Obsessive-compulsive disorder (OCD) is a chronic and debilitating psychiatric condition that affects an estimated 1.3% to 3% of the global population over a lifetime¹. It is characterized by the presence of obsessions, compulsions, or both resulting in persistent, intrusive thoughts, urges, or images and repetitive behaviors or mental acts performed to cope with the distress. These symptoms often center around themes such as contamination, harm, symmetry, taboos, or religious matters, which significantly interfere with daily functioning and quality of life². According to the Global Burden of Disease (GBD) Study, OCD ranks among the top ten leading causes of non-fatal disability (years lived with disability, YLDs) in many high-income countries, contributing to ~0.5–1.3% of total YLDs globally and an estimated 3 million disability-adjusted life years (DALYs) worldwide³.

The pathophysiology of OCD involves a multifaceted interplay of genetic, neurobiological, and environmental factors⁴. Traditional neurochemical models have emphasized the role of serotonergic, dopaminergic, noradrenergic, and cholinergic systems, leading to for first-line treatments such as selective serotonin reuptake inhibitors (SSRIs)⁵, clomipramine⁶, and cognitive behavioral therapy (CBT)⁷, particularly with exposure and response prevention (ERP)⁸. Sertraline is one of several first-line, FDA-approved SSRIs for the treatment of OCD in adults⁹. However, it is not the only drug with a formal indication for OCD; other SSRIs such as fluoxetine, fluvoxamine and paroxetine¹⁰. Current guidelines therefore recommend SSRIs (including sertraline) or clomipramine as first‑line pharmacotherapy, with the specific agent chosen largely on the basis of tolerability, comorbidities and patient preference rather than clear superiority in anti‑obsessional efficacy¹⁰.

Despite these established therapies, 40–60% of patients fail to achieve sufficient symptom relief or experience limiting side effects, highlighting the need for alternative or adjunctive treatment strategies¹¹.

In addition to the core symptoms of obsessions and compulsions, individuals with OCD often display persistent cognitive deficits¹². Especially in domains of executive functioning, such as set-shifting, response inhibition, and attention control¹³. Prior reviews indicated that these impairments are not only common but also trait-like, suggesting that they may reflect underlying neurobiological vulnerabilities rather than being exclusively secondary to symptom severity¹⁴. Importantly, these cognitive deficits are linked to poorer functional outcomes and may persist even after symptomatic improvement with SSRIs¹⁵. Despite their clinical relevance, cognitive impairments in OCD have received limited attention in treatment trials.

In recent years, glutamate dysregulation has emerged as a critical mechanism in OCD, particularly within the cortico-striato-thalamo-cortical (CSTC) circuits¹⁶. Elevated glutamatergic activity and impaired glutamate clearance have been observed in neuroimaging and genetic studies, supporting the exploration of glutamate-modulating agents such as N-acetylcysteine (NAC)¹⁷. NAC, a cysteine prodrug with antioxidant and neuroprotective properties, modulates extracellular glutamate levels via the cystine–glutamate antiporter and helps restoring redox balance, thereby targeting both glutamatergic dysfunction and oxidative stress^18,19. Neuroinflammatory pathways, evident by increased pro-inflammatory cytokines and microglial activation also contribute to CSTC circuit abnormalities and may be modulated by NAC²⁰.

Preclinical studies have shown that NAC reduces compulsive and impulsive behaviors in animal models, including excessive grooming and perseverative responses²¹. These findings parallel core OCD symptom dimensions and are accompanied by improvements in inhibitory control and reduced oxidative markers²². Early clinical trials in humans with OCD and related disorders (e.g., trichotillomania, excoriation disorder, and substance use disorders) suggest that NAC augmentation can reduce symptom severity, with a favorable safety and tolerability profile²³. Randomized controlled trials (RCTs) of NAC in OCD and related conditions have shown mixed but encouraging results¹⁸. Some demonstrating significant symptom improvement, others reporting more modest or nonsignificant effects, potentially due to variability in study design, dosage, or patient characteristics.

Beyond its role in symptom reduction, NAC has shown potential cognitive benefits²⁴. Preliminary evidence suggests it may enhance executive functioning, reduce cognitive rigidity, and improve attention²⁵, all of which are frequently impaired in individuals with OCD. These multidimensional effects, likely mediated through glutamatergic modulation and antioxidative action, position NAC as a promising adjunctive treatment with both symptomatic and cognitive therapeutic potential in OCD²⁶.

Despite the growing interest in NAC as an adjunctive treatment for OCD, prior studies often suffer from key methodological limitations. Many lack cognitive outcome measures, making it difficult to assess the broader neuropsychological impact of NAC. In addition, heterogeneity in SSRI regimens across studies complicates the interpretation of NAC’s additive effects¹⁸. The absence of standardized pharmacotherapy, along with limited use of rigorous methodologies (such as double-blind, placebo-controlled designs) further limits the generalizability of the findings. Moreover, further trials are needed to evaluate the effectiveness of treatment across diverse genetic and cultural backgrounds.

This pilot study reveals that adjunctive NAC enhances early clinical response in obsessive-compulsive disorder, particularly for obsessive symptoms, when added to high-dose sertraline. The findings demonstrate clinical response rates of 100% in the NAC group versus 92.6% in the placebo group, with maximal symptom separation evident at week eight. However, no cognitive advantages emerge within the 12-week treatment period. These results support NAC’s safety profile and suggest potential utility as a short-term adjunctive treatment in OCD, though the magnitude of benefit and its specific mechanisms warrant further investigation in larger, longer-duration trials.

Method

Participant selection and study setting

Participants were recruited between Sept2023-Dec2024 from outpatient psychiatric clinics affiliated with Iran University of Medical Sciences, Tehran, Iran. Eligibility was determined using the Structured Clinical Interview for DSM-5 (SCID-5)²⁷, conducted by trained psychiatrists or supervised psychiatry residents.

Participants were included if they were between 18 and 60 years old, had a primary diagnosis of OCD based on DSM-5 criteria, exhibited a Yale-Brown Obsessive Compulsive Scale (Y-BOCS) score of ≥21²⁸, had not received psychiatric medication or psychotherapy for at least six weeks prior to enrollment, and were able to provide written informed consent.

Participants with comorbid psychiatric disorders (except for mild depressive or anxiety symptoms not meeting diagnostic criteria), a history of neurological disorders, substance use disorders except for cigarette smoking, acute suicidal ideation, or serious medical conditions such as cardiac, renal, or hepatic disease in addition to individuals with intellectual disability, significant cognitive impairment, or current use of psychoactive substances were excluded.

Adverse events and dropout rates were tracked throughout the trial systematically using a brief self-designed checklist based on common treatment-related side effects; participants who developed moderate to severe side effects or experienced significant adverse medical reactions during the trial were to be withdrawn and provided with standard clinical care which none took place throughout the trial.

Study design and randomization

This study employed a randomized, double-blind, placebo-controlled clinical trial design with fixed dosing. Participants were randomly assigned to either the experimental condition or control condition group using block randomization (block size = 4), stratified by study site. Randomization sequences were computer-generated, and allocation concealment was maintained using sequentially numbered, opaque, sealed envelopes. Participants and assessors were blinded to group assignments. No patient or public involvement occurred in the design, conduct, or reporting of the trial.

Both groups received a standardized, fixed-dose regimen of sertraline to minimize confounding effects from variable dosing. Sertraline was initiated at 25 mg/day in week one, increased to 50 mg/day in week two, 75 mg/day in week three, 100 mg/day in week four, 150 mg/day in week five, and reached the target fixed dose of 200 mg/day by week six. This dose was maintained at 200 mg/day for the remainder of the 12-week trial. No individualized dose adjustments were allowed; All participants who complete the trial achieved and maintained the 200 mg/day target dose. Adherence to the sertraline titration schedule was monitored at each visit. No concurrent psychological treatments or additional psychiatric interventions were allowed during the study period. NAC was initiated as adjunctive therapy at 1200 mg every 12 h from week one and sustained throughout the trial, with medications taken daily at home.

Assessments and outcome measures

The first outcome measure was the change in the Y-BOCS Total Score from baseline to Week 12, and the secondary clinical outcomes included changes in the Y-BOCS Obsession and Compulsion subscales to evaluate specific symptom domains. Cognitive endpoints included changes in executive functioning and attention from baseline to Week 12.

Assessments were conducted at strict time intervals: baseline, week 4, week 8, and week 12 for symptom evaluation; and two timepoints for cognitive assessments: baseline, week 12. To minimize heterogeneity, all follow-up visits were scheduled within a window of ±3 days of the target date; and all evaluations were conducted by trained assessors who were blinded to treatment allocation. Assessors underwent standardized training, and reliability among them is established prior to data collection. Any diagnostic discrepancies were resolved by consulting with a senior faculty psychiatrist.

Clinical and cognitive measures

The study included both clinical and neurocognitive outcome measures:

1. Obsessive-compulsive symptoms were assessed using the Y-BOCS, a semi-structured clinician-administered scale with validated Persian translation²⁹.

2. Executive functioning, cognitive flexibility, and problem-solving were measured using the Wisconsin Card Sorting Test (WCST)³⁰.

3. Sustained and selective attention, processing speed, and accuracy were evaluated using the D2 Test of Selective Attention³¹.

Yale-Brown obsessive compulsive scale

The Y-BOCS was used to assess the severity of obsessive-compulsive symptoms at baseline and at weeks 0, 4, 8, and 12 of treatment. The Y-BOCS is a clinician-rated instrument consisting of 10 items, each scored on a 5-point scale from 0 (no symptoms) to 4 (extreme symptoms), resulting in a total score ranging from 0 to 40³². It provides a standardized measure of symptom severity and is widely employed in both clinical and research settings³³. The Persian version of the Y-BOCS has demonstrated strong psychometric properties in previous studies, including high internal consistency and test-retest reliability, supporting its validity for use in Persian-speaking populations²⁹.

Wisconsin card sorting test

The WCST was used to assess participants’ executive functioning at baseline (week 0) and at the end of the intervention (week 12). Originally developed by Berg and Grant, the WCST is designed to evaluate cognitive flexibility, particularly the ability to shift cognitive strategies in response to changing environmental contingencies³⁰. It is widely recognized as a sensitive measure of executive function, including abstract reasoning, set-shifting, and problem-solving abilities³⁰.

The WCST involves four stimulus cards and two sets of 64 response cards. Participants were instructed to match each response card to one of the stimulus cards and received feedback indicating whether their choice was correct or incorrect. Key outcome measures included the number of categories completed, perseverative errors, and set-loss errors, which collectively reflect various components of executive functioning³⁰.

The change in each WCST scale from baseline to week 12 between the two study groups was analyzed to evaluate the effects of the intervention on executive function.

D2 test of selective attention

The D2 Test evaluates visual attention and concentration³¹. Participants scan rows of letters to identify instances of the target letter “d” with two dashes. The test reports the following indices:

• Total Number of Items Processed (TNIP)

• Omission Errors (OE)

• Commission Errors (CE)

• Total Errors (TE = OE + CE)

• Concentration Performance (CP)

• Fluctuation Rate (FR)

• Functional Index (FI = TNIP − TE)

The D2 is completed in ~10 min.

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Iran University of Medical Sciences (ID: IR.IUMS.REC.1402.119). All participants provided written informed consent after receiving a full explanation of the study’s purpose and means. Participants were informed of their right to withdraw at any point without consequence. Individuals experiencing side effects or who wished to discontinue were withdrawn and offered standard clinical care. The trial was registered at the Iranian Registry of Clinical Trials on 2023-06-08 (www.irct.ir; IRCT ID: IRCT20170123032145N7).

Sample size calculation

A priori power analysis (G*Power 3.1.9.7) was conducted for the outcome, Y-BOCS total score, assuming a mixed repeated-measures design with 2 groups (NAC vs placebo) and 4 assessments (baseline, Weeks 4, 8, 12).

The target effect was a group × time interaction of medium magnitude (Cohen’s $f=0.25$, approximately partial ${\eta }_p^2\approx 0.059$), with $\alpha =0.05$, power = 0.80, correlation among repeated measures = 0.05, and nonsphericity correction = 1.0³⁴. This yielded a minimum total sample size of 26 (13 per arm).

Cognitive outcomes (WCST, D2) were exploratory and were not used to determine the sample size.

Blinding and compliance monitoring

The trial employed a double-blind design. The primary psychiatrist, and participants were all blinded to group assignments. Blinding was maintained using capsules identical in appearance for NAC and placebo and sequentially numbered, opaque, sealed envelopes for allocation concealment.

Treatment compliance was monitored through self-report at each visit. Participants were asked to report daily medication use, and adherence was discussed during each follow-up appointment.

Statistical analysis

Baseline demographic and clinical characteristics were summarized for the total sample and by treatment group. Categorical variables (e.g., sex) were reported as counts and percentages, while continuous variables (e.g., age, Y-BOCS scores, WCST, and D2) were described using means and standard deviations.

To assess baseline equivalence, the distribution of each variable was examined using the Shapiro-Wilk test and visual inspection of histograms and Q-Q plots³⁵. Variables that were non-normally distributed were Y-BOCS total score, WCST non-perseverative errors, WCST categories completed, D2 CE, and D2 FR. These were analyzed using the Mann–Whitney U test. All other continuous variables were normally distributed and were compared using independent samples t-tests³⁶. Categorical variables were assessed with chi-square tests or Fisher’s exact tests where appropriate. Assumptions of sphericity were tested using Mauchly’s test³⁷. Greenhouse-Geisser corrections were applied when sphericity was violated³⁸. Where appropriate, repeated-measures General Linear Models were conducted using pooled multiply imputed datasets³⁹.

All statistical analyses were conducted using SPSS version 27⁴⁰. Statistical significance was set at p <0.05 (two-tailed). Analyses follow the intention-to-treat (ITT) principle. Missing data were handled using multiple imputations (five datasets) assuming that data were missing at random. In addition, a mixed model for repeated measures (MMRM) on the observed case data (without imputation) was conducted as sensitivity analysis⁴¹, to assess the robustness of findings. Rubin’s Rules were applied to pool parameter estimates and standard errors⁴².

Missing outcome data were handled under a missing‑at‑random (MAR) assumption using multiple imputation with five imputed datasets, and a MMRM was also fitted on the observed cases to assess robustness. To further evaluate the MAR assumption, a sensitivity analysis approximating a missing‑not‑at‑random (MNAR) mechanism was conducted for Y‑BOCS Total by applying last‑observation‑carried‑forward (LOCF) to participants with missing post‑baseline values and re‑estimating the Group×Time interaction in a linear mixed‑effects model.

Multiplicity/type I error control

The primary confirmatory hypothesis was the Group × Time interaction for the Y‑BOCS Total score, tested two‑sided at α = 0.05. To control the overall type I error across multiple outcomes and contrasts, a prespecified hierarchical testing strategy was applied. First, the Group × Time interaction for Y‑BOCS Total was evaluated at the 0.05 significance level. Only if this primary test was statistically significant did we proceed to evaluate the Group × Time interactions for the three related Y‑BOCS outcomes (Total, Obsession, and Compulsion). For these three tests, the Hochberg step‑up procedure was used to control the family‑wise error rate (family α = 0.05; m = 3). Conditional on statistical significance within this family of tests, post‑hoc pairwise comparisons of estimated marginal means between treatment groups at Weeks 4, 8, and 12 were performed. These three time‑point contrasts within each outcome were adjusted using Bonferroni correction (m = 3 comparisons per outcome; adjusted α = 0.0167). This hierarchical gatekeeping approach was used to maintain control of the overall family‑wise type I error rate.

Outcome analyses

General linear models (GLM)

Repeated measures GLMs were used to assess treatment effects over time in variables with normally distributed residuals. In these analyses, time was treated as a within-subject factor, with four levels for the Y-BOCS subscales and two levels for WCST and D2 outcomes. The adjunctive therapy (NAC vs. placebo) was the between-subject factor. Bonferroni adjustment was applied to post-hoc estimated marginal mean comparisons (between-group contrasts at Weeks 4, 8, and 12; m = 3 per outcome). Wilks’ Lambda was used to evaluate multivariate within-subject effects³⁹.

Linear mixed models (LLM)

For outcomes with non-normally distributed residuals (Y-BOCS Total, WCST Categories Completed, CE, and FR), LMMs were used. Fixed effects include time, group, and interaction. Random intercepts were included for each participant. Covariance structures (unstructured or compound symmetry) were chosen according to timepoint count. Models were estimated by Restricted Maximum Likelihood (REML) to better handle non-normality and missing data⁴³.

Post-trial care

Upon completion of the 12-week trial, the investigational adjunctive intervention (NAC/placebo) was discontinued for all participants. Participants who demonstrated a satisfactory clinical response to sertraline continued receiving sertraline at the discretion of their treating psychiatrist. For participants with partial or no response, medication regimens were modified according to standard clinical practice guidelines. No further data collection happened after the endpoint.

Results

Participants characteristics

A total of 35 participants are enrolled in the study, with 19 randomized to the experimental group and 16 to the control group as presented in Fig. 1. The sample consists of of 10 males (28.6%) and 25 females (71.4%), with no significant difference in sex distribution between groups (χ²(1) = 0.10, p = 0.75). The mean age of participants is 39.3 years (SD = 9.9), with the experimental condition group averaging 38.26 years (SD = 10.4) and the control group 40.68 years (SD = 9.53); this difference is not statistically significant (p = 0.47).

Fig. 1.

Flowchart of the participants enrolled in the trial.

A total of nine participants (25.7%) discontinued the treatment before the final assessment all due to personal reasons, six from the experimental condition group and three from the control group. The overall dropout rate does not significantly differ between groups (χ²1) = 0.75, p = 0.39; Fisher’s exact p = 0.46).

Baseline comparison between groups

No significant differences are observed between the NAC and control groups at baseline for any demographic or clinical variable (p > 0.05 for all comparisons), indicating successful randomization and baseline comparability. A full summary of baseline characteristics is provided in Table 1.

Table 1.

Baseline demographic and clinical characteristics of participants with between-group comparisons

Variable	Total (N = 35)	NAC + Sertraline (n = 19)	Placebo + Sertraline (n = 16)	p-value
Sex (n, % male)	10 (28.6%)	5 (26.3%)	5 (31.3%)	0.75a
Age (years), Mean ± SD	39.3 ± 9.9	38.26 ± 10.4	40.68 ± 9.53	0.47b
Sex (Male/Female)	–	5/14	5/11	1.0a
Y-BOCS Obsession, Mean ± SD	14.51 ± 3.74	14.61 ± 3.22	14.68 ± 4.31	0.82c
Y-BOCS Compulsion, Mean ± SD	14.65 ± 3.74	12.77 ± 3.63	12.56 ± 3.14	0.92c
Y-BOCS Total, Mean ± SD	27.11 ± 3.58	27.88 ± 3.72	26.62 ± 3.15	0.55c
WCST Non-perseverative Errors, Mean ± SD	24.17 ± 5.90	23.22 ± 4.55	25.25 ± 7.12	0.42c
WCST Perseverative Errors, Mean ± SD	7.32 ± 3.81	6.33 ± 2.86	8.43 ± 4.48	0.21b
WCST Categories Completed, Mean ± SD	3.05 ± 1.25	3.33 ± 1.02	2.75 ± 1.43	0.29c
D2 TNIP, Mean ± SD	240.25 ± 45.10	235.57 ± 49.09	245.81 ± 40.84	0.52b
D2 Omission Errors, Mean ± SD	62.45 ± 44.60	61.83 ± 46.69	57.50 ± 38.36	0.59b
D2 Commission Errors, Mean ± SD	4.94 ± 4.02	4.38 ± 3.79	5.25 ± 4.26	0.68c
D2 Total Errors, Mean ± SD	67.40 ± 43.40	66.22 ± 45.20	62.75 ± 36.55	0.60b
D2 Concentration Performance, Mean ± SD	233.26 ± 42.41	231.21 ± 47.23	235.31 ± 36.77	0.56b
D2 Fluctuation Rate, Mean ± SD	8.51 ± 5.14	8.05 ± 4.67	9.06 ± 5.76	0.62c
D2 Functional Index, Mean ± SD	172.85 ± 88.01	173.77 ± 91.10	183.06 ± 76.98	0.53b

^aChi-square test (for categorical variables).

^bIndependent samples t-test for normally distributed variables.

^cMann–Whitney U test for non-normally distributed variables.

Yale-Brown obsessive compulsive scale

Y-BOCS total

The linear mixed-effects model reveals significant Group × Time interaction (p ≤0.005), indicating that the reduction in overall OCD symptoms over time is significantly greater in the NAC group compared to placebo. Estimated marginal means demonstrated that the NAC group has significantly lower Y-BOCS total scores compared to placebo at weeks four (MD = −2.52, 95% CI: −3.48 to −1.56), eight (MD = −2.78, 95% CI: −3.80 to −1.76), and 12 (MD = −1.49, 95% CI: −2.19 to −0.79), with maximal separation evident at week eight. These findings demonstrate that NAC augmentation provided clinically meaningful treatment effects, with the NAC group showing a steeper and more sustained decline in overall OCD symptom severity (Fig. 2).

Fig. 2. Mean Y-BOCS total score over time by treatment group.

The NAC group showed significantly greater reduction in symptoms compared to Placebo (Interaction p = 0.001).

According to the prespecified hierarchical testing strategy, the Group × Time interaction for Y‑BOCS Total (the primary endpoint) was first evaluated at the two‑sided 0.05 level. As this test was statistically significant, subsequent analyses of the Y‑BOCS outcomes were performed with multiplicity control.

Error bars indicate standard errors of the mean. Baseline sample sizes were NAC+sertraline n = 19 and placebo+sertraline n = 16, with n at subsequent visits reflecting observed data after dropout.

Y-BOCS obsession

The repeated measures general linear model reveals a significant main effect of time, (F(3, 615) = 1118.55, p < 0.001), indicating substantial reduction in obsession severity over the 12-week treatment period across all participants. More importantly, there is a significant Time × Group interaction, (F(3, 615) = 22.13, p = 0.001), demonstrating that the pattern and rate of symptom reduction differ significantly between the NAC and placebo groups. Contrasts reveal a significant quadratic Time × Group interaction, (F(1, 205) = 69.56, p = 0.001), indicating that the NAC group experience a steeper trajectory of improvement in the early phase of treatment, with faster response evident at Weeks 4 and 8, whereas the placebo group showed more gradual improvement across the treatment period.

Estimated marginal means reveal that the NAC group have significantly lower obsession scores compared to placebo at weeks four (MD = −2.22, 95% CI: −3.00 to −1.43), eight (MD = −2.60, 95% CI: −3.30 to −1.89), and 12 (MD = −1.32, 95% CI: −1.83 to −0.81), with maximal separation evident at week eight. These findings demonstrate clinically meaningful treatment effects favoring NAC augmentation (Fig. 3).

Fig. 3. Mean Y-BOCS obsession scores over time by treatment group.

Error bars represent standard errors. The NAC + Sertraline group shows significantly lower scores compared to placebo at Weeks 4, 8, and 12.

After Hochberg adjustment across the three prespecified Y‑BOCS Group × Time interaction tests (Total, Obsession, Compulsion), the Group × Time interaction remained significant for the Obsession subscale.

Asterisks represent mean Y-BOCS Obsession subscale scores for each treatment group at each time point that were plotted as outliers; error bars indicate standard errors of the mean. Baseline sample sizes were NAC+sertraline n = 19 and placebo+sertraline n = 16

Y-BOCS compulsion

The Compulsion Group × Time interaction is not significant and remained non‑significant after Hochberg multiplicity control across the Y‑BOCS interaction tests (p = 0.597), indicating both groups improve similarly on compulsive symptoms (Supplementary Fig. 1).

Clinical response rate

A responder analysis based on a ≥35% reduction in total Y-BOCS scores shows that 100% of participants in the NAC group and 92.6% in the placebo group achieved clinical response. This difference is statistically significant (χ²(1) = 8.71, p = 0.003; Fisher’s Exact p = 0.004), indicating that NAC augmentation may increase the probability of meaningful improvement in OCD symptomatology.

Table 2 demonstrates the mean of Y-BOCS scores at each timepoints and their comparison with the baseline scores. Within‑group baseline comparisons in this table are descriptive; confirmatory inference is based on the Group × Time interaction models described above.

Table 2.

The mean (±standard deviation) Y-BOCS scores of weeks 4, 8, and 12, and their comparison with the baseline scores (95% CI)

		NAC + Sertraline			Placebo + Sertraline
		Mean ± SD	MD	Cohen’s d effect size	Mean ± SD	MD	Cohen’s d effect size
Obsession	Week 4	10.8 ± 3.29	3.57 ± 2.24	1.59	12.9 ± 3.81	1.76 ± 1.43	1.42
	Week 8	7.95 ± 2.50	6.42 ± 2.40	2.67	10.47 ± 3.58	4.22 ± 1.79	2.36
	Week 12	5.8 ± 1.89	8.57 ± 2.60	3.30	7.09 ± 2.19	7.60 ± 2.91	2.61
Compulsion	Week 4	10.35 ± 2.93	2.34 ± 2.35	1.00	10.6 ± 2.92	1.96 ± 1.66	1.18
	Week 8	8.28 ± 2.93	4.41 ± 2.59	1.70	8.4 ± 2.13	4.17 ± 2.07	1.95
	Week 12	5.65 ± 2.47	7.03 ± 3.18	2.21	5.64 ± 1.94	6.92 ± 2.39	2.89
Total	Week 4	21.08 ± 3.78	6.44 ± 3.41	1.89	23.5 ± 3.94	3.12 ± 2.91	1.07
	Week 8	16.21 ± 4.13	11.32 ± 3.57	3.17	18.92 ± 4.21	7.71 3.25	2.37
	Week 12	11.38 2.93	16.15 3.72	4.34	12.87 ± 2.74	13.75 ± 2.77	4.97

Cognitive outcomes

Changes in executive functioning (WCST) and attention (D2 Test) are summarized in Supplementary information.

For the WCST, there are no significant main effects of Time or Group, nor any significant Time × Group interactions in categories completed, perseverative errors, or non-perseverative errors (p = 0.24, p = 0.26, p = 0.64 respectively). Both groups show stable performance over the 12-week period (Supplementary Figs. 2–4 and Supplementary Table 1).

For the D2 Test of Attention, there is a significant main effect of Time for TNIP (p = 0.020) and Omission Errors (p = 0.022), indicating that all participants, regardless of treatment assignment, improve their processing speed and sustained attention over the course of the study. However, no significant Time × Group interactions are observed for any D2 index, suggesting that NAC augmentation do not confer additional cognitive benefits beyond the practice effects or general clinical improvement seen in the placebo group (Supplementary Figs. 5–11 and Supplementary Table 1).

Sensitivity analysis

Sensitivity analysis using a Mixed Model for Repeated Measures (MMRM) on the observed case data (without imputation) yields consistent results, confirming the robustness of the primary findings.

In the additional MNAR-oriented sensitivity analysis using LOCF for missing Y-BOCS Total values, the Group × Time interaction remains non-significant (p = 0.23), and the direction and magnitude of the estimated treatment effects closely parallel those from the MAR-based models. This supports the robustness of the efficacy findings to plausible deviations from the MAR assumption (Table 3).

Table 3.

Proportion of missing Y-BOCS Total scores at each visit by treatment group from baseline

Visit	NAC sertraline (n = 19), n missing (%)	Placebo sertraline (n = 16), n missing (%)	Total (N = 35), n missing (%)
Week 4	1 (5.3%)	1 (6.3%)	2 (5.7%)
Week 8	1 (5.3%)	1 (6.3%)	2 (5.7%)
Week 12	1 (5.3%)	2 (12.5%)	3 (8.6%)

Adverse drug reaction (ADR)

ADRs are transient and non-severe, without any cases necessitating treatment discontinuation. Using Fisher’s Exact Test, the frequency of side effects is not significantly different between the groups (Table 4)

Table 4.

Comparison of the frequency of adverse events between the two groups (N, %)

ADR	NAC + Sertraline (n = 19)	Placebo + Sertraline (n = 16)	p-value
Decreased Libido	2 (10.5%)	2 (12.5%)	1.0
Decreased appetite	1 (5.2%)	1 (6.25%)	1.0
Headache	1 (5.2%)	–	1.0
Dyspepsia	–	1 (6.25%)	0.4

Discussion

This randomized, double-blind clinical trial investigated the efficacy of N-acetylcysteine (NAC) as an adjunctive treatment to sertraline over a 12-week period, with parallel evaluation of clinical and cognitive outcomes in an Iranian population. This pilot study was designed to assess whether glutamatergic modulation via NAC could provide additional benefits beyond optimized SSRI therapy.

Our results demonstrated that NAC combined with sertraline was associated with a faster and numerically greater reduction in Y-BOCS obsession and total scores, reflected by a significant Time × Group interaction. Responder rates (defined as ≥35% reduction in Y-BOCS total score) were high in both groups, reaching 100% in the NAC+ sertraline arm and 92.6% in the placebo + sertraline arm. In contrast, aside from a small, non-significant trend toward improved attentional consistency, NAC did not confer statistically significant advantages over placebo on executive function or attention as measured by the WCST and D2 tests. Both treatment arms were well tolerated, with only mild and transient adverse events reported, no treatment discontinuations, and no significant between-group differences in side-effect profiles.

Taken together, these findings suggest that adjunctive NAC may be associated with an earlier reduction in obsessive symptoms. However, given the modest effect sizes and the absence of consistent between-group differences across all outcomes, this interpretation should be considered preliminary. This pattern is consistent with hypotheses suggesting that glutamatergic modulation may influence intrusive thoughts earlier in treatment, whereas serotonergic mechanisms are generally associated with more gradual clinical improvement. Nonetheless, the present design does not allow definitive separation of these temporal effects^44,45.

Additionally, a degree of symptom-dimension specificity was observed. The apparent benefits of NAC were largely confined to obsessive symptoms, while compulsive behaviors showed less pronounced change. This distinction is clinically plausible, as compulsions are often more rooted behavioral patterns that may require longer treatment durations, targeted behavioral interventions such as ERP, or potentially higher NAC doses to achieve meaningful change^18,46.

Despite this, the overall response rate across both groups was unusually high (93–100%). This ceiling effect may reflect: I) the robust efficacy of sertraline at 200 mg/day⁴⁷, II) expectancy or placebo effects in a structured academic clinical trial setting⁴⁸, or III) statistical fluctuation due to small sample size⁴⁹. Such high overall response rates may have limited the statistical detectability of incremental between-group differences attributable specifically to NAC augmentation.

The present results are broadly consistent with prior research. A recent meta-analysis of six randomized controlled trials (n = 195) reported that NAC augmentation significantly reduced total Y-BOCS scores between weeks 5 and 8¹⁸. However, that analysis did not identify significant improvements on obsession or compulsion subscales individually, and the observed benefits did not persist beyond 12 weeks, which is similar to the temporal pattern reported in the current study¹⁸.

These findings contribute to the growing body of evidence implicating glutamatergic dysregulation within CSTC circuits in obsessive-compulsive related disorders (OCRDs)^50,51. Elevated glutamate signaling in these circuits has been linked to intrusive thoughts and impaired inhibitory control in neuroimaging and genetic studies⁵⁰. NAC may target these disruptions by modulating extracellular glutamate through the cystine–glutamate antiporter by restoring redox balance⁵². The earlier reduction in obsessive symptoms observed in the NAC group is compatible with this mechanistic framework, although causal inferences should remain conservative⁵².

Importantly, the present findings align with conclusions from umbrella reviews examining glutamate-modulating agents (including NAC, memantine, riluzole, and ketamine) in OCRDs^53–56. These reviews suggest that such agents are effective in reducing obsessions, especially early in treatment or among partial SSRI responders^53–56. They further emphasize the CSTC specificity of glutamatergic interventions, with benefits concentrated on cortico-striatal hyperactivity rather than on broader cognitive domains⁵⁷. This framework is consistent with the lack of significant NAC-related improvements in executive function or attention observed in the present study. Accordingly, the current findings extend prior reviews by providing controlled evidence of obsession-focused and temporally early clinical effects of NAC under fixed-dose SSRI adjunction, alongside complementary cognitive outcome data.

An alternative explanation for the observed early improvement in obsessive symptoms is the delayed therapeutic effect of sertraline itself. Selective serotonin reuptake inhibitors, particularly at higher doses, are well known to exert their full anti-obsessional effects after a latency period of ~8–12 weeks, reflecting gradual serotonergic neuroadaptation and downstream cortical–striatal modulation⁵⁸. Consequently, the early symptom changes observed cannot be unequivocally attributed to NAC augmentation alone.

The absence of statistically significant cognitive benefits may be attributable to several factors. One important consideration is the practice effect, as both the WCST and D2 were administered only 12 weeks apart. Repeated exposure to these tasks can improve performance through familiarity with test structure, reduced performance anxiety, or recall of task demands rather than true cognitive change⁵⁹. Meta-analytic evidence indicates that re-administration of such tests within two to three months can yield improvements of 0.5 to 1 standard deviation even in the absence of intervention⁶⁰.

Although NAC did not significantly improve executive function or attention in our findings, inclusion of cognitive outcomes was justified by increasing recognition of persistent cognitive impairments in OCD⁶¹. Deficits in set-shifting and sustained attention have been linked to dysfunction within frontostriatal and cingulo-opercular networks⁶¹. While robust cognitive benefits were not observed over the 12-week period, subtle trends favoring NAC in attentional consistency suggest that this domain warrants further investigation.

Treatment duration may also have limited detection of cognitive effects. Glutamatergic and antioxidant mechanisms involved in cognitive plasticity may require longer than 12 weeks to manifest measurable changes⁶², particularly in executive functioning and sustained attention⁶³. Additionally, ceiling effects may have constrained detectability, as several participants entered the trial with high baseline cognitive scores, leaving limited room for improvement on measures such as WCST categories completed and D2 concentration indices⁶⁴. The modest sample size further raises the possibility of Type II error, reducing power to detect subtle between-group differences⁶⁵.

Measurement sensitivity represents another consideration. While the WCST and D2 are well-validated instruments, they may lack sensitivity to detect nuanced changes in cognitive control or flexibility⁶⁶. More refined assessments, such as stop-signal or emotional Stroop tasks, may be better suited to capture subtle neurocognitive effects of NAC^67,68.

Finally, the use of high-dose sertraline (200 mg/day) may have produced a therapeutic ceiling, limiting the observable magnitude of NAC augmentation⁶⁹. It is also plausible that glutamatergic modulation preferentially benefits individuals with underlying CSTC hyperactivity or redox imbalance⁷⁰. Without baseline biomarker stratification (such as magnetic resonance spectroscopy or peripheral glutathione levels), potential responders and non-responders were likely pooled, limiting observable effects⁶⁹.

Despite these limitations, this study has several strengths. The fixed-dose SSRI regimen ensured pharmacological consistency and minimized heterogeneity common in prior augmentation trials. Concurrent assessment of clinical and neurocognitive outcomes allowed exploration of symptom–cognition dissociation. Analytically, the study adhered to intention-to-treat principles, employed multiple imputation for missing data, and confirmed findings using MMRM analyses, all of which led to convergent results and support the robustness of the conclusions.

Future studies should employ larger samples to detect moderate adjunctive effects and extend treatment duration to assess longer-term clinical and cognitive outcomes. Exploration of higher NAC doses, if tolerable, may further clarify dose-response relationships. Biomarker-based stratification could help identify individuals most likely to benefit from glutamatergic modulation. Finally, incorporation of dimensional symptom measures and behavioral tasks (such as obsessive belief inventories, response inhibition paradigms, or ecologically valid assessments of compulsivity)^71,72 may provide deeper insight into NAC’s specific therapeutic mechanisms.

Conclusion

The results of this pilot study suggest N-acetyl cysteine seems to act as an accelerator rather than a booster in SSRI-treated OCD, speeding the alleviation of obsessions, raising early responder rates, and doing so with great tolerability. Cognitive benefits are not demonstrated, likely owing to methodological ceilings. For clinicians, NAC remains a low-risk option worth practicing in patients seeking quicker obsession relief or partial responders. For researchers, larger, longer, biomarker enhanced studies are needed to assess who truly benefits and to clarify if NAC can add value beyond robust SSRI therapy.

Author contributions

S.E., S.V.S., M.T., A.G.J., E.S., and M.S. made substantial contributions to the conception and design of the work. S.A., A.G.J., M.Y., M.T., S.T.E. and M.S. had substantial contribution in data gathering. S.E. and M.S. analyzed and interpreted the data. S.E., M.S. had major contribution in writing the manuscript. All authors read and approved of the final manuscript.

Peer review

Peer review information

Communications Medicine thanks Dalei Chen and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

The authors disclose receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Iran University of Medical Sciences (Grant no: 1401-4-99-25244).

Data availability

The individual de-identified participant data and the source data for the figures and tables can be found in the Supplementary Data.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s43856-026-01711-6.