Recovery in stroke patients treated with uoxetine versus placebo: A pooled analysis of 7,165 patients

Purpose We conducted this study to build clear evidence regarding the outcomes of recovery in stroke patients treated with uoxetine versus placebo. Methods Seventeen randomized clinical trials were identied by searching PubMed, Cochrane, Scopus, and Web of Science until June 2021. Fluoxetine enhances the National Institutes of Health Stroke Scale (NIHSS) score [MD -0.67, 95% CI (-1.19 to -0.15), P = 0.01] and the Fugl-Meyer Motor Scale (FMMS) score [MD 17.36, 95% CI (12.12 to 22.61), P <0.00001] at three months follow up. However, the NIHSS score showed no signicant difference between the two groups at two weeks [MD -0.32 ,95% CI (-0.72 to 0.07), p=0.11] or at six months [MD -0.17 ,95% CI (-0.47 to 0.14), p=0.29]. Fluoxetine and placebo-treated patients had the same overall impact on FMMS scores at one month [MD 7.77, 95% CI (-10.57 to 26.11), P = 0.41]. The uoxetine had a higher risk of broken bone (RR = 2.30, 95% CI [1.59, 3.32], p < 0.001) and hyponatremia (RR = 2.12, 95% CI [1.19, 3.76], p= 0.01) with lower risk of new depression (RR = 0.72, 95% CI [0.61, 0.84], p< 0.001) in comparison to placebo. The ecacy of uoxetine on the NIHSS and FMMS is likely to take time to emerge and is expected to be transient. The BI score did not differ between the uoxetine and placebo groups. The use of uoxetine increased the incidence of hyponatremia and bone fractures while decreasing the risk of new-onset depression.


Introduction
Stroke is one of the predominant causes of permanent disability, mortality, signi cant cognitive, physical, and psychosocial morbidity in the world [1], [2]. It has three signi cant impacts on patients and their families: impairment, disability, and handicap. Furthermore, patients frequently experience depression following injury to either the right or left side of the brain [3]. The most prevalent and well-known disability induced by stroke is motor impairment [4]. At this time, stroke research has put its attention on therapeutic strategies that minimize neurological damage and improve stroke patients' functionality and recovery [5].
Regrettably, medication therapy designed to reduce the neurological damage had little effect [6]. Post-stroke depression and emotional impairment are treated with uoxetine, an antidepressant, and selective serotonin reuptake inhibitor (SSRI) [7].
It is still unclear exactly how SSRIs impact rehabilitation from motor impairment [8]. SSRIs have been shown to enhance outcomes following a stroke in numerous clinical trials [9]- [11]. In 2011, the results of the FLAME trial showed that uoxetine may enhance motor recovery and reduce dependence after three months through Modulation of spontaneous brain plasticity [12]. In animal stroke models, there was a signi cant bene t of therapy with SSRIs on neurobehavioral outcomes, but there was no detectable impact on infarct volume and there was no discernible in uence affecting structural or functional outcomes as a result of the type of SSRI usage [13]. After 6 months of uoxetine treatment, there is no improvement in the patients' modi ed Rankin Scale (mRS) as reported by the FOCUS trial [7]. Furthermore, the AFFINITY trial [14] reported that taking uoxetine daily for six months didn't lead to better functional outcomes. They also observed an increased risk of bone fractures and epileptic episodes [7], [14]. In the EFFECTS trial, uoxetine given orally showed limited bene t in functional outcomes, reduced the signs of depression, raised the chance of bone fractures, and caused hyponatremia [15]. To summarize, their research shows that uoxetine should not be used in patients who have had an acute stroke [15]. Marquez-Romero et al, [8]suggested that adding uoxetine to standard treatment early on was safe and helpful to speed up motor recovery 3 months following an intracerebral haemorrhage (ICH) [8].
Currently, the evidence regarding the effectiveness of SSRIs in improving speci c cognitive domains, or dependency after stroke, is uncertain [16]. Also, if similar bene ts in outcomes may be detected across populations are sustained following treatment cessation [16]. It was shown by Previous meta-analyses that uoxetine could potentially improve the defective neurologic function if administered to stroke patients early after their accident [17]- [19]. But these studies used a limited number of patients and have low methodological quality in most included studies [17]- [19]. A meta-analysis by Liu et al,[20] detected an improvement in Fugl-Meyer Motor Scale (FMMS) and BI scores with no gain in mRS and National Institutes of Health Stroke Scale (NIHSS) scores which could be caused by the heterogeneity across the trials. Mead et al, [21] found that uoxetine had no effect on disability or dependency following a stroke. This metaanalysis comprised just four high-quality studies with a signi cant degree of heterogeneity [21]. There is no subgroup analysis for treatment duration previous meta-analyses [20], [21], despite the fact that uoxetine was used for three months in some trials and six months in others which could potentially in uence the outcome. To elucidate the existing research gap and resolve the con icting evidence regarding the safety and e cacy of uoxetine for stroke patients, this comprehensive systematic review and meta-analysis (SR-MA) were conducted.

Materials And Methods
When performing this SR-MA, we followed the procedures outlined in the Cochrane Handbook for Systematic Reviews of Interventions [22]. This SR-MA was prepared in accordance with the PRISMA declaration requirements [23].
The criteria for exclusion were as follows: (1) study type: observational studies, reviews, editorials, abstracts, thesis, letters, books, and chapters; (2) Studies whose data are not extractable or analysable.

Data extraction:
In the data extraction, all study authors participated equally. The following data items were extracted: 1) summary of the included studies: list of authors, site, study design, study arms, sample size, doses, length of follow up, outcomes, and results; 2) baseline characters of the included studies' population: age, sex, hypertension, dyslipidaemia, smoking, coronary artery disease, and diabetes mellitus; 3) e cacy outcomes: poststroke depression, excessive/ inappropriate crying, excessive/ inappropriate laughing, post-stroke anger proneness, Modi ed Rankin Scale, Montgomery-Asberg Depression Rating Scale (MADRS); 4) safety outcomes: insomnia, nausea, diarrhoea, recurrent stroke, acute coronary events, ischaemic stroke, other major bleeds, upper gastrointestinal bleed, haemorrhagic stroke, epileptic seizures, fall with injury, and death.

Risk of bias assessment:
It was assessed using the Cochrane Risk of Bias tool for each of the included studies [24]. Selection, performance, detection, attrition bias, reporting and other possible biases are assessed by this tool. Third-party reviewers assigned each bias domain a value of "low," "high" or "unclear".
Data synthesis: For statistical analysis, we use Review Manager 5.4.1. All analyses used a xed-effects model except if there is substantial heterogeneity in which we used the random-effects model. We calculated mean difference (MD) and 95 percent con dence interval (CI) for continuous data and risk ratios (RRs) and 95% CI for dichotomous data in order to pool the data between the uoxetine and control arms. A p-value of less than 0.05 was regarded as statistically signi cant. To determine heterogeneity, the chi-square and I-square tests were applied. When the chi-square P-value was less than 0.1 and the I-value was larger than 50%, heterogeneity was considered.

Characteristics of included studies:
We were able to include 17 RCTs with 7365 participants in our analysis after we accepted all of the studies that ful lled our inclusion criteria. We included all placebo-controlled studies assessing uoxetine's e cacy and safety in poststroke patients. There were 16 studies [7], [8], [12], [14], [15] Table 1 shows a summary of the design and principal outcomes of included studies, whereas Table 2 contains baseline characteristics of included studies participants.

Risk of Bias Assessment Results:
Based on the Cochrane risk of bias assessment tool, reviewed studies were of moderate to high quality. Fig. 2 summarises the quality assessment domains for all of the papers that were considered. In Table 5 Online Resource 1, the authors' opinions and explanations are both presented. The mean difference and 95% CI of e cacy outcomes are presented in Online Resource 1: Table 3. E cacy analysis: NIHSS score: Fluoxetine was found to have a greater overall effect than placebo at three months [MD -0.67, 95 % CI (-1.19 to -0.15), P = 0.01], the results were homogenous (p =0.40; I 2 = 0%), but there was no difference between the two arms at two weeks [MD -0.32 ,95% CI (-0.72 to 0.07), p=0.11], and six months [MD -0.17 ,95% CI (-0.47 to 0.14), p=0.29]; pooled analysis were homogeneous at 2 weeks (p =056; I 2 = 0%), and at 6 months (p =0.06; I 2 = 65%) respectively (Fig. 3).

Discussion
In this meta-analysis of usage of uoxetine 20 mg for stroke patients, we identi ed 17 RCTs [7], [8], [12], [14], [15], The NIHSS is commonly applied to provide an objective estimation of the severity of a stroke [37]. It calculates an overall score for each ability, ranging from 0 to 4, A normal state is scored 0, while a completely impaired state is scored 4. The NIHSS score is the sum of all of the scale elements, and the maximum score is 42. A higher NIHSS score is associated with a greater level of impairment in stroke patients [37]. Our meta-analysis showed signi cant improvement in NIHSS score at 3 months but without any signi cance at 2 weeks or 6 months respectively, suggesting that uoxetine effect needs time to occur and is considered as temporary from the pooled results of 6 RCTs.
As for the FMMS score which is an index designed for assessing the motor and joint function, balance, and sensation after stroke-induced hemiplegia[38], [39]. There were no signi cant differences between the Fluoxetine and placebo groups one month later. At three months, we exhibited a greater improvement, but the outcomes were heterogeneous. Patients and healthcare circumstances vary by trial, which could explain this disparity.
Disability or independence for daily tasks is evaluated by BI, in which higher scores suggest better functioning [40]. Neither the uoxetine nor the placebo groups showed signi cant differences at 3 or 6 months.
Fluoxetine showed better improvement in the Hamilton depression scale, which is the most used scale for detecting depression, Fluoxetine speci city resides in its effect in treating depression since it belongs to the SSRI family, and mostly used for its anti-depressive and emotional lability after stroke and also used for the treatment of the obsessive-compulsive disorder, and other mental disorders. The question that needs to be addressed is how uoxetine affects the neurological scale and induces its post-stroke recovery effect. Some animal experiments have shown how uoxetine could help with motor recovery [13]. The fact that SSRIs stimulate hippocampal and subventricular zone neurogenesis is quite intriguing [41]. Moreover, through the processes of neurogenesis, this could be a factor in a subsequent recovery following an episode of cerebral ischemia, with the possibility of new neurons migrating to the affected region [42], [43]. Fluoxetine enhances the secretion of the brain-derived neurotrophic factor, a neurotrophin strongly implicated in neuronal growth and survival and brain plasticity within Two suggested mechanisms: 5-HTT blockade and 5-HTT-independent direct activation of tropomyosin receptor kinase B (TrkB), which is the high-a nity receptor of the neurotrophin BDNF [44]. Furthermore, SSRIs can inhibit the in ammatory response the latter is the main cause of brain damage[18], There is considerable brain damage area reduction and improved neurological function in mice treated with SSRIs 9 hours following the onset of stroke [47].
This reinforces the hypothesis that SSRIs have a bene cial effect on mice's neurological function. Other molecules are upregulated by SSRIs such as heme oxygenase-1 (HO-1) and hypoxia-inducible factor-1alpha which play a particular role in regulating cerebral blood ow [48]. Carbon monoxide is produced as a result of the HO-1 expression, which regulates vascular tone promoting better blood ow to the intended brain regions [49], and also SSRIs have been shown to increase the expression of β-1 adrenergic receptors in the brain of mice. [18]. Expression of these factors may improve functional recovery. Our results go hand in hand with animal studies conclusions and showed better amelioration in the function recovery predicted by the NIHSS score at 3 months, but the effect didn't last for too long and also by the improvement in FMMS score showed in the 3 rd month.
In terms of Safety outcomes, our analysis showed a higher risk for bone fracture and hyponatremia among the participants in the uoxetine group. It is unknown how uoxetine causes hyponatremia. According to a suggested mechanism, SIADH would be induced when there is an increase in the production of antidiuretic hormone (ADH) (serotonin has an effect on the 5-HT2 and 5-HT1C receptors) or an increased renal sensitivity to ADH . The included RCTs were of high quality. We provide a subgroup analysis according to the duration of the follow-up to provide reliable results. A three-month treatment is not equivalent to a six-month treatment in terms of pharmacology, and the results may vary. This option was not provided in the previously published Meta-analysis. Our primary limitation is that we covered both types of strokes (ischemic and haemorrhagic). Knowing that different compositions of the participant groups may affect the results and induce slight heterogeneity which was observed in our analysis.

Conclusion
Short-term uoxetine treatment failed to show any substantial improvement in the NIHSS and FMMS scores. There is a positive impact on NIHSS and FMMS scores at three months following therapy, however, at six months there is no signi cant change in NIHSS score. These ndings demonstrate that uoxetine's effect may take time to manifest and is assumed to be transient. There were no signi cant variations in BI score between the uoxetine and placebo groups. Other than vitality, no signi cant differences were found in the SIS score between the uoxetine and placebo groups. The only exception to this rule was for vitality, which favoured the placebo. The use of uoxetine caused an increase in the risk of hyponatremia and bone fracture, but it lowered the possibility of developing a new-onset depression.

Recommendations
Much more research is required to pinpoint other therapies that may bring benefit and quantify such benefit in a clinical setting. Trials will have to be largesized, of high quality, and their target population must be well de ned in order to reduce heterogeneity.    Figure 1 The study search, selection, and inclusion process Figure 2 Risk of bias: a summary table for each risk of bias item for each study Figure 3 The pooled mean difference of the National Institutes of Health Stroke Scale (NIHSS) Figure 4 The pooled mean difference of the Barthel Index (BI) Figure 5 The pooled mean difference of the Fugl-Meyer Motor Scale (FMMS)