Effect of a saffron extract on sleep quality in adults with moderate insomnia: A decentralized, randomized, double-blind, placebo-controlled trial

Abstract

Aim: Natural interventions for sleep disturbances, such as saffron extract, are gaining scientific and clinical interest. This 3-arm, randomized, double-blind, placebo-controlled trial examined the effect of a standardized saffron extract (Safr'Inside™) on sleep, stress, and other associated psychological outcomes in 165 adults reporting moderate insomnia.

Methods: Participants received 30 mg, 20 mg saffron extract, or placebo for 4 weeks. The primary endpoint was the change in insomnia symptoms (Athens Insomnia Scale, AIS). Secondary outcomes were the Single-Item Sleep Quality Scale (SQS), Perceived Stress Scale (PSS), Patient Health Questionnaire-4 (PHQ-4), Positive and Negative Affect Schedule (PANAS), Epworth Sleepiness Scale (ESS), and World Health Organization Quality of Life (WHOQOL). Analyses followed an intention-to-treat (ITT) approach, with per-protocol (PP) confirmation.

Results: Among 150 completers, saffron extract led to a greater reduction in insomnia symptoms (AIS) than the placebo (between-group adjusted mean difference β = -0.95 [95 % CI: -1.79, -0.11], P < .05). In secondary analyses, sleep quality (SQS) improved significantly after 3 weeks and was sustained at week 4 in both saffron groups compared to placebo (30 mg vs placebo: β = 0.82 [95 % CI: 0.22, 1.41], P = .004; 20 mg vs placebo: β = 1.02 [0.43, 1.62], P < .001). Perceived stress (PSS) was significantly reduced with 30 mg or 20 mg saffron extract compared to placebo (30 mg vs placebo: β = -1.87 [95 % CI: -3.23, -0.53], P = .01; 20 mg vs placebo: β = -1.89 [95 % CI: -3.22, -0.52], P = .04). Some improvement in psychological symptoms (PHQ-4) was also observed with 30 mg saffron extract compared to placebo (β = -0.79 [-1.40, -0.18], P = .03). All other measures showed no significant differences. No serious adverse events occurred.

Conclusions: Four weeks of 20 or 30 mg saffron extract may reduce insomnia and stress in middle-aged adults. Future research should assess longer interventions and explore which subgroups benefit most from saffron extract.

Keywords: Athens insomnia scale; Decentralized trial; Insomnia; Mental health; Randomized controlled trial; Saffron extract; Sleep.

Fig. 1

Participant flow diagram according to CONSORT guidelines. The diagram illustrates participant progression through the stages of enrollment, randomization, allocation, follow-up, and analysis.

Fig. 2

Effect of saffron extract on insomnia symptoms and self-reported sleep quality compared with placebo. (A) Mean change in Athens Insomnia Scale (AIS) scores. Scatter points represent individual participant AIS score changes from baseline (T0) to day 28 (T28) in the placebo group (n = 48) and saffron intervention group (n = 102; 20 mg, n = 50; 30 mg, n = 52; combined for analysis). Bars represent group means ± standard error of the mean (SEM). The horizontal dashed line indicates no change in AIS score (mean change = 0). Statistical analysis was conducted using an analysis of covariance (ANCOVA) within an intention-to-treat (ITT) framework, adjusted for age, sex, body weight, and AIS baseline score. Helmert contrasts were used to compare mean AIS score changes. Asterisks indicate statistically significant differences (P < .05). (B) Mean scores of the single-item sleep quality scale (SQS). Scores from baseline (T0) to day 28 (T28) in the placebo group (n = 49), 20 mg saffron extract group (n = 55), and 30 mg saffron extract group (n = 54) are shown. Scores represent observed means ± SEM. Statistical analysis was conducted using a linear mixed model (LMM) within an ITT framework, adjusted for age, occupation, and baseline SQS scores. Pairwise comparisons were conducted with Tukey adjustments. Asterisks at specific time points indicate significant post hoc pairwise comparisons, while those to the right of brackets represent overall group differences over time (simple slope analysis), with P values Tukey-adjusted (∗P < .05, ∗∗P < .01, ∗∗∗P < .001).

Fig. 3

Change in Perceived Stress Scale (PSS) Scores from Baseline to Day 28. Scatter points represent individual participant PSS score changes in the placebo group (n = 48), 20 mg saffron extract group (n = 55), and 30 mg saffron extract group (n = 54); points are jittered slightly for better visibility. Bars represent changes in group means ± standard error of the mean (SEM). The horizontal dashed line indicates no change in PSS score (mean change = 0). Statistical analysis was conducted using the Kruskal-Wallis test due to non-normal data distribution. Pairwise comparisons were performed using Dunn’s test, and the analysis followed an intention-to-treat (ITT) approach. Asterisks indicate statistically significant between-group differences (P < .05) compared with the placebo group.

Fig. 4

Mean Patient Health Questionnaire-4 (PHQ-4) scores from baseline to day 28. Scores are shown for the placebo group (n = 49), 20 mg saffron extract group (n = 55), and 30 mg saffron extract group (n = 54). Values represent observed means ± standard error of the mean (SEM). Statistical analysis was conducted using a linear mixed model (LMM) within an intention-to-treat (ITT) framework, adjusted for age, occupation, and baseline PHQ-4 scores. Asterisk indicates a statistically significant between-group difference in change scores (P < .05) based on Tukey-adjusted pairwise comparisons.