Screening and identification
Post-stroke depressive (PSD) disorders are the most commonly reported and widely investigated among all types of NDP in the literature. PSD is the most frequent treatable neuropsychiatric complication of stroke at any one time after onset. A prospective study showed PSD can occur from 1 to 18 mo after the onset of stroke, and prevalence of PSD was not found to vary considerably over time (the prevalence at 1, 3, 6, 12, and 18 mo were 24.5%, 27.1%, 28.3%, 19.8%, and 26.3% respectively).
A previous meta-analysis of 61 studies with 25488 patients indicated that the pooled frequency of PSD was 34% in 32 stroke cohorts, consistent with a meta-analysis of 32 studies with 8938 patients receiving antidepressant therapy where the pooled frequency was 31%.
The proportional frequency of depression reported ranged from 5%-84%, which varied considerably across studies because of different PSD identification criteria, threshold time points of assessment during follow-up, and clinical setting. A national register-based cohort study in Denmark consecutively recruited 157243 first-time hospitalized patients with new-onset PSD and 160236 local healthy residents as a reference population during 2 years of follow-up between 2001 and 2011. The total incidence of depressive disorders after stroke was 25.4%, compared with 7.8% in the control population.
Compared with orthopedic patients with similar degrees of motor disability, myocardial infarction, and similar predisposing factors of cardiovascular disease, patients with stroke are more likely to suffer from depressive disorders, suggesting that there is a more complex neurobiological mechanism for the etiology of PSD.
Generally, PSD occurs in the range from 25%-35% of stroke survivors with the frequency estimated to be the highest in the 1st year after onset, with gradually declining prevalence thereafter.
Optimal screening for and identification of PSD is vital for following treatment and management; however, there is currently no established diagnostic criterion for PSD. The DSM-5 classifies PSD as a “depressive disorder due to another medical condition”. The identification and diagnosis of PSD is usually based on the combination of detailed clinical assessment and screening scale tools in the clinical practice. For example, clinicians diagnose PSD using a structured clinical interview for DSM-5 combined with a screening scale before initiating PSD treatment. There is no universally accepted screening tool for PSD. The following psychiatric scales are frequently used to measure PSD symptoms in clinical study and practice: The center for epidemiological studies depression scale, hospital anxiety depression scale, Hamilton depression rating scale, beck depression inventory, geriatric depression scale, and nine-item patient health questionnaire (PHQ-9). In 2017, the American Heart Association and American Stroke Association jointly issued the first scientific consensus statement for healthcare, which comprehensively discussed the epidemiology, pathophysiology, screening, management, and prevention of PSD. Based on the results of a meta-analysis with 2907 participants, the center for epidemiological studies depression scale, Hamilton depression rating scale, and PHQ-9 scores have proven to have higher sensitivities for identifying PSD, using the international classification of disease or DSM diagnosis of depression as the reference standard. The PHQ-9 is one of the most commonly used tools for screening for PSD with high validity and reliability in primary care. One individual patient’s data meta-analysis showed that a cutoff of a score of 10 on the PHQ-9 yielded a maximum diagnostic performance. Considering the structured interview for DSM as the reference standard for PSD, the sensitivity and specificity of PHQ-9 were 0.82 and 0.97, respectively. The overall diagnostic performance of the PHQ-9 was better than the hospital anxiety depression scale-D and geriatric depression scale. Another individual participant’s data meta-analysis of PHQ-9 reported the existence of selective cutoff reporting bias when estimating sensitivity in most studies. In addition, a systematic review concluded that the results regarding the sensitivity and specificity of the PHQ-9 for PSD screening and identification were uncertain.
Remarkably, unlike depressive disorders caused by other diseases, screening for PSD faces many challenges; particular attention should be paid to the actual condition of patients with stroke. Neurological symptoms resulting from stroke such as aphasia, alexia, or agnosia may lead to expressive or receptive dysfunction. Cognitive impairment such as loss of concern, anosognosia, abulia, or lack of insight may develop similar depressive symptoms. The above adverse factors for screening could hinder the identification and diagnosis of PSD. Therefore, screening and identification procedures of PSD should be performed following protocols tailored to the individual.
Management and therapeutic implication
PSD is associated with worsened functional outcomes after stroke. A meta-analysis including 14 studies before May 2018 with 17609 PSD patients evaluating the association between PSD and the mortality of different follow-up times revealed that PSD showed a negative impact on survival rates; the effect of PSD on short-term mortality was slightly higher than its effect on long-term mortality. A recent case-control study showed PSD increased disability severity in ischemic stroke survivors, whose Barthel index and Rivermead mobility index scores were both lower than stroke survivors without PSD at both admission and discharge.
In theory, the early and prophylactic use of PSD may reduce the risk of PSD in stroke survivors. A meta-analysis with eight prospective randomized controlled trials published from 1990 to 2011 revealed that antidepressant prophylaxis (mianserin, fluoxetine, nortriptyline, sertraline, escitalopram, milnacipran) reduced the odds of developing PSD, and pooled results uncovered the benefit of early initiation of pharmacotherapy in stroke patients; however, the final conclusion of this review was based on eight studies with four classes of antidepressants [selective serotonin reuptake inhibitor (SSRI), tetracyclic antidepressant, tricyclic antidepressant, serotonin-norepinephrine reuptake inhibitor (SNRI)]. Therefore, there may be a high relative heterogeneity among the studies. A systematic review just published in November 2019, which retrieved data from 2009 to 2018, suggested that the use of SSRIs, psychological intervention [e.g., cognitive behavioral therapy (CBT)], as well as mental and physical exercise could relieve most mood symptoms of PSD, but the level of evidence quality of the included studies were low to moderate. A meta-analysis including 20 studies with 1485 patients indicated that both SSRIs and SNRIs had favorable therapeutic effects on PSD, and furthermore that citalopram may improve depressed moods faster than other SSRIs.
Fluoxetine for motor recovery after acute ischemic stroke is a randomized placebo-controlled trial conducted in France, which included 118 patients with ischemic stroke and moderate-to-severe motor deficits, found that the early use of fluoxetine with physiotherapy promoted motor recovery after 3 mo. Similar to the conclusion of fluoxetine for motor recovery after acute ischemic stroke, most meta-analyses and systematic reviews published before 2019 supported that, if given early, fluoxetine could alleviate neurological deficits and disability and allow patients to recover independently through rehabilitation after stroke[41-43].
With the release in December 2018 of results on the effect of fluoxetine on functional outcomes after acute stroke (FOCUS), SSRI-modulated neuroplasticity that could enhance neurological recovery began to be questioned. The FOCUS trial is a multicenter randomized double blind and parallel control, collaborative study held at 103 hospitals through the National Health Service, United Kingdom, which focuses on the effect of fluoxetine on neurological functional outcomes after acute stroke.
In FOCUS, from 2 to 15 d after onset, there were 3,127 eligible patients with stroke (not patients with PSD) that were recruited and randomly allocated fluoxetine (20 mg daily) or placebo for 6 mo. After an extended follow-up period of up to 12 mo, only the neuropsychological scale questionnaire showed statistically significant differences between the two groups, although results of the clinical trial indicated that fluoxetine would enable the improvement of depression symptoms rather than clinical outcomes, and even increase the risk of bone fractures. The results from the FOCUS trial do not support the routine use of fluoxetine in prophylactic treatment for PSD or to promote the recovery of neurological function. The TALOS study (the Efficacy of Citalopram Treatment in Acute Stroke) was a placebo-controlled, randomized, double-blind study with 642 stroke patients in Denmark. Similarly to the FOCUS results, the TALOS study also did not show that citalopram could promote functional recovery, reduce the dependence on activities of daily living, or decrease the risk of recurrent cardiovascular events in acute ischemic stroke.
Since the FOCUS study has the largest number of patients among similar studies so far, the results from the FOCUS study undoubtedly carry a higher weight in the present meta-analysis. Both a recent systematic review and a meta-analysis that encompassed FOCUS data did not support the routine prescription of fluoxetine or other SSRIs to reduce and promote function recovery early after stroke without PSD. Instead, they suggested that fluoxetine or other SSRIs might be used to treat depressive disorders in patients with PSD. Nonetheless, the present result may not be the final conclusion, and the therapeutic effect of SSRIs and SNRIs for PSD functional rehabilitation remains controversial. Consequently, the two big ongoing trials, assessment of Fluoxetine in Stroke recovery (participants are being recruited from Australia, New Zealand and Vietnam), and efficacy of fluoxetine, a randomized Controlled Trial in Stroke (participants are being recruited from Sweden) will provide further information regarding fluoxetine for stroke recovery. In addition, a meta-analysis using individual participant data will be needed.
SSRIs and antithrombotics are always simultaneously prescribed for patients with PSD in clinical practice. Clopidogrel is one of the commonly used anti-platelet medications that prevent and treat ischemic stroke. Clopidogrel can be metabolized into active products with therapeutic properties by cytochrome P450 (CYP) enzymes. A cohort study and meta-analysis (which included 72020 participants) have shown that CYP2C19-inhibiting SSRIs (fluoxetine and fluvoxamine) can decrease the therapeutic efficacy of clopidogrel. Patients using clopidogrel who were co-prescribed CYP2C19-inhibiting SSRIs had an 11% higher risk of developing ischemic disease than patients using clopidogrel who were treated with non-inhibiting SSRIs. Serotonin could be released from platelets in the blood during the coagulation process. Aspirin is another important prescription medication for treating and preventing ischemic stroke and TIA. In theory, SSRI and SNRI reuptake serotonin in platelets as well as they do in the central nervous system, which reduce platelet serotonin and may be associated with aspirin-related bleeding.
Therefore, there are also growing concerns on the relationship of SSRIs with abnormal bleeding events. Mortensen’s study demonstrated that prestroke SSRI exposure was significantly associated among the severity and mortality of patients with hemorrhagic stroke. In a large collaborative study, the preadmission use of SSRI alone did not increase the risk of spontaneous intracerebral hemorrhage after intravenous thrombolytic therapy for acute ischemic stroke. While there was a significant interaction between the concurrent preadmission use of SSRIs and oral anticoagulants on the occurrence of intracerebral hemorrhage related to thrombolysis, this condition can be seen in PSD patients with recurrent acute ischemic stroke that are treated with SSRIs. Moreover, fluoxetine and fluvoxamine are reported to have potential interactions with warfarin, and inhibit warfarin metabolism by competitively binding plasma protein and interfering with CYP isoenzymes, which are more likely to strengthen the anticoagulant effects of warfarin. Paroxetine also seems to have a low-to-moderate risk of enhancing the pharmacological effects of warfarin; however, other SSRIs and SNRIs do not appear to interact with warfarin.
A recent systematic review suggested that there is no high quality evidence to support that SSRIs used alone can increase the risk of spontaneous intracerebral hemorrhage. In addition, the association between SSRIs and intracerebral hemorrhage as previously reported was partly accounted for by biases and methodological limitations. Neurologists and psychiatrists need to be well aware of the pharmacological interaction profiles when co-prescribing antidepressants and antithrombotics to patients with PSD and other NDS, monitor the possible adverse events during follow-up, and provide tailored therapeutic strategies for treating PSD and other NDS.
Psychotherapy is also an important intervention for PSD. CBT may be the most effective psychotherapeutic intervention. A meta-analysis on the efficacy of psychotherapy for PSD concluded that the evidence for the benefit of CBT in PSD remains inconclusive due to the high degree of heterogeneity and low quality across the majority of included studies. Neuromodulation, such as transcranial magnetic stimulation and transcranial direct current stimulation, are promising adjunctive therapies. However, high quality randomized controlled trials using psychotherapy or neuromodulation are limited, and further research is needed.
In summary, antidepressant therapy should be used early once the definitive diagnosis of PSD has been made. SSRIs and SNRIs are recommended as a first-line pharmacotherapy for mitigating depression. Other treatment approaches, i.e., psychotherapy, neuromodulation, and psychosocial interventions, should also be considered. No reliable evidence exists to show that the use of SSRIs and other antidepressants can improve neurological function outcomes for patients with PSD.