Efficacy and Safety of Olanzapine Long-acting Injection in Schizophrenia and Schizoaffective Disorder: A Systematic Review and Meta-analysis

Abstract

Olanzapine long-acting injection (LAI) is a second-generation antipsychotic used in patients with schizophrenia-spectrum disorders, particularly those with poor adherence. However, evidence on its comparative efficacy, safety, and tolerability remains limited. We conducted a systematic review and meta-analysis following PRISMA 2020 guidelines. Six databases (PubMed, Embase, Scopus, Web of Science, Cochrane CENTRAL, ProQuest) were searched through May 2025. We included randomized controlled trials (RCTs), quasi-experimental, and observational studies evaluating olanzapine LAI in adults with schizophrenia or schizoaffective disorder. Risk of bias was assessed using Revised Cochrane Risk-of-Bias Tool for Randomized Trials (Version 2), Risk Of Bias In Non-randomized Studies - of Interventions, and the Newcastle–Ottawa Scale. Data were synthesized using RevMan 5.4. Nine studies (n = 2,915) met inclusion criteria; only two RCTs were eligible for meta-analysis. Olanzapine LAI showed no significant difference from oral olanzapine in reducing Positive and Negative Syndrome Scale scores (mean difference = 1.41; 95% CI: −0.22 to 3.04; p = 0.09) or Clinical Global Impression–Severity scores (mean difference = 0.03; 95% CI: −0.15 to 0.21; p = 0.73). Relapse (OR = 1.15; 95% CI: 0.82 to 1.61), adverse events (OR = 1.12; 95% CI: 0.88 to 1.42), and weight gain were also similar. Across included studies, post-injection delirium/sedation syndrome was reported in 47 participants and resolved without hospitalization or long-term sequelae. Based on low-certainty evidence from limited RCTs, olanzapine LAI appears comparable in efficacy and tolerability to oral olanzapine. It may be considered for patients with adherence difficulties, but careful safety monitoring is required and further head-to-head trials are needed.

INTRODUCTION

Schizophrenia and schizoaffective disorders are chronic psychiatric illnesses that impair social, occupational, and physical functioning [1,2]. With a global prevalence of approximately 1%, schizophrenia contributes significantly to personal and societal burden [3]. Antipsychotics are the first-line treatment for both disorders [4,5], but medication nonadherence is reported in 40−60% of patients, leading to relapse, hospitalization, and poor long-term outcomes [6].

Long-acting injectable (LAI) antipsychotics were developed to improve adherence by ensuring sustained drug levels and reducing the need for daily oral dosing. Studies show LAIs reduce relapse and hospitalization rates and support better functional outcomes in schizophrenia [7,8]. They are also effective and well tolerated in schizoaffective disorder maintenance therapy [9].

Olanzapine, a second-generation antipsychotic, is effective for positive, negative, and mood symptoms in both disorders [10,11]. Its LAI form has shown benefits in relapse prevention and reduced hospitalizations [2,12,13]. Standard dosing regimens range from 150−405 mg every 2 to 4 weeks [14]. However, concerns remain, particularly the rare but serious risk of post-injection delirium/sedation syndrome (PDSS), which requires 3-hour post-injection monitoring [13]. Metabolic side effects are also notable [15]. Given the mixed findings across studies, a comprehensive review is needed. Previous reviews in this area had focused on comparing different antipsychotic LAIs together with oral or placebo, and reviews summarizing the evidence strength of olanzapine LAI alone are absent.

This systematic review evaluates the efficacy and safety of olanzapine LAI in adults with schizophrenia or schizoaffective disorder, focusing on symptom control, relapse prevention, functional outcomes, and adverse effects.

METHODS

The systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) statement [16]. The protocol for this systematic review was registered prospectively in the PROSPERO database (Registration ID: CRD42021275382) and is available at https://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42021275382. We included randomized controlled trials (RCTs), quasi-experimental studies, and observational studies evaluating olanzapine LAI in adults with schizophrenia or schizoaffective disorder. Eligible studies reported efficacy using validated clinical rating scales pre- and post-treatment. We excluded non-English studies, conference abstracts, unpublished data, case reports/series, studies with nonhuman subjects, and those involving patients with substance dependence (except tobacco) or concurrent use of oral and depot medications. Studies were grouped as: (1) olanzapine LAI vs. placebo, (2) olanzapine LAI vs. oral olanzapine, and (3) olanzapine LAI vs. other LAIs for quantitative synthesis. Only studies using standard olanzapine LAI doses (150−405 mg every 2−4 weeks) were included.

We searched six electronic databases, PubMed, Web of Science, Embase, Scopus, ProQuest, and Cochrane Central Register of Controlled Trials (CENTRAL), for relevant studies published up to May 20, 2025, and also screened the reference lists of included articles. Two independent reviewers (SS, RV) conducted the searches using a combination of keywords and MeSH terms related to: schizophrenia, mental disorders, personality disorders, psychotic disorders, schizophrenic disorder, schizoaffective disorder, mental illness, olanzapine, olanzapine long-acting injection, olanzapine depot, olanzapine pamoate, Zyprexa, LY-170053, Zolafren, and outcome terms including adverse effects, quality of life (QoL), Brief Psychiatric Rating Scale (BPRS), response rate, side effects, toxicity, functioning, Positive and Negative Syndrome Scale (PANSS). Boolean operators AND and OR were used to structure the search, and filters for human studies in English were applied. No age-based restrictions were applied during the database searches. The complete search strategy for each database is provided in Supplementary Table 1 (available online).

Search results were deduplicated using the Deduplicator tool from the Systematic Review (SR) Accelerator tool. Automatically excluded records were reviewed to avoid erroneous exclusions [17]. Titles and abstracts were screened independently by both reviewers using the Screenatron tool, and obvious exclusions were removed [18]. Conflicts were identified using the Disputatron tool and resolved through discussion, with final decisions made by a third reviewer (JTJ) when necessary. Full texts of potentially eligible articles were retrieved and assessed against predefined inclusion and exclusion criteria. Studies that met eligibility criteria were included in the final systematic review.

Two reviewers (SS and RV) independently extracted data from the included studies using a predesigned form, resolving discrepancies through discussion. Extracted information included study design, sample size, dosing and schedule of olanzapine LAI, treatment duration, comparator interventions, clinical outcomes such as changes in PANSS, BPRS, or Clinical Global Impression–Severity (CGI-S) scores, incidence and type of adverse effects including PDSS syndrome, treatment response rates, relapse and hospitalization data, and dropout rates with reasons for discontinuation. Outcomes were classified into efficacy (changes in symptom severity and number of responders), safety (adverse effects and weight changes), and feasibility (dropouts). The primary outcome was the change in PANSS or BPRS scores from baseline to end of treatment, while overall improvement was assessed using the CGI-S scale. Response rates were estimated from studies reporting a percentage reduction in symptom scores.

Risk of bias was assessed using the Revised Cochrane Risk-of-Bias Tool for Randomized Trials (Version 2) (RoB 2) for randomized and crossover trials, the Risk Of Bias In Non-randomized Studies - of Interventions (ROBINS-I) for nonrandomized studies, and the Newcastle–Ottawa Scale (NOS) for observational studies [19-21]. Both reviewers rated studies independently across all relevant domains, with disagreements resolved through discussion or by a third investigator (JTJ). The results were visualized using the robvis tool [22]. For continuous outcomes, standardized mean differences (SMDs) or mean differences (MDs) with 95% confidence intervals (CIs) were calculated. For dichotomous outcomes, such as response, adverse effects, and dropouts, pooled odds ratios (ORs) with 95% CIs were used.

Evidence Synthesis

Statistical synthesis was conducted when at least two studies reported comparable outcomes. Mean changes in symptom severity from baseline to endpoint, along with standard deviations (SD), were extracted for continuous outcomes. When SDs were not reported, they were calculated using baseline and final SDs following Cochrane methods, with a correlation coefficient of 0.7 imputed where unavailable [23]. For studies providing standard error (SE) instead of SD, the latter was derived using the formula SD = SE × √n. When only graphical data were available, numerical values were extracted using PlotDigitizer [24].

Meta-analysis was performed using Review Manager (RevMan) 5.4. Pooled SMDs and MD with 95% CIs were calculated for continuous outcomes using an inverse-variance method. Dichotomous outcomes (e.g., response, adverse effects, dropouts) were analyzed using Mantel-Haenszel ORs. A random-effects model was used to account for methodological heterogeneity. Statistical significance was set at p < 0.05. Forest plots were generated, and heterogeneity was assessed using Cochran’s Q and the I² statistic, interpreted as low (0−40%), moderate (30−60%), substantial (50−90%), or considerable (75−100%). Subgroup or sensitivity analyses were not conducted due to the limited number of studies. Publication bias was also not assessed, as only two studies were included in the meta-analysis.

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE approach) [25] was used to evaluate the certainty of evidence across five domains: study limitations, consistency, imprecision, indirectness, and publication bias. Certainty was rated as high, moderate, low, or very low, and a summary of findings table was created using GRADEpro GDT software.

RESULTS

We identified 9,177 records from six databases and removed 605 duplicates using the SR Accelerator tool. After screening 8,572 titles and abstracts, 8,419 records were excluded. The full texts of 151 articles were reviewed, and nine studies [2,12-14,26-30] were included in the systematic review, with two eligible for meta-analysis [12,13]. One RCT (Lauriello et al. [26], 2008) was identified based on abstract data; however, attempts to obtain the full manuscript through journal archives and author contact were unsuccessful, and the study was therefore included only in the qualitative synthesis and excluded from quantitative analysis and formal risk-of-bias assessment. The PRISMA flowchart of the screening and selection is presented in Figure 1.

Fig. 1.

PRISMA chart of the study selection.

PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Study Characteristics

The review included five clinical trials (four multicenter) and four observational studies (total n = 2,915). Sample sizes ranged from 9 to 1,065, with a mean participant age of 39.2 ± 11.1 years. Female representation was 33.5% (n = 976). Ethical approval was reported in all but one study; four were industry-funded. Full methodological details for Lauriello et al. [26], were unavailable due to abstract-only data. Additional characteristics are provided in Supplementary Table 2 (available online).

Interventions and Outcome Measures

Details of interventions and outcomes are summarized in Table 1. In the four observational studies, olanzapine LAI was administered at doses ranging from 210 mg every 2 or 4 weeks to 405 mg every 4 weeks. Some studies used fixed regimens (e.g., 210 mg every 4 weeks in Fefeu et al. [2] and Mauri et al. [28]), while others applied flexible dosing. None included formal comparator groups. In the five clinical trials, doses ranged from 45 to 405 mg at 2−4-week intervals, generally adjusted to match 10−20 mg/day of oral olanzapine. Oral olanzapine served as the comparator in two trials (Kane et al. [12], Detke et al. [13]), a placebo was used in one (Lauriello et al. [26]), and two trials (McDonnell et al. [27], Mitchell et al. [14]) were open-label without control arms.

Table 1.

Summary of included studies

Study ID and study design	Diagnosis and sample size	Intervention and comparison	Treatment duration and outcome time point	Primary outcomes	Secondary outcomes	Adverse effects and dropouts
Grover et al. [30], 2020 Retrospective study	Schizophrenia (n = 40)	Intervention: olanzapine LAI Dose: 210 mg/2 weeks, 300 mg/2 weeks, 300 mg/4 weeks, 405 mg/4 weeks Comparison: NIL	Duration: 3 years Outcome time points: not mentioned	Changes in symptom severity were measured using CGI-S.CGI score reduced from 5.8 (0.7) to 2.7 (1.1) after initiation of olanzapine LAI (t = 16.41; p < 0.001) CGI-S scores decreased in both patients with treatment-resistant schizophrenia (TRS) (t = 8.51, p < 0.001) and without TRS (t = 14.29, p < 0.001).	Hospitalization: reduction in the number of hospitalizations	PDSS: 1 patient Dropouts = 9 Patient reluctance, 5; lack of efficacy, 3; affordability, 1
Fefeu et al. [2], 2018 Retrospective study	Schizophrenia and schizoaffective disorder (n = 40)	Intervention: olanzapine LAI Dose: 210 mg/4 weeks Comparison: any other antipsychotic	Duration: 1 year Outcome time points: 1 year before and after starting LAI	Hospitalization and bed days Hospitalization decreased by 65.7% after initiating olanzapine LAI, from 2.10 ± 1.76 to 0.72 ± 1.35 (p < 0.0001) Bed days decreased by 86.2% after initiating olanzapine LAIs, from 58.00 ± 52.45 to 7.97 ± 15.78 days (p < 0.0001)	BPRS scores decreased by 47.0% (p < 0.0001) and CGI-SCH scores decreased by 63.2% (p < 0.0001). Average costs dropped by 86.2% after starting olanzapine LAIs treatment, while outpatient visits didn’t show a significant change.	PDSS: 5 patients No change in BMI and other metabolic parameters Dropouts: not mentioned
Kasinathan et al. [29], 2016 Retrospective study	Schizophrenia and schizoaffective disorder (n = 9)	Intervention: olanzapine LAI Dose: not given Comparison: NIL	Duration: 1 year Outcome time points: 1 year before and after starting LAI	Changes in BPRS scores BPRS reduction from 58 to 37 (Z = −2.67, p = 0.008)	Hospital admission days were significantly reduced (Z = −2.37, p = 0.018). Days incarcerated in prison were also significantly reduced (Z = −2.023, p = 0.043).	PDSS: 3 patients No significant increase in body weight Dropouts: Nil
Mauri et al. [28], 2015 Prospective study	Schizophrenia and schizoaffective disorder (n = 25) Final (n = 21)	Intervention: olanzapine LAI Dose: 210 mg/4 weeks, 300 mg/4 weeks, 405 mg/4 weeks Comparison: NIL	Duration: 9 months Outcome time points: 12 weeks (T2) and 36 weeks	BPRS total score improved by 19.57% from 44.44 to 37.52 at 12 weeks (p < 0.001) and by 26.96% from 44.44 to 32.22 at 36 weeks.	The PANSS total score showed significant improvement, reducing from 76.73 to 62.94 (18.06% improvement) at 12 weeks and to 58.22 (23.86% improvement) at 36 weeks (p < 0.001).	Four patients out of 21 developed mild sedation Dropouts: 4 Sedation: 1 Patient refusal: 3
McDonnell et al. [27], 2014 Open label	Schizophrenia and schizoaffective disorder (n = 931)	Intervention: olanzapine LAI Comparison: Nil Dose of LAI: 45 to 405 mg Frequency: every 2, 3, or 4 weeks	Treatment duration: 6 years Primary outcome: Positive and Negative Syndrome Scale (PANSS) and CGI-S Outcome time points: every 6 months	PANSS total score: mean change from baseline to endpoint was 0.30 (SD = 16.4, p = 0.590) CGI-S score: mean change: −0.17 (significant improvement starting at week 3, p < 0.001)	Most patients were satisfied with olanzapine LAI (73.2%), preferred it over oral medications (66.8%), and reported fewer side effects (73.3%).	Common AEs: weight gain, anxiety, insomnia, somnolence. Serious AEs: 18.3% of patients. Deaths: 11 (1 possibly related) Injection-site AEs: 3.9% PDSS events: 36 patients, all recovered. Dropouts: 561 (60.3%), 77 (8.3%) due to AEs Other AEs: no significant difference between groups in metabolic profile (extrapyramidal symptoms [EPS])
Detke et al. [13], 2014 Open label	Schizophrenia (n = 524) Intervention = 264 Comparison = 260	Intervention: olanzapine LAI Comparison: oral olanzapine Dose of LAI: 150, 210, 255, 300, 345, or 405 mg Frequency: 4 weeks Dose of oral olanzapine: 5−20 mg	Treatment duration: 2 years Primary outcome: time to all-cause discontinuation Outcome time points: weekly	All-cause discontinuation: median time was 645 days for LAI and 678 days for oral, with no significant difference (p = 0.612). Relapse: relapse rates were 20.1% for LAI and 18.5% for oral, no significant difference (p = 0.659). PANSS total score change: LAI, −0.82 (SE = 1.2); oral, −1.14 (SE = 1.2); no significant difference (p = 0.834).	Caregiver burden: significant improvement in both groups, no significant group difference (p = 0.558). Patient satisfaction: no difference Mean weight gain: LAI: 2.0 kg (SD = 7.2); oral: 2.2 kg (SD = 7.3); no significant difference (p = 0.777).	Common AEs: increased weight, insomnia, anxiety, somnolence. Serious AEs: 11.3% Deaths: 2 (1 in each group, unrelated) Injection-site AEs: 6.4% PDSS events: none Dropouts (2 years): LAI 53.8%, oral 51.2% Other AEs: no significant difference between groups in metabolic profile, electrocardiogram, and EPS
Mitchell et al. [14], 2013 Open label	Schizophrenia (n = 281)	Intervention: olanzapine LAI Comparison: Nil Dose of LAI: biweekly (100−300 mg), 3-week (210−255 mg), and 4-week (210−405 mg)	Treatment duration: 24 weeks Primary outcome: safety and efficacy of single and multiple doses of LAI Outcome time points: every 4 weeks Blood level: 0, 6, 24, 48, 72, and 96 hours after injection, and then weekly for 4 weeks	Single-dose olanzapine LAI (50−450 mg) showed dose-proportional plasma concentrations remaining measurable for 4 weeks. Multiple-dose LAI (150−310 mg/ 2 weeks; 405 mg/4 weeks): achieved steady-state concentrations comparable to 5−20 mg/day oral olanzapine.	BPRS total score: Decreased significantly by a mean of 2.68 points from baseline (17.27) at endpoint (p < 0.001). CGI-S score: showed significant improvement with a mean decrease of 0.23 points from baseline (3.39) at endpoint (p < 0.001).	Common AEs: 77.7% (at least one AE), injection site pain, anxiety, insomnia, headache, somnolence, and sedation Serious AEs:7.7% Deaths: none Injection-site AEs: (20.2%) PDSS events: 1, recovered in 24 hours Dropouts: 7(27.8%) Other AEs: significant weight gain and changes in blood glucose levels
Kane et al. [12], 2010 RCT	Schizophrenia (n = 1,065) Intervention = 743 Comparison = 322	Intervention: olanzapine LAI Comparison: oral olanzapine Dose of LAI: low dose: 150 mg/2 weeks; medium dose: 405 mg/ 4 weeks; high dose: 300 mg/2 weeks; very low dose: 45 mg/4 weeks; oral dose: 10, 15, or 20 mg/day	Treatment duration: 24 weeks Primary outcome: rates of and time to psychotic exacerbation Outcome time points: weekly for the first 12 weeks and every 2 weeks thereafter during the 24-week double-masked period	Psychotic exacerbation: at 24 weeks, most patients (oral 93%, therapeutic LAI 84−95%) remained exacerbation-free, with standard LAI doses superior to very low dose (69%; p < 0.01) and comparable to oral. PANSS total score: the very low LAI dose worsened PANSS total (+7.2), while therapeutic LAI doses (−0.1 to +2.7) and oral (−2.5) maintained or slightly improved scores.	BPRS total score: score worsened with very low LAI (+4.6) but showed improvement or minimal change across therapeutic LAI (−1.05 to +2.3) and oral (1.1). CGI-S score: very low LAI worsened CGI-S (+0.3), while therapeutic LAI (−0.0 to +0.1) and oral (−0.1) groups maintained or slightly improved severity.	Common AEs: insomnia, weight increase, anxiety, nasopharyngitis, and somnolence. Serious AEs: 57 patients (5%) experienced at least one serious AE Deaths: none Injection-site AEs: 30 (3%) PDSS events: 2, fully recovered Dropouts: no information Other AEs: Weight: weight gain was greater for the high-dose group than the low-dose group (p = 0.04). No change in lab values and metabolic profile
Lauriello et al. [26], 2008 RCT	Schizophrenia (n = 404)	Intervention: olanzapine LAI Comparison: placebo LAI Dose of LAI: 210 mg/2 weeks, 300 mg/2 weeks, 405 mg/4 weeks Control group: placebo/2 weeks	Treatment duration: 8 weeks Primary outcome: mean baseline-to-endpoint change in PANSS total score. Outcome time points: weekly, until 8 weeks	PANSS total score: all olanzapine LAI regimens resulted in significantly greater mean baseline-to-endpoint decreases in PANSS total scores relative to placebo (all pvalues < 0.001), with separation from placebo observed as early as 3−7 days.	Clinical improvement (CGI-I): rates of clinical improvement (≤ 3) were significantly higher for all olanzapine LAI groups compared to placebo (p < 0.001).	Common AEs: sedation and increased appetite. Serious AEs: no information Deaths: not mentioned Injection-site AEs: no information PDSS events: no information Dropouts: no information Other AEs: olanzapine LAI significantly increased weight gain, cholesterol, and triglycerides compared to placebo.

This table details the study characteristics, intervention, and outcome assessment methods, and primary and secondary outcomes for each included study.

LAI, long-acting injection; CGI-S, Clinical Global Impression–Severity; PDSS, post-injection delirium/sedation syndrome; BPRS, Brief Psychiatric Rating Scale; CGI-SCH, Clinical Global Impression–Schizophrenia; BMI, body mass index; SD, standard deviation; SE, standard error; AEs, adverse events; RCT, randomized controlled trial.

All studies used validated scales, such as the PANSS, BPRS, and CGI to evaluate symptom severity and clinical improvement. Functional outcomes, including hospitalization rates, inpatient stay duration, and incarceration days, were reported in observational studies. Fefeu et al. [2] also assessed healthcare utilization. Follow-up durations ranged from 12 weeks to 3 years. Clinical trials additionally evaluated relapse, treatment discontinuation, caregiver burden, and pharmacokinetics. Adverse events (AEs), including sedation, PDSS, weight gain, and metabolic disturbances, were systematically monitored across all studies

Risk of Bias

Risk of bias assessment using the NOS revealed moderate to high risk across all observational studies. Grover et al. [30], Fefeu et al. [2], and Kasinathan et al. [29] scored 2/9, indicating serious limitations in selection, absence of comparator groups, and unblinded outcome assessment. Mauri et al. [28] scored 4/9, reflecting moderate risk with better outcome reporting but similar methodological concerns. Among clinical trials, the open-label studies by McDonnell et al. [27] and Mitchell et al. [14] were rated as high risk due to a lack of blinding and comparators. Kane et al. [12] and Detke et al. [13] were assessed as having “some concerns,” reflecting moderate risk. Risk of bias for Lauriello et al. [26] was judged as unclear/not assessable because only abstract data were available and the full text could not be retrieved. Detailed assessments are presented in Table 2 (NOS Scale) and visualized in Figure 2A (RoB 2) and 2B (ROBINS-I).

Table 2.

Risk of bias for the observational studies assessed using Newcastle Ottawa Scale (NOS)

Domains of bias	Selection					Comparability			Exposure			Total score
Study ID	Case definition	Representativeness	Selection of controls	Definition of controls		Important factor	Other factors		Ascertainment of exposure	The same methods for cases and controls	Non-response rate
Grover et al. [30], 2020	*	*	nil	nil		nil	nil		c	nil	b	2/9
Fefeu et al. [2], 2018	*	*	nil	nil		nil	nil		d	nil	c	2/9
Kasinathan et al. [29], 2016	*	*	nil	nil		nil	nil		c	nil	c	2/9
Mauri et al. [28], 2014	*	*	nil	nil		nil	nil		*	nil	*	4/9

*, domain meets NOS criteria; nil, criterion not met; b, partial compliance (e.g., structured interview, limited quality); c, weak method (e.g., not blinded, unclear description); d, very limited evidence or self-report.

Fig. 2.

(A) Risk of bias of included randomized control trials assessed using RoB 2 Tool. (B) Risk of bias of included non randomized trials assessed using ROBINS-I Tool.

RoB 2, Revised Cochrane Risk-of-Bias Tool for Randomized Trials (Version 2); ROBINS-I, Risk Of Bias In Non-randomized Studies - of Interventions.

Treatment Efficacy and Adverse Effects

Olanzapine LAI consistently demonstrated efficacy across both observational studies and clinical trials. Grover et al. [30] reported a significant reduction in CGI-S scores (from 5.8 ± 0.7 to 2.7 ± 1.1; p < 0.001), while Fefeu et al. [2] noted 47% and 63.2% reductions in BPRS and CGI-S scores, respectively (p < 0.0001). Mauri et al. [28] observed 27% and 24% reductions in BPRS and PANSS scores over 9 months, and Kasinathan et al. [29] reported significant symptom reduction (p = 0.008) alongside fewer hospitalizations and incarcerations. Among clinical trials, Lauriello et al. [26] found significantly greater PANSS reductions with LAI compared to placebo (p < 0.001), evident within one week. Kane et al. [12] showed comparable efficacy between LAI and oral olanzapine, though lower LAI doses were less effective. Mitchell et al. [14] reported significant improvements on BPRS and CGI-S (p < 0.001), while McDonnell et al. [27] showed non-significant PANSS changes but modest CGI-S gains. Detke et al. [13] found no significant difference between LAI and oral olanzapine in PANSS, relapse, or caregiver burden.

Olanzapine LAI was generally well tolerated. In observational studies, AEs were mild and infrequent. PDSS was reported in nine cases, all resolving without hospitalization. Mild sedation and no significant weight or metabolic changes were noted. Dropout rates ranged from 0−16%, mostly due to logistical reasons. In clinical trials, AEs aligned with known olanzapine effects, weight gain, sedation, insomnia, and injection-site reactions. PDSS occurred in 38 participants across three trials, resolving without long-term effects. Serious AEs (5−18%) included two deaths (one per arm in Detke et al. [13]), neither related to treatment. Dropouts ranged from 27.8% to 60.3%, often due to side effects or nonadherence.

Quantitative Synthesis

Two RCTs [12,13] comparing olanzapine LAI with oral olanzapine were included in the meta-analysis, comprising 1,305 patients (723 LAI; 582 oral). Both studies were conducted in outpatient settings and assessed as having “some concerns” in risk of bias. Lauriello et al. [26] were excluded due to a lack of full-text data and unassessable bias. Given that only two RCTs were eligible for pooling, the meta-analysis was exploratory in nature and intended to summarize available comparative data rather than provide definitive estimates of treatment effect.

Efficacy Outcomes

Meta-analysis of PANSS score reduction showed no significant difference between olanzapine LAI and oral olanzapine (MD = 1.41; 95% CI: −0.22 to 3.04; p = 0.09), with no observed heterogeneity (I² = 0%, p = 0.46) (Fig. 3A). Similarly, pooled analysis of CGI-S scores showed no significant difference (MD = 0.03; 95% CI: −0.15 to 0.21; p = 0.73), with low heterogeneity (I² = 17%, p = 0.27) (Fig. 3B). One study marginally favored LAI and the other oral olanzapine, but neither reached statistical significance. Ninety two relapses occurred in the LAI group and 71 occurred in the oral group. The pooled OR was 1.15 (95% CI: 0.82 to 1.61; p = 0.43), indicating no significant difference in relapse risk (Fig. 3C). Heterogeneity was minimal (I² = 0%, p = 0.81), suggesting consistency between studies.

Fig. 3.

Forest plots for the efficacy outcomes comparing olanzapine long-acting injection (LAI) with oral olanzapine. (A) The figure displays, for each study included in the meta-analysis, the summary statistics (mean, standard deviation [SD], and sample size) for the olanzapine LAI and olanzapine oral groups, and the mean difference and its 95% confidence interval (CI) for the continuous outcome, PANSS total score change. (B) The figure displays for each study included in the meta-analysis the summary statistics (mean, SD, and sample size) for the olanzapine LAI and olanzapine oral groups, and the mean difference and its 95% CI for the continuous outcome, CGI-S score change. (C) The figure displays for each study included in the meta-analysis the number of events and sample size for the olanzapine LAI and olanzapine oral groups, and the odds ratio (OR) and its 95% CI for the dichotomous outcome, relapse rate.

IV, inverse variance; M-H, Mantel-Haenszel; PANSS, Positive and Negative Syndrome Scale; CGI-S, Clinical Global Impression–Severity.

Tolerability and Safety of Olanzapine Long-acting Injection

Tolerability was evaluated in two RCTs [12,13] using three outcomes: incidence of AEs, treatment discontinuation, and weight gain. 314 AEs were reported in the LAI group and 259 in the oral group. The pooled OR was 1.12 (95% CI: 0.88 to 1.42; p = 0.37), indicating no significant difference between formulations, with low heterogeneity (I² = 0%; p = 0.71) (Fig. 4A). Discontinuation occurred in 199 LAI-treated patients and 129 oral-treated patients. The pooled OR was 1.32 (95% CI: 0.88 to 1.96; p = 0.17), which was also not statistically significant. Moderate heterogeneity was observed (I² = 57%; p = 0.13), possibly due to study design differences (Fig. 4B). Weight gain did not differ significantly between groups, with a pooled MD of −0.18 kg (95% CI: −1.05 to 0.69; p = 0.69), and negligible heterogeneity (I² = 0%; p = 0.96) (Fig. 4C).

Fig. 4.

Forest plots for the safety outcomes comapring olanzapine long-acting injection (LAI) with oral olanzapine. (A) The figure displays for each study included in the meta-analysis the summary statistics (number of events and sample size) for the olanzapine LAI and olanzapine oral groups, and the odds ratio and its 95% confidence interval (CI) for the dichotomous outcome, any adverse effects. (B) The figure displays for each study included in the meta-analysis the summary statistics (number of events and sample size) for the olanzapine LAI and olanzapine oral groups, and the odds ratio and its 95% CI for the dichotomous outcome, all cause discontinuation. (C) The figure displays for each study included in the meta-analysis the summary statistics (mean, standard deviation [SD], and sample size) for the olanzapine LAI and olanzapine oral groups, and the mean difference and its 95% CI for the continuous outcome, weight gain.

M-H, Mantel-Haenszel; IV, inverse variance.

Sensitivity and subgroup analyses were not performed due to the limited number of included trials, which precluded robust exploration of heterogeneity or moderator effects. Also, we did not look at publication bias using a funnel plot, as two studies were included.

Certainty of Evidence

The certainty of evidence for all outcomes was rated as low, primarily due to concerns about risk of bias (both studies had “some concerns”) and imprecision (wide CIs overlapping the null effect). Specifically, the evidence for symptom reduction measured by PANSS and CGI-S, relapse prevention, weight gain, AEs, and all-cause discontinuation was graded as low-certainty, indicating that the true effect may differ substantially from the observed estimates. The GRADE summary of findings is depicted in Table 3.

Table 3.

GRADE evidence profile: efficacy and safety of olanzapine long-acting injection (LAI) compared to oral olanzapine for schizophrenia and schizoaffective disorders

Outcome	Certainty assessment								No. of patients (%)			Effect		Certainty
	No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations		Efficacy of olanzapine LAI	Placebo		Relative (95% CI)	Absolute (95% CI)
PANSS total reduction	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousb	None		723	582		-	MD 1.41 higher (0.22 lower to 3.04 higher)	⊖⊖◯◯ Lowa,b
CGI-S reduction	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousb	None		723	582		-	MD 0.03 higher (0.15 lower to 0.21 higher)	⊖⊖◯◯ Lowa,b
Relapse rate	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousc	None		92/723 (12.7)	71/582 (12.2)		OR 1.15 (0.82 to 1.61)	16 more per 1,000 (from 20 fewer to 61 more)	⊖⊖◯◯ Lowa.c
Weight gain	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousb	None		723	582		-	MD 0.18 lower (1.05 lower to 0.69 higher)	⊖⊖◯◯ Lowa,b
Any adverse events	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousc	None		314/723 (43.4)	259/582 (44.5)		OR 1.12 (0.88 to 1.42)	28 more per 1,000 (from 31 fewer to 87 more)	⊖⊖◯◯ Lowa,c
All cause discontinuation	2	Randomised trials	Seriousa	Not serious	Not serious	Seriousc	None		199/723 (27.5)	129/582 (22.2)		OR 1.32 (0.88 to 1.96)	52 more per 1,000 (from 21 fewer to 137 more)	⊖⊖◯◯ Lowa,c

PANSS, Positive and Negative Syndrome Scale; CGI-S, Clinical Global Impression–Severity; CI, confidence interval; MD, mean difference; OR, odds ratio; -, not available.

^aBoth studies had some concerns in the overall risk of bias assessment. ^bThe CI for pooled MD crossed zero, showing both harm and benefit. ^cThe CI for the pooled OR crossed 1, thus showing no clear benefit.

DISCUSSION

While previous systematic reviews have compared various LAI antipsychotics with oral formulations and placebo, the present review is the first to exclusively evaluate and synthesize evidence on the efficacy and safety of olanzapine LAI in the treatment of schizophrenia and schizoaffective disorder.

Efficacy of Olanzapine Long-acting Injection

Olanzapine LAI was associated with improvements in psychotic symptoms in both RCTs and observational studies. In RCTs, olanzapine LAI was shown to be non-inferior to oral olanzapine in improving core symptoms of schizophrenia, as measured by PANSS and CGI-S scores. The pooled meta-analysis of two RCTs revealed no statistically significant difference between the LAI and oral formulations in PANSS reduction (MD = 1.41; 95% CI −0.22 to 3.04) or overall clinical improvement on CGI-S (MD = 0.03; 95% CI −0.15 to 0.21). Relapse rates were also similar (OR = 1.15; 95% CI 0.82 to 1.61); however, these non-significant findings were graded as low certainty, indicating that the true effect may be substantially different. Given that the quantitative synthesis was based on only two trials, these pooled estimates should be interpreted cautiously as supportive summaries rather than definitive measures of comparative treatment effect.

Observational studies supported these findings and further highlighted functional improvements with olanzapine LAI, including significant reductions in hospitalization rates, duration of illness episodes, and severity of symptoms. Studies by Grover et al. [30], Fefeu et al. [2], and Mauri et al. [28] reported marked improvements in BPRS, PANSS, and CGI scores over follow-up periods ranging from several months to three years. These real-world findings suggest that olanzapine LAI may offer sustained symptom control and improve treatment continuity, particularly in patients with adherence difficulties. In interpreting these findings, it is important to recognize that RCTs provide comparative estimates of efficacy and safety, whereas observational studies primarily inform long-term tolerability, adherence, and real-world functioning; consequently, evidence from nonrandomized studies should be regarded as hypothesis-generating rather than confirmatory.

Safety and Tolerability of Olanzapine Long-acting Injection

The overall safety profile of olanzapine LAI was comparable to oral olanzapine. Across the studies included in this review, commonly reported adverse effects included sedation, weight gain, and metabolic disturbances, consistent with the known side effect profile of olanzapine. Importantly, the pooled analysis of RCTs revealed no statistically significant difference in the incidence of AEs (OR = 1.12; 95% CI, 0.88 to 1.42), discontinuation due to any cause (OR = 1.32; 95% CI, 0.88 to 1.96), or weight gain (MD = −0.18 kg; 95% CI, −1.05 to 0.69) between the LAI and oral formulations. A unique safety concern with olanzapine LAI is PDSS, a rare but potentially serious AE. Although PDSS cases reported across included studies were self-limiting and resolved without long-term sequelae, its clinical significance lies in the unpredictability of onset and the requirement for mandatory post-injection observation. The estimated incidence of PDSS is approximately 0.07% per injection, necessitating a minimum three-hour monitoring period after each injection. While this requirement is effective in mitigating risk, it introduces practical and logistical challenges in outpatient settings and may limit broader implementation. These requirements may affect clinic workflow, staffing, resource allocation, and patient acceptance, and should be considered when selecting patients and planning service delivery for olanzapine LAI.

Our review findings align with the existing evidence for the safety and tolerability of olanzapine LAI. A recent network meta-analysis of 115 RCTs found that LAI antipsychotics are better than placebo, but not superior to their oral preparation in terms of efficacy and safety [31]. In particular, they found comparable efficacy for olanzapine LAI and oral, consistent with our findings. A large network meta-analysis by Ostuzzi et al. [32] found that olanzapine LAI and oral olanzapine had similarly high efficacy in preventing relapse in schizophrenia-spectrum disorders. Although LAIs did not consistently outperform oral antipsychotics, the findings support their use in patients where adherence is a concern. Importantly, the study also reported no significant safety disadvantages with LAIs, aligning with our review’s findings that olanzapine LAI is both effective and well-tolerated [32]. Detke et al. [33] reported that olanzapine LAI was comparable to oral olanzapine in efficacy and onset of action for acute schizophrenia, with symptom improvement seen by day 3 and moderate-to-large effect sizes on BPRS and PANSS. The safety profile was similar between groups, with low rates of extrapyramidal symptoms and no cases of PDSS reported.

Strengths and Limitations of the Study

The major strength of the review is that we included both RCTs and observational studies. The review used a comprehensive search strategy across six databases with PRISMA-compliant reporting. We performed a meta-analysis on the quantitative synthesis of the key outcomes across the studies. A key limitation is the very small number of RCTs eligible for meta-analysis. With only two trials available for pooling, the quantitative synthesis provides limited incremental inferential power beyond individual studies and precluded the conduct of subgroup or sensitivity analyses. In addition, clinical heterogeneity related to dosing schedules, follow-up duration, and patient populations, along with methodological heterogeneity across study designs, further limits the interpretability and generalizability of the findings. A moderate to high risk of bias was observed in the observational studies due to the lack of blinding and retrospective study designs, and the strength of evidence was ‘low’ for most comparisons. These methodological limitations increase the likelihood of overestimation of treatment effects in nonrandomized studies and necessitate cautious interpretation of observed benefits. One RCT [26] could not be fully analyzed due to the unavailability of the full text. We also could not assess publication bias and bias due to missing data, thus reducing the evidence strength. Finally, most included studies involved patients with schizophrenia, with limited representation of schizoaffective disorder, and we were unable to perform quantitative comparisons between olanzapine LAI and placebo or other LAI antipsychotics, limiting generalizability and direct comparative conclusions regarding efficacy and safety.

Conclusion

This systematic review and meta-analysis found no statistically significant differences between olanzapine long-acting injection and oral olanzapine across pooled efficacy and safety outcomes. All meta-analytic estimates were graded as low certainty, indicating that the true effect may be substantially different from the observed results and that conclusions should be interpreted with caution. Evidence from observational studies should be regarded as supportive and hypothesis-generating rather than confirmatory.

Implications for Practice and Policy, and Future Research

Olanzapine LAI represents a viable treatment option for patients with schizophrenia and schizoaffective disorder who experience adherence challenges. It offers comparable efficacy and safety to oral olanzapine with added pharmacokinetic benefits of depot administration. However, the post-injection three-hour observation for the risk of PDSS may cause practical challenges in outpatient settings, requiring careful infrastructure assessment for monitoring and patient education. Future research comparing head-to-head trials of olanzapine LAI with other antipsychotic LAIs is needed to see the comparative effectiveness. Studies on patients with schizoaffective disorder with olanzapine LAI would be beneficial.