NIH Public Access

Author Manuscript
Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
Published in final edited form as:
NIH-PA Author Manuscript

Int J Geriatr Psychiatry. 2012 August ; 27(8): 777–784. doi:10.1002/gps.2783.

Change in Cognitive Functioning in Depressed Older Adults

Following Treatment with Sertraline or Nortriptyline
Michelle E. Culang-Reinlieb1,2, Joel R. Sneed2,3, John G. Keilp3, and Steven P. Roose3
1Graduate Center, City University of New York

2Queens College, City University of New York

3Columbia University and the New York State Psychiatric Institute

Abstract
Objective—To compare the impact of nortriptyline to sertraline on change in cognitive
functioning in depressed older adults.
NIH-PA Author Manuscript

Methods—We used pre-post neuropsychological data collected as part of a 12-week medication

trial comparing sertraline to nortriptyline in the treatment of older adults with non-psychotic,
unipolar major depression to examine change in cognitive functioning. Neuropsychological
assessments included mental status (Mini-Mental Status Exam), psychomotor speed (Purdue
Pegboard), attention (Continuous Performance Test; Trail Making Test A), executive functioning
(Stroop Color/Word Test; Trail Making Test B), and memory (Buschke Selective Reminding
Test).
Results—Within treatment groups, patients treated with sertraline improved only on verbal
learning. This change did not depend on responder status. Between treatment groups, patients
treated with sertraline improved more in verbal learning compared to patients treated with
nortriptyline. Looking at change in cognition as a function of medication condition and responder
status revealed that sertraline responders improved more in verbal learning compared to
nortriptyline responders but not more than sertraline non-responders or nortriptyline non-
responders. Nortriptyline responders were the only treatment by responder status group to show no
improvement in verbal learning from baseline to endpoint.
Conclusions—Unexpectedly, nortriptyline responders showed no improvement in verbal
learning as compared to patients treated with sertraline or nortriptyline non-responders. However,
NIH-PA Author Manuscript

given the small sample sizes and number of statistical tests (potential for type 1 error), replication
is warranted.

Keywords
cognitive functioning; cognitive impairment; depression; nortriptyline; sertraline

Cognitive impairment is common in geriatric depression. Cognitive domains reported to be

affected include memory (Gallassi, et al. 2006; Kramer-Ginsberg, et al. 1999; Salloway, et
al. 1996), visuospatial functioning (Butters, et al. 2004; Kramer-Ginsberg et al. 1999),
information processing speed (Lesser, et al. 1996; Nebes, et al. 2000), and executive
functioning (Lesser et al. 1996; Lockwood, et al. 2002). The conversion rate of late-life

Corresponding Author: Michelle Culang-Reinlieb, Queens College, City University of New York, Department of Psychology, 65-30
Kissena Blvd, Flushing, NY, 11367, USA. Michelle.Culang@qc.cuny.edu.
Disclosures: SPR has received consultant fees from Medtronics and Orexigen
 Culang-Reinlieb et al. Page 2

depression into dementia among those with cognitive impairment is much higher than those
without cognitive impairment (Modrego and Ferrandez 2004). Antidepressant medication is
the first line of treatment for depression, particularly in the older adult community where
NIH-PA Author Manuscript

primary care doctors provide the majority of treatment. Therefore, it is important to

understand the effects of antidepressant medication on cognition among the depressed
elderly, a cognitively vulnerable population.

When examining the effect of antidepressant medication on cognition in depressed older

adults, two interrelated sets of questions arise. The first set of questions focuses on within-
treatment changes in cognitive functioning: a) does cognitive functioning change from pre-
to post-treatment? and b) is there a differential impact of medication on cognition depending
on whether or not the patient responds? Within the second set of questions, the focus is on
between-treatment changes in cognitive functioning: a) is there a differential impact of
antidepressant medication on cognitive functioning depending on medication class (e.g.,
SSRIs, SNRIs, TCAs, MAOs)? b) does change in cognitive functioning between treatment
conditions depend on responder status? Of course, the concept of cognitive functioning is
broad and consists of a number of functions including but not limited to memory, attention,
processing speed, executive functioning, and visuospatial/visuoconstructional skills.
Therefore, it is important to examine each of the above questions within the context of a
multi-domain assessment to address how antidepressant medication (including whether or
not the patient responds) differentially affects each of these cognitive domains.
NIH-PA Author Manuscript

A number of studies have examined the impact of antidepressant medication on cognition

but have not fully addressed the two interrelated sets of questions we have outlined (Bhalla,
et al. 2006; Bondareff, et al. 2000; Butters, et al. 2000; Doraiswamy, et al. 2003; Gallassi et
al. 2006; Nebes, et al. 2003; Raskin, et al. 2007). For example, some studies did not address
whether change in cognition depends on response to treatment (Bondareff et al. 2000;
Raskin et al. 2007). Other studies included only responders in their analyses rather than
comparing change in cognition among responders and non-responders, which would have
helped to differentiate the impact of medication versus response on change in cognitive
functioning (Bhalla et al. 2006; Butters et al. 2000; Gallassi et al. 2006; Nebes et al. 2003).
Finally, another study used a composite cognitive score in its analyses and thus did not
address the differential impact of medication on different cognitive domains (Doraiswamy et
al. 2003).

To our knowledge, few studies have fully tested the impact of medication on cognitive
function in depressed older adults. For example, in an 8-week randomized placebo-
controlled trial of citalopram (Culang, et al. 2009), antidepressant non-response was
associated with cognitive decline in verbal learning and psychomotor speed. Although
NIH-PA Author Manuscript

responders on medication improved on some domains (psychomotor speed and visuospatial

functioning), their improvement did not exceed the expected practice effect observed among
patients randomized to placebo. This study, therefore, shows decline in some domains and
improvement in other domains, and is an example of a differential impact of medication on
cognition. In another study, nortriptyline and phenelzine produced no change in any
cognitive domain assessed (verbal learning, psychomotor speed, visual memory) in
depressed older adults when compared to placebo, and this effect did not depend on
responder status (Georgotas, et al. 1989). In a third study, an uncontrolled trial in depressed
older adults showed no change in any cognitive domain assessed (processing speed,
executive functioning, verbal and visual memory, visuoperceptive functioning, and
attention) over 12 months of treatment when compared to age-matched controls (Portella, et
al. 2003). Furthermore, there were no differences in change among remitted and non-
remitted patients.

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 3

The purpose of this study was to examine the impact of a TCA and SSRI on cognition and to
determine whether change in cognition depends on response to treatment and cognitive
domain. To accomplish this aim, we used pre-post neuropsychological data on global
NIH-PA Author Manuscript

cognitive functioning, verbal learning, attention, psychomotor speed, and executive

functioning (the switching and response inhibition components), collected as part of a 12-
week, randomized, double-blind, parallel-group design comparing sertraline and
nortriptyline in the treatment of depressed older adults. In the context of this multi-domain
assessment, we addressed the two sets of questions outlined above: 1.a) whether there were
pre- to post-treatment changes in cognitive functioning within the sertraline and nortriptyline
conditions, 1.b) whether change in cognitive functioning within each treatment condition
depended on responder status, 2.a) whether there were differences in change in cognitive
functioning between the nortriptyline and sertraline conditions, and 2.b) whether change in
cognitive functioning between treatment conditions depended on responder status.

METHOD
Study Procedures
This study was a double-blind, randomized, 12-week clinical trial comparing nortriptyline to
sertraline in depressed older adults. Patients were recruited by radio and newspaper
advertisements and/or through referral from other physicians. At the initial visit, a
comprehensive psychiatric evaluation, including a Structured Clinical Interview for DSM-
NIH-PA Author Manuscript

IV, 24-item Hamilton Rating Scale for Depression (HRSD), Mini-Mental Status
Examination (MMSE), Newcastle I scale for the assessment of melancholia, and a medical
history were performed. If the patient met inclusion criteria and signed informed consent, a
physical examination, ECG, CBC, chemistries, electrolytes, and thyroid panel were
performed.

Inclusion criteria were 1) age > 45; 2) unipolar depression, single or recurrent, nonpsychotic,
by DSM-IV criteria; 3) HRSD ≥ 16 at the initial visit and at the end of 1 week of placebo; 4)
MMSE score ≥ 24; and 5) willing and able to give informed consent. Exclusion criteria were
1) current or history of obsessive-compulsive disorder, psychotic disorder, or substance
dependence within the past year (other than nicotine) by DSM-IV criteria; 2) judged to be a
current suicide risk or serious suicide attempt within the past year; 3) patients status post
myocardial infarction, coronary artery bypass, or angioplasty, or with a positive history of
angina or positive stress test; 4) QRS interval greater than 0.12 sec or QTc interval ≥ 46
msec; 5) treatment with coumadin, heparin or type 1 antiarrhythmic medications; 6)
diagnosis of narrow angle glaucoma; 7) stroke, epilepsy or Parkinson’s disease; 8) acute,
severe or unstable medical condition; 9) positive urine toxicology screen for drugs of abuse
including amphetamine, barbiturates, cocaine, marijuana, methadone, methaqualone,
NIH-PA Author Manuscript

opioids, and PCP; 10) treatment in the current episode of depression with either nortriptyline
with a plasma level between 50 and 150 ng/ml, desipramine or imipramine with a plasma
level of 250 ng/ml or greater, paroxetine 40mg, fluoxetine 40mg, or sertraline 200mg for at
least 4 weeks.

Patients who met inclusion/exclusion criteria and signed informed consent were given one
week of single-blind placebo. If patients still met inclusion/exclusion criteria at the end of
the placebo week and did not reduce their HRSD score by 25%, they were randomized. The
assessments performed at the end of the placebo week and every visit thereafter included the
HRSD, the Montgomery-Åsberg Depression Rating Scale (MADRS), the Beck Depression
Inventory (BDI), and the Clinical Global Impression (CGI) of severity and improvement.
The Hamilton Anxiety Rating Scale was performed at baseline and at the end of weeks 2, 4,
and 8 of treatment; the Medical Outcomes Study 36-Item Short-Form Health Survey and the
MMSE were performed at baseline and at the end of week 12 or upon early termination. The

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 4

Cumulative Illness Rating Scale for Geriatrics (CIRS-G) was also administered at baseline.
Stratification of the sample was based on diagnosis of melancholia by DSM-IV criteria
(questions resolved by case conference). Randomization was done using permuted blocks of
NIH-PA Author Manuscript

ten.

Participants randomized to sertraline received 50 mg for one week and then 100 mg for the
next 4 weeks. If the patient did not meet criteria for remission (HRSD < 10) by week 5, the
dose was increased to 150 mg/day. If the patient did not show evidence of response by week
9, the dose was increased to 200 mg/day. The nortriptyline dose was calculated at 1 mg/kg;
1/3 of that dose was given days 1 through 3, 2/3 on days 3 through 6, and the full dose of
medication was given on day 7. A plasma level was drawn 7 days later and the dose of
nortriptyline was adjusted so that the plasma level was within 80-120 ng/ml. The New York
State Psychiatric Institute IRB approved this study.

Neuropsychological Test Battery

The test battery was designed to assess a number of cognitive functions pertinent to aging
and major depression including mental status, psychomotor speed, attention, and memory.
The tests included the 30-item Folstein MMSE (Folstein, et al. 1975) to estimate global
cognitive functioning, the Purdue Pegboard (both hands) (Tiffin and Asher 1948) as a
measure of psychomotor speed, the 4-digits fast condition of the Continuous Performance
Test (CPT) – Identical Pairs (Cornblatt, et al. 1988) and Trail Making Test A (TMT A)
NIH-PA Author Manuscript

(Reitan and Wolfson 1985) to assess attention, the Stroop Color/Word Test (MacLeod 1991)
and Trail Making Test B (TMT B) (Reitan and Wolfson 1985) to assess the response
inhibition and switching components of executive functioning, respectively, and the
Buschke Selective Reminding Test (SRT) (Buschke and Fuld 1974) as a measure of verbal
learning. Two of the tests (CPT and Stroop) were presented on a Macintosh laptop computer
and were written in the PsyScope programming language (Cohen, et al. 1993). Performance
on the CPT was summarized by d-prime, a sensitivity index that represents the standardized
difference between hit and false alarm rates. Percent interference (percent change in median
reaction time to color/word versus color responses) was used as the outcome measure on the
Stroop. The other five tests (MMSE, SRT, TMT A and B, and Purdue Pegboard) were
administered by hand. Alternate forms of the CPT and SRT were used in an attempt to
eliminate the problem of practice effects.

Missing Data
One hundred and twelve patients were randomized to treatment with either nortriptyline or
sertraline. Forty-nine patients were missing all neuropsychological data across the two time-
points (baseline and week 12) and were excluded from this study. No differences between
NIH-PA Author Manuscript

those patients with neuropsychological data (n=63) and those without neuropsychological
data (n=49) were detected on any clinical or demographic variable. To accommodate
missing data for the remaining sample (n=63), we used the multiple imputation (Schafer and
Olsen 1998) procedure in SPSS. Multiple imputation replaces missing data with a set of
plausible values based on all variables in the working dataset, which includes demographic,
clinical outcome, and neuropsychological test variables. To capture the uncertainty in the
estimated values, multiple imputation is conducted several times yielding similar but not
identical datasets. This report is based on five imputed data sets, which is sufficient to obtain
excellent results unless rates of missing data are exceptionally high (Schafer 1999). The five
imputed data sets are analyzed separately using standard statistical analyses. Results from
the analyses are then combined using Rubin’s rules (Schafer and Graham 2002; Schafer and
Olsen 1998) to generate valid statistical inferences that reflect uncertainty due to missing
values and improve the accuracy of the results.

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 5

Statistical Analyses
Prior to testing for differences in change in neuropsychological test performance, we used
simple and logistic regression in SPSS to test for differences at baseline between the two
NIH-PA Author Manuscript

treatment conditions as well as the four treatment condition by responder status groups (see
below). There were no differences on age, education, gender, baseline depression severity,
responder status, or on any of the neuropsychological tests when comparing the two
(treatment condition) and four (responder status by treatment condition) patient groups.
Therefore, we did not adjust for demographic, clinical, or neuropsychological tests in the
subsequent analyses.

To test whether antidepressant medication (sertraline or nortriptyline) had an impact on

cognition, pre- and post-treatment performance on each of the neuropsychological tests were
compared using paired t-tests. Independent samples t-tests were used to determine whether
change in cognitive performance within each treatment condition depended on responder
status; in this model, change scores were computed for each neuropsychological test (t2-t1)
and used as the outcome variable and responder status was treated as the independent
variable. We next tested for differences in change in neuropsychological test performance
between the two treatment conditions using an independent samples t-test. In these analyses,
treatment condition (nortriptyline=0, sertraline=1) was treated as the independent variable
and the change score as the outcome variable. Finally, to test whether change in
neuropsychological test performance between treatment conditions depended on responder
NIH-PA Author Manuscript

status, we used a dummy-coded variable to designate the four patient groups (sertraline
responders, sertraline non-responders, nortriptyline responders, and nortriptyline non-
responders). Multiple regression was used and the neuropsychological test change scores
were again treated as the outcome variable.

The partial or regressed change approach to two time point data is often recommended
(Cohen, et al. 2003). In this procedure, the endpoint neuropsychological test score is treated
as the outcome variable and the baseline test score is treated as a covariate. This effectively
removes all correlation of the endpoint score from the baseline score and represents an
improvement over simple change scores (subtracting baseline from endpoint) which tend to
overcorrect the endpoint score by the baseline score due to unreliability of measurement
(Cohen et al. 2003). We used a change score model in order to be consistent throughout our
statistical analyses and to facilitate the presentation of results. Furthermore, conducting the
analyses using both strategies did not yield substantively different findings. Throughout our
analyses, significance tests were evaluated at the 5% level.

RESULTS
NIH-PA Author Manuscript

Descriptive Statistics
Table 1 presents baseline demographic and clinical characteristics of the total sample (n=63)
and sertraline and nortriptyline subgroups. The average study participant was 64 years old
and completed about 4 years of college. Approximately 60% of the sample were women,
average baseline depression severity was 24.37 on the 24-item HRSD, and 43% of the
sample were classified as responders. The average MMSE score of the sample at baseline
was 27.71.

Table 2 presents complete case pre- and post-treatment data for all neuropsychological tests
by medication group and responder status. As can be seen from Table 2, test scores
remained relatively stable in both treatment conditions and in the treatment condition by
responder status groups with the exception of scores on the Buschke SRT, CPT and TMT B.
Qualitatively, nortriptyline responders showed no improvement on the Buschke SRT as
compared to patients treated with sertraline or nortriptyline non-responders. On the TMT B,

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 6

nortriptyline responders seemed to decline as compared to the improvement seen among

nortriptyline non-responders. Finally, both nortriptyline and sertraline responders appeared
to improve on the CPT whereas non-responders across treatment conditions declined. To
NIH-PA Author Manuscript

formally test these apparent differences in change based on complete case data, we ran a
series of analyses using change scores on each test as the outcome variable and
accommodated missing data using multiple imputation.

Hypothesis Testing
We first compared pre- and post-treatment neuropsychological test scores within each
treatment condition to address whether antidepressant medication has an impact on cognitive
functioning. Within the sertraline condition, significant change occurred on the Buschke
SRT [t(2082)=-6.30, p=.001]. No change was observed on the other neuropsychological
tests. Within the nortriptyline condition, no significant change was observed on any of the
neuropsychological tests. These results are graphically displayed in Figure 1, which depicts
pre- to post-treatment change in cognitive performance within each treatment condition
across the seven neuropsychological tests.

Next, we examined whether change in cognitive functioning within treatment depended on

response. In the sertraline condition, 33% of patients responded compared to 53% in the
nortriptyline condition (χ2(1)=2.57, p=.11). Change in cognitive performance from baseline
to endpoint on the Buschke SRT did not depend on response to sertraline [t(40)=-0.54, p=.
NIH-PA Author Manuscript

60]. Although there were no statistically significant changes in cognition within the
nortriptyline condition, we nevertheless examined whether there were differences depending
on response; however, no significant findings were observed.

We next examined whether there were differences in change in cognitive functioning

between the nortriptyline and sertraline conditions. There was a statistically significant
difference between the sertraline and nortriptyline conditions at endpoint on the Buschke
SRT. Although both treatment conditions improved on the Buschke SRT from baseline to
endpoint, patients treated with sertraline improved significantly more than patients treated
with nortriptyline [t(8803)=-2.44, p=0.02]. No other comparisons between sertraline and
nortriptyline were statistically significant.

Finally, we examined whether change in cognitive functioning between treatment conditions

depended on responder status. Patients who responded on sertraline improved significantly
more on the Buschke SRT than patients who responded on nortriptyline [B=-15.73,
SE=6.14, t=-2.56, p=0.01] but no more than sertraline non-responders or nortriptyline non-
responders. No other comparisons were statistically significant.
NIH-PA Author Manuscript

DISCUSSION
The purpose of the present study was to examine the impact of nortriptyline and sertraline
on change in cognitive functioning of depressed older adults using data from a twelve-week,
double-blind, randomized clinical trial. Within this multi-domain assessment, we addressed
two interrelated sets of questions: 1) Within treatment condition, does cognitive functioning
change from pre- to post-treatment and does it depend on medication response? 2) Between
treatment conditions, is there a differential effect of medication on change in cognition and
does it depend on medication response?

We found that patients treated with sertraline showed a significant change in verbal learning
from pre- to post-treatment, but this change did not depend on responder status. Therefore,
taking sertraline improved memory regardless of whether the patient responded to the
medication. Of course, this effect could be because of sertraline or non-specific factors

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 7

associated with participating in a clinical trial. However, improvement in memory (or any
other cognitive domain) was not observed in the nortriptyline condition. Therefore, we can
infer that memory improvement is likely to be associated with taking sertraline because we
NIH-PA Author Manuscript

did not see a similar effect in the nortriptyline condition, which shared the same non-specific
factors. This finding is consistent with previous reports showing improvement in memory on
sertraline in non-depressed older adults (Furlan, et al. 2001; Schmitt, et al. 2001).

We next compared change in cognitive functioning between the nortriptyline and sertraline
conditions and found that patients treated with sertraline showed significantly more
improvement in verbal learning compared to patients treated with nortriptyline. This finding
is consistent with previous studies comparing the impact of sertraline to nortriptyline on the
cognitive functioning of older adults. For example, treatment with sertraline in the geriatric
depressed led to greater improvement in verbal learning (as measured by the Shopping List
Task (SLT)) when compared to treatment with nortriptyline (Doraiswamy et al. 2003). In
another study of depressed older adults, sertraline treatment led to improvement in verbal
learning (as measured by the SLT) whereas nortriptyline treatment led to a mild decline over
12 weeks of treatment (Finkel, et al. 1999).

Finally, we compared change in cognition as a function of medication condition and

responder status and found that sertraline responders showed significantly more
improvement in verbal learning compared to nortriptyline responders but no more than
NIH-PA Author Manuscript

sertraline non-responders or nortriptyline non-responders. To our surprise, nortriptyline

responders were the only treatment by responder status group to show no improvement in
verbal learning from baseline to endpoint.

The most cogent explanation for this unexpected finding is that memory improvement is
blocked by the anticholinergic effect of nortriptyline. Tricyclic antidepressants have five
times the anticholinergicity of SSRIs in older adults (Pollock, et al. 1998). More specifically,
sertraline has been found to produce no anticholinergic activity at therapeutic doses whereas
nortriptyline demonstrates a moderate anticholinergic activity (5-15 pmol/ml) (Chew, et al.
2008). Furthermore, drug-induced anticholinergic activity has been associated with
cognitive impairment in older adults (Oxman 1996); greater anticholinergic effect was
significantly (negatively) associated with endpoint cognitive improvement (in verbal
learning and processing speed) in depressed older adults (Doraiswamy et al. 2003). In
another study of the geriatric depressed, higher plasma nortriptyline concentration over 6
weeks of treatment was associated with poorer free recall but better affective outcome
(Young, et al. 1991) indicating that the therapeutic and cognitive effects of nortriptyline may
have different mechanisms. Even very low anticholinergic activity has been associated with
specific cognitive deficits. In one study, depressed elderly subjects with serum
NIH-PA Author Manuscript

anticholinergic activity performed more poorly in verbal learning than did those without
anticholinergic activity (Nebes, et al. 1997). However, plasma drug levels of nortriptyline
were blood-controlled in the present study, and it is unlikely that the anticholinergic effect
differentially impacted the cognitive functioning of responders and non-responders on
nortriptyline.

Another possible explanation for the unexpected finding is that the nortriptyline responder
group had a disproportionately high number of cognitively impaired patients. However, the
average MMSE score at baseline for the sample was 27.01, which is within normal limits
and there was no significant difference in MMSE score between the four treatment by
responder status groups at baseline. There were also no differences between the four groups
in age or education. It is also possible that the overall medical burden was higher among
nortriptyline responders compared to the other three groups. The interaction between
medical illness and antidepressant medication could adversely affect cognitive functioning.

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 8

However, there were no significant differences in medical burden (as assessed by the
Cumulative Illness Rating Scale for Geriatrics) between the four groups.
NIH-PA Author Manuscript

This study should be interpreted in the context of several limitations. First, the sample size
was relatively small and this study was not specifically powered to detect between drug
differences in cognitive function by responder status. The responder analyses should
therefore be interpreted with caution. Furthermore, the small sample size did not allow for a
test of the interaction between depressive subtype, responder status, and treatment condition
on change in cognitive function nor did it allow us to take into account differences due to
depressive subtype (melancholia vs. non-melancholia). Second, the findings of this study
may have been only a statistical anomaly. These were post-hoc analyses involving a
multiplicity of statistical tests. Therefore, the findings are intended to be hypothesis
generating only and are clearly in need of replication. Third, although multiple cognitive
domains were examined, the assessment within each domain was relatively limited. Fourth,
there was no placebo control group making it difficult to determine whether the observed
improvement in verbal learning was nothing more than a practice effect. However, treatment
with sertraline led to an improvement on the Buschke SRT that exceeded the improvement
observed in patients randomized to placebo in our previous study (Culang et al. 2009) (¾
and ¼ of a standard deviation, respectively), suggesting there was significant change from
pre- to post-treatment beyond a practice effect.
NIH-PA Author Manuscript

Acknowledgments
This research was supported by a grant from National Institute of Mental Health grants R01 MH55716 (Steven P.
Roose) and K23 MH075006 (Joel R. Sneed).

References
Bhalla RK, Butters MA, Mulsant BH, Begley AE, Zmuda MD, Schoderbek B, Pollock BG, Reynolds
CF, Becker JT. Persistence of Neuropsychologic Deficits in the Remitted State of Late-Life
Depression. American Journal of Geriatric Psychiatry. 2006; 14:419–427. [PubMed: 16670246]
Bondareff W, Alpert M, Friedhoff AJ, Richter EM, Clary CM, Batzar E. Comparison of sertraline and
nortriptyline in the treatment of major depressive disorder in late life. American Journal of
Psychiatry. 2000; 157:729–736. [PubMed: 10784465]
Buschke H, Fuld P. Evaluating storage, retention, and retrieval in disordered memory and learning.
Neurology. 1974; 24:1019–1025. [PubMed: 4473151]
Butters MA, Becker JT, Nebes RD, Zmuda MD, Mulsant BH, Pollock BG, Reynolds CF III. Changes
in Cognitive Functioning Following Treatment of Late-Life Depression. Am J Psychiatry. 2000;
157:1949–1954. [PubMed: 11097959]
Butters MA, Whyte EM, Nebes RD, Begley AE, Dew MA, Mulsant BH, Zmuda MD, Bhalla R,
NIH-PA Author Manuscript

Meltzer CC, Pollack BG, et al. The nature and determinants of neuropsychological functioning in
late-life depression. Archives of General Psychiatry. 2004; 61:587–595. [PubMed: 15184238]
Chew ML, Mulsant BH, Pollock BG, Lehman ME, Greenspan A, Mahmoud RA, Kirshner MA,
Sorisio DA, Bies RR, Gharabawi G. Anticholinergic activity of 107 medications commonly used by
older adults. Journal of the American Geriatrics Society. 2008; 56:1333–1341. [PubMed: 18510583]
Cohen, J.; Cohen, P.; West, SG.; Aiken, LS. Applied multiple regression/correlation for the behavioral
sciences. Mahwah, NJ: Lawrence Erlbaum Publishers; 2003.
Cohen JD, MacWhinney B, Flatt M, Provost J. PsyScope: An interactive graphic system for designing
and controlling experiments in the psychology laboratory using Macintosh computers. Behavioral
Research, Methods, Instruments, and Computers. 1993; 25:257–271.
Cornblatt BA, Risch NJ, Faris G, Friedman D, Erlenmeyer-Kimling L. The Continuous Performance
Test, identical pairs version (CPT-IP): I. New findings about sustained attention in normal families.
Psychiatry Res. 1988; 26:223–238. [PubMed: 3237915]

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 9

Culang M, Sneed J, Keilp J, Rutherford B, Pelton GH, Devanand DP, Roose SP. Change in cognitive
functioning following acute antidepressant treatment in late-life depression. Am J Geriatr
Psychiatry. 2009; 17:881–888. [PubMed: 19916207]
NIH-PA Author Manuscript

Doraiswamy PM, Krishnan KR, Oxman T, Jenkyn LR, Coffey DJ, Burt T, Clary CM. Does
antidepressant therapy improve cognition in elderly depressed patients? J Gerontol A Biol Sci Med
Sci. 2003; 58:M1137–1144. [PubMed: 14684712]
Finkel SI, Richter EM, Clary CM, Batzar E. Comparative efficacy of sertraline vs. fluoxetine in
patients age 70 or over with major depression. American Journal of Geriatric Psychiatry. 1999;
7:221–227. [PubMed: 10438693]
Folstein MF, Folstein SE, McHugh PR. Mini-mental state: A practical method for grading the state of
patients for the clincian. Journal of Psychiatric Research. 1975; 12:189–198. [PubMed: 1202204]
Furlan PM, Kallan MJ, Ten Have T, Pollock BG, Katz I, Lucki I. Cognitive and psychomotor effects
of paroxetine and sertraline on healthy elderly volunteers. Geriatr Psychiatry. 2001; 9
Gallassi R, Di Sarro R, Morreale A, Amorec M. Memory impairment in patients with late-onset major
depression: The effect of antidepressant therapy. Journal of Affective Disorders. 2006; 91:243–
250. [PubMed: 16499971]
Georgotas A, McCue RE, Reisberg B, Ferris SH, Nagachandran N, Chang I, Mir P. The Effects of
Mood Changes and Antidepressants on the Cognitive Capacity of Elderly Depressed Patients.
International Psychogeriatrics. 1989; 1:135–143. [PubMed: 2491140]
Kramer-Ginsberg E, Greenwald BS, Krishnan KRR, Christiansen B, Hu J, Ashtari M, Patel M, Pollack
S. Neuropsychological Functioning and MRI Signal Hyperintensities in Geriatric Depression. Am
NIH-PA Author Manuscript

J Psychiatry. 1999; 156:438–444. [PubMed: 10080561]

Lesser I, Boone K, Mehringer C, Wohl M, Miller B, Berman N. Cognition and white matter
hyperintensities in older depressed patients. American Journal of Psychiatry. 1996; 153:1280–
1287. [PubMed: 8831435]
Lockwood KA, Alexopoulos GS, van Gorp WG. Executive dysfunction in geriatric depression.
American Journal of Psychiatry. 2002; 159:1119–1126. see comment. [PubMed: 12091189]
MacLeod C. A half-century of research on the Stroop effect: An integrative review. Psychological
Bulletin. 1991; 109:163–203. [PubMed: 2034749]
Modrego PJ, Ferrandez J. Depression in patients with mild cognitive impairment increases the risk of
developing dementia of Alzheimer type: a prospective cohort study. Arch Neurol. 2004; 61:1290–
1293. [PubMed: 15313849]
Nebes RD, Butters MA, Mulsant BH, G PB, Zmuda MD, Houck PR, Reynolds CF III. Decreased
working memory and processing speed mediate cognitive impairment in geriatric depression.
Psychological Medicine. 2000; 30:679–691. [PubMed: 10883722]
Nebes RD, Pollock BG, Houck PR, Butters MA, Mulsant BH, Zmuda MD, Reynolds CF III.
Persistence of cognitive impairment in geriatric patients following antidepressant treatment: a
randomized, double-blind clinical trial with nortriptyline and paroxetine. Journal of Psychiatric
Research. 2003; 37:99–108. [PubMed: 12842163]
Nebes RD, Pollock BG, Mulsant BH, Kirshner MA, Halligan E, Zmuda M, Reynolds CFr. Low-level
NIH-PA Author Manuscript

serum anticholinergicity as a source of baseline cognitive heterogeneity in geriatric depressed

patients. Psychopharmacology Bulletin. 1997; 33:715–720. [PubMed: 9493484]
Oxman TE. Antidepressants and cognitive impairment in the elderly. Journal of Clinical Psychiatry.
1996; 57:38–44. [PubMed: 8647791]
Pollock BG, Mulsant BH, Nebes R, Kirshner MA, Begley AE, Mazumdar S, Reynolds CFr. Serum
anticholinergicity in elderly depressed patients treated with paroxetine or nortriptyline. American
Journal of Psychiatry. 1998; 155:1110–1112. [PubMed: 9699704]
Portella MJ, Marcos T, Rami L, Navarro V, Gasto C, Salamero M. Residual cognitive impairment in
late-life depression after a 12-month period follow-up. International J Geriatric Psychiatry. 2003;
18:571–576.
Raskin J, Wiltse CG, Siegal A, Sheikh J, Xu J, Dinkel JJ, Rotz BT, Mohs RC. Efficacy of duloxetine
on cognition, depression, and pain in elderly patients with major depressive disorder: an 8-week,
double-blind, placebo-controlled trial. Am J Psychiatry. 2007; 164:900–909. [PubMed: 17541049]

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 10

Reitan, RM.; Wolfson, D. The Halstead-Reitan neuropsychological test battery: Theory and clinical
interpretation. Tucson: Neuropsychology Press; 1985.
Salloway S, Malloy P, Kohn R, Gillard E, Duffy J, Rogg J, Tung G, Richardson E, Thomas C,
NIH-PA Author Manuscript

Westlake R. MRI and neuropsychological differences in early- and late-life-onset geriatric

depression. Neurology. 1996; 46:1567–1574. [PubMed: 8649550]
Schafer JL. Multiple imputation: A primer. Statistical Methods in Medical Research. 1999; 8:3–15.
[PubMed: 10347857]
Schafer JL, Graham JW. Missing data: Our view of the state of the art. Psychological Methods. 2002;
7:147–177. [PubMed: 12090408]
Schafer JL, Olsen MK. Multiple imputation for multivariate missing-data problems: A data analyst’s
perspective. Multivariate Behavioral Research. 1998; 33:545–571.
Schmitt JAJ, Kruizinga MJ, Riedel WJ. Non-serotonergic pharmacological profiles and associated
cognitive effects of serotonin reuptake inhibitors. Journal of Psychopharmacology. 2001; 15:173–
179. [PubMed: 11565624]
Tiffin J, Asher EJ. The purdue pegboard; norms and studies of reliability and validity. J Appl Psychol.
1948; 32:234–247. [PubMed: 18867059]
Young RC, Mattis S, Alexopoulos GS, Meyers BS, Shindledecker RD, Dhar AK. Verbal memory and
plasma drug concentrations in elderly depressives treated with nortriptyline. Psychopharmacology
Bulletin. 1991; 27:291–294. [PubMed: 1775601]
NIH-PA Author Manuscript
NIH-PA Author Manuscript

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 Culang-Reinlieb et al. Page 11
NIH-PA Author Manuscript

Figure 1.
Change in cognitive performance from pre to post-treatment in the (a) sertraline condition
and (b) nortriptyline condition across seven neuropsychological tests.
NIH-PA Author Manuscript
NIH-PA Author Manuscript

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Table 1
Baseline Clinical and Demographic Characteristics of the total sample, the sertraline and nortriptyline conditions, and the four patients groups classified by treatment condition and responder status (complete
case data).

Variable Total Sample Sertraline Nortriptyline Sertraline Sertraline Nortriptyline Nortriptyline

(n=63) (n= 33) (n= 30) Responders Non-responders Responders Non-responders
(n=11) (n=22) (n=16) (n=14)
Culang-Reinlieb et al.

Age 64.19 (8.47) 64.85 (8.83) 63.47 (8.15) 65.82 (9.19) 64.36 (8.82) 63.25 (8.94) 63.71 (7.47)
Women (%) 60 61 60 45 68 63 57
Education, Yr 16.17 (2.42) 16.29 (2.14) 16.04 (2.73) 17.40 (.52) 15.67 (2.45) 16.23 (2.35) 15.83 (3.19)
HRSD Baseline 24.37 (4.87) 23.91 (4.38) 24.87 (5.39) 23.18 (4.56) 24.27 (4.36) 24.81 (4.45) 24.93 (6.47)
CGI-Severity Baseline 4.35 (.93) 4.38 (.71) 4.33 (1.12) 4.30 (.82) 4.41 (.67) 4.63 (.72) 4.00 (1.41)
CIRS-G 3.11 (2.14) 3.52 (2.22) 2.70 (2.0) 3.30 (1.83) 3.60 (2.48) 2.25 (1.69) 3.21 (2.26)

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
Page 12
 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Table 2
Unadjusted neuropsychological test performance scores at baseline and endpoint for the sertraline and nortriptyline subsamples and the four patient groups classified according to treatment condition and
responder status (complete case data)

NP Test Sertraline Nortriptyline Sertraline Sertraline Nortriptyline Nortriptyline

(n=33) (n=30) Responders Non-responders Responders Non-responders
(n=11) (n=22) (n=16) (n=14)
Culang-Reinlieb et al.

PRE POST PRE POST PRE POST PRE POST PRE POST PRE POST
MMSE 27.85 (2.56) 28.19 (2.16) 27.55 (2.87) 28.32 (2.50) 28.45 (2.07) 28.55 (2.34) 27.55 (2.77) 27.80 (1.99) 27.07 (3.37) 28.00 (2.85) 28.07 (2.24) 29.17 (.75)

TMT A 48.16 (36.39) 48.55 (24.27) 49.86 (20.64) 51.01 (20.65) 38.30 (9.83) 36.75 (12.09) 53.10 (43.54) 61.54 (28.08) 50.23 (21.53) 55.09 (21.68) 49.46 (20.44) 40.13 (13.68)

TMT B 106.37 (52.53) 101.80 (49.38) 109.49 (63.89) 126.15 (80.73) 89.55 (48.96) 90.24 (39.45) 114.79 (53.31) 115.92 (58.62) 118.16 (72.68) 142.95 (87.37) 100.20 (54.08) 81.35 (34.06)

CPT 1.74 (1.02) 1.57 (.97) 1.51 (.56) 1.48 (.90) 1.75 (1.11) 1.96 (1.05) 1.72 (.98) 1.12 (.68) 1.49 (.52) 1.71 (.89) 1.56 (.74) .95 (.78)

Purdue Pegboard 10.11 (3.11) 10.02 (3.13) 10.66 (3.06) 11.18 (3.84) 10.91 (3.30) 10.14 (3.48) 9.70 (3.00) 9.90 (2.88) 11.00 (3.36) 10.94 (4.02) 10.29 (2.79) 11.83 (3.54)

Buschke SRT 102.16 (16.38) 121.55 (16.35) 108.96 (23.18) 114.27 (21.43) 105.36 (15.14) 122.70 (19.85) 100.48 (17.11) 120.40 (12.93) 112.07 (17.96) 112.19 (24.13) 105.62 (28.13) 119.83 (11.44)

Stroop .54 (.34) .55 (.29) .45 (.34) .43 (.27) .45 (.30) .43 (.23) .59 (.35) .67 (.31) .43 (.41) .43 (.30) .47 (.27) .43 (.18)

Note. Table values are means and standard deviations.

MMSE=Mini Mental Status Exam, total number correct; Stroop=Color/Word Test, interference effect; CPT=Continuous Performance Test, d-prime; TMT=Trail Making Test, seconds; Buschke SRT=Buschke Selective Reminding Test, immediate recall, total number correct

Int J Geriatr Psychiatry. Author manuscript; available in PMC 2013 August 01.
Page 13