Cardiovascular Effects of Non-Selective Monoamine Oxidase Inhibitors and Intranasal Esketamine Combination in Depression – A Quasi-Experimental Design with Bayesian Analyses

• Abstract
• Introduction
• Methods
• Data preparation
• Descriptive and demographic analysis
• Evolution of cardiovascular parameters during esketamine session and comparison between groups
• Bayesian random model
• Markov Chain Monte Carlo (MCMC) network random model
• Estimation of factors related to blood pressure elevation and high-risk situations
• Markov chain Monte Carlo mixed model
• Results
• Descriptive and demographic data
• Evolution of cardiovascular parameters during esketamine sessions and comparison between groups
• Bayesian random model
• Bayesian network random model
• Estimation of factors related to blood pressure elevation and high-risk situations - Markov Chain Monte Carlo mixed model
• Mild BP elevation
• Moderate blood pressure elevation
• Hypertensive crisis
• Discussion
• Funding
• CRediT authorship contribution statement
• Data availability
• References

Abstract

Introduction

Ketamine and esketamine (ESK) offer new treatment options for resistant depression. Unlike traditional antidepressants, they can be used in combination with non-selective monoamine oxidase inhibitors (NS-MAOI) without the risk of serotonergic syndrome. However, potential sympathomimetic synergy may lead to elevated blood pressure (BP). This series investigates whether cardiovascular parameters (heart rate, systolic [SP], and diastolic [DP] pressures) increase during ESK sessions and whether the ESK+NS-MAOI combination is associated with BP elevations.

Methods

We collected cardiovascular parameters for ESK sessions conducted between 2018 and 2022. These parameters were measured at baseline and every 30 min for 2 h. Patients were categorized into two non-equivalent groups: those receiving ESK alone and those receiving ESK+NS-MAOI. A Bayesian random model was used to estimate the evolution of these parameters, while a Bayesian hierarchical model assessed factors contributing to BP elevation.

Results

ESK sessions (n=193), of which 116 involved NS-MAOI, were performed in 13 patients. SP, DP, and heart rate showed peak increases during sessions, but these changes were not clinically significant (SP+8.68 mmHg, DP+6.57 mmHg, and heart rate+3.5 bpm). No significant differences were found between the ESK-alone and ESK+NS-MAOI groups. The combination was not identified as a factor linked to BP elevations.

Discussion

These findings align with previous research on ketamine derivatives and suggest minimal peripheric sympathomimetic synergy with NS-MAOI. Bayesian models were used to account for biases intrinsically related to these ecological data and provide a foundation for future open adversarial collaborations.
Registration NCT05530668

Introduction

Glutamatergic modulators such as ketamine and esketamine (ESK) have recently renewed the therapeutic armamentarium for treatment-resistant depression (TRD), offering new strategic opportunities [1]. They pave the way for the combination of antidepressants with non-selective monoamine oxidase inhibitors (NS-MAOI). Historically, NS-MAOIs are one of the few triaminergic antidepressants, but their use has become marginal [2] [3]. This is primarily related to dietary restrictions (e. g., tyramine-free diet) and pharmacological limitations they impose, including incompatibility with classical antidepressants inhibiting serotonin reuptake [4] [5] [6]. This creates a significant challenge for psychiatrists, who may feel “trapped” if patients show only partial improvement with NS-MAOI alone. By combining ESK and NS-MAOI, one can expect at least a dual antidepressant effect—ESK through the glutamatergic system, and NS-MAOI through monoaminergic neurotransmission (NS-MAOIs inhibit the degradation of serotonin, norepinephrine, and dopamine). Some authors have even hypothesized a potential synergistic effect between ESK and dopaminergic agents, including NS-MAOIs [7]. Several theoretical mechanisms could support this assumption: on the one hand, the increased dopaminergic tone may enhance glutamatergic activity at both the synaptic and astrocytic levels and, on the other hand, a body of evidence suggests that ketamine derivatives, particularly ESK, may induce a dopamine release [8] [9] [10] [11]. This latter may be further enhanced by inhibition of dopamine degradation through MAOIs – and even by the low amphetaminic properties of certain MAOIs’ (e. g., tranylcypromine) [12] [13].

However, combining these two drugs could also have a synergistic effect on the cardiovascular system, particularly blood pressure (BP). Ketamine, especially ESK, even at low doses, is known to induce a release of catecholamines and to inhibit their reuptake, leading to tachycardia and elevated BP [14] [15] [16] [17]. This elevation is typically moderate (usually<10 mmHg for systolic and diastolic pressure [SP and DP, respectively]); however, it raises concerns among regulatory authorities, which contraindicate its use in “patients for whom an increase in blood pressure or intracranial pressure poses a serious risk” [18]. On their side, NS-MAOIs inhibit catecholamine modulation, and NS-MAOI-facilitated hypertensive crises are thought to result from sympathomimetic mechanisms [19] [20] (for reviews, see [6] [21]). Due to the risk of sympathomimetic synergy on BP, regulatory authorities have issued special warnings regarding their combined use [18].

While some case reports and previous series offer reassuring data, they remain relatively scarce and inconsistent (a recent overview is available in Ludwig et al. [22]). A retrospective study showed no increase in the frequency of hemodynamic events when NS-MAOIs were combined with anesthetic drugs, including ketamine, in surgical settings [23]. In psychiatric settings (i. e., repeated ketamine or esketamine sessions at sub-anesthetic doses), most studies report no significant changes in hemodynamic parameters, although some describe mild increases in both systolic pressure (SP) and diastolic pressure (DP) [24] [25] [26] [27]. Only one series described a severe hypertension crisis (up to 180/110 mmHg) [26]. The largest series to date reported a mild (3 mmHg) yet significant increase in SP following intravenous or subcutaneous ESK administration [22]. Only one case report involves intranasal ESK in combination with tranylcypromine and reports a mild BP increase that did not reach the threshold for a hypertensive crisis [28].

Exploring the real-world use and short-term effects of ESK surpasses conventional analytical tools, as the treatment is delivered through iterative sessions, often with varying session numbers per patient and no carry-over effects. Analyzing individual sessions may increase statistical power, but it could underestimate the fixed effects (i. e., patient-related effects), making the results non-generalizable beyond the exact recruited population. Conversely, averaging data per individual may skew the analysis (e. g., data from patients with fewer sessions could outweigh others) and reduce statistical power for rare events like NS-MAOI combinations. Bayesian models can help address these challenges, as they allow beliefs to be updated and can be adapted through productive adversarial collaboration.

This study focuses on two main questions: 1) Does combining NS-MAOI and intranasal ESK lead to significant changes in hemodynamic parameters during an ESK session? and 2) To what extent does the ESK+NS-MAOI combination, along with other baseline factors, affect the frequency and/or severity of BP elevation during an ESK session? We propose using random and hierarchical Markov chain Monte Carlo (MCMC) models to address these questions and discuss how these tools handle statistical constraints, with a focus on the generalizability of the results.

Methods

All patients treated with intranasal ESK at the non-invasive neuromodulation center of Strasbourg (CEMNIS), the outpatient clinic of the TRD recourse center of the University Hospitals of Strasbourg (“centre expert dépression résistante d’Alsace” – CEDRA), between January 1, 2018, and March 1, 2022, were included in the analysis. Data were retrospectively collected from the patients’ medical files unless they expressed opposition to the use of their data.

The data of interest included:

• Medical and demographic data: sex, age, episode duration, psychiatric and medical comorbidities, previous and current treatments (with duration and doses when available). NS-MAOI doses were converted into tranylcypromine equivalents when applicable [29].

• ESK session data: number of sessions, overall treatment duration, ESK dose, hemodynamic parameters (systolic pressure [SP], blood pressure [BP], and heart rate [HR]) before ESK administration and at 30, 60, 90, and 120 min post-administration, and adverse effects (including hypertensive crises, defined as SP or DP above 180 and 120 mmHg, respectively) [30].

• Follow-up data: Montgomery and Asberg Depression Rating Scale (MADRS) and Patient Health Questionnaire 9-item version (PHQ-9) scores before treatment and at 7, 15, and 30 days after the introduction of ESK or combined treatment [31] [32].

Since these patients were managed under routine care, ESK was openly administered and adjusted according to standard protocols in the department and the European Medicines Agency recommendations for ESK [18]. According to the French regulation on retrospective studies, a written notice explaining the study was given to all patients, precising their right to oppose the use of their data. This study was approved by the local ethics committee (CE-2022–47) and was registered on ClinicalTrials.gov (NCT05530668).

Data preparation

For each patient, data from all ESK sessions were averaged to obtain mean variables for statistical analysis. The variables obtained included SP variation at 30 (ΔSP+30), 60 (ΔSP+60), 90 (ΔSP+90), and 120 (ΔSP+120) minutes after ESK administration. Two global variables were also calculated: mean SP variation (meanΔSP) and the maximum SP increase observed during each session, averaged for each patient (maxΔSP). The same procedures were applied for diastolic pressure (i. e., ΔDP+30, ΔDP+60, ΔDP+90, ΔDP+120, meanΔDP, and maxΔDP) and heart rate (i. e., ΔHR+30, ΔHR+60, ΔHR+90, ΔHR+120, meanΔHR, and maxΔHR).

Each session was assigned to either the NS-MAOI+ESK group or the ESK-alone group. If a patient had received both treatments, they were considered two distinct patients in the descriptive analysis and Bayesian random model but as one patient when analyses allowed for it (i. e., in MCMC network random and mixed models).

Descriptive and demographic analysis

Quantitative variables were expressed in median [first quartile, third quartile]. Groups were compared by Bayesian Mann-Whitney U-test for independent samples with a data augmentation algorithm with five chains of 10,000 random iterations. Qualitative variables were expressed in N (proportion). Comparisons were made using the Bayesian multinominal test for contingency tables with JASP 0.16.4 [33].

We reported the plausibility of results using Bayes factors (BF₁₀) that indicates the ratio of the amount of evidence supporting the plausibility of a bilateral H1 hypothesis (alternative to the null H0 hypothesis [no change]). The evidence supporting H1 was considered possible if BF₁₀≥3, as strong if BF₁₀≥10, very strong if BF₁₀≥30, and extreme if BF10≥100 [34].

Evolution of cardiovascular parameters during esketamine session and comparison between groups

Bayesian random model

To estimate the evolution of cardiovascular parameters during ESK sessions in each group, considering that patients underwent a different number of ESK sessions (resulting in a statistical fixed effect), we introduced averaged variables per patient (i. e., the mean and 95% confidence interval of all sessions for each patient) into a random-effects Bayesian model using JASP 0.16.4.[33]. In other words, we treated each patient as we would with studies in a meta-analysis, comparing baseline hemodynamic parameter values with those recorded during the session. This model used a standard Cauchy prior (location=0, scale=0.707). Results were expressed as Effect Size [95% credibility interval] and BF₁₀ .

Markov Chain Monte Carlo (MCMC) network random model

To compare the two groups, where data for most patients only allowed for pre-administration versus post-administration comparisons, we used a network analysis. We introduced averaged variable per patient (i. e., the mean and 95% confidence interval of all sessions of each patient) into a Bayesian network random model using MCMC in R 4.2.1, with RStudio 2022.07.2–576, GEMTC 1.0–2 and JAGS 4.3.1 [35] [36] [37] (see R-code in Supplementary Material 1 – part 1). For patients who underwent both ESK-alone and ESK+NS-MAOI sessions, variables were introduced as a direct comparison between treatments. A difference was considered plausible if the 95% credibility interval did not contain 0. Consistency was assessed using the nodesplit function.

Estimation of factors related to blood pressure elevation and high-risk situations

Markov chain Monte Carlo mixed model

For each ESK session, we arbitrarily defined three dichotomic variables according to the occurrence of a mild BP elevation (i. e., SP≥140 mmHg and/ or DP≥90 mmHg), a moderate BP elevation (i. e., SP≥160 mmHg and/ or DP≥100 mmHg) and a hypertensive crisis (i. e., SP≥180 mmHg and/ or DP≥110 mmHg).

Each of the three variables was independently introduced, considering all sessions individually, in a Bayesian MCMC hierarchical mixed model using R, R-studio, and JAGS (see R-code in Supplementary Material 1 – part 2) with a null prior. The fact that the session involved a particular patient, age, use of antihypertensive medication, use of NS-MAOI+ESK combination, and baseline hemodynamic parameters for each ESK session were introduced as covariates (the first one adjusted the analysis for the patient-related fixed effect, while the latter were potential explicative variables). Interactions between the use of NS-MAOI+ESK combination and baseline hemodynamic parameters were estimated. To facilitate interpretation, results were expressed in odds ratio (OR) [95% plausibility interval] and with the probability that OR is positive (i. e.,>1) (interOR), expressed in percentage.

A similar model was used for each factor identified as relevant in the previous analysis. Clinically relevant subcategories were proposed for these factors (e. g., 5-year intervals for age, 10 mmHg for SP and DP, 10 bpm for HR) and were introduced as covariates. The fact that a particular patient was involved in the session was also introduced as a covariate.

Results

Descriptive and demographic data

All 13 patients (median age 56.3 [39.8, 66.5] years) who received intranasal ESK were included in the analysis. In total, 193 intranasal ESK sessions were conducted, with 116 of those sessions involving NS-MAOI. The median number of ESK sessions was 8 [3] [10]. Five patients (38.5%) received NS-MAOI (four patients [30%] received tranylcypromine, and one patient [7%] received phenelzine), with a median dose of 50 [30,75] mg/day. Three patients (23%) started treatment with NS-MAOI during the period covered by ESK sessions (each was treated as two distinct patients in the analysis). General characteristics and demographic features are provided in [Table 1], and there were no significant differences between the groups.

There were reductions in the median MADRS scores at 30 days in both groups: (22 [12] [29] vs. 29 [25.5, 33.5] for the overall population, 16.5 [10.25, 25] vs. 23 [22] [29] in the NS-MAOI+ESK group, and 23 [21] [29] vs. 29 [26.5, 34] in the ESK-alone group). Similar trends were observed with PHQ-9 scores (9.25 [14, 16.75] vs. 18 [15, 20.5] in the overall population, 13 [10] [16] vs. 16 [14] [18] in the NS-MAOI+ESK group, and 15 [10] [16] vs. 20 [18.25, 21.75] in the ESK-alone group). Group comparisons did not reveal any significant difference at 7 and 30 days (see [Table 1]).

No patient from the ESK+NS-MAOI group, but one patient from the ESK-alone group, experienced a hypertensive crisis (occurring in only one session, representing<1% of the total number of sessions). No patient exhibited symptoms suggestive of serotonin syndrome. Further details regarding the hypertensive crisis and adverse events (AEs) are reported in Supplementary Material 2 – Box S1 and Table S1.

Evolution of cardiovascular parameters during esketamine sessions and comparison between groups

For this analysis, data of only 12 patients (2 in the NS-MAOI+ESK group, 7 in the ESK-alone group, and 3 patients who received both treatments) were used because of the lack of variance in the data of one patient who had only benefited from one ESK session.

Bayesian random model

Variations in cardiovascular parameters following ESK administration for each group are illustrated in [Fig. 1] (detailed results are provided in Supplementary Material 2 – Table S2). Overall, the mean trends indicated stability in cardiovascular parameters, with only a mild meanΔDP observed during ESK sessions in the overall population (+2.22 [0.04, 4.36] mmHg, BF10=7.58) and in the NS-MAOI+ESK group (+1.89 [0.74, 3] mmHg, BF10=49.91). When considering only the maximum values reached during each session, plausible increases were observed, including:

• SP augmentation (maxΔSP) in the overall population (+8.68 [2.37, 13.28] mmHg, BF10=51.08) and in the NS-MAOI+ESK group (+7.97 [0.99, 11.13] mmHg, BF10=9.3).

• DP augmentation (maxΔDP) in the overall population (+6.57 [5.58, 7.80] mmHg, BF10=206711), in the ESK-alone group (+6.70 [4.68, 9.31] mmHg, BF10=219.07), and in the NS-MAOI+ESK group (+6.46 [5.24, 7.69] mmHg, BF10=1388.82).

• HR augmentation (maxΔHR) in the overall population (+3.5 [0.67, 5.87] bpm, BF10=21.12) and in the ESK-alone group (+3.22 [0.13, 5.42] bpm, BF10=11.63).

Bayesian network random model

Differences between the ESK+NS-MAOI group and the two others are detailed in [Fig. 2]. There was no plausible difference. The most important difference (11 [-11; 37] mmHg) was between maxSP observed during sessions in the NS-MAOI+ESK group and baseline. No model presented inconsistency (detailed results in Supplementary Material 2 – Table S3).

Estimation of factors related to blood pressure elevation and high-risk situations - Markov Chain Monte Carlo mixed model

Data from only 183 sessions, including all 13 patients, were analyzed due to missing data.

Mild BP elevation

It was unlikely that the combination of NS-MAOI+esketamine influenced the risk of mild BP elevation (OR=0.66 [0.04, 12.34]) with a 39% probability of a positive OR. However, baseline SP and DP were plausibly related to mild BP elevation (respectively: OR=9.45 [1.72, 56.97], 100% probability of a positive OR, and OR=1.29 [1.16, 1.48], 100% probability of a positive OR). Baseline HR was plausibly negatively related to mild BP elevation (OR=0.93 [0.86, 0.996], 2% probability of a positive OR). No interaction between the NS-MAOI+ESK combination and any other factors studied was plausibly related to the occurrence of mild BP elevation. Detailed results are provided in [Table 2A].

We established five levels of baseline SP (i. e., level 1:<100 mmHg, level 2: [100; 109] mmHg, level 3: [110; 119] mmHg, level 4: [120; 129] mmHg, level 5:≥130 mmHg). Only level 5 was plausibly associated with an increased risk of mild BP elevation (OR=5.97 [1.98, 58.89], 100% probability of a positive OR). We set four levels for baseline DP (i. e., level 1:<60 mmHg, level 2: [60; 69] mmHg, level 3: [70; 79] mmHg, level 4: [80; 89] mmHg, level 5:≥90 mmHg). Only level 4 was plausibly associated with an increased risk of mild BP elevation (OR=11.94 [1.99, 73.88], 100% probability of a positive OR). We also established five levels for baseline HR (i. e., level 1:<60 bpm, level 2: [60; 69] bpm, level 3: [70; 79] bpm, level 4: [80; 89] bpm, level 5:≥90 bpm). Only level 1 was plausibly associated with an increased risk of mild BP elevation (OR=15.93 [1.88, 151.84], 99.5% probability of a positive OR). Detailed results are provided in [Table 2B].

Moderate blood pressure elevation

It was unlikely that the combination of NS-MAOI+esketamine influenced the risk of moderate BP elevation (OR=0.87 [0.04, 17.51] with a 46% probability of a positive OR). A high baseline HR negatively influenced the risk of moderate BP elevation (OR=0.61 [0.24, 0.95] with a 1% probability of a positive OR). An interaction between the NS-MAOI+esketamine combination and baseline DP was plausibly related to the risk of moderate BP elevation (OR=3.163 [1.09, 19.39] with a 99% probability of a positive OR). Detailed results are provided in [Table 3A].

No level of baseline HR was associated with moderate BP elevation, and no level of baseline DP was related to moderate BP elevation in each group. However, level 4 DP presented a high probability of being positive only in the NS-MAOI+ESK group (OR=5.96 [0.56, 68.4] with a 92.9% probability of a positive OR). Detailed results are provided in [Table 3B].

Hypertensive crisis

The analysis could not be realized because of the lack of events of interest (only one hypertensive crisis in the ESK-alone group).

Discussion

The main results of this series suggest that although cardiovascular parameters are likely to peak during ESK sessions, these changes are of little clinical significance and appear independent of the combination with NS-MAOI. The influence of the NS-MAOI+ESK combination on BP elevation events, which might require clinical attention, is likely negligible or marginal. This is consistent with previous case reports involving this combination [28]. Our findings are comparable to those obtained with intravenous and subcutaneous ESK, as well as with ESK combined with traditional antidepressants [22] [38]. Relatively to mean maximum BP changes observed by Popova and colleagues during their RCT on intranasal esketamine in depression, our results are comparable to those in patients receiving esketamine (i. e.,+11.6 mmHg for SP and+8.1 mmHg) and slightly higher than those in patients receiving placebo (i. e.,+5 mmHg for SP and+4.5 mmHg for DP) [39]. All these variations are also consistent with physiological variability, particularly in the case of stressful events, and measure-related variability [40]. Moreover, we found no evidence that NS-MAOIs played a role in BP elevation in our patients. These results suggest there is no clinically significant sympathetic synergy between NS-MAOIs and ESK, supporting the cardiovascular safety of the combination.

Counterintuitively, patients with low baseline HR were more likely to exceed the mild BP threshold. This altered HR/BP relationship may imply baroreflex impairment, although disruption in BP regulation is commonly associated with higher HR in depression [41]. Alternative hypotheses may involve an allostatic override of an effective baroreflex, the effects of age and treatment on baroreflex and the cardiovascular system, or even a reduction in the adaptability of the hypothalamic-pituitary-adrenal axis in certain depression subtypes [41] [42] [43] [44]. While this result is not generalizable, it highlights the need for more comprehensive physiological assessments before ESK treatment in future studies.

Interestingly, age was not associated with a higher risk of BP elevation, despite previous reports suggesting that the elderly have reduced cardiovascular adaptability during sessions [45] [46]. Similarly, prior diagnosis of hypertensive disease did not influence cardiovascular parameters, but all our hypertensive patients were treated with antihypertensive drugs, and their condition was controlled in all except one patient who exceeded the hypertensive crisis threshold once. This unique event obscured the fact that this patient frequently approached the threshold (with SP peaks above 174 mmHg in 50% of her sessions, evenly distributed across her nine sessions). However, all hypertensive patients were assigned to the ESK-alone group. It is worth noting that the most vulnerable patients were excluded since our procedure requires a cardiological evaluation before considering treatment with ESK or NS-MAOIs. As a result, the NS-MAOI+ESK combination was not proposed for patients with cardiovascular disease or fragile conditions.

Like many similar reports, our series fits a quasi-experimental design, with non-equivalent groups and interrupted time-series. In such series, the population often reflects real-world clinical practice, where treatment decisions are influenced by practical considerations and the preferences of practitioners and patients [22]. Additionally, a portion of the observed variance was due to individual characteristics and shared covariates, which could be adjusted by most regression models. Intra-patient variance may also have been affected by the number of observations. The coexistence of intra- and inter-individual heterogeneity justified the use of a random-effects model when possible [47].

The main question may be: are these models, typically used in meta-analyses, suitable for analyzing data at the patient level? The difference between this approach and a meta-analysis is not mathematical but conceptual, as data are grouped by patients instead of studies. The specific configuration of our series, with baseline comparisons for both groups and between-condition comparisons for some patients, creates a loop of evidence—a fundamental structure for network analysis [48]. transitivity must be maintained to ensure the validity of these models; in other words, patients must not be too heterogeneous to be compared [48]. The use of standardized treatment procedures limits this heterogeneity, but differences between patients should still be accounted for. Our consistency analysis, along with the coherence between the random model and the network random model results, suggests that the data are sufficiently homogeneous for the results to be reliable.

The hierarchical MCMC mixed model provided notable flexibility, allowing for the consideration of both session-related data for each session and parameters shared across multiple sessions (i. e., individual-related features). It also accommodated the fact that a single patient could be in one group, the other group, or move between both groups without splitting the data. These models may be more appropriate when the ratio of observations per patient to the number of patients is high, making them suitable for small series with frequently repeated events. However, introducing too many covariates to control for biases may weaken the model, whereas quasi-experimental designs gain accuracy and validity by considering as many biases as possible. We chose to consider baseline values, age, and pre-existing hypertensive disease, as these factors were identified as being related to BP increase in previous studies and were key in our medical decision to combine NS-MAOIs and ESK. Other potential factors, such as anxiety and the intensity of dissociative effects, were not considered [45] [46] [49]. The hierarchical MCMC mixed model remains a mixed-effects model; the fixed effect is only partially corrected, and the results are not easily generalizable. This model may be useful for further exploration of specific factors by repeating the analysis while refining hypotheses to guide future study designs.

Bayesian analysis relies on the core principle of Bayesian statistics—updating probability based on prior knowledge and beliefs in clinical research. These analyses provide models and results that can be shared and progressively enriched by the research community. Both results and models can be challenged with additional datasets, used as priors, and contribute to refining new probabilities and hypotheses. This is especially valuable for redefining adversarial collaboration in experimental research, as well as in observational studies on current care practices [50] [51]. In unusual clinical situations, such as the combination of ESK and NS-MAOI, data may be scarce and the value of investing resources in randomized or double-blind trials may be limited. In this context, the goal is not to provide definitive results, which would be difficult to obtain, but to contribute to the refinement of current hypotheses and models while keeping them generalizable and usable to inform current practice.

Our series has some limitations. The extrapolation of these results is constrained by the retrospective and non-randomized design, and the small sample size. Some advanced analyses like meta-regression would have required higher-quality data, particularly multiple assessments before each ESK session. There was no difference in clinical outcome, potentially related to the specific clinical profile of our patients – likely chronic, highly resistant, and often associated with comorbidities – commonly linked to poor outcomes, including with ESK [52]. Only three patients were treated with ESK alone and then with ESK+NS-MAOI, although this subgroup may be of interest as these patients may be under their own control. The detailed data from these patients highlights that individual trajectories can be heterogeneous even if there was no plausible intra-individual difference between the two treatment modalities (see Supplementary Material 2 - Figure S1). These latest results remind us that reassuring population-level results can mask individual outliers, and that close monitoring of these procedures by trained and vigilant professionals remains necessary.

Our results suggest that combining ESK and NS-MAOI is unrelated to a greater increase in BP during the ESK session, nor with hypertensive crises. These findings align with previous research on ketamine derivatives and suggest minimal peripheric sympathomimetic synergy between ESK and NS-MAOI. Bayesian models were used to account for biases intrinsically related to this type of real-world data. The latter are likely to become more common in psychiatry, with the emergence of treatments with similar modes of administration and relevant immediate effects, e. g., electroconvulsive therapy, neuromodulation or psychedelic drugs. The Bayesian framework also allows improvement and informs further analyses in different populations, offering a way to rethink future open adversarial collaborations, particularly in fields where reproducibility is challenging.

Funding

This work was supported by the University Hospital of Strasbourg (DRCI – Direction de la Recherche Clinique et de l’Innovation).

CRediT authorship contribution statement

LD-J: conceptualization, project administration, methodology, investigation, data curation, formal analysis, writing- original draft preparation, writing- reviewing and editing, visualization. SL: methodology, investigation, data curation, formal analysis, writing- reviewing and editing. CB, OM, AO, IH: investigation, writing- reviewing and editing. BM, GM, HJ, BG: formal analysis, writing- reviewing and editing. JF: supervision, methodology, investigation, formal analysis, writing- reviewing and editing.

Data availability

The data that support the findings of this study are not openly available due to legal limitations related to the use of clinical data. Anonymized data used for this work may be available from the corresponding author upon reasonable request and only with permission from the University Hospitals of Strasbourg Ethics Board. All participants must be informed and express their non-opposition to the new use – documented in their medical file – before any sharing or re-use of anonymized data.

Acknowledgments

The authors express their gratitude to the personnel of the CEMNIS (Noninvasive Neuromodulation Center of Strasbourg): Mrs. Haumesser, Schilt, Marzak, Schneider, Berentz, and Mr. Ricka and Prudhomme, and to the DRCI – Direction de la Recherche Clinique et de l’Innovation of the University Hospital of Strasbourg: Mrs. Soavelo and Mr. Chayer.

• References

• 1 McIntyre RS, Rosenblat JD, Nemeroff CB. et al. Synthesizing the evidence for ketamine and esketamine in treatment-resistant depression: An international expert opinion on the available evidence and implementation. Am J Psychiatry 2021; 178: 383-399
  CrossrefPubMedSearch in Google Scholar
• 2 Culpepper L. Reducing the burden of difficult-to-treat major depressive disorder: Revisiting monoamine oxidase inhibitor therapy. Prim Care Companion CNS Disord 2013; 15
  CrossrefPubMedSearch in Google Scholar
• 3 Asnis GM, Henderson MA. EMSAM (deprenyl patch): How a promising antidepressant was underutilized. Neuropsychiatr Dis Treat 2014; 10: 1911-1923
  CrossrefPubMedSearch in Google Scholar
• 4 Van den Eynde V, Abdelmoemin WR, Abraham MM. et al. The prescriber’s guide to classic MAO inhibitors (phenelzine, tranylcypromine, isocarboxazid) for treatment-resistant depression. CNS Spectr 2022; in press 1-14
  CrossrefPubMedSearch in Google Scholar
• 5 Gillman PK. A review of serotonin toxicity data: Implications for the mechanisms of antidepressant drug action. Biol Psychiatry 2006; 59: 1046-1051
  CrossrefPubMedSearch in Google Scholar
• 6 Gillman PK. A reassessment of the safety profile of monoamine oxidase inhibitors: Elucidating tired old tyramine myths. J Neural Transm (Vienna) 2018; 125: 1707-1717
  CrossrefPubMedSearch in Google Scholar
• 7 Cook J, Halaris A. Adjunctive dopaminergic enhancement of esketamine in treatment-resistant depression. Prog Neuropsychopharmacol Biol Psychiatry 2022; 119: 110603
  CrossrefPubMedSearch in Google Scholar
• 8 Di Maio V, Ventriglia F, Santillo S. A model of dopamine modulated glutamatergic synapse. Biosystems 2015; 136: 59-65
  CrossrefPubMedSearch in Google Scholar
• 9 Bezerra TO, Roque AC. Dopamine facilitates the response to glutamatergic inputs in astrocyte cell models. PLoS Comput Biol 2024; 20: e1012688
  CrossrefPubMedSearch in Google Scholar
• 10 Kokkinou M, Ashok AH, Howes OD. The effects of ketamine on dopaminergic function: Meta-analysis and review of the implications for neuropsychiatric disorders. Mol Psychiatry 2018; 23: 59-69
  CrossrefPubMedSearch in Google Scholar
• 11 Hashimoto K, Kakiuchi T, Ohba H. et al. Reduction of dopamine D2/3 receptor binding in the striatum after a single administration of esketamine, but not R-ketamine: A PET study in conscious monkeys. Eur Arch Psychiatry Clin Neurosci 2017; 267: 173-176
  CrossrefPubMedSearch in Google Scholar
• 12 Ulrich S, Ricken R, Adli M. Tranylcypromine in mind (Part I): Review of pharmacology. Eur Neuropsychopharmacol 2017; 27: 697-713
  CrossrefPubMedSearch in Google Scholar
• 13 Kegeles LS, Abi-Dargham A, Zea-Ponce Y. et al. Modulation of amphetamine-induced striatal dopamine release by ketamine in humans: Implications for schizophrenia. Biol Psychiatry 2000; 48: 627-640
  CrossrefPubMedSearch in Google Scholar
• 14 Timm C, Linstedt U, Weiss T. et al. Sympathomimetic effects of low-dose S(+)-ketamine. Effect of propofol dosage. Anaesthesist 2008; 57: 338-346
  CrossrefPubMedSearch in Google Scholar
• 15 Lundy PM, Lockwood PA, Thompson G. et al. Differential effects of ketamine isomers on neuronal and extraneuronal catecholamine uptake mechanisms. Anesthesiology 1986; 64: 359-363
  CrossrefPubMedSearch in Google Scholar
• 16 Zheng W, Cai DB, Xiang YQ. et al. Adjunctive intranasal esketamine for major depressive disorder: A systematic review of randomized double-blind controlled-placebo studies. J Affect Disord 2020; 265: 63-70
  CrossrefPubMedSearch in Google Scholar
• 17 Kryst J, Kawalec P, Pilc A. Efficacy and safety of intranasal esketamine for the treatment of major depressive disorder. Expert Opin Pharmacother 2020; 21: 9-20
  CrossrefPubMedSearch in Google Scholar
• 18 European Medicine Agency. Spravato - authorization for use and product information [Internet]. 2019 Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/spravato
  PubMedSearch in Google Scholar
• 19 Bovill JG. Adverse drug interactions in anesthesia. J Clin Anesth 1997; 9: 3s-13s
  CrossrefPubMedSearch in Google Scholar
• 20 Cuthbert MF, Vere DW. Potentiation of the cardiovascular effects of some catecholamines by a monoamine oxidase inhibitor. BrJ Pharmacol 1971; 43: 471p-472
  PubMedSearch in Google Scholar
• 21 Trendelenburg U, Langeloh A, Bönisch H. Mechanism of action of indirectly acting sympathomimetic amines. Blood Vessels 1987; 24: 261-270
  CrossrefPubMedSearch in Google Scholar
• 22 Ludwig VM, Sauer C, Young AH. et al. Cardiovascular effects of combining subcutaneous or intravenous esketamine and the MAO inhibitor tranylcypromine for the treatment of depression: A retrospective cohort study. CNS Drugs 2021; 35: 881-892
  CrossrefPubMedSearch in Google Scholar
• 23 van Haelst IM, van Klei WA, Doodeman HJ. et al. Antidepressive treatment with monoamine oxidase inhibitors and the occurrence of intraoperative hemodynamic events: A retrospective observational cohort study. J Clin Psychiatry 2012; 73: 1103-1109
  CrossrefPubMedSearch in Google Scholar
• 24 Bartova L, Vogl SE, Stamenkovic M. et al. Combination of intravenous S-ketamine and oral tranylcypromine in treatment-resistant depression: A report of two cases. Eur Neuropsychopharmacol 2015; 25: 2183-2184
  CrossrefPubMedSearch in Google Scholar
• 25 Bottemanne H, Bonnard E, Claret A. et al. Ketamine and monoamine oxidase inhibitor combination: Utility, safety, efficacy?. J Clin Psychopharmacol. 2020 40.
  PubMedSearch in Google Scholar
• 26 Katz RB, Toprak M, Wilkinson ST. et al. Concurrent use of ketamine and monoamine oxidase inhibitors in the treatment of depression: A letter to the editor. Gen Hosp Psychiatry 2018; 54: 62-64
  CrossrefPubMedSearch in Google Scholar
• 27 Wang JCC, Swainson J. The concurrent treatment with intravenous ketamine and an irreversible monoamine oxidase inhibitor for treatment-resistant depression without hypertensive crises. J Clin Psychopharmacol 2020; 40: 515-517
  CrossrefPubMedSearch in Google Scholar
• 28 Dunner DL, Fugate RM, Demopulos CM. Safety and efficacy of esketamine nasal spray in a depressed patient who was being treated with tranylcypromine: A case report. Neurol Psychiat Br Research 2020; 36: 30-31
  CrossrefPubMedSearch in Google Scholar
• 29 Fiedorowicz JG, Swartz KL. The role of monoamine oxidase inhibitors in current psychiatric practice. J Psychiatr Pract 2004; 10: 239-248
  CrossrefPubMedSearch in Google Scholar
• 30 Chobanian AV, Bakris GL, Black HR. et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 2003; 42: 1206-1252
  CrossrefPubMedSearch in Google Scholar
• 31 Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med 2001; 16: 606-613
  CrossrefPubMedSearch in Google Scholar
• 32 Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry 1979; 134: 382-389
  CrossrefPubMedSearch in Google Scholar
• 33 JASP Team. JASP [Internet]. 2022 Available from: https://jasp-stats.org/
  PubMed
• 34 Kelter R. Bayesian alternatives to null hypothesis significance testing in biomedical research: A non-technical introduction to Bayesian inference with JASP. BMC Med Res Methodol 2020; 20: 142
  CrossrefPubMedSearch in Google Scholar
• 35 Plummer M. JAGS: A Program for Analysis of Bayesian Graphical Models using Gibbs Sampling. 3rd International Workshop on Distributed Statistical Computing (DSC 2003). Vienna, Austria 4 n. Chr.; 124
  PubMed
• 36 R_Core_Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. 2020
  PubMed
• 37 RStudio Team. RStudio: Integrated Development for R. Boston, MA; 2019 Available from: http://www.rstudio.com
  PubMed
• 38 Takahashi N, Yamada A, Shiraishi A. et al. Efficacy and safety of fixed doses of intranasal esketamine as an add-on therapy to oral antidepressants in Japanese patients with treatment-resistant depression: A phase 2b randomized clinical study. BMC Psychiatry 2021; 21: 526
  CrossrefPubMedSearch in Google Scholar
• 39 Popova V, Daly EJ, Trivedi M. et al. Efficacy and safety of flexibly dosed esketamine nasal spray combined with a newly initiated oral antidepressant in treatment-resistant depression: A randomized double-blind active-controlled study. Am J Psychiatry 2019; 176: 428-438
  CrossrefPubMedSearch in Google Scholar
• 40 Schutte AE, Kollias A, Stergiou GS. Blood pressure and its variability: Classic and novel measurement techniques. Nat Rev Cardiol 2022; 19: 643
  CrossrefPubMedSearch in Google Scholar
• 41 Davydov DM, Shapiro D, Cook IA. et al. Baroreflex mechanisms in major depression. Prog Neuropsychopharmacol Biol Psychiatry 2007; 31: 164-177
  CrossrefPubMedSearch in Google Scholar
• 42 Duval F, Mokrani MC, Erb A. et al. Dopaminergic, noradrenergic, adrenal, and thyroid abnormalities in psychotic and affective disorders. Front Psychiatry 2020; 11: 533872
  CrossrefPubMedSearch in Google Scholar
• 43 Agorastos A, Heinig A, Stiedl O. et al. Vagal effects of endocrine HPA axis challenges on resting autonomic activity assessed by heart rate variability measures in healthy humans. Psychoneuroendocrinology 2019; 102: 196-203
  CrossrefPubMedSearch in Google Scholar
• 44 Vaz M, Sucharita S, Bharathi AV. Heart rate and systolic blood pressure variability: The impact of thinness and aging in human male subjects. J Nutr Health Aging 2005; 9: 341-345
  PubMedSearch in Google Scholar
• 45 Ochs-Ross R, Wajs E, Daly EJ. et al. Comparison of long-term efficacy and safety of esketamine nasal spray plus oral antidepressant in younger versus older patients with treatment-resistant depression: Post-hoc analysis of SUSTAIN-2, a long-term open-label phase 3 safety and efficacy study. Am J Geriatr Psychiatry 2022; 30: 541-556
  CrossrefPubMedSearch in Google Scholar
• 46 Zhou Y-L, Liu W-J, Wang C-Y. et al. Cardiovascular effects of repeated subanaesthetic ketamine infusion in depression. J Psychopharmacol 2021; 35: 159-167
  CrossrefPubMedSearch in Google Scholar
• 47 Kanters S. Fixed- and random-effects models. Methods Mol Biol 2022; 2345: 41-65
  CrossrefPubMedSearch in Google Scholar
• 48 Cipriani A, Higgins JPT, Geddes JR. et al. Conceptual and technical challenges in network meta-analysis. Ann Intern Med 2013; 159: 130-137
  CrossrefPubMedSearch in Google Scholar
• 49 Riva-Posse P, Reiff CM, Edwards JA. et al. Blood pressure safety of subanesthetic ketamine for depression: A report on 684 infusions. J Affect Disord 2018; 236: 291-297
  CrossrefPubMedSearch in Google Scholar
• 50 Corcoran AW, Hohwy J, Friston KJ. Accelerating scientific progress through Bayesian adversarial collaboration. Neuron 2023; 111: 3505-3516
  CrossrefPubMedSearch in Google Scholar
• 51 Olson WH, Crivera C, Ma Y-W. et al. Bayesian data analysis in observational comparative effectiveness research: Rationale and examples. J Comp Eff Res 2013; 2: 563-571
  CrossrefPubMedSearch in Google Scholar
• 52 Lucchese AC, Sarin LM, Magalhães EJM. et al. Repeated subcutaneous esketamine for treatment-resistant depression: Impact of the degree of treatment resistance and anxiety comorbidity. J Psychopharmacol 2021; 35: 142-149
  CrossrefPubMedSearch in Google Scholar

Correspondence

Publication History

Received: 04 December 2024

Accepted: 08 February 2025

Article published online:
02 June 2025

© 2025. Thieme. All rights reserved.

Georg Thieme Verlag KG
Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany

• References

• 1 McIntyre RS, Rosenblat JD, Nemeroff CB. et al. Synthesizing the evidence for ketamine and esketamine in treatment-resistant depression: An international expert opinion on the available evidence and implementation. Am J Psychiatry 2021; 178: 383-399
  CrossrefPubMedSearch in Google Scholar
• 2 Culpepper L. Reducing the burden of difficult-to-treat major depressive disorder: Revisiting monoamine oxidase inhibitor therapy. Prim Care Companion CNS Disord 2013; 15
  CrossrefPubMedSearch in Google Scholar
• 3 Asnis GM, Henderson MA. EMSAM (deprenyl patch): How a promising antidepressant was underutilized. Neuropsychiatr Dis Treat 2014; 10: 1911-1923
  CrossrefPubMedSearch in Google Scholar
• 4 Van den Eynde V, Abdelmoemin WR, Abraham MM. et al. The prescriber’s guide to classic MAO inhibitors (phenelzine, tranylcypromine, isocarboxazid) for treatment-resistant depression. CNS Spectr 2022; in press 1-14
  CrossrefPubMedSearch in Google Scholar
• 5 Gillman PK. A review of serotonin toxicity data: Implications for the mechanisms of antidepressant drug action. Biol Psychiatry 2006; 59: 1046-1051
  CrossrefPubMedSearch in Google Scholar
• 6 Gillman PK. A reassessment of the safety profile of monoamine oxidase inhibitors: Elucidating tired old tyramine myths. J Neural Transm (Vienna) 2018; 125: 1707-1717
  CrossrefPubMedSearch in Google Scholar
• 7 Cook J, Halaris A. Adjunctive dopaminergic enhancement of esketamine in treatment-resistant depression. Prog Neuropsychopharmacol Biol Psychiatry 2022; 119: 110603
  CrossrefPubMedSearch in Google Scholar
• 8 Di Maio V, Ventriglia F, Santillo S. A model of dopamine modulated glutamatergic synapse. Biosystems 2015; 136: 59-65
  CrossrefPubMedSearch in Google Scholar
• 9 Bezerra TO, Roque AC. Dopamine facilitates the response to glutamatergic inputs in astrocyte cell models. PLoS Comput Biol 2024; 20: e1012688
  CrossrefPubMedSearch in Google Scholar
• 10 Kokkinou M, Ashok AH, Howes OD. The effects of ketamine on dopaminergic function: Meta-analysis and review of the implications for neuropsychiatric disorders. Mol Psychiatry 2018; 23: 59-69
  CrossrefPubMedSearch in Google Scholar
• 11 Hashimoto K, Kakiuchi T, Ohba H. et al. Reduction of dopamine D2/3 receptor binding in the striatum after a single administration of esketamine, but not R-ketamine: A PET study in conscious monkeys. Eur Arch Psychiatry Clin Neurosci 2017; 267: 173-176
  CrossrefPubMedSearch in Google Scholar
• 12 Ulrich S, Ricken R, Adli M. Tranylcypromine in mind (Part I): Review of pharmacology. Eur Neuropsychopharmacol 2017; 27: 697-713
  CrossrefPubMedSearch in Google Scholar
• 13 Kegeles LS, Abi-Dargham A, Zea-Ponce Y. et al. Modulation of amphetamine-induced striatal dopamine release by ketamine in humans: Implications for schizophrenia. Biol Psychiatry 2000; 48: 627-640
  CrossrefPubMedSearch in Google Scholar
• 14 Timm C, Linstedt U, Weiss T. et al. Sympathomimetic effects of low-dose S(+)-ketamine. Effect of propofol dosage. Anaesthesist 2008; 57: 338-346
  CrossrefPubMedSearch in Google Scholar
• 15 Lundy PM, Lockwood PA, Thompson G. et al. Differential effects of ketamine isomers on neuronal and extraneuronal catecholamine uptake mechanisms. Anesthesiology 1986; 64: 359-363
  CrossrefPubMedSearch in Google Scholar
• 16 Zheng W, Cai DB, Xiang YQ. et al. Adjunctive intranasal esketamine for major depressive disorder: A systematic review of randomized double-blind controlled-placebo studies. J Affect Disord 2020; 265: 63-70
  CrossrefPubMedSearch in Google Scholar
• 17 Kryst J, Kawalec P, Pilc A. Efficacy and safety of intranasal esketamine for the treatment of major depressive disorder. Expert Opin Pharmacother 2020; 21: 9-20
  CrossrefPubMedSearch in Google Scholar
• 18 European Medicine Agency. Spravato - authorization for use and product information [Internet]. 2019 Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/spravato
  PubMedSearch in Google Scholar
• 19 Bovill JG. Adverse drug interactions in anesthesia. J Clin Anesth 1997; 9: 3s-13s
  CrossrefPubMedSearch in Google Scholar
• 20 Cuthbert MF, Vere DW. Potentiation of the cardiovascular effects of some catecholamines by a monoamine oxidase inhibitor. BrJ Pharmacol 1971; 43: 471p-472
  PubMedSearch in Google Scholar
• 21 Trendelenburg U, Langeloh A, Bönisch H. Mechanism of action of indirectly acting sympathomimetic amines. Blood Vessels 1987; 24: 261-270
  CrossrefPubMedSearch in Google Scholar
• 22 Ludwig VM, Sauer C, Young AH. et al. Cardiovascular effects of combining subcutaneous or intravenous esketamine and the MAO inhibitor tranylcypromine for the treatment of depression: A retrospective cohort study. CNS Drugs 2021; 35: 881-892
  CrossrefPubMedSearch in Google Scholar
• 23 van Haelst IM, van Klei WA, Doodeman HJ. et al. Antidepressive treatment with monoamine oxidase inhibitors and the occurrence of intraoperative hemodynamic events: A retrospective observational cohort study. J Clin Psychiatry 2012; 73: 1103-1109
  CrossrefPubMedSearch in Google Scholar
• 24 Bartova L, Vogl SE, Stamenkovic M. et al. Combination of intravenous S-ketamine and oral tranylcypromine in treatment-resistant depression: A report of two cases. Eur Neuropsychopharmacol 2015; 25: 2183-2184
  CrossrefPubMedSearch in Google Scholar
• 25 Bottemanne H, Bonnard E, Claret A. et al. Ketamine and monoamine oxidase inhibitor combination: Utility, safety, efficacy?. J Clin Psychopharmacol. 2020 40.
  PubMedSearch in Google Scholar
• 26 Katz RB, Toprak M, Wilkinson ST. et al. Concurrent use of ketamine and monoamine oxidase inhibitors in the treatment of depression: A letter to the editor. Gen Hosp Psychiatry 2018; 54: 62-64
  CrossrefPubMedSearch in Google Scholar
• 27 Wang JCC, Swainson J. The concurrent treatment with intravenous ketamine and an irreversible monoamine oxidase inhibitor for treatment-resistant depression without hypertensive crises. J Clin Psychopharmacol 2020; 40: 515-517
  CrossrefPubMedSearch in Google Scholar
• 28 Dunner DL, Fugate RM, Demopulos CM. Safety and efficacy of esketamine nasal spray in a depressed patient who was being treated with tranylcypromine: A case report. Neurol Psychiat Br Research 2020; 36: 30-31
  CrossrefPubMedSearch in Google Scholar
• 29 Fiedorowicz JG, Swartz KL. The role of monoamine oxidase inhibitors in current psychiatric practice. J Psychiatr Pract 2004; 10: 239-248
  CrossrefPubMedSearch in Google Scholar
• 30 Chobanian AV, Bakris GL, Black HR. et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 2003; 42: 1206-1252
  CrossrefPubMedSearch in Google Scholar
• 31 Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med 2001; 16: 606-613
  CrossrefPubMedSearch in Google Scholar
• 32 Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry 1979; 134: 382-389
  CrossrefPubMedSearch in Google Scholar
• 33 JASP Team. JASP [Internet]. 2022 Available from: https://jasp-stats.org/
  PubMed
• 34 Kelter R. Bayesian alternatives to null hypothesis significance testing in biomedical research: A non-technical introduction to Bayesian inference with JASP. BMC Med Res Methodol 2020; 20: 142
  CrossrefPubMedSearch in Google Scholar
• 35 Plummer M. JAGS: A Program for Analysis of Bayesian Graphical Models using Gibbs Sampling. 3rd International Workshop on Distributed Statistical Computing (DSC 2003). Vienna, Austria 4 n. Chr.; 124
  PubMed
• 36 R_Core_Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. 2020
  PubMed
• 37 RStudio Team. RStudio: Integrated Development for R. Boston, MA; 2019 Available from: http://www.rstudio.com
  PubMed
• 38 Takahashi N, Yamada A, Shiraishi A. et al. Efficacy and safety of fixed doses of intranasal esketamine as an add-on therapy to oral antidepressants in Japanese patients with treatment-resistant depression: A phase 2b randomized clinical study. BMC Psychiatry 2021; 21: 526
  CrossrefPubMedSearch in Google Scholar
• 39 Popova V, Daly EJ, Trivedi M. et al. Efficacy and safety of flexibly dosed esketamine nasal spray combined with a newly initiated oral antidepressant in treatment-resistant depression: A randomized double-blind active-controlled study. Am J Psychiatry 2019; 176: 428-438
  CrossrefPubMedSearch in Google Scholar
• 40 Schutte AE, Kollias A, Stergiou GS. Blood pressure and its variability: Classic and novel measurement techniques. Nat Rev Cardiol 2022; 19: 643
  CrossrefPubMedSearch in Google Scholar
• 41 Davydov DM, Shapiro D, Cook IA. et al. Baroreflex mechanisms in major depression. Prog Neuropsychopharmacol Biol Psychiatry 2007; 31: 164-177
  CrossrefPubMedSearch in Google Scholar
• 42 Duval F, Mokrani MC, Erb A. et al. Dopaminergic, noradrenergic, adrenal, and thyroid abnormalities in psychotic and affective disorders. Front Psychiatry 2020; 11: 533872
  CrossrefPubMedSearch in Google Scholar
• 43 Agorastos A, Heinig A, Stiedl O. et al. Vagal effects of endocrine HPA axis challenges on resting autonomic activity assessed by heart rate variability measures in healthy humans. Psychoneuroendocrinology 2019; 102: 196-203
  CrossrefPubMedSearch in Google Scholar
• 44 Vaz M, Sucharita S, Bharathi AV. Heart rate and systolic blood pressure variability: The impact of thinness and aging in human male subjects. J Nutr Health Aging 2005; 9: 341-345
  PubMedSearch in Google Scholar
• 45 Ochs-Ross R, Wajs E, Daly EJ. et al. Comparison of long-term efficacy and safety of esketamine nasal spray plus oral antidepressant in younger versus older patients with treatment-resistant depression: Post-hoc analysis of SUSTAIN-2, a long-term open-label phase 3 safety and efficacy study. Am J Geriatr Psychiatry 2022; 30: 541-556
  CrossrefPubMedSearch in Google Scholar
• 46 Zhou Y-L, Liu W-J, Wang C-Y. et al. Cardiovascular effects of repeated subanaesthetic ketamine infusion in depression. J Psychopharmacol 2021; 35: 159-167
  CrossrefPubMedSearch in Google Scholar
• 47 Kanters S. Fixed- and random-effects models. Methods Mol Biol 2022; 2345: 41-65
  CrossrefPubMedSearch in Google Scholar
• 48 Cipriani A, Higgins JPT, Geddes JR. et al. Conceptual and technical challenges in network meta-analysis. Ann Intern Med 2013; 159: 130-137
  CrossrefPubMedSearch in Google Scholar
• 49 Riva-Posse P, Reiff CM, Edwards JA. et al. Blood pressure safety of subanesthetic ketamine for depression: A report on 684 infusions. J Affect Disord 2018; 236: 291-297
  CrossrefPubMedSearch in Google Scholar
• 50 Corcoran AW, Hohwy J, Friston KJ. Accelerating scientific progress through Bayesian adversarial collaboration. Neuron 2023; 111: 3505-3516
  CrossrefPubMedSearch in Google Scholar
• 51 Olson WH, Crivera C, Ma Y-W. et al. Bayesian data analysis in observational comparative effectiveness research: Rationale and examples. J Comp Eff Res 2013; 2: 563-571
  CrossrefPubMedSearch in Google Scholar
• 52 Lucchese AC, Sarin LM, Magalhães EJM. et al. Repeated subcutaneous esketamine for treatment-resistant depression: Impact of the degree of treatment resistance and anxiety comorbidity. J Psychopharmacol 2021; 35: 142-149
  CrossrefPubMedSearch in Google Scholar

• Supplementary Material