Results
Baseline Characteristics
Before visiting each trial site, 34 of 210 candidates did not meet the eligibility criteria based on background information, 21 did not meet the criteria for Social Responsiveness Scale[48] and/or Autism Spectrum Quotient[49] scores (Supplementary material) and 11 declined to consent to participate. Of the 144 participants assessed for eligibility, consent was obtained from 127 adult males with ASD. Of these, 18 did not meet the eligibility criteria and 109 males with high-functioning ASD were enrolled between 25 June 2018 and 26 December 2019. End point assessments were performed between 27 September 2018 and 19 March 2020. The actual rate of withdrawal (7 of 109 cases, 6.4%) was lower than expected (i.e. 20%). Because the number of cases needed for analyses was calculated as ~120, we completed recruitment of 127 participants before registering a target sample size of 144 participants, which was calculated with an expected withdrawal rate of 20%.
Six of the 109 participants did not complete the trial because they withdrew their consent. Of these, three withdrew before the first administered period and three withdrew during the washout period. Another participant in the placebo-TTA-121 10 U group did not complete the trial because of an adverse event (a mild emergence of blasts in the blood) during the washout period (Figure 1).
Consequently, we analysed 106 participants as the full analysis set (FAS; Table 1). The FAS was the analysis target population, consisting of all participants except for those who met the following criteria among the participants who had been randomized: participants who were not diagnosed with ASD (n = 0); participants who have never taken the investigational drug (n = 3); participants with no data evaluating efficacy after randomization (n = 0). The per protocol set (PPS) was the analysis target population, consisting of all FAS participants except those that met the following criteria: participants who did not meet the inclusion criteria; participants who fulfilled the exclusion criteria; participants who violated the rules for prohibited drugs; participants with an investigational drug compliance rate of 85% or less. For PPS of the first period, nine participants with poor (<85%) drug adherence (Supplementary Table 3), one with visit out of period, one with more than one psychotropic drug and one with deviations from the protocol potentially influencing the assessment of efficacy were excluded; 94 participants were included. For the PPS of the second period, eight participants with poor drug adherence, one with visit out of period, one with more than one psychotropic drug and one with deviations from the protocol potentially influencing the assessment of efficacy were excluded; 91 participants were included. Plasma oxytocin levels were measured in peripheral blood samples collected just before the first and 60 min after the last double-blind administration in each period, using liquid chromatography–mass spectrometry (details in the Supplementary material). These levels were elevated relative to baseline following TTA-121 administration [means (pg/ml): 3 U: from 0.00 to 1.63; 6 U: 0.00 to 1.84; 10 U: 0.04 to 6.92; 20 U: 0.00 to 6.47]; such elevations were not observed following placebo administration (placebo: 0.02 to 0.09; Supplementary Table 4).
Primary Outcome. The contrast with the smallest P-value, judged as the dose–response relationship, was [Placebo, TTA-121 3 U, 6 U, 10 U, 20 U: 4, −1, −6, −1, 4]. The contrast of inverted U-shape had its peak at TTA-121 6 U (i.e. 3 U/morning and evening) for both the FAS (P = 0.182) and PPS (P = 0.073; Table 2). Next, differences between TTA-121 6 U and placebo were calculated using paired t-tests; the ADOS reciprocity score was reduced from baseline to end point in the TTA-121-administered period compared with the placebo period (FAS: P = 0.118, mean difference = −0.5; 95% CI: −1.1 to 0.1; PPS: P = 0.012, mean difference = −0.8; 95% CI: −1.3 to −0.2; Figure 2 and Table 3). No significant design and hangover effects in crossover design were revealed for any doses (3 U: P = 0.215; 6 U: P = 0.365; 10 U: P = 0.476; 20 U: P = 0.868; Supplementary Tables 5 and 6).
Figure 2.
The effects of oxytocin on the primary outcome: ADOS reciprocity. (A and B) Plots illustrating changes in the means and error bars representing the standard deviations of ADOS reciprocity scores (range of possible scores, 0–14: higher scores indicate greater severity) in the FAS (A) and PPS (B). (C and D) Plots illustrating the mean differences between oxytocin and placebo, and error bars representing their 95% CIs, in ADOS reciprocity scores in the FAS (C) and PPS (D). *Statistically significant as indicated by P < 0.05 from paired t-test.
Secondary and Exploratory Outcomes
No secondary outcomes were significantly improved by TTA-121 compared with placebo (Supplementary Tables 7–12). Among the exploratory outcomes, the mean of log F0 (P = 0.015, mean difference = −0.066; 95% CI: −0.117 to −0.014) and correlation of blockwise mean of log F0 (P = 0.039, mean difference = −0.379; 95% CI: −0.736 to −0.023) were reduced from baseline to end point in the oxytocin administration period compared with placebo, while the log pause-to-turn ratio was increased during the oxytocin administration period (P = 0.028, mean difference = 0.085; 95% CI: 0.010 to 0.159; Supplementary Table 13). Fixation times on eye regions in the movies of human faces without lip movement and while blinking were reduced from baseline to end point in the oxytocin administration period compared with placebo (P = 0.042, mean difference = −0.217; 95% CI: −0.426 to −0.008; P = 0.046, mean difference = −0.219; 95% CI: −0.433 to −0.004, respectively; Supplementary Table 14). No other exploratory outcomes were significantly improved by oxytocin administration compared with placebo (Supplementary Table 15).
Safety
No severe adverse events were observed in any doses (Supplementary Tables 16–19). Although one participant discontinued the trial because of an adverse event (mild emergence of blasts in blood), the event was observed during washout after the placebo administration period and recovered after observation. The emergence ratios of adverse events in administration periods were 24.0% in TTA-121 3 U, 46.4% in 6 U, 40.7% in 10 U, 40.0% in 20 U and 36.9% in placebo (Supplementary Table 16).
Brain. 2022;145(2):490-499. © 2022 Oxford University Press
