VOLUME 18 NUMBER 8 | AUgUst 2010 | www.obesityjournal.org
Methods and techniques
nature publishing group
Many, but not all, of the atypical antipsychotic drugs increase
body weight (1–4). Depending on the atypical antipsy-
chotic, the magnitude of the average weight gain can be as
high as 10 kg (5) with individual patients gaining 40 kg (6).
Among the atypical antipsychotic drugs, clozapine and olan-
zapine (OLZ) provoke the greatest weight gain (2). The gain
in weight has significant clinical impact in a patient popula-
tion already prone to diabetes mellitus and coronary artery
The most likely explanation for the increase in weight is
increased appetite (4). Increased food intake (FI) has been
documented in patients with schizophrenia (9), although
some report decreases in energy expenditure (EE) (10).
As feeding behavior and weight are reflections of a com-
plex set of multiple, interacting variables, it is particularly
difficult to define the reasons for the weight changes in
patients with schizophrenia. For example, schizophrenia is
often accompanied by agitated behavior, whereas atypical
antipsychotic drugs are typically sedating. Do patients gain
weight because agitation has decreased and/or they are more
sedated? Although different antipsychotic drugs do not all
yield an increase in weight, separating the components of
relative efficacy, altered behavior, and environment in dispa-
rate groups of patients is very difficult. Thus, the aim of the
present study is to specifically focus on the pharmacology of
OLZ, the prototypic and highly prescribed atypical antipsy-
chotic that reproducibly induces weight gain. The intent is
to define the changes in FI and EE in the context of measur-
ing activity in healthy men to determine whether there is any
evidence of sedation. In the same study, and as suggested
by others (11), we assessed measures of insulin sensitiv-
ity (IS) as well as explored the mechanistic basis for these
changes within the same subjects uncomplicated by schizo-
phrenia. However, as there is high interindividual variability
in many of these measures, we studied subjects using a ran-
domized, double-blind, placebo (PBO) controlled, crossover
Increased Food Intake and Energy
Expenditure Following Administration
of Olanzapine to Healthy Men
Robert J. Fountaine1, Ann E. Taylor1, James P. Mancuso1, Frank L. Greenway2, Lauri O. Byerley3,
Steven R. Smith2, Marlene M. Most2 and David A. Fryburg1
Atypical antipsychotic medications like olanzapine (OLZ) induce weight gain and increase the risk of diabetes in
patients with schizophrenia. The goal of this study was to assess potential mechanisms of OLZ-induced weight
gain and accompanying metabolic effects. Healthy, lean, male volunteers received OLZ and placebo (PBO) in a
randomized, double-blind, crossover study. In periods 1 and 2, subjects received OLZ (5 mg for 3 days then OLZ 10 mg
for 12 days) or matching PBO separated by a minimum 12-day washout. Twenty-four hour food intake (FI), resting
energy expenditure (REE), activity level, metabolic markers, and insulin sensitivity (IS) were assessed. In total, 30
subjects were enrolled and 21 completed both periods. Mean age and BMI were 27 years (range: 18–49 years) and
22.6 ± 2.2 kg/m2, respectively. Relative to PBO, OLZ resulted in a 2.62 vs. 0.08 kg increase in body weight (P < 0.001)
and 18% (P = 0.052 or 345 kcal) increase in FI. Excluding one subject with nausea and dizziness on the day of OLZ FI
measurement, the increase in FI was 547 kcal, (P < 0.05). OLZ increased REE relative to PBO (113 kcal/day, P = 0.003).
Significant increases in triglycerides, plasminogen activator inhibitor-I (PAI-I), leptin, and tumor necrosis factor-α
(TNF-α) were observed. No significant differences in activity level or IS were observed. This study provides evidence
that OLZ pharmacology drives the early increase in weight through increased FI, without evidence of decreased
energy expenditure (EE), activity level, or short-term perturbations in IS.
Obesity (2010) 18, 1646–1651. doi:10.1038/oby.2010.6
1Pfizer Global Research and Development, New London, Connecticut, USA; 2Pennington Biomedical Research Center, Baton Rouge, Louisiana, USA; 3Department of
Physiology, Louisiana State University, New Orleans, Louisiana, USA. Correspondence: Robert J. Fountaine (email@example.com)
Received 27 August 2009; accepted 6 January 2010; published online 4 February 2010. doi:10.1038/oby.2010.6
obesity | VOLUME 18 NUMBER 8 | AUgUst 2010 1647
Methods and techniques
Methods and Procedures
Thirty healthy, lean, male volunteers (mean age 27 (range 18–49) years;
BMI 22.6 ± 2.2 kg/m2 (range: 18.1–25.7 kg/m2)) were enrolled in the
study. Subjects with (i) a history of eating disorders, (ii) a first-degree
relative with type 1 or type 2 diabetes mellitus, (iii) restricted or special
diet treatment within the month before screening, (iv) participation in
any formal or informal weight loss program within 3 months before
screening or during the study, or (v) a dislike of or inability to eat >20%
of the test foods, were excluded from the study. This study was per-
formed in accordance with the Declaration of Helsinki, and approved by
the Institutional Review Board of the Pennington Biomedical Research
Center of the Louisiana State University System. Informed consent was
obtained from all participants prior to study participation. This study
was conducted prior to institution of the clinical trials registry process.
This was a randomized, PBO controlled, two-treatment (OLZ and
PBO), crossover study. Study treatments (OLZ or PBO) were only
administered on study days 1–15 of each study period. Subjects were
randomized to one of two possible treatment sequences, using a com-
puter-generated randomization schedule.
In sequence 1, subjects were treated with PBO for 15 days (period 1),
and then with OLZ (period 2; 5 mg daily for 3 days, then 10 mg daily for
12 days). In sequence 2, the order was reversed. Each treatment period
was separated by a minimum 12-day washout.
FI was measured at all meals and snacks on day 14 of study periods 1
and 2 and the same food choices were offered on each occasion. At least
two standardized menu choices were available for each course. Excess
calories were provided on each tray to avoid limitation of FI based on
availability, and subjects were advised to eat as much or as little as they
desired at each meal. Each portion was weighed and recorded before
and after each meal following the method of Rolls et al. (12). The total
daily macronutrient composition was 50% carbohydrate, 35% fat, and
15% protein. Non-caffeinated, nonalcoholic, and noncaloric beverages
were provided ad lib and all fluid ingestion was recorded on day 14.
Body weight was measured predose and in the fasting state on days 1,
4, 8, 12, 14, and 16 in light indoor clothing without shoes and items
removed from pockets using the same calibrated scale each time.
dual-energy X-ray absorptiometry
Dual-energy X-ray absorptiometry scans were performed on study
day 1 of period 1 and then on study day 16 of periods 1–2 using
(QDR4500 Acclaim Series model 010-0667 with Hologic QDR for
Windows version 11.2.3 software; Hologic, Bedford, MA). Each dual-
energy X-ray absorptiometry scan was performed using standardized
ree (ventilated hood indirect calorimeter)
Resting energy expenditure (REE) and respiratory quotient were meas-
ured using a ventilated hood system (Deltatrac Metabolic Monitor;
DATEX, Helsinki, Finland) on study day 14 and 15 of each study period.
Before each measurement period, the calorimeter was calibrated with a
reference gas containing 5% CO2 and 95% O2. Subjects remained fast-
ing from the evenings of study days 13 and 14 through completion of
morning calorimetry measurements on days 14 and 15. During each
calorimetry session, O2 and CO2 sampling was conducted for 45 min.
Data from the first 15 min of each sampling interval were discarded.
Volumes for CO2 and O2 (VCO2 and VO2) for the remaining 30 min
of each sampling interval were collected, averaged and then converted
to kcal using the Weir (14) equation and expressed on a kcal/day basis.
The calorimetry measurement obtained on study day 14 was used as the
primary assessment of REE.
Activity levels were continually assessed from prior to dosing on day
8 through day 16 of each study period. Accelerometers (Motionlogger
Actigraphs, model AMA-32; Precision Control Devices, Ft Walton
Beach, FL) were affixed around the wrist in a standardized fashion
across study subjects and study periods. The accelerometers were
placed just before dosing on day 8 and were to remain in place until
day 16. The principle end points of interest were daily mean activity
(counts/min) and mean scored sleep (min) (15).
euglycemic, hyperinsulinemic glucose clamp
On day 15 of each study period, IS and hepatic glucose output were
assessed using a two-step euglycemic, hyperinsulinemic clamp proce-
dure (16) that was conducted concurrently with the assessment of REE.
The clamp consisted of the following: baseline, step 1, and step 2. OLZ
or PBO dose was administered at 8:00 am, 1 h before the start of the
clamp after an overnight fast. Two baseline blood samples were drawn
to determine glucose and insulin concentrations, and background glu-
cose deuterium enrichment. A priming dose of 6,6-D2 glucose (2.5 mg)
was then administered followed by a constant infusion of 6,6-D2
glucose (2.0 mg/kg/h). Between 75 and 120 min after the start of the
infusion, three blood samples were drawn 15 min apart to quantitate
glucose and insulin concentrations, and glucose deuterium enrichment
during the baseline step. Following the acquisition of baseline samples,
an infusion of insulin at 0.25 mU/min/kg (10 mU/min/m2) was started.
Plasma glucose was maintained at 94 ± 5 mg/dl by varying the infusion
rate of a 20% dextrose solution premixed with 6,6-D2 glucose (10 mg/
kg). At 85, 100, and 115 min after the start of the insulin infusion, a
blood sample was collected to quantitate glucose and insulin concentra-
tion, and glucose deuterium enrichment. The insulin infusion was then
increased to 1 mU/min/kg (40 mU/min/m2). and repeat blood samples
were collected 85, 100, and 115 min after the start of step 2 for the quan-
titation of glucose and insulin concentrations, and glucose deuterium
For determination of isotope enrichment and estimation of hepatic glu-
cose production, plasma glucose was isolated and converted to its respec-
tive pentaacetate derivative (17). Isotope enrichment was determined by
gas chromatograph-mass spectrometry. Plasma glucose concentration
was determined enzymatically on a Yellow Springs Instrument (Yellow
Springs, OH) (18).
The biochemical markers of interest were triglycerides, cortisol, insu-
lin, prolactin, C-reactive protein, interleukin-6, plasminogen activator
inhibitor (PAI), leptin, tumor necrosis factor-α (TNF-α), adiponec-
tin, and ghrelin. Triglycerides were analyzed using the Beckman
Coulter DXC-600 (Beckman Coulter, Brea, CA). Cortisol, insulin,
and C-reactive protein were analyzed using the Diagnostic Products
Corporation 2800. The Luminex Labmap 100 (Luminex, Austin, TX)
was used to analyze interleukin-6 and TNF-α. Adiponectin, ghrelin,
and leptin were measured using radioimmunoassay and the Packard
Riastar gamma counter (Perkin Elmer, Shelton, CT) whereas PAI was
measured by enzyme-linked immunosorbent assay technology.
In all study periods the following samples were collected for assessment
of biochemical markers. On day 1 and day 14, blood was drawn before
dosing in the fasted condition. Following collection, samples were rapidly
separated and frozen at either −25 °C or below within 1 h of collection.
Euglycemic, hyperinsulinemic clamp. Whole body glucose uptake
(M) and the IS index (M/I) were calculated from the euglycemic, hyper-
insulinemic clamp procedure. M/I was expressed as the steady-state
ratio of M (mg/kg/min) to the steady-state plasma insulin concentration
VOLUME 18 NUMBER 8 | AUgUst 2010 | www.obesityjournal.org
Methods and techniques
(mU/ml). Hepatic glucose output was calculated as the rate of glucose
appearance (Ra) during the last 30 min of the equilibrium period at each
stage of the clamp procedure (i.e., baseline, step 1, and step 2). Ra was
determined by dividing the 6,6-D2 glucose infusion rate by the steady-
state concentration of blood glucose that was achieved in the last 30 min
of each step. Nonsteady state equations of Steele were used (19).
The primary end point was total FI (kcal) on day 14 of periods 1 and
2. The FI was compared between treatments (OLZ vs. PBO) with PBO
acting as the reference. The percent difference in FI relative to PBO
was calculated for each individual subject, and a one-sided, one-sam-
ple t-test was performed at the 0.05 α-level to test the null hypothesis
that the true mean percent difference was zero against the alternative
hypothesis that the true mean percent difference was >0. The P value
corresponding to this hypothesis test was reported as well as the 90%
confidence interval for the true mean percent difference. In order to
assess possible period and first-order carryover effects in the two-by-
two crossover of periods 1 and 2, the day 14 FI data from these periods
was modeled with a mixed effects ANOVA including period, treatment,
and first-order carryover as fixed effects and subject (within sequence)
as a random effect. The period and first-order carryover effects were
assessed according to the methods given in Chow and Liu (20). Other
secondary end points included REE, activity level, and body weight.
The analysis of these end points was similar to that of the primary end
point. As the secondary analysis was intended to be of an exploratory
nature, no adjustment for multiple comparisons was utilized. A post hoc
analysis of the ratio of REE to lean body mass (REE/LBM) was con-
ducted using the same mixed effects ANOVA method that was used for
the primary and secondary analyses. Except where otherwise indicated,
results are reported as mean ± s.d.
Of the 30 healthy, male subjects enrolled, 21 subjects com-
pleted the two randomized crossover periods. None of the
discontinuations were related to study treatment. The mean
age of the subjects was 27 years (range: 18–49 years) with a
mean BMI of 22.6 ± 2.2 kg/m2 (range: 18.1–25.7). The mean
weight at screening was 69.8 ± 5.4 kg.
FI and body weight
Mean total FI for PBO on day 14 was 3,860 kcal. Administration
of OLZ for 2 weeks resulted in a mean FI of 4,230 kcal, an
increase of 18% (90% confidence interval −0.2, 35.9%; P =
0.052) or 370 kcal relative to PBO (Figure 1a). A single sub-
ject who had adverse event reports of nausea, dizziness, and
diaphoresis on the day of FI measurement during the OLZ
treatment arm had a very low 24 h FI of 920 kcal. A sensitiv-
ity analysis excluding data from this subject was conducted to
assess the effect of this data point on the results. Excluding this
subject, the mean FI for the OLZ treatment arm was estimated
to be 23% (90% confidence interval 5.7, 39.6%, P < 0.05) or
547 kcal greater than that of the PBO treatment arm.
The administration of OLZ for 2 weeks resulted in an aver-
age increase in body weight of 2.62 vs. 0.08 kg for PBO, a mean
difference between treatments of 2.55 kg (90% confidence
interval 1.71, 3.38 kg, P < 0.001).
Relative to PBO, OLZ increased total lean body mass slightly,
by 0.83 kg or 1.4% (P < 0.05). Total body fat increased slightly
in response to OLZ (0.24 kg), but insignificantly (P > 0.2). No
significant changes were observed in the relative composition
of fat to lean body mass in response to dosing with OLZ.
Following OLZ dosing, mean REE for OLZ was 1,630 ±
187.3 kcal/24 h, in contrast to 1,520 ± 157.1 kcal/24 h for PBO
(Figure 1b). The adjusted mean difference reflected a signifi-
cant rise of 113 kcal/24 h (90% confidence interval of 53.05,
173.29 kcal/24 h; P = 0.003) or 7.4% over PBO. Respiratory
quotient rose significantly following 2 weeks of OLZ from
0.874 during PBO to 0.912 during OLZ (P < 0.005).
After treatment with OLZ, the mean REE/kg of lean body
mass was 26.34 as compared to 25.38 for PBO, a difference of
0.96 units (90% confidence interval: 0.04–1.89; P = 0.089).
Relative to PBO, there was a nonsignificant trend for OLZ
administration to result in an increase in daily mean activity
score by ~12 counts/min (120 vs. 110 counts/min, 90% confi-
dence interval of −2.51, 26.15 counts/min; P = 0.171). No sig-
nificant difference was observed in scored sleep between PBO
(590 min) and OLZ (580 min) administration. On average,
OLZ was associated with ~15 min less scored sleep than PBO
(90% confidence interval −111.63, 82.39; P = 0.799).
euglycemic, hyperinsulinemic glucose clamp
During the two-step euglycemic, hyperinsulinemic glucose
clamp procedure mean insulin levels at 0.25 mU/min/kg
increased in step 1 to 23.0 ± 5.85 and 22.7 ± 5.54 μU/ml at
steady state in the OLZ and PBO treatment periods, respec-
tively (P = n.s.). At 1.0 mU/min/kg in step 2, mean insulin levels
Change=370 kcal/day P =0.052
Day 14 total food intake (kcal)
Change=113 kcal/day P =0.003
Resting energy expenditure (kcal/day)
Figure 1 Statistical contrast (OLZ vs. PBO) of day 14 end
points. (a) 24-h food intake and (b) resting energy expenditure.
OLZ, olanzapine; PBO, placebo.
obesity | VOLUME 18 NUMBER 8 | AUgUst 2010 1649
Methods and techniques
increased to 61.8 ± 13.62 and 63.3 ± 10.45 μU/ml at steady
state in the OLZ and PBO treatment periods, respectively
(P = n.s.). The estimated glucose disposal rates (M) for either
step of the insulin infusion did not differ between treatment
groups. At the 0.25 mU/min/kg infusion rate, total glucose dis-
posal was 182.1 ± 78.18 and 173.1 ± 77.22 mg/min for OLZ
and PBO, respectively. At 1.0 mU/min/kg of insulin, the esti-
mated M value was 626.2 ± 157.88 and 629.3 ± 159.34 mg/ min
for OLZ and PBO, respectively. The estimated rate of glucose
appearance (Ra) also did not differ between treatment condi-
tions. At 0.25 mU/min/kg, glucose Ra was 211.5 ± 88.96 and
235.2 ± 97.19 mg/min for OLZ and PBO, respectively. At
1.0 mU/ min/ kg, the estimated Ra values were 665.6 ± 145.30
and 710.4 ± 179.18 mg/min for OLZ and PBO, respectively.
Table 1 provides the day 1 and day 14 mean values for each
marker along with the statistical contrast of day 14 OLZ and
PBO data. In response to OLZ, there were statistically signifi-
cant increases in triglycerides (49%), PAI-I (40%), leptin (16%),
and TNF-α (15%) over the PBO-treated state. Although fast-
ing serum insulin tended to increase (15%), this change was
not statistically significant. No significant differences in fasting
glucose concentrations were observed between treatments.
Ghrelin, conversely, decreased by ~11%.
This study demonstrates that, in healthy men, OLZ at clinical
doses increases body weight solely by increasing appetite and
FI. There was no decrease in activity or increase in sleep con-
tributing to the increase in body weight. In addition, instead of
any evidence for a decrease in EE, we observed increases in EE
and significant changes in respiratory quotient, suggesting a
change in substrate metabolism or a more general effect related
to over feeding healthy subjects (21). Finally, within the short
timeframe of OLZ dosing, significant shifts in several circulat-
ing markers of the metabolic syndrome, including increases in
triglycerides and PAI-I, became manifest.
The increase in body weight due to atypical antipsychot-
ics drugs is well documented in both patients (1–4,22) and
healthy subjects (23–25). The increase is fairly reproducible,
with both healthy subjects and patients gaining 2–4 kg within
the first few weeks of dosing (2,24,25). In the present study,
a similar magnitude of weight gain was observed. Two weeks
of titrated OLZ dosing gave a mean 2.6 kg weight gain com-
pared to PBO.
table 1 summary of day 1 and day 14 biochemical marker data and statistical contrast of day 14 olZ and PBo data
BiomarkerDayOLZ PBO Mean difference (OLZ–PBO)90% Confidence interval and P valuea
Triglycerides (mg/dl)1 83.0 ± 29.44 85.0 ± 36.49
117.7 ± 47.4478.7 ± 27.013922.21, 55.79; P < 0.001
Cortisol (μg/ml)1 12.0 ± 3.90 12.0 ± 5.98
14 11.7 ± 5.7512.1 ± 5.06−0.36 −1.79, 1.07; P = 0.674
Insulin (μU/ml)1 6.0 ± 3.00 7.3 ± 3.83
14 8.5 ± 4.04 7.4 ± 2.491.13 −0.276, 2.536; P = 0.183
Prolactin (ng/ml)1 12.0 ± 4.92 12.2 ± 3.85
22.1 ± 8.26 13.5 ± 6.048.545.81, 11.27; P< 0.001
C-reactive protein (mg/dl)1 0.3 ± 0.620.1 ± 0.07
140.2 ± 0.19 0.1 ± 0.260.05−0.656, 0.755; P = 0.906
Interleukin-6 (pg/ml)1 36.7 ± 82.1235.3 ± 81.02
1431.3 ± 69.43 28.7 ± 61.552.52 −3.32, 8.37; P = 0.471
PAI (ng/ml)1 27.9 ± 18.3430.4 ± 17.95
37.5 ± 29.7525.6 ± 13.0611.882.28, 21.48; P = 0.044
Leptin (ng/ml)1 3.5 ± 1.84 3.8 ± 2.16
5.0 ± 2.64 4.3 ± 2.07 0.730.119, 1.341; P = 0.051
TNF-α (pg/ml)14.9 ± 2.82 4.9 ± 2.60
5.3 ± 2.954.5 ± 2.01 0.770.224, 1.317; P = 0.023
Adiponectin (μg/ml)1 7.5 ± 3.267.2 ± 4.08
9.2 ± 4.99 7.1 ± 3.582.1450.966, 3.324; P = 0.004
Ghrelin (pg/ml)11,078.2 ± 303.09992.3 ± 288.77
864.3 ± 266.62971.5 ± 360.52 −107.15 −186.28, −28.02; P = 0.028
CI, confidence interval; OLZ, olanzapine; PAI, plasminogen activator inhibitor; PBO, placebo; TNF-α, tumor necrosis factor-α.
Boldface values indicate statistically significant differences.
aP value is a two-sided P value based on the comparison difference. 90% CI of the difference between OLZ and PBO at day 14.
VOLUME 18 NUMBER 8 | AUgUst 2010 | www.obesityjournal.org
Methods and techniques
The clearest contributor to the gain in weight was increased
FI. After removing the subject who had unexpected nausea on
the day of study, average FI increased by 547 kcal/day (23%
increase of OLZ over PBO). This observation is in accord with
that of Roerig and colleagues, who studied healthy subjects
before and after 2 weeks of OLZ or risperidone and observed
short-term weight gain similar in magnitude to the present
study (25). Although there was an increase in FI of ~150 kcal
(as measured at a dinner test meal in a feeding laboratory), this
result was statistically insignificant (25). Finally, the increase
in feeding is also consistent with observations of patients with
schizophrenia (9). After 4 weeks of treatment, adolescent
patients treated with OLZ gained 3.8 kg and had an average
27% increase in FI of 587 kcal/day (9).
There is no evidence from the present study that a decrease
in EE contributed to the weight gain. Although Virkkunen
et al. reported a 6% decrease in EE in a single subject (10),
we observed an increase in EE—by 7.5%, as estimated by
indirect calorimetry. These results are in contradistinction
to an earlier report in healthy subjects, in which EE was
reported to have not changed, although no data were shown
(25). Similar to Roerig et al., Gothelf et al. did not observe
a change in REE as estimated by indirect calorimetry in
patients with schizophrenia (9). It is possible that these stud-
ies did not detect the rise in EE due to the relative insensi-
tivity to indirect calorimetry. Alternatively, a combination of
factors, including the effects of treatment, per se, on smoking
(26,27), fidgeting (28), and agitation/sedation may explain
this observation in patients with schizophrenia that might not
be present in healthy volunteers. In addition, we employed a
crossover design to minimize the effects of intersubject vari-
ability which likely enhanced our ability to detect differences
An increase in EE in response to increased caloric intake has
been observed in a nonpharmacologic experiment of obese
and nonobese individuals (29). In that experiment, subjects
were given excess calories to drive weight up to 10% of their
initial body weight. In response to the excess calories and
associated weight gain, TEE rose significantly. In a subsequent
report, weight gain was also associated with an increase in uri-
nary norepinephrine and a rise in total T3 (30). Thus, the cause
of the increased EE in the present study may be secondary to
the increased caloric intake and not a direct effect of OLZ.
Additional investigation would be necessary to determine
whether this is true.
Changes in body composition do not seem to account for
the observed 7.4% increase in EE. These changes in body com-
position were small and relatively proportional increases in
lean body (1.4%) as well fat mass (2.1%). Although the percent
changes in body composition were proportionately compara-
ble, only the increase in lean body mass due to OLZ was signif-
icantly different, likely due to the short timeframe of this study
as well as the measurement variability of dual-energy X-ray
absorptiometry. Furthermore, our post hoc analysis results
suggest that REE normalized per kg of lean body mass may
have been increased by OLZ as compared to PBO.
Changes in circulating markers of insulin resistance, in
general, align with expected outcomes in the metabolic syn-
drome. Many of these changes are consistent with published
observations in patients with schizophrenia (31). However, the
changes in ghrelin and adiponectin in response to OLZ are,
at first glance, contrary to expectations. In the schizophre-
nia literature, ghrelin appears to be elevated and adiponectin
depressed in patients with schizophrenia (31–33). Yet careful
examination of the results suggests that the result may depend
on the circumstances for the measurement of these hormones.
That is, in response to chronic therapy, circulating adiponectin
concentrations are decreased, an effect consistent with chronic
obesity and insulin resistance (33). In the short-term, however,
and when compared to haloperidol, OLZ has been reported
to increase adiponectin (33). A parallel observation has been
reported for ghrelin (31). Taken together, there may be a
time-dependent response to OLZ. As metabolic homeostasis
changes over time, secondary effects on ghrelin and adiponec-
tin may become manifest.
Time may be important to the development of insulin
resistance. In the present study, no significant shift in IS was
observed, as assessed by the euglycemic clamp technique.
Similarly, Sowell et al. (23) did not observe a change in IS
in healthy subjects dosed with OLZ for a similar duration.
These concordant observations suggest that if OLZ induces
insulin resistance in healthy subjects, it is not a direct effect
and requires other changes to occur first. Recently however,
Vidarsdottir et al. showed that OLZ reduces insulin action
on glucose disposal in normal weight men who do not gain
weight after 8 days of treatment with OLZ 10 mg/day (34). The
dysregulation in glucose metabolism in schizophrenic patients
may be due to an underlying predisposition to diabetes and
related metabolic disorders that decompensates when treated
with OLZ. This effect appears to be reversible upon withdrawal
of the drug and may be independent of weight gain (35).
There is some evidence in the literature suggesting a direct
effect of OLZ on adipose tissue metabolism. Vestri et al. showed
that high concentrations of OLZ (100 µmol/l) decreased lipoly-
sis and increased lipogenesis (36). It is unclear whether these
concentrations are relevant given the in vivo concentrations of
OLZ are substantially lower. Nonetheless, alterations in adi-
pose tissue function, whether due to direct effects or the indi-
rect effects via weight gain, could contribute to the observed
This study provides fairly conclusive evidence that the early
increase in weight due to OLZ is through increased FI and is
not confounded by changes in activity level or EE. The accom-
panying exacerbation of factors such as PAI-I and TNF-α,
within 2 weeks of dosing, suggests that OLZ induces a clinical
picture very consistent with metabolic syndrome. This should
reflect a significant concern for prescribers given the baseline
high risk for diabetes and cardiovascular disease in patients
with schizophrenia (37). Beyond implications for schizophre-
nia treatment, results from this study might suggest a potential
human model to better understand human feeding mecha-
nisms and induction of metabolic syndrome.
obesity | VOLUME 18 NUMBER 8 | AUgUst 2010 1651 Download full-text
Methods and techniques
We thank the entire staff of the Pennington Biomedical Research Center
and the study volunteers for their efforts during conduct of this intensive
study. this research was funded by Pfizer.
this trial was sponsored by Pfizer, R.J.F., A.E.t., J.P.M., and D.A.F. are
currently, or were employed by Pfizer during conduct of this study.
© 2010 The Obesity Society
1. Baptista T, Zárate J, Joober R et al. Drug induced weight gain, an
impediment to successful pharmacotherapy: focus on antipsychotics.
Curr Drug Targets 2004;5:279–299.
2. Allison DB, Mentore JL, Heo M et al. Antipsychotic-induced weight gain: a
comprehensive research synthesis. Am J Psychiatry 1999;156:1686–1696.
3. Lieberman JA, Stroup TS, McEvoy JP et al. Effectiveness of antipsychotic
drugs in patients with chronic schizophrenia. N Engl J Med 2005;353:
4. Basson BR, Kinon BJ, Taylor CC et al. Factors influencing acute weight
change in patients with schizophrenia treated with olanzapine, haloperidol,
or risperidone. J Clin Psychiatry 2001;62:231–238.
5. Kinon BJ, Basson BR, Gilmore JA, Tollefson GD. Long-term olanzapine
treatment: weight change and weight-related health factors in schizophrenia.
J Clin Psychiatry 2001;62:92–100.
6. Theisen FM, Cichon S, Linden A et al. Clozapine and weight gain. Am J
7. Daumit GL, Goff DC, Meyer JM et al. Antipsychotic effects on estimated
10-year coronary heart disease risk in the CATIE schizophrenia study.
Schizophr Res 2008;105:175–187.
8. Newcomer JW. Metabolic considerations in the use of antipsychotic
medications: a review of recent evidence. J Clin Psychiatry 2007;68
9. Gothelf D, Falk B, Singer P et al. Weight gain associated with increased
food intake and low habitual activity levels in male adolescent schizophrenic
inpatients treated with olanzapine. Am J Psychiatry 2002;159:1055–1057.
10. Virkkunen M, Wahlbeck K, Rissanen A, Naukkarinen H, Franssila-Kallunki
A. Decrease of energy expenditure causes weight increase in olanzapine
treatment—a case study. Pharmacopsychiatry 2002;35:124–126.
11. Bergman RN, Ader M. Atypical antipsychotics and glucose homeostasis.
J Clin Psychiatry 2005;66:504–514.
12. Rolls BJ, Shide DJ, Thorwart ML, Ulbrecht JS. Sibutramine reduces food
intake in non-dieting women with obesity. Obes Res 1998;6:1–11.
13. Greenway FL. The safety and efficacy of pharmaceutical and herbal caffeine
and ephedrine use as a weight loss agent. Obes Rev 2001;2:199–211.
14. Weir JB. New methods for calculating metabolic rate with special reference
to protein metabolism. 1949. Nutrition 1990;6:213–221.
15. Tonetti L, Pasquini F, Fabbri M, Belluzzi M, Natale V. Comparison of two
different actigraphs with polysomnography in healthy young subjects.
Chronobiol Int 2008;25:145–153.
16. DeFronzo RA, Ferrannini E, Hendler R, Felig P, Wahren J. Regulation of
splanchnic and peripheral glucose uptake by insulin and hyperglycemia in
man. Diabetes 1983;32:35–45.
17. Horowitz JF, Mora-Rodriguez R, Byerley LO, Coyle EF. Substrate
metabolism when subjects are fed carbohydrate during exercise. Am J
18. Twomey PJ. Plasma glucose measurement with the Yellow Springs Glucose
2300 STAT and the Olympus AU640. J Clin Pathol 2004;57:752–754.
19. Steele R. Influences of glucose loading and of injected insulin on hepatic
glucose output. Ann N Y Acad Sci 1959;82:420–430.
20. Chow S, Liu J. Design and Analysis of Bioavailability and Bioequivalence
Studies. 2nd edn. Marcel Dekker: New York, 2000.
21. Joosen AM, Bakker AH, Westerterp KR. Metabolic efficiency and energy
expenditure during short-term overfeeding. Physiol Behav 2005;85:
22. Henderson DC, Cagliero E, Gray C et al. Clozapine, diabetes mellitus, weight
gain, and lipid abnormalities: a five-year naturalistic study. Am J Psychiatry
23. Sowell M, Mukhopadhyay N, Cavazzoni P et al. Evaluation of insulin
sensitivity in healthy volunteers treated with olanzapine, risperidone,
or placebo: a prospective, randomized study using the two-step
hyperinsulinemic, euglycemic clamp. J Clin Endocrinol Metab
24. Sowell MO, Mukhopadhyay N, Cavazzoni P et al. Hyperglycemic clamp
assessment of insulin secretory responses in normal subjects treated with
olanzapine, risperidone, or placebo. J Clin Endocrinol Metab 2002;87:
25. Roerig JL, Mitchell JE, de Zwaan M et al. A comparison of the effects of
olanzapine and risperidone versus placebo on eating behaviors. J Clin
26. Audrain JE, Klesges RC, DePue K, Klesges LM. The individual and
combined effects of cigarette smoking and food on resting energy
expenditure. Int J Obes 1991;15:813–821.
27. Collins LC, Cornelius MF, Vogel RL, Walker JF, Stamford BA. Effect of
caffeine and/or cigarette smoking on resting energy expenditure. Int J Obes
Relat Metab Disord 1994;18:551–556.
28. Levine JA, Schleusner SJ, Jensen MD. Energy expenditure of nonexercise
activity. Am J Clin Nutr 2000;72:1451–1454.
29. Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting
from altered body weight. N Engl J Med 1995;332:621–628.
30. Rosenbaum M, Hirsch J, Murphy E, Leibel RL. Effects of changes in body
weight on carbohydrate metabolism, catecholamine excretion, and thyroid
function. Am J Clin Nutr 2000;71:1421–1432.
31. Sentissi O, Epelbaum J, Olié JP, Poirier MF. Leptin and ghrelin levels in
patients with schizophrenia during different antipsychotics treatment: a
review. Schizophr Bull 2008;34:1189–1199.
32. Perez IR, Vazquez BJL, Amado JA et al. Effect of antipsychotics on peptides
involved in energy balance in drug-naive psychotic patients after 1 year of
treatment. J Clin Psychopharmacol 2008;28:289–295.
33. Oriot P, Feys JL, Mertens de Wilmars S et al. Insulin sensitivity, adjusted
β-cell function and adiponectinaemia among lean drug-naive schizophrenic
patients treated with atypical antipsychotic drugs: a nine-month prospective
study. Diabetes Metab 2008;34:490–496.
34. Vidarsdottir S, de Leeuw van Weenen JE, Frölich M et al. Effects of
olanzapine and haloperidol on the metabolic status of healthy men. J Clin
Endocrinol Metab 2010;95:118–125.
35. Popli AP, Konicki PE, Jurjus GJ, Fuller MA, Jaskiw GE. Clozapine and
associated diabetes mellitus. J Clin Psychiatry 1997;58:108–111.
36. Vestri HS, Maianu L, Moellering DR, Garvey WT. Atypical antipsychotic
drugs directly impair insulin action in adipocytes: effects on glucose
transport, lipogenesis, and antilipolysis. Neuropsychopharmacology
37. Newcomer JW. Metabolic risk during antipsychotic treatment. Clin Ther