Antidepressant Medication and Stability in Seniors Case Study


Data Source: 

Dr. Brian Maki of Sunnybrook and Women’s College Health Sciences Centre, University of Toronto and his co-investigators Barbara Liu, Nicole Mittmann and Nathan Herrmann.


Alison Gibbs, Department of Statistics, University of Toronto


Falls and hip fracture are important sources of morbidity and mortality in the geriatric population. Several studies have identified psychotropic drugs as a risk factor for both falls and hip fracture. Selective serotonin reuptake inhibitors (SSRIs) are a relatively new class of antidepressant which have become the standard of therapy in geriatric depression. Their efficacy is similar to older tricyclic antidepressants; however SSRIs reportedly have a preferential side-effect profile in older patients. Recently, however, the principal investigator published the first large scale trial (n=8,239 cases and 41,195 controls) which demonstrated that SSRI-use is associated with a risk for hip fracture (adjusted odds ratio 2.4, 95% confidence interval 2.0-2.7) similar to tricyclic antidepressants (Liu et al., Lancet 1998). The relative contribution of medication (SSRI) versus disease (depression)-related effects to the risk of hip fracture remains unclear.

Static and dynamic posturography measures (related to control of postural sway and dynamic balancing reactions) have been shown to be useful clinical predictors of falls in older subjects. There is evidence that some SSRIs increase sway in depressed subjects. In the present study, the effects of two of the most widely recommended SSRIs (sertraline and paroxetine) on postural stability of healthy older subjects were assessed and comparisons with the most widely recommenede tricyclic antidepressant, nortriptyline, were made. The study used sensitive dynamic measures of balance, and a range of functionally-important balancing tasks, not assessed in previously published studies of antidepressant effects on balance. By assessing the pharmacodynamic effects in healthy older subjects, it will be possible to evaluate the drug-related effects separately from the possible disease-related effects that depressive illness itself may have on postural control.

Study design:

Using a randomized, double-blind, parallel-group design, the study examined four groups of subjects. One group of subjects received sertraline (25 mg daily for one week, then 50 mg daily for one week), a second group received paroxetine (10 mg daily, then 20 mg daily), a third group received nortriptyline (25 mg daily, then 50 mg daily) and a fourth group received a placebo (for two weeks). Healthy, non-depressed, older adult (>65 years) volunteers were recruited from the community. Although the original study design called for 10 subjects per group, recruitment difficulties and time constraints limited the actual sample to four subjects per group with an addition fifth subject in one of the groups (total n=17). Subjects underwent balance testing at baseline and at the end of weeks 1 and 2 (total of three testing sessions). Serum levels of sertraline and paroxetine were determined at the end of week 1 and 2. The balance testing included measurements of: 1) spontaneous postural sway (during quiet unperturbed stance); 2) maximal capacity to lean in forward, backward, left and right directions; 3) induced sway reactions evoked by continuous pseudorandom platform motion (antero-posterior or medio-lateral); 4) rapid multi-directional stepping balance-recovery reactions evoked by unpredictable transient platform motion (forward, backward, left and right directions of platform motion; moderate and large magnitudes); 5) rapid forward stepping and grasping balance-recovery reactions cued by transient platform motion; and 6) rapid forward volitional stepping and grasping movements cued by a visual cue. Secondary cognitive tasks were performed, during a subset of balance tests, in order to heighten the demands placed on the central nervous system. To further heighten demands, some of the tasks were performed without vision (wearing a blindfold) or while standing on a compliant (foam rubber) surface. The primary outcome measures included: 1) anterior-posterior and medial-lateral centre of pressure displacement (spontaneous sway, maximal lean and induced-sway tasks), 2) spatial-temporal characteristics of the limb movements (stepping and grasping tasks) and 3) the relative frequency of specific patterns of limb movement (multi-directional stepping task).

Statistical issues:

  • Small sample size.
  • Possible evidence of non-compliance in some subjects.
  • Very large number of variables (many of which may be correlated to varying degrees) available to characterize postural stability.
  • Possibility of learning effects in some of the balance measures.
  • Missing values for some of the balance tests.
  • Although most of the balance measures are continuous variables, some are categorical. The continuous variables may or may not have a normal distribution.

Frequently Asked Questions

Please check this section regularly for updates.

Research Question: 


The primary objective of the study was to evaluate and compare the effect of three antidepressant medications (sertraline, paroxetine and nortriptyline) on the control of static and dynamic postural balance in healthy older adults. A secondary objective was to identify the presence of a dose-response relationship between serum drug levels and postural control measures.


Sertraline and paroxetine increase measures of postural instability compared to placebo. Nortriptyline increases measures of postural instability to a greater extent than sertraline and paroxetine.


Each subject participated in three four-hour data-collection sessions, conducted at weekly intervals over a period of 2 weeks.



Total of 51 observations (17 subjects x 3 sessions per subject)


Subject group: variable=drug (A, B, C, or D) 
Subject identifier: idsub=1578-1594 
Session: variable=session (1=baseline, 2=after taking drug for one week, 3=after taking the drug for two weeks)

Possible covariates: 

age (variable=age) 
gender (variable=gender) 
drug serum level (variables=serumA, serumB, serumC, serumD) 
height (variable=height) 
body mass (variable=bodymass) 
length of base-of-support (variable=lbos) 
width of base-of-support (variable=wbos)

Outcome measures: There are more than 1500 outcome measures potentially available to characterize postural stability. See Naming Conventions for outcome variables.


  • B. Liu, G. Anderson, N. Mittmann, T. To, T. Axcell, and N. Shear (1998) Use of selective serotonin-reuptake inhibitors or tricyclic antidepressants and risk of hip fractures in elderly people. Lancet 351, 1303-7.
  • B. Efron and D. Feldman (1991) Compliance as an explanatory variable in clinical trials. JASA 86, 9-17.