One-Year Outcomes Following Intravenous Ketamine Plus Digital Training Among Patients with Treatment-Resistant Depression

Price and colleagues previously reported, in a randomised clinical trial, that a brief digital intervention based on Pavlovian conditioning—automated self-association training (ASAT)—appeared to prolong the antidepressant effect of a single ketamine infusion in patients with treatment-resistant depression. Those earlier results showed an effect on the primary clinician-rated depression outcome that persisted through the last in-person visit at 1 month postinfusion. Given that no waning was observed by the end of the acute-phase protocol, the investigators collected self-report data over a 1-year naturalistic follow-up to characterise the longer-term durability of the combined pharmacological and digital intervention. This paper reports a secondary analysis of that randomised trial, using the Quick Inventory for Depressive Symptoms, self-report version (QIDS-SR), to examine symptom trajectories across multiple scheduled follow-up time points up to 360 days after infusion. The aim was to determine whether the early synergistic benefit of ketamine plus ASAT persisted beyond the acute phase under naturalistic conditions and to identify when, if at all, group differences converged over the year-long follow-up.

This secondary analysis followed CONSORT guidelines and used data from the parent randomised clinical trial (ClinicalTrials.gov identifier: NCT03237286). A total of 154 adult patients with treatment-resistant depression who were on stable psychiatric treatments were randomised to one of three conditions: a single intravenous infusion of ketamine (0.5 mg/kg) followed by 4 days of active ASAT (about 30–40 minutes per day); ketamine followed by sham ASAT (computer exercises lacking therapeutic and self-relevant content); or saline infusion followed by active ASAT. The ASAT intervention was designed as a brief computer-based programme using positive words and images to target implicit self-referential patterns during a putative window of plasticity after ketamine. The prespecified follow-up used the QIDS-SR as the primary outcome during the naturalistic 1-year period. Self-report surveys were solicited at 30, 60, 90, 120, 150, 180, and 360 days after the infusion date. Analyses were conducted on an intent-to-treat basis using a hierarchical linear model with time point and group (saline plus ASAT as the reference) included as categorical factors; preinfusion baseline QIDS-SR scores were included as a covariate. The hierarchical model automatically handles missing data through its estimation procedure. Because the investigators did not have a priori expectations about how long any synergistic benefit might last, they inspected symptom trajectories visually and identified a data-driven breakpoint at greater than or equal to 120 days postinfusion where group error bars appeared to overlap. To limit multiple comparisons, planned pairwise contrasts were then used to compare group means in the early period (defined post hoc as less than 120 days) and the late period (greater than or equal to 120 days).

In the full hierarchical model, a significant time point × group interaction was observed (P for interaction < .001), indicating differing symptom trajectories across the three treatment arms over the follow-up period. During the early timeframe (post hoc defined as <120 days), ketamine plus active ASAT produced a statistically significant reduction in QIDS-SR scores compared with saline plus ASAT (β = -0.37; 95% CI, -0.71 to -0.04; P = .03). By contrast, ketamine followed by sham ASAT did not differ significantly from saline plus ASAT in that early period (β = -0.20; 95% CI, -0.54 to 0.14; P = .25). In the late timeframe (≥120 days), group differences were no longer statistically significant. Specifically, ketamine plus ASAT versus saline plus ASAT yielded β = 0.18 (95% CI, -0.15 to 0.51; P = .29), and ketamine plus sham versus saline plus ASAT yielded β = 0.11 (95% CI, -0.23 to 0.45; P = .52). Thus, the early benefit associated with the combination of ketamine and ASAT was evident for approximately the first three months postinfusion but was not maintained from months 4 through 12 under the naturalistic follow-up.

The investigators interpret these findings to mean that a single 40-minute ketamine infusion followed by 4 days of brief, portable digital training (ASAT) produced a statistically detectable augmentation of antidepressant effect that endured for about 3 months in a real-world follow-up context. Price and colleagues note that this durability occurred despite naturalistic treatment changes across groups during the follow-up period. However, the effect did not persist through months 4 to 12, as symptom trajectories converged. Several limitations acknowledged by the study team temper the conclusions. The trial did not include a true no-treatment control (saline plus sham), so the design cannot isolate the standalone effect of ASAT because every participant received at least one active component. Low base rates of clinically critical outcomes, such as suicidal events, limit inference about safety and impacts on those outcomes. The authors also highlight the persistence of mild-to-moderate average depressive symptoms across all groups, indicating the initial effect size was modest. For future work, the authors suggest strategies to increase both magnitude and durability of benefit: repeated ketamine administrations, booster sessions of the digital therapy (including at-home delivery), and additional conditioning modules targeting other implicit cognitive patterns. They further propose testing similar low-resource digital techniques, paired with rapid-acting treatments, for other common psychiatric conditions such as suicidality and anxiety.