Long-acting injectable (LAI) antipsychotics ensure sustained medication exposure, improve adherence and reduce relapse risk compared with oral formulations. Reference Kishimoto, Hagi, Kurokawa, Kane and Correll1 Olanzapine pamoate (Zypadhera) is widely used in Europe and offers efficacy comparable with oral olanzapine while addressing adherence problems, Reference Citrome, McEvoy, Todtenkopf, McDonnell and Weiden2 with potentially lower risks of anticholinergic effects and extrapyramidal symptoms. Reference Wang, Schneider-Thoma, Siafis, Burschinski, Dong and Wu3 However, LAIs are vulnerable to supply chain disruptions because substitution is not always straightforward.
In early 2025, Spain experienced a nationwide shortage of olanzapine pamoate (Zypadhera), forcing clinicians to switch treatments, often under suboptimal conditions and without the possibility of gradual cross-titration. Although shortages of psychiatric medications are reported worldwide, systematic data on their clinical consequences remain scarce, Reference Moerkens, Jacobsen and TeStroet4 particularly regarding LAI antipsychotics.
Here, we examined the therapeutic alternatives chosen after the Zypadhera shortage, and subsequent relapse and service disengagement. This natural experiment provided real-world evidence on the impact of drug shortages.
Method
We conducted a retrospective cohort study at the Mental Health Day Hospital of Mataró (Catalonia, Spain), which treats patients from a catchment area of about 400 000 inhabitants. All consecutive adult patients (≥18 years) receiving Zypadhera who were affected by the shortage between February and June 2025 were included.
Data were extracted from electronic health records by treating clinicians and anonymised before analysis. Sociodemographic characteristics, psychiatric diagnosis, substance use, concomitant psychotropic medications, time on Zypadhera before discontinuation, mean weekly dose in the previous 6 months, and therapeutic alternative chosen after the shortage (oral olanzapine, LAI aripiprazole or risperidone in situ microparticles (ISM)) were collected. Alternatives were selected by the treating psychiatrist according to routine clinical practice, considering previous treatment response, tolerability and adverse effects. Baseline illness severity was assessed using the CGI-Schizophrenia (CGI-SCH) overall severity score.
The primary outcome was a composite of relapse or disengagement from follow-up. Relapse was defined as acute psychotic decompensation requiring hospital admission or urgent intervention by crisis teams. Disengagement was defined as complete interruption of follow-up without agreement with the treatment team, operationalised as failure to attend three consecutive scheduled visits within a 6-month period. Time to event was calculated from the date of the last Zypadhera injection until either event or censoring at 6 months.
Statistical analyses included Kaplan–Meier survival analysis and log-rank test for comparison of alternatives, and Cox proportional hazards regression adjusted for age, sex, tobacco use, other substance use and duration of prior Zypadhera treatment. For descriptive comparisons, Mann–Whitney U-tests were used for continuous variables and chi-squared or Fisher’s exact test for categorical variables. Given the small sample size, analyses were considered exploratory.
Results
A total of 23 patients were included. The median age was 46 years (interquartile range: 37–60), and nine participants (39%) were women. Thirteen patients (57%) were smokers, and five (22%) reported cannabis use, three (13%) cocaine use and two (9%) alcohol use. Most patients (48%) had a diagnosis of schizophrenia; the remainder were diagnosed with schizoaffective disorder (26%), bipolar disorder (13%) or psychosis not otherwise specified (13%). Ten patients (43%) were not receiving any additional psychotropic medication, whereas the others were on concomitant antipsychotics, mood stabilisers or antidepressants. During the 6 months preceding the last Zypadhera injection, five patients (21.7%) had experienced a relapse (three during treatment with Zypadhera and two before Zypadhera had been started).
Therapeutic alternatives chosen during the shortage were: oral olanzapine in 18 patients (78%), LAI aripiprazole in four patients (17%) and risperidone ISM in one patient (4%). Baseline characteristics according to replacement treatment are presented in Supplementary Table 1 available at https://doi.org/10.1192/bjo.2026.12119. Baseline illness severity (CGI-SCH) and other baseline characteristics did not differ significantly between treatment groups. During follow-up, nine patients (39%) experienced relapse or disengagement, whereas 14 (61%) remained stable. Baseline characteristics according to outcome status are shown in Supplementary Table 2. Previous relapse was observed in three patients (33.3%) who subsequently experienced relapse or disengagement, compared with two of the 14 patients (14.3%) who remained stable. Kaplan–Meier survival analysis (Fig. 1) demonstrated significant differences between alternatives (log-rank χ 2 = 9.57, P = 0.008). All four patients who were switched to LAI aripiprazole experienced relapse or disengagement within 145 days. By contrast, five of the 18 patients switched to oral olanzapine experienced events (27.8%), and the single patient on risperidone ISM remained event-free.
Kaplan–Meier survival curves by therapeutic alternative following dicontinuation of Zypadhera. Time to event was defined as the number of days since the last Zypadhera injection until relapse or service disengagement. The composite outcome included acute psychotic relapse requiring admission or disengagement from follow-up. Patients were switched to oral olanzapine (n = 18), long-acting injectable (LAI) aripiprazole (n = 4), or risperidone in situ microparticles (ISM) (n = 1). Survival distributions differed significantly across groups (log-rank χ 2 = 9.57, P = 0.008). All patients switched to LAI aripiprazole experienced relapse or disengagement, whereas the majority of those switched to oral olanzapine remained stable. The single patient receiving risperidone ISM remained event-free throughout follow-up.
In Cox regression, female sex (hazard ratio = 11.8, P = 0.023), younger age (hazard ratio = 0.89 per year, P = 0.031) and tobacco use (hazard ratio = 12.5, P = 0.016) were independently associated with increased risk of relapse or disengagement. Other substance use and prior time on Zypadhera were not significantly associated with outcome. Patients with events tended to be younger (median 42.0 v. 49.5 years), had received higher mean weekly doses of Zypadhera before discontinuation (median 135.0 v. 87.5 mg) and were more frequently on concomitant mood stabilisers. These comparisons did not always reach statistical significance, probably owing to the limited sample size.
Discussion
This study provides real-world evidence on the clinical consequences of a nationwide shortage of olanzapine pamoate in Spain. Almost 40% of patients experienced relapse or disengagement during 6-month follow-up, highlighting their vulnerability when continuity of LAI treatment is disrupted.
Fewer relapse or disengagement events were observed among patients switched to oral olanzapine than among those switched to aripiprazole LAI, although treatment allocation was not random, and clinicians may have preferentially selected alternative LAIs for patients with perceived greater illness severity, poorer adherence or higher relapse risk. Nevertheless, baseline illness severity (CGI-SCH overall severity score) did not differ significantly between treatment groups. The observed differences may reflect baseline clinical characteristics rather than treatment effects, particularly given the small number of patients switched to LAI aripiprazole (n = 4). Approximately one-third of patients switched to oral olanzapine experienced relapse or disengagement within 6 months, a pattern similar to that reported during the COVID-19 pandemic. Reference Miron, Teodorescu, Ifteni, Irimie, Dima and Petric5
We observed that female sex, smoking and younger age were associated with higher risk of relapse or disengagement, consistent with evidence linking younger age and substance use to poorer adherence and outcomes in psychotic disorders. Reference Malla, Norman, Bechard-Evans, Schmitz, Manchanda and Cassidy6,Reference Pelayo-Terán, Gajardo Galán, de la Ortiz-García de la Foz, Martínez-García, Tabarés-Seisdedos and Crespo-Facorro7 The association with female sex is less well established and may reflect confounding by indication. Reference Sommer, Tiihonen, van Mourik, Tanskanen and Taipale8 Such confounding could not be excluded, Reference Groenendaal, Lynch, Dornbush, Klepacz and Ferrando9 as treatment allocation may have been influenced by unmeasured clinical factors despite no significant differences in baseline illness severity (CGI-SCH) between treatment groups. Other limitations of the study include its small sample size, single-centre design and short follow-up.
Our results have practical implications, as drug shortages can compromise patient care, increase risk of medication errors and consume substantial healthcare resources. 10 Psychiatric medications are particularly critical, as abrupt interruptions may precipitate relapse, hospital admission and patient distress. Health authorities should anticipate and mitigate shortages through coordinated policies and collaborative management strategies, 10,Reference Chen, Goold, Harrison, Ali, Makki and Kent11 and future studies should examine the clinical and economic consequences of shortages and identify system-level interventions to strengthen resilience in the psychiatric care supply chain.
Supplementary material
The supplementary material is available online at https://doi.org/10.1192/bjo.2026.12119
Acknowledgement
The authors used ChatGPT (OpenAI) to assist with language editing.
Author contributions
A.T., C.D.-A. and J.L.: study conception and design. A.T., C.D.-A., M.R., M.N., Y.M., V.G. and S.S.: data collection. J.L.: statistical analyses. A.T.: drafting of the manuscript. All authors contributed to data interpretation, critically revised the manuscript for important intellectual content, approved the final version and agree to be accountable for all aspects of the work.
Declaration of interest
J.L. and his research group (2021SGR-00789) are part of Red Temática Española de Investigación en Estrés, funded by the Ministerio de Ciencia e Innovación (grant RED2022–134191-T). J.L. has received honoraria for lectures or advisory boards from Janssen-Cilag, Otsuka, Lundbeck, Laboratorios Rovi, Angelini, Idorsia, Casen Recordati and Bausch Health.
Ethical standards
The study was approved by the local ethics committee (code 71/25); given the retrospective design, the requirement for informed consent was waived.
Open access
eLetters
No eLetters have been published for this article.