Pharmacological Management of Anxiety in End‐of‐Life Care: A Systematic Review of Benzodiazepines, Opioids, and Psilocybin

The inclusion criteria were studies involving adult patients (18 years or older) receiving end‐of‐life care due to terminal conditions, including cancer, heart failure, advanced neurological diseases or other terminal conditions. Eligible studies evaluated the efficacy and/or safety of any pharmacological intervention for managing anxiety in these patients, included a comparator, such as placebo, another anxiolytic medication, or usual care, and reported at least one outcome related to the efficacy or safety of the intervention. Suitable study designs included randomized controlled trials (RCTs), non‐randomized controlled studies, cohort studies, and case series providing comparative information.

The exclusion criteria for this review were studies that focused on pediatric patients or those that did not clearly differentiate between general palliative care and end‐of‐life care. Studies exclusively centered on populations with primary mental health conditions, non‐pharmacological interventions, or those that solely addressed other symptoms like pain or depression without focusing on anxiety were also excluded. Specifically, we defined 'primary mental health conditions' as those in which the main focus of treatment was chronic psychiatric disorders—such as recurring depression, severe anorexia leading to organ failure, or other conditions where the mental illness was the principal driver of the terminal status—rather than anxiety secondary to a terminal physical illness. Consequently, studies in which patients were treated primarily for long‐standing mental health issues, even if they also had a terminal illness, were not considered eligible. Additionally, studies lacking a comparator, single‐case studies, narrative reviews, editorials, and letters to the editor were excluded from the review.

To identify relevant studies, comprehensive searches were conducted in electronic databases including PubMed (at 12/06/2024), EMBASE (at 12/06/2024), Cochrane Library (at 16/06/2024) and ClínicalTrials.gov (at 24/06/2024) using the following search terms. "End of life care" OR "Terminal care" OR "Hospice care" OR "Terminally ill" OR "End‐stage disease" OR "Advanced disease" "Anxiety" OR "anxiousness" OR "nervousness" OR "panic" OR "fear" OR "worry" AND ("Pharmacological treatment" OR "Drug therapy" OR "Medication" OR "Pharmacotherapy" OR "Benzodiazepines" OR "Antidepressants" OR "SSRIs" OR "SNRIs" OR "MAOIs" OR "Tricyclic antidepressants" OR "Antipsychotics" OR "Anxiolytics" OR "Sedatives" OR "Tranquilizers") "Placebo" OR "Usual care" OR "standard care" OR "Control group" OR "Comparative" "Efficacy" OR "Effectiveness" OR "Outcomes" OR "Symptom relief" OR "Symptom reduction" OR "Quality of life" OR "QOL" OR "Side effects" OR "Adverse effects" OR "Safety"

Specific approaches were used for each search tool (Supporting Information : Appendix). S1

We also hand searched the reference lists of relevant studies, particularly those from previously published systematic reviews in related fields. This strategy aimed to identify additional articles that might have been missed in our initial database searches.

The literature search was limited to published original articles written in English language. No date limits were used.

The selection process for our systematic review began with one researcher extracting reports from MEDLINE via PubMed, EMBASE, Cochrane Library and ClinicalTrials. No automation tools were used beside Zotero's duplicate identifier. Titles and abstracts were screened manually to exclude irrelevant subjects.

Next, two independent researchers assessed the remaining reports for eligibility. Discrepancies were resolved through discussion, and if necessary, a third researcher was consulted to reach a consensus. This thorough process ensured the inclusion of relevant studies.

Data from the selected articles were collected by one researcher, and the results of this collection were reviewed and supervised by two additional researchers. The interventions were divided into two groups due to their similarities.

Bias assessment was conducted using RoB‐2 (Sterne et al. 2019) for randomized trials, RoB‐2 adapted for crossover trials and ROBINS‐I (Sterne et al. 2016) for non‐randomized studies in order to ensure the inclusion of relevant and high‐quality studies.

A total of 5 studies were included in this review (Figure 1). The systematic database search resulted in 392 records identified with 145 being removed before screening for being duplicates.

In the initial screening, 247 studies were analyzed for its title and abstract. After removing 210 studies with irrelevant subjects, 37 reports were assessed for eligibility by two independent review authors. Following the exclusion of 31 studies that did not fulfill our eligibility criteria (reasons of exclusion explained in Supporting Information S1: Appendix), 6 studies remained for inclusion. One of them was identified to be an ongoing study after contacting the authors, with the recruitment already completed but no data available (Atkin et al. 2023).

To assess the quality and risk of bias of the selected studies, tools appropriate to the design of each study were used to evaluate their specific characteristics as accurately as possible.

The study by Navigante et al. (2006) despite having randomized comparison groups and representing the largest sample size among the selected studies, raised some concerns, especially because it is a single‐blinded study where both the intervention administrator and the outcome assessors were aware of the interventions performed, which could introduce some risk of bias into the results. Additionally, there was no information in the paper or its registration about the planned statistical analyses (Figure 2).

In the study by Clemens and Klaschik (2011), in addition to the absence of randomization and lack of group division for better comparison, the fact that the outcome assessors were aware of the intervention also raises concerns, introducing concerns about a risk of bias to the study (Figure 2).

In the pilot study published by Grob et al. (2011) the small sample size is noteworthy. Participants acted as their own control, meaning that it lacks a control group and randomization. It was not clarified how much time elapsed between the first and second intervention sessions, raising the possibility of a carry‐over effect. The results were not reported numerically, only graphically. Additionally, measures of dispersion were not provided. We contacted the authors but were unable to obtain additional data. These factors raise concerns about the risk of bias in the selection of the presented results. Furthermore, there was also no information about prior plans for statistical analysis (Figure 2).

Ross et al. (2016) was a double‐blind study. However, the information contained in the supplementary material suggests that the psychoactive effects characteristic of the intervention may have compromised ideal blinding in the study, leading to some concern about bias in the results (Figure 2).

In the study by Griffiths et al. (2016), a control group that also received the psychoactive substance, but at a lower dose, was used. This approach appears to have ensured greater reliability in the results (Figure 2).

Traditionally, benzodiazepines and opioids have been employed to manage anxiety in end‐of‐life care, especially within the context of palliative sedation (Slawnych 2020). Despite their common usage, only two studies met the inclusion criteria for this review, highlighting the scarcity of research focused specifically on anxiety management in this setting with these agents. This gap points to a pressing need for more targeted investigations into their efficacy and safety, particularly for the management of anxiety symptoms. Both studies demonstrated consistent symptom relief with minimal adverse effects, suggesting these medications remain valuable therapeutic options.

In contrast to these conventional treatments, psychedelic substances, particularly psilocybin, have emerged in the last decade as promising alternatives in various psychiatric fields, including the treatment of depression and post‐traumatic stress disorder (Nutt 2019). In this review, psilocybin surfaced as a potential treatment option for anxiety in end‐of‐life care, offering a novel and potentially transformative approach to therapy in this context.

Psilocybin's appeal is further enhanced by its rapid onset of action and its apparently favorable safety profile (Whinkin et al. 2023), making it especially suitable for patients in end‐of‐life care who require fast and effective relief from anxiety. These characteristics differentiate it from traditional medications like antidepressants, which often take weeks to demonstrate efficacy.

Three of the studies selected in this systematic review reveal a convergence of data supporting the use of psilocybin in reducing anxiety in end‐of‐life care. These findings suggest a promising and consistent therapeutic potential for psilocybin, even with a small number of studies available, and highlight the need for further investigation in this area.

However, despite the encouraging results surrounding psilocybin, further robust clinical trials are essential to firmly establish its efficacy and safety in end‐of‐life anxiety treatment. The evidence remains preliminary and larger, well‐designed studies are needed to validate its therapeutic potential and guide clinical practice.

One notable challenge in psychedelic research, including psilocybin trials, is the difficulty in achieving effective blinding, which raises concerns about potential bias in the reported outcomes (Van Elk and Fried 2023). This methodological issue has been a point of concern in psychedelic research more broadly, and it underscores the need for innovative trial designs to ensure rigorous evaluation of these treatments.

The risk‐of‐bias assessments across the included studies showed significant methodological limitations, such as insufficient blinding, small sample sizes, and inconsistent reporting practices. These factors may lead to an overestimation of treatment effects, thus necessitating a cautious interpretation of the findings. Furthermore, the study by Grob et al. (2011) did not provide clear numerical data for its anxiety outcomes, relying instead on graphical representations, which adds another layer of uncertainty to our synthesis. Overall, while the convergence of results is encouraging, these issues highlight the need for more robust, well‐designed trials to firmly establish the efficacy and safety of these interventions.

Although the current evidence base is limited, the findings from this review offer some guidance for clinical decision‐making in end‐of‐life care, suggesting a potential benefit in stratifying treatment according to the predominant clinical presentation.

For anxiety that is closely intertwined with physical symptoms—particularly dyspnea—the studies by Navigante et al. (2006) and Clemens & Klaschik (2011) suggest that the concurrent administration of opioids (such as morphine) and benzodiazepines (e.g., midazolam or lorazepam) appears to be a viable and generally well‐tolerated strategy. Despite common concerns regarding respiratory depression, the reviewed data indicate that, when carefully titrated and monitored, this combination may alleviate dyspnea‐associated anxiety without significant compromise to oxygen saturation. Clinicians might consider this approach particularly for patients where 'air hunger' exacerbates anxious distress.

Conversely, for anxiety characterized primarily by existential distress, demoralization, and fear of death—symptoms often less responsive to traditional anxiolytics (Niles et al. 2021)—preliminary evidence suggests psilocybin could be a promising alternative. The studies reviewed indicate that psilocybin may facilitate rapid and sustained symptomatic relief. While regulatory frameworks currently limit its routine use, these findings support the consideration of psilocybin within the context of clinical trials or, where legal structures permit, compassionate use programs for patients with refractory existential suffering.

Our review presents important strengths and limitations that must be considered. A notable strength is the rigorous methodology, including adherence to PRISMA guidelines and prospective registration on PROSPERO, ensuring transparency and reproducibility. However, several limitations are also evident. The limited number of included studies, with relatively small sample sizes, may reduce the robustness and generalizability of the conclusions. Additionally, restricting the search to English‐language publications might have led to exclusion of relevant studies published in other languages. Furthermore, the psilocybin studies identified originated exclusively from the United States (Table 3) and predominantly included cancer patients, which may limit generalizability to other terminal conditions, such as advanced heart failure or neurological diseases, and to diverse cultural contexts. Lastly, regulatory and legal restrictions surrounding psilocybin pose significant challenges for the broader implementation of this treatment in clinical practice, highlighting the need for further context‐sensitive studies.

To strengthen the quality of evidence in this field, future research would benefit from addressing specific challenges related to diagnostic heterogeneity and study design.

One significant hurdle identified is the variability in the definition of 'end‐of‐life anxiety'. As noted during the review process, generic inclusion criteria may dilute observed treatment effects. We propose that future studies could be enhanced by adopting more precise diagnostic frameworks, such as those outlined in the ICD‐11. Specifically, differentiating between Generalized Anxiety Disorder and Adjustment Disorders or Disorders specifically associated with stress could help researchers recruit more homogeneous cohorts. Furthermore, establishing 'Existential Distress' as a validated inclusion criterion may improve the targeting of interventions like psychedelic‐assisted therapy.

Additionally, the challenge of maintaining effective blinding in psychedelic research remains a concern, as suggested by the risk‐of‐bias assessment in the included studies. To mitigate the potential for expectancy bias, future trial designs should ideally incorporate 'active placebos'—such as niacin or lower doses of the active compound, as utilized in some of the reviewed studies —to mimic physiological arousal without inducing the full psychedelic experience. Addressing these nosological and methodological factors may be crucial for facilitating high‐quality trials in this vulnerable population.

This systematic review provides an updated synthesis of current evidence on pharmacological interventions specifically targeting anxiety in end‐of‐life care, highlighting both traditional approaches and emerging treatments. Our findings suggest that benzodiazepines and opioids continue to represent important therapeutic options, although supported by limited empirical data, while psilocybin has emerged as a promising alternative, demonstrating rapid, sustained effects and a favorable safety profile. Nevertheless, significant methodological limitations and the restricted scope of available studies highlight considerable gaps in the evidence base, emphasizing the urgent need for further research in this area.

Future research should address critical unanswered questions, such as. How effective and safe are benzodiazepines and opioids specifically for anxiety, independent of their analgesic or sedative effects, in broader end‐of‐life populations beyond oncology? Could the promising effects observed with psilocybin in cancer patients generalize to individuals with other terminal illnesses, such as advanced cardiac or neurological conditions? What are the optimal dosing strategies, frequency of administration, and ideal timing for initiating psilocybin therapy to manage anxiety at the end of life? How do regulatory, cultural, and healthcare‐system barriers affect the feasibility and acceptability of implementing psilocybin‐based treatments in diverse global contexts?

Addressing these questions through robust future studies could significantly advance clinical practice and inform evidence‐based guidelines for managing anxiety in end‐of‐life care, enhancing both patient‐centered approaches and the quality of remaining life.