Systematic Review of the Association between Serotonin Affinity of Anti-Depressant Agents and the Occurrence of Fractures, Falls, and Bone Mineral Density Change in Patients With Depression

1. EXECUTIVE SUMMARY\n1.1 Background\nIn 2012, 5.4% of Canadians aged 15 years or more had symptoms consistent with a mood disorder, and 4.7% were documented as having a major depressive episode. Pharmacotherapy with antidepressants including selective serotonin reuptake inhibitors (SSRIs), trycyclic antidepressants (TCAs), and other non-SSRI, non-TCA agents represent a long-standing component of treatment to manage patients’ symptoms on a daily basis. Major depressive disorder is associated with increased morbidity and mortality and also with higher healthcare costs and more severe functional impairment with a large impact on workforce productivity. Currently in Canada, clinicians select from amongst more than a dozen agents to prescribe to their patients to manage their depression.\nRecent systematic reviews of observational studies have suggested the presence of a potential association between use of antidepressant treatments and an increased risk of fractures. While suggestive, the observational studies are susceptible to a high risk of residual confounding and bias. As such, establishing a stronger causal link, and the underlying mechanism of this association currently remains unclear. One existing theory is based on the evidence of a pathway by which the serotonin system impacts bone structure. The Drug Safety and Effectiveness Network was approached to explore if the serotonin affinity level (i.e. a pharmacokinetic characteristic of drugs derived from a drug’s measured dissociation constant which can be classified as low or moderate or high, and which is a measure of how strongly serotonin receptors in the body uptake the drug) of anti-depressant agents used by patients might explain a long-term association between antidepressant use and an increased risk of fractures .and adverse changes in bone mineral density. This is considered to be one part of a bi-modal relationship between antidepressant use and fractures. A second concern is a short-term effect of antidepressant agents wherein there may exist differential risks between drugs for the occurrence of outcomes which may cause patients to fall more frequently (e.g. dizziness, insomnia, blurred vision, hypotension, fatigue). A rigorous evaluation of the risks of these agents can also play an important role in evaluating the benefit/risk balance, thereby helping to guide treatment choices for patients suffering form depression. Given these knowledge gaps, this systematic review was designed to explore the hypothesized bi-modal association between the serotonin affinity level of different antidepressant agents and the occurrence of fractures, adverse changes in bone mineral density, and the occurrence of fall-related outcomes.\n1.2 Research Questions\nThis systematic review was performed to address a DSEN query by addressing the following research questions:\n1. Does the choice of antidepressant agent increase the risk of falls? Is there a differential risk in fall-related events between individual anti-depressant agents? (i.e.vision problems, hypotension, dizziness, insomnia, fatigue, drowsiness?)\n2. Does the serotonin affinity level of currently available anti-depressant agents (i.e. SSRIs, TCAs, non-SSRI and TCA agents) impact the risk of fractures?\n3. Does the serotonin affinity level of currently available anti-depressant agents (i.e. SSRIs, TCAs, non-SSRI and TCA agents) modify the risk of negative changes in bone mineral density?\n8\n1.3 Methods\nWe collaborated with an Oxford-based research group (authors of a previous network meta-analysis evaluating the efficacy and tolerability of antidepressants) to collect information from a large number of randomized controlled trials reporting on side effects of interest chosen for this review that could place patients at an increased risk of falls: vision problems, dizziness, hypotension, insomnia, fatigue, and drowsiness. A systematic literature search was designed and performed by an information specialist to identify relevant observational studies as outlined in an a priori designed study protocol. Observational studies performed in patients diagnosed with depression were selected for inclusion if they included relevant antidepressant treatments which could also be classified into serotonin affinity classes (low, moderate or high based on dissociation constant value), and if relevant outcomes [fractures, change in bone mineral density, falls] were reported.\nNetwork meta-analyses of randomized controlled trials were used to explore the relative frequencies of fall-related outcomes; agent-level comparisons were the focus, with a secondary focus on drug class. Data from observational studies related to the occurrence of fractures and changes in bone mineral density were summarized and critically appraised with consideration of the serotonin affinity of agents studies. Given the presence of varied outcome reporting and clinical heterogeneity, no meta-analyses of these studies were performed.\n1.4 Results\n1.4.1 Summary of Key Clinical Findings\n A total of 202 randomized controlled trials meeting the eligibility criteria for our review were included. Amongst them, 62 reported on vision problems, 119 reported on insomnia, 98 reported on fatigue, 141 reported on dizziness, 46 reported on hypotension, and 145 reported on drowsiness. The collection of included trials was composed of trials from a series of systematic reviews for antidepressant pharmacotherapies across which studies were selected based on analogous criteria. Overall 194 (96%) trials were conducted in patients with moderate to severe depression, while the remaining studies were conducted in patients with severe depression.\n A total of 5 observational studies meeting the inclusion criteria for our review were located. These studies were heterogeneous in terms of outcomes studied, patient population characteristics, and approaches to statistical analysis.\n Based on network meta-analyses of actively controlled randomized trials, there is evidence to support an increased risk between certain therapies for the outcomes reviewed, many of which are current considerations in current clinical practice. Across most outcomes, all classes of agents showed increased risks of harms compared to placebo. There was a general lack of clinically relevant differences for comparisons between agents in the same class. The reputation of increased risk of TCAs was reflected for several agents in this class for several outcomes including dizziness. Several SSRIs and non-SSRI/TCA agents were associated with an increased risk of insomnia compared to TCAs, though SSRIs were associated with less drowsiness relative to the other classes.\n9\nFor some outcomes, a high level of uncertainty as reflected by wide credible intervals led to the identification of few differences.\n Based on clinical review of observational studies, there is not sufficient or consistent evidence to confirm a long-term association between the serotonin affinity level of antidepressants and fracture risk or changes in BMD in the body.\n1.5 Strengths and Limitations\n Our network meta-analyses to address the short-term association between anti-depressant agents and adverse effects which could contribute to the risk of falls were based on a large number of trials derived from a collaboration with an academic research group who has performed a series of systematic reviews of antidepressant agents, and thus represents a considerable amount of important imformation. However, the notable variation in the number of studies included for analysis for each outcome suggests considerable variability in the transparency of reporting of harms information. When dealing with rare harms outcomes, this can lead to wide uncertainty around summary estimates.\n As noted above, few observational studies meeting the inclusion criteria for our review were identified. Further research incorporating analyses whose primary focus is serotonin affinity are still needed.\n Many observational studies were excluded during screening because reports presented interventions at the antidepressant class level, and serotonin affinity level varies within classes of antidepressants. Thus, clear classifications in these studies with regard to exposure could not be made.\n Evaluation of the long-term association with serotonin affinity is limited by the quality of the included observational studies as a consequence of their residual confounding, as well as variation in the outcomes reported and manner in which exposure groups were defined.\n1.6 Conclusions\n Based on findings from network meta-analyses of randomized trials of short duration, many important differences exist between antidepressant agents in terms of the risks of adverse effects which may lead to the occurrence of falls, however it is not known how often such harms may lead to falls. These differences are often associated with differences between drug classes. As was anticipated, no discernable association between the serotonin affinity of antidepressants and the occurrence of these harms was evident.\n Little observational information directly addressing the serotonin affinity level of antidepressants on fractures and bone mineral density was found. As serotonin affinity does not directly correlate to drug class and it was common to encounter studies only evaluating antidepressant users at the class level, the ability to draw interpretations based on serotonin affinity was limited. Findings from observational research did not identify a consistent pattern of findings for fracture risk or bone mineral density change in relation to serotonin affinity. Interpretations are also limited by the complications of confounding by patient age and depression severity. Future primary research studies focusing on affinity could potentially improve knowledge gaps in this area of study.