Significant improvement in overall survival among patients diagnosed with low‐risk myelodysplastic syndrome treated with selective serotonin reuptake inhibitors
Year: 2016
Type: article
Abstract: Low risk myelodysplaastic syndrome (LR-MDS) is observed in elderly patients and characterized by symptomatic cytopenias. Molecular studies highlight a central role for apoptosis and immune deregulation. The capacity of cytokines, such as IL6, TNF- α and INF-ү to sustain ineffective haematopoiesis suggests that intervention in this pathway could provide benefit. Similar to LR-MDS, deregulated cytokine signalling is observed in both depression and rheumatoid arthritis and selective serotonin reuptake inhibitors (SSRIs) abrogate the symptoms of these disorders in part by modulating cytokine abnormalities (Sutcigil et al, 2007; Baharav et al, 2012; Taraz et al, 2013). We hypothesized that SSRIs could delay disease progression and prolong survival in LR-MDS patients. After Institutional Review Board approval, 118 primary MDS patients aged 18 years and older were identified from the Michael E. DeBakey VA Medical Center cancer registry. Eighty-seven patients had a low or intermediate-1 International Prognostic Scoring System score (i.e., lower-risk). LR-MDS patients were considered to be in the SSRI+ group if they had received at least 6 months of an SSRI (including fluoxetine, sertraline, paroxetine or citalopram) during the 12 months prior to and/or after MDS diagnosis. The primary outcome was defined as overall survival (OS), in days, from bone marrow diagnosis to death from any cause. Univariate and multivariate analysis of SSRI effect were performed using the Cox proportional hazards regression model. Statistical analyses were performed with spss 22.0 (IBM Corp., Armonk, NY, USA). P values were considered statistically significant when <0·05. Among the 87 patients with low risk disease, balanced baseline characteristics were observed (15 SSRI+ vs. 72 SSRI−). The risk of death in the non-SSRI group was almost 2·4 times higher than that observed in the treated group [Hazard ratio (HR) = 2·38, 95% confidence interval (CI) = 1·20–4·71, P = 0·0128]. The estimated median OS for patients with lower-risk disease was 57·5 months in the SSRI+ group and 23·9 months in the SSRI− group (Fig 1A). Multivariate analysis identified SSRI use, white blood cell (WBC) count and haemoglobin (Hb) as independent predictors of survival. A Cox multivariate model containing all 3 significant predictors demonstrated that patients not taking SSRIs were twice as likely to die, after controlling for the effect of Hb and WBC [HR = 2·082, 95% CI (1·042, 4·158), P = 0·0378]. A similar analysis was repeated excluding patients who had received disease-modifying agents. Demographics and patients characteristics were balanced among 8 SSRI+ and 47 SSRI− patients (Table 1), except for Hb level (Mann–Whitney P = 0·027). To prevent potential bias from difference in Hb levels between groups, only patients with Hb >70 g/l were included. As shown in Fig 1B, among patients with best supportive care only (8 SSRI+ vs. 41 SSRI−), the risk of death in non-SSRI treated patients was almost 2·8 times higher than that of the treated group (HR = 2·81, 95% CI = 1·05–7·50, P = 0·0390). The estimated median OS for patients with lower-risk disease was 63·1 months in the SSRI+ group and 31·6 months in the SSRI− group. Multivariate analysis including Hb and WBC resulted in an almost identical estimate of death risk due to SSRI status, but, due to small sample size, it was not statistically significant. LR-MDS remains a challenge due to the high frequency of primary and secondary drug failure. Many MDS patients are elderly and unsuitable for intensive treatment. In this retrospective study, we detected a significant improvement in OS for LR-MDS patients treated with SSRIs. This translated into more than a doubling of the risk of dying for patients not taking SSRIs. An interesting observation is that SSRIs interfere with abnormal cytokine signalling associated with inflammation in patients with depression and rheumatoid arthritis. IL6, IL12 and TNF-α levels are known to be abnormal in these disorders. Sertraline decreases TNF-α release from human T-cell lymphocytes and (IFN)-γ-activated microglia (Taler et al, 2007), suggesting that this mechanism might also facilitate resolution of cytokine deregulation in MDS. Nautiyal et al (2014) recently reported that sertraline treatment was associated with transfusion and growth factor independence in a LR-MDS patient with refractory cytopenia with multilineage dysplasia. Erythroid haematological improvement (HI-E) was sustained for 11 months after discontinuing an erythropoiesis-stimulating agent (ESA), and endogenous erythropoietin level normalized, suggesting that sertraline cooperated with ESA plus granulocyte colony-stimulating factor to resolve severe manifestations of ineffective haematopoeisis (Nautiyal et al, 2014). It is possible that SSRI administration exerts a cytokine regulatory effect leading to improved ineffective haematopoiesis. Excessive apoptosis is associated with amplification of bone marrow inflammatory response. In recent years, abnormal signalling via high expression of Toll-like receptor (TLR) 3, 4 and 7 in patients with major depression (Hung et al, 2014) and TLR1, TLR2 and TLR6 overexpression in CD34+ cells of patients with MDS (Wei et al, 2013) highlights the role of deregulated innate immunity in disease pathogenesis and suggests possibility for therapeutic intervention. Although endosomal TLRs retain affinity for pathogens-associated molecular patterns, increasing evidence supports the recognition of self-ligands by innate immune receptors as effector of autoimmune disorders (de Alba, 2009; Lee & Barton, 2014), a concept that intuitively links potential defective processing of host-derived nucleic acids in inflammatory amplification of MDS pathogenesis. High-throughput screening for TLR3-INF regulatory factor 3 signalling pathway modulators identified antipsychotic medications, such as sertraline and trifluoperazine, as potential TLR inhibitors (Zhu et al, 2010). This observation suggests that sertraline could execute anti-inflammatory effect via direct inhibition of the innate immune signaling (Zhu et al, 2010). In our study, SSRIs influenced survival outcome whether or not subjects had received disease-modifying agents. After adjustment for multiple potential covariates, a substantial favourable survival was retained for patients taking SSRIs. Acknowledging our small sample size, and after controlling for potential bias by removing patients with co-administration of MDS-modifying agents, a similar favourable significant survival was maintained in univariate, but not multivariate analysis. The present study has limitations, which include the highly variable Hb level in the SSRI+ group, making it difficult to distinguish the effect of SSRI from Hb on survival in a small sample. In addition, it is possible that SSRIs improve depression and influence the physiological well-being of LR-MDS patients resulting in direct impact on survival. That said, this provocative observation provides a preliminary framework linking the use of SSRIs to survival benefit in patients with myelodysplasia. Financial support for this study was provided by the Texas Golfers Against cancer grant funding. GR conceived the study; GR, AL, SY, MM designed the research; AL, SY, GR, AF performed the research; GR and AF analysed and interpreted the results; GR, AL, MM, SY, AF wrote, reviewed and approved the manuscript. All authors reported no relevant conflict of interest to disclose.
Source: British Journal of Haematology
DOI: 10.1111/bjh.14031
Language: English
Subfield: Hematology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
Funder: Golfers Against Cancer
OpenAlex ID: W2316515744