Trazodone Bone Health and Density Impact

Does Trazodone Lower Bone Density?

The most direct human data come from the Osteoporotic Fractures in Men study, a cohort of 5,995 men aged 65 and older [1]. In a cross-sectional analysis, mean bone mineral density among the 160 men taking an SSRI was 3.9% lower at the total hip and 5.9% lower at the lumbar spine than in the 5,708 men taking no antidepressant [1]. The 52 men taking trazodone showed no significant difference, and neither did the 99 men taking a tricyclic [1].

That is one snapshot of density rather than a measure of loss over time, and 52 users is a small group. A separate analysis of the same cohort reported that current SSRI use, but not other antidepressants, was associated with at least 4% lower bone density [7]. A 2014 systematic review of 19 observational studies in adults 60 and older found conflicting results for serotonin reuptake inhibitors and concluded there was insufficient evidence that they harm bone [8].

What the Serotonin-Bone Biology Says

Bone cells carry the machinery that serotonergic drugs act on. In a murine osteocyte cell line, researchers found the serotonin transporter plus 5-HT1A and 5-HT2A receptors, and the antidepressants fluoxetine and imipramine blocked serotonin uptake into those cells [9].

Where the serotonin comes from appears to matter. In rodent work, serotonin made in the gut acts as a circulating hormone that restrains bone formation, while serotonin made in the brain increases bone mass accrual [10]. The gut signal was traced to duodenal serotonin acting on the osteoblast Htr1b receptor and CREB to slow osteoblast proliferation in mice [11].

Trazodone's label describes it as both a selective serotonin reuptake inhibitor and a 5-HT2 receptor antagonist, with preclinical Ki values of 367 nM for serotonin reuptake and 35.6 nM at 5-HT2A [2]. These are receptor-binding figures, not bone data.

Falls: What the Label and Class Data Show

Falls are one way an antidepressant could matter for fractures. The label states that trazodone blocks alpha-1 adrenergic receptors and links this to postural (orthostatic) hypotension [2]. In the label's controlled MDD outpatient data (trazodone n=157, placebo n=158), hypotension occurred in 4% versus 0% and syncope in 5% versus 1% [2]. Drowsiness was 41% versus 20%, dizziness or lightheadedness 28% versus 15%, and blurred vision 15% versus 4% [2].

The label also warns that trazodone has the potential to impair judgment, thinking and motor skills, and advises caution when operating machinery [2]. It says drowsiness may require giving a major portion of the daily dose at bedtime or reducing the dose, and that taking trazodone with blood pressure medicines may require lowering the blood pressure drug dose [2]. The label gives no half-life; a three-way crossover study in 8 healthy volunteers given 100 mg measured a terminal half-life of 7.3 plus or minus 0.8 hours [2][5].

A meta-analysis of 22 studies covering 79,081 participants estimated an unadjusted odds ratio for falling of 1.68 (95% credible interval 1.47 to 1.91) with antidepressants, and 1.36 (1.13 to 1.76) after adjustment, while sedatives and hypnotics came in at 1.47 unadjusted [3]. Those figures are for drug classes, not for trazodone alone.

Hip Fracture Data

A Canadian case-control study included 8,239 adults aged 66 or older who were hospitalized for hip fracture, with each case matched to five controls. The adjusted odds ratio was 2.4 for SSRIs and ranged from 1.5 to 2.2 for tricyclic antidepressants [4]. The abstract gives estimates by drug class and none for trazodone.

When Another Drug Raises Trazodone Levels

Trazodone is metabolized by CYP3A4 [2][12]. The label says to consider a lower trazodone dose with a strong CYP3A4 inhibitor [2]. In a crossover study of 10 healthy volunteers, adding low-dose ritonavir (200 mg four times) to 50 mg of trazodone cut apparent oral clearance from 155 to 75 mL/min, extended half-life from 6.7 to 14.9 hours, and raised peak concentration from 842 to 1,125 ng/mL [6]. Sedation and fatigue increased, three participants had nausea, dizziness or hypotension, and one of them fainted; among the performance tests, only the digit-symbol substitution test differed significantly [6].