Prolonged Grief Symptom Outcomes During At-Home Ketamine-Assisted Therapy: A Real-World Retrospective Analysis of 503 Adults

Abstract Background Prolonged grief disorder (PGD) is a clinically distinguishable bereavement-related condition characterized by persistent yearning, identity disruption, and impaired functioning, formalized in DSM-5-TR and ICD-11 (6L72) and empirically distinguishable from but frequently co-occurring with major depressive disorder. Approximately 7 to 10 percent of bereaved adults meet criteria for PGD, with elevated suicide risk and functional impairment. Complicated grief therapy is the targeted psychotherapy most often referenced for prolonged grief, and existing PGD treatments have been studied primarily in specialty academic and clinical settings without examining at-home telehealth delivery. A small psychedelic-grief literature has emerged in psilocybin and ayahuasca observational and pilot studies. Prior real-world ketamine cohorts have characterized depression outcomes without focusing on prolonged grief.

Objective: This exploratory, hypothesis-generating analysis describes prolonged grief symptom outcomes among adults with clinically elevated baseline grief who received at-home ketamine-assisted therapy through Mindbloom, a telehealth ketamine therapy platform.

Methods: We conducted a retrospective cohort analysis of 503 bereaved adults with clinically elevated prolonged grief symptoms (PGD-13 Total grief score ≥ 30 at baseline, a pragmatic severity threshold consistent with the score-based caseness cut used in the Shear randomized trials of complicated grief therapy on the predecessor Inventory of Complicated Grief) who confirmed the loss of a significant person on the PGD-13 gating question and received guided at-home sublingual or subcutaneous ketamine-assisted therapy through Mindbloom. Prolonged grief symptoms were assessed using the PGD-13 (mean baseline = 37.18) at post-session 2 (n = 282), post-session 4 (n = 196), and post-session 6 (n = 121). The primary study-defined response was a ≥ 5-point improvement in PGD-13 Total grief score from baseline. We additionally applied an adapted version of a formal Prolonged Grief Disorder caseness algorithm at baseline and post-session 6 to characterize diagnostic remission.

Results: Mean PGD-13 Total grief score declined from 37.18 at baseline (95% bootstrap CI 36.72–37.63) to 31.07 at post-session 2 (mean change − 6.11), 27.61 at post-session 4 (− 9.57), and 25.81 at post-session 6 (− 11.37). Among the 121 post-session 6 completers (24% of the eligible cohort), study-defined response (≥ 5-point improvement) was achieved by 51.8% at post-session 2 (95% Wilson CI 46.0–57.5%), 68.4% at post-session 4 (61.6–74.5%), and 76.0% at post-session 6 (67.7–82.8%); any improvement (≥ 1-point) was achieved by 90.1% of post-session 6 completers (83.5–94.2%). The proportion of post-session 6 completers below the score-based threshold of 30 was 63.6% (95% CI 54.8–71.7%); 0.8% reported no change and 9.1% reported any worsening. Applying the adapted PGD caseness algorithm, 42.7% of baseline participants met full diagnostic criteria for Prolonged Grief Disorder under the ICD-11 duration threshold (95% CI 38.5–47.1%); 35.6% met DSM-5-TR criteria (95% CI 31.5–39.9%). Among completers who met caseness at baseline, 73.2% no longer met ICD-11 criteria at post-session 6 (n = 41/56; 95% CI 60.4–83.0%); 71.1% no longer met DSM-5-TR criteria (n = 32/45; 95% CI 56.6–82.3%). In a worst-case sensitivity analysis in which all non-completers were assumed to retain full caseness, the confirmed diagnostic remission rates were 19.1% (ICD-11; 95% CI 14.4–24.8%) and 17.9% (DSM-5-TR; 95% CI 13.0–24.1%). Item-level analysis revealed that the single largest mean change across the 10 PGD-13 symptom items occurred on the identity/role confusion item (mean 4.12 to 2.64; change − 1.48, 95% bootstrap CI − 1.70 to − 1.24, corresponding to a 35.9% reduction from baseline to post-session 6). In contrast, the emotional pain item most closely overlapping major depressive disorder phenomenology improved 26.4% over the same interval, ranking sixth of ten. Side effects of any type were uncommon (7.4% at post-session 2, 5.6% at post-session 4, 4.1% at post-session 6). Subgroup patterns of larger mean change in more recently bereaved patients were observed (largest at post-session 6 in the 6-to-11-months-since-loss band, mean change − 13.83 points; smallest in the 12-months-or-longer band, − 10.40 points).

Conclusions: In this retrospective cohort of 503 adults with PGD-13 Total grief score ≥ 30 at baseline receiving at-home ketamine-assisted therapy through Mindbloom, prolonged grief symptoms declined monotonically across post-session timepoints; among the 121 post-session 6 completers, 76% achieved the primary study-defined response (≥ 5-point improvement) and 64% dropped below the score-based threshold of 30. Among completers meeting formal PGD caseness at baseline, the majority (73% under ICD-11, 71% under DSM-5-TR) no longer met diagnostic criteria at post-session 6, and item-level patterns showed the largest improvement on a non-depression-overlapping grief-specific symptom. Observed PGD-13 changes cannot be cleanly disentangled from concurrent depression-symptom changes in this uncontrolled design, because the cohort was treated for depression and the PGD-13 shares symptom content with depression instruments. The observed subgroup gradient runs opposite to what a strict mood-mediation account would predict, though concurrent depression improvement remains a compatible explanation. Randomized controlled trials with PGD-13 (or comparable independent grief instrument) as primary outcome, depression-independent comparator arms, and longer follow-up are needed to confirm these findings and to characterize the mechanism and durability of any observed effects.