S P E C I A L
C l i n i c a l
F E A T U R E
R e v i e w
Drugs Commonly Associated With Weight Change:
A Systematic Review and Meta-analysis
Juan Pablo Domecq, Gabriela Prutsky, Aaron Leppin, M. Bassam Sonbol,
Osama Altayar, Chaitanya Undavalli, Zhen Wang, Tarig Elraiyah, Juan Pablo Brito,
Karen F. Mauck, Mohammed H. Lababidi, Larry J. Prokop, Noor Asi, Justin Wei,
Salman Fidahussein, Victor M. Montori, and Mohammad Hassan Murad
Knowledge and Evaluation Research Unit (J.P.D., G.P., A.L., M.B.S., O.A., C.U., Z.W., T.E., J.P.B., K.F.M.,
M.H.L., N.A., J.W., S.F., V.M.M., M.H.M.), Mayo Clinic, Rochester, Minnesota 55905; Unidad de
Conocimiento y Evidencia (J.P.D., G.P., V.M.M.), Universidad Peruana Cayetano Heredia, Lima 31, Peru;
and Division of Preventive, Occupational, and Aerospace Medicine (N,A., M.H.M.), Division of
Endocrinology, Diabetes, Metabolism, and Nutrition (J.P.B., V.M.M.), Division of General Internal
Medicine (K.F.M.), and Mayo Clinic Libraries (L.J.P.), Mayo Clinic, Rochester, Minnesota 55905
Context: Various drugs affect body weight as a side effect.
Objective: We conducted this systematic review and meta-analysis to summarize the evidence
about commonly prescribed drugs and their association with weight change.
Data Sources: MEDLINE, DARE, and the Cochrane Database of Systematic Reviews were searched
to identify published systematic reviews as a source for trials.
Study Selection: We included randomized trials that compared an a priori selected list of drugs to
placebo and measured weight change.
Data Extraction: We extracted data in duplicate and assessed the methodological quality using the
Cochrane risk of bias tool.
Results: We included 257 randomized trials (54 different drugs; 84 696 patients enrolled). Weight
gain was associated with the use of amitriptyline (1.8 kg), mirtazapine (1.5 kg), olanzapine (2.4 kg),
quetiapine (1.1 kg), risperidone (0.8 kg), gabapentin ( 2.2 kg), tolbutamide (2.8 kg), pioglitazone
(2.6 kg), glimepiride (2.1 kg), gliclazide (1.8 kg), glyburide (2.6 kg), glipizide (2.2 kg), sitagliptin
(0.55 kg), and nateglinide (0.3 kg). Weight loss was associated with the use of metformin (1.1 kg),
acarbose (0.4 kg), miglitol (0.7 kg), pramlintide (2.3 kg), liraglutide (1.7 kg), exenatide (1.2 kg),
zonisamide (7.7 kg), topiramate (3.8 kg), bupropion (1.3 kg), and fluoxetine (1.3 kg). For many
other remaining drugs (including antihypertensives and antihistamines), the weight change was
either statistically nonsignificant or supported by very low-quality evidence.
Conclusions: Several drugs are associated with weight change of varying magnitude. Data are provided
to guide the choice of drug when several options exist and institute preemptive weight loss strategies
when obesogenic drugs are prescribed. (J Clin Endocrinol Metab 100: 363–370, 2015)
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2015 by the Endocrine Society
Received September 4, 2014. Accepted December 4, 2014.
First Published Online January 15, 2015
Abbreviations: BMI, body mass index; CI, confidence interval; GLP-1, glucagon-like peptide-1; RCT, randomized controlled trial; SR, systematic review.
For article see page 342
doi: 10.1210/jc.2014-3421
J Clin Endocrinol Metab, February 2015, 100(2):363–370
jcem.endojournals.org
363
364
Domecq et al
Drugs Commonly Associated With Weight Change
Background
besity is a worldwide epidemic and one of the most
important public health concerns in developed nations (1). The World Health Organization estimates that,
globally, over 200 million men and nearly 300 million
women were obese in 2008 (2). In the United States, the
lifetime risk of obesity has increased to approximately 25%
(3), making it a major cause of morbidity and mortality
(4 – 6).
Efforts to control this epidemic are dependent on stakeholder understanding of the factors that contribute to
weight change and how these might be open to modification. According to the World Health Organization, over
75% of all cardiovascular disease mortality may be prevented with adequate changes in lifestyle (7). However,
clinicians who prescribe medications that can impact
weight can also modify these outcomes, if possible. Clinically, it would be useful to know the absolute effects on
weight of various medications so that certain drugs can be
avoided, replaced, or sought out as appropriate to a given
situation. Current treatment guidelines are often limited
by a lack of comparative data, such that it is impossible to
determine the magnitude of the difference in weight change
caused by two otherwise acceptable drugs.
An expert panel from The Endocrine Society is charged
with developing clinical practice guidelines for the management of obesity. As part of a comprehensive management approach, medication choice can play a role. To aid
in the development of the Society guidelines, we conducted
this systematic review and meta-analysis aimed at summarizing the evidence about commonly prescribed drugs
and their associations with weight loss or weight gain. The
goal is to provide patients and clinicians with information
that can inform and individualize treatment decisions.
O
Methods
Search and analysis methods, eligibility criteria, and the outcomes of interest were specified in advance in a protocol developed by study investigators with input from the expert panel
from The Endocrine Society. This protocol is described in detail
and has been published elsewhere (8).
To make this review of a large number of weight-affecting
drugs feasible, and due to the availability of multiple systematic
reviews (SRs) of these drugs, we conducted an umbrella (9)
search strategy to identify eligible randomized controlled trials
(RCTs). The umbrella strategy differs from a regular systematic
search approach in that the umbrella strategy identifies studies
from previous systematic reviews as opposed to the primary literature. To this extent, we searched for any systematic review
that included RCTs comparing the drugs of interest to placebo.
The list of the most relevant drugs was developed by members
from The Endocrine Society. This list consisted of commonly
J Clin Endocrinol Metab, February 2015, 100(2):363–370
prescribed drug families and specific drugs that have been associated with weight gain (obesogenic) or weight loss (leptogenic)
(Supplemental Table 1, interventions list). Eligible SRs were used
as a source to identify relevant RCTs. Considering that multiple
published systematic reviews and meta-analyses (10 –13) have
already summarized and appraised the evidence supporting the
efficacy of antiobesity drugs such as orlistat and phentermine,
these drugs were not included in this summary.
Search methods and selection of SRs
The first author (J.P.D.) searched MEDLINE, DARE, and the
Cochrane Database of SRs for at least two SRs per drug through
January 2013. An expert reference librarian (L.J.P.) provided
assistance throughout the process.
When multiple SRs evaluated the same drug, we chose the one
with the most recent search date. When more than two SRs
shared a similar search date (⬍1 y apart), we chose the one with
the largest number of included RCTs that were most clinically
relevant to the typical application of the drug. For example,
although sertraline has a therapeutic indication for eating disorders, its major clinical use is in depression and obsessive-compulsive disorder. When there was no clear difference in the frequency of the use of drugs by a specific condition, we included the
SRs without considering this criterion (for example, beta blockers for myocardial infarction or for essential hypertension).
When we found more than two SRs with no clear rationale to
select one over the other, we included all of them (Supplemental
Table 2, included SRs).
Eligibility criteria for RCT
We included parallel or crossover RCTs that enrolled adults
(ⱖ18 y old) and evaluated any drug listed in Supplemental Table
1 as an intervention as long as the length of treatment was at least
30 days. Studies that investigated combinations of drugs (except
for the listed ones) were excluded. We also excluded studies that
reported only subjective outcome measures (self-reported weight
change). We elected not to include observational or quasi-randomized studies.
Agreement among the reviewers was measured using the ␬
statistic. A reference management system (DistillerSR) was used
for study selection, providing real-time agreement statistics. The
first author (J.P.D.) monitored the agreement between evaluators during the trial selection in order to discuss disagreements
and clarify the protocol and selection criteria when needed. The
whole selection team met with the first author and the senior
author (M.H.M.) five times during the selection process.
Data extraction and management
Using a standardized, piloted, and web-based data extraction
form and working in duplicate, we abstracted the following descriptive data from each study: full description of participants
enrolled, the interventions received (dose, frequency, route), the
monitoring for efficacy or adherence, and the measure of outcome (specifically defined as event rate or continuous measure
and time frame). For studies with more than one follow-up period, we selected the longest.
When necessary, we calculated needed data elements from
other reported statistics such as confidence intervals (CIs), P, or
t values (14). When this was not possible, we imputed the missing
values, such as standard deviation, from one large study (another
RCT or a SR) with a similar population and intervention (15).
doi: 10.1210/jc.2014-3421
Assessment of risk of bias in included studies and
confidence on the estimates
We assessed the methodological quality of RCTs using the
Cochrane risk of bias tool to determine: how the randomization
sequence was generated; how allocation was concealed; whether
there were important imbalances at baseline; which groups were
blinded (patients, caregivers, data collectors, outcome assessors,
data analysts); what was the loss to follow-up; whether the analyses were by intention to treat; and how missing outcome data
were handled. We also analyzed the adequacy of the outcome
measurement process, assigning higher confidence to the RCTs
that evaluated weight changes using a specific predefined protocol. No scoring system was derived for risk of bias assessment.
We used the GRADE framework (Grading of Recommendations, Assessment, Development and Evaluation) (16) to rate the
confidence on the evidence supporting the weight change effect
associated with each drug. This rating reflects our confidence in
the pooled estimate. We rated the confidence downward based
on methodological limitations, imprecision, indirectness, inconsistency, and reporting and publication biases. Factors that led to
rating the confidence upward were large magnitude of effect and
the presence of a dose-duration response gradient. The confidence in the estimates was rated as high (QQQQ), moderate
(QQQE), low (QQEE), or very low (QEEE).
Meta-analysis
We defined a clinically important weight change (for either
weight gain or loss) as a change ⱖ 2 kg or ⱖ 5% from the
baseline, defined by Stevens et al (17). The outcomes of this
meta-analysis were: 1) absolute weight change (Abs WC)—
weight loss or gain assessed as a continuous outcome expressed
as a mean difference in the absolute weight change in kilograms
or as a body mass index (BMI) change in kilograms/(meter)2; 2)
percentage weight change (Per WC)—weight loss or gain assessed as a continuous outcome expressed as a mean difference
in the relative weight change in kilograms or as a BMI change in
kilograms/(meter)2 from the baseline weight or BMI; 3) weight
gain/weight loss (WG/WL) ⱖ 5%—an important change in
weight of 5% or more as described by Stevens et al (17) (for
example, 7 to 10%), either for weight gain or loss, from the
baseline weight (this outcome is presented as a relative risk); and
4) any WG/WL—weight change rate assessed as a dichotomous
outcome, defined as a number of patients with increased or decreased weight over the total number of patients in each group
(this outcome is presented as a relative risk).
The precision of estimates of drug effects on weight is reflected in 95% CIs around such estimates. We extracted and
evaluated outcomes by analyzing participants in the groups in
which they were randomized (ie, intention to treat). For studies
with loss to follow-up, we used the number of patients randomized as a denominator for the risk estimate, preserving randomization benefits in balancing prognosis of trial arms, realizing
that this may underestimate the effect size (18).
We conducted random-effects meta-analysis using the DerSimonian and Laird (19) method to pool treatment effects from
included studies. We used the I2 statistic (20) and Cochran’s Q
test to assess heterogeneity across studies. We assessed publication bias by the Begg adjusted rank correlation test and visual
examination of funnel plots whenever there were at least 10
included studies with no considerable heterogeneity (defined as
less than 70% of I2 according to the Cochrane collaboration)
jcem.endojournals.org
365
(14, 21, 22). Analysis was conducted using STATA version 12.0
(StataCorp).
Subgroup and sensitivity analysis
We planned to explore a few subgroup interactions to explain
inconsistency in results across trials including subgroup analyses
based on: 1) BMI status— obese (BMI ⱖ 30 kg/m2) vs nonobese
(BMI ⬍ 30 kg/m2); 2) risk of bias of the included studies (low and
unclear risk of bias vs high risk of bias); and 3) gender (male vs
female). Due to the lack of data about gender variation and BMI
status within each included RCT, we could not perform these
subgroup analyses. Most of the included studies presented serious or very serious methodological limitations; therefore, we did
not perform a subgroup analysis based on the risk of bias.
We conducted a meta-regression to test whether the effect size
(weight change) was affected by daily drug dose.
This review is reported in accordance with the recommendations set forth by the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) work groups (23).
Results
Search results and study description
We identified 99 relevant SRs that included 3548
RCTs, which we reviewed and from which we selected 257
eligible RCTs (Figure 1). These RCTs evaluated weight
change associated with 54 different drugs and enrolled
more than 84 696 patients. During the full-text screening,
reviewers had excellent agreement (average ␬ coefficient,
0.87). The main reasons for exclusion were nonrandomized study design and lack of weight change assessment.
In these trials, 46% of participants were men. Ages
ranged from 22 to 88 years. Most RCTs (63%) exclusively
enrolled obese patients; 23% enrolled only overweight
patients. We found an underrepresentation of nonwhite
populations in the included trials (Supplemental Table 3,
study characteristics).
Methodological quality of included studies and
confidence in the estimates
We found at least one trial evaluating the effect on
weight for 54 different drugs. The risk of bias was moderate or high in 191 of 257 (74%) of the trials. This was
mainly due to the failure to report randomization methods
or allocation concealment or the lack of following the
intention-to-treat principle. Only 17 of 257 trials (7%)
were open-label studies and did not blind the patients or
investigator. The risk of bias for each individual trial is
described in Supplemental Table 4 (methodological quality of included studies). The risk of bias was assessed overall for each comparison and was used to rate the quality of
evidence. The confidence in the estimates was rated downward for most comparisons due to the risk of bias. The size
and precision of the effect varied across drug classes. A few
366
Domecq et al
Drugs Commonly Associated With Weight Change
99 included systemac
reviews
3548 RCTs for tle and
abstract screening
Exclusion Criteria
Language other than English, Spanish,
Portuguese or Arabic
Non RCTs
RCT with combined intervenons
Single dose
Length of intervenon =< than 30 days
1060 RCTs for full text
screening
Exclusion Criteria
Weight changes not reported or analyzable
Uncommon evaluated condion (e.g. Tays
Sacks Disease)
Acve control
Trials of an-obesity (weight loss) drugs
257 included RCTs for Data
extracon and analysis
Figure 1. The process of study selection.
estimates were rated upward based on a large effect size
and an exposure-response gradient. The complete data
including the weight change and confidence rating process
for each drug are shown in Supplemental Table 5.
Weight changes per drug class
Atypical antipsychotics
This drug class was associated with the most weight
gain. Olanzapine was associated with the largest gain
(2.4 kg), followed by quetiapine (1.1 kg) and risperidone (0.8 kg).
These three atypical antipsychotics reached their effect
on weight in a median duration of 3 months. Quetiapine
showed a dose response gradient suggesting higher weight
gain with doses above 450 mg/d.
None of the drugs in this class were associated with
weight loss. Aripiprazole was associated with weight gain
(0.6 kg); however, the confidence in this estimate was very
low. Ziprasidone may be a weight-neutral atypical antipsychotic; however, this estimate was associated with very
low confidence (rated downward due to imprecision and
risk of bias).
J Clin Endocrinol Metab, February 2015, 100(2):363–370
Anticonvulsants and mood
stabilizers
Gabapentin was associated with
a weight gain of 2.2 kg after 1.5
months of use. Divalproex was associated with increased risk of weight
gain (of any magnitude), but not of
clinically significant (ⱖ5%) weight
gain (relative risk, 2.8; 95% CI, 1.30,
6.02; and relative risk, 1.39; 95% CI,
0.38, 5.14, respectively; data insufficient to determine absolute weight
difference). Carbamazepine was associated with weight gain of 1.0 kg;
however, the confidence in this estimate is low.
Significant weight loss was noted
with zonisamide (7.7 kg) and topiramate (3.8 kg).
Lithium and lamotrigine were not
associated with a statistically significant effect on weight; therefore,
they may be weight neutral.
Hypoglycemic agents
Agents associated with statistically significant weight gain were
tolbutamide (2.8 kg), pioglitazone
(2.6 kg), glimepiride (2.1 kg), gliclazide (1.8 kg), glyburide (2.6 kg),
glipizide (2.2 kg), sitagliptin (0.55
kg), and nateglinide (0.3 kg).
Agents associated with statistically significant weight
loss were metformin (1.1 kg; estimate heterogeneity was
partially explained by the dose and duration of therapy),
acarbose (0.4 kg), miglitol (0.7 kg), pramlintide (2.3 kg),
and glucagon-like peptide-1 (GLP-1) agonists. Liraglutide
doses of 1.2 ␮g or above used for a median of 6 months
(range, 3– 6.5 mo) demonstrated a weight loss of 1.7 kg.
Exenatide was associated with a significant weight loss of
1.2 kg after 4 months of usage (range, 1–7 mo). Similar
weight change was noted with exenatide 10, 15, and 20
␮g. The weight loss with exenatide weekly dosing was 0.9
kg, and with daily dosing it was 1.3 kg (P value for difference ⫽ .7). Weight loss with GLP-1 agonists used ⬍ 3
months was 0.3 kg, 3– 6 months was 1.3 kg, and ⬎ 6
months was 0.9 kg.
Data were also insufficient for other hypoglycemic
agents such as rosiglitazone and repaglinide.
Antihypertensive agents
We found no evidence of a significant effect on weight
for any of the listed antihypertensive agents. The estimates
doi: 10.1210/jc.2014-3421
for all the included drugs under this class were inconclusive, with low or very low confidence.
Hormones
A pooled estimate from four RCTs of glucocorticoids
used in rheumatoid arthritis suggested a weight increase of
4 to 8% (24). We also found one RCT (25) that compared
glucocorticoids against sulfasalazine and demonstrated a
weight gain of 1.7 kg after 1 year of treatment.
GH use was associated with a reduction of BMI by 0.59
kg/m2; however, data were too imprecise and heterogeneous to estimate a specific absolute weight change. The
quality of evidence supporting weight change for T, leuprolide, and medroxyprogesterone was low; hence, they
may be weight neutral.
Antihistamines
Data were only available on cyproheptadine, which
was likely weight neutral.
Antidepressants
Weight gain was associated with the use of amitriptyline (1.8 kg) and mirtazapine (1.5 kg).
Weight loss was associated with the use of bupropion
(1.3 kg) and fluoxetine (1.3 kg).
There was a nonclinically significant weight loss supported with low-quality evidence for sertraline, venlafaxine, and duloxetine and a nonsignificant weight change
with the use of citalopram, escitalopram, paroxetine, and
nortriptyline.
Discussion
Main findings
We conducted a systematic review and meta-analysis of
257 RCTs (54 different drugs; 84 696 patients enrolled).
Our goal was to evaluate the weight change associated
with the use of an a priori selected list of commonly used
drugs.
We that found weight gain was associated with the use
of amitriptyline, mirtazapine, olanzapine, quetiapine,
risperidone, gabapentin, tolbutamide, pioglitazone,
glimepiride, gliclazide, glyburide, sitagliptin, and nateglinide. Weight loss was associated with the use of metformin, acarbose, miglitol, pramlintide, liraglutide, exenatide, zonisamide, topiramate, bupropion, and
fluoxetine. For many other remaining drugs (including
antihypertensives and antihistamines), the weight change
was either statistically nonsignificant or supported by
low-quality evidence.
A key issue in interpreting data on drugs with estimates
that are imprecise (not statistically significant) or esti-
jcem.endojournals.org
367
mates associated with low confidence rating is that there
are two possibilities of inference: 1) these drugs may be
weight neutral; or 2) these drugs cause a weight change
that the current published literature cannot document
with confidence. Thus, the absence of evidence (for a
weight change) is not necessarily an evidence of absence
(of a weight effect).
Limitations and strengths
The strengths of this review include the rigorous and
comprehensive nature of the literature search. We attempted to identify at least two high-quality SRs for each
evaluated drug. This umbrella approach (9) also made this
systematic review feasible, a key concern that was realized
from the outset.
Our findings are limited by the short-term follow-up of
most of the included RCTs. Observational studies have a
longer follow-up, but they are subject to a higher risk of
bias typically, particularly when the outcome is weight
change. These outcomes are subject to many cointerventions, confounders, and baseline prognostic imbalances.
Therefore, we made the decision to consider only RCTs for
this review.
We recognize that the evaluated drugs were chosen by
the experts of The Endocrine Society based on their knowledge of the field. Although this process has its own limitations, it was felt to be an approach that closely met the
needs of the guideline developers and clinicians in daily
practice. To make this evidence synthesis feasible, The
Endocrine Society task force had to choose a discrete list
of drugs that had certain characteristics (suspected to affect weight, to have randomized evidence available, and to
be commonly encountered in the practice). This list of
drugs was chosen a priori. It is certainly plausible that
other drugs not included here exert important weight
changes.
Comparison with other reviews
Our findings are in general consistent with other systematic reviews that evaluated single drugs or drug classes
(26 –30) and faced similar challenges. The Agency for
Healthcare Research and Quality commissioned a technology assessment report (29) comparing second-generation antidepressants, but a quantitative meta-analysis was
not performed on the outcome of weight change. Our systematic review is conducted as a companion document to
The Endocrine Society guideline on the pharmacological
management of obesity. Analysis of weight change is challenging and can be complicated by nonlinear weight
change during drug use. For example, our analysis of antidepressants showed weight gain with amitriptyline and
mirtazapine and weight loss with bupropion and fluox-
368
Domecq et al
Drugs Commonly Associated With Weight Change
etine. However, the weight change is unclear for most
other agents. Some RCTs showed statistically significant
weight loss with duloxetine, (31), venlafaxine (32, 33),
and sertraline (34), and other trials did not. A pooled analysis (35) of 10 RCTs showed that duloxetine-treated patients experienced weight loss after short-term treatment,
followed by modest weight gain on longer-term treatment,
and suggested similar weight changes among duloxetine,
fluoxetine, and paroxetine. Therefore, for paroxetine, duloxetine, venlafaxine, and sertraline, the effect on weight
remains unclear, and if differences between these three
drugs exist, they are unlikely to be clinically significant.
An example of other nuances is observed when evaluating the association between GLP-1 agonists and weight
loss. A recent RCT (36) suggested that long-acting GLP-1
agonists appear to be more effective on glycemic control
given their effect on insulin production and suppression of
glucagon, whereas shorter-acting ones may have a bigger
effect on appetite and gastric emptying, and thus weight
loss. We used such hypotheses to explore the observed
heterogeneity in GLP-1 meta-analysis by conducting subgroup analyses based on dose, dosing interval, and duration of use. We found low-quality evidence suggesting a
small amount of weight gain with sitagliptin. A network
meta-analysis (27) showed no significant weight loss with
sitagliptin (0.20 kg; 95% CI, ⫺0.18, 60), and another
meta-analysis of the head-to-head trials showed that patients on GLP-1 analogs lost 1.55 kg (95% CI, 1.12, 1.98)
compared to those on sitagliptin (30).
A meta-analysis of RCTs of antipsychotics followed a
Bayesian approach and provided rankings of these drugs
in terms of weight gain. The ranking order, starting with
the largest weight gain, was olanzapine, clozapine, risperidone, haloperidol, and aripiprazole (28). We opted to
avoid ranking because it hides the quality of evidence and
has multiple other limitations (37) and opted for presenting weight change estimates associated with each drug to
facilitate shared decision making.
This systematic review is clinically relevant particularly
because of the impact of weight on cardiovascular morbidity and mortality (38) and overall mortality (4), and it
can help inform decisions about whether or not to use a
specific drug (39). This systematic review represents a significant effort and is the product of an adequate balance
between rigor and feasibility; it provides the best comparative evidence on weight effect for many different drugs.
These data will be central for guideline developers, clinicians, and patients when choosing between different available therapies.
J Clin Endocrinol Metab, February 2015, 100(2):363–370
Implications for practice
The summarized evidence herein will help in treatment
choice when several options are available. In addition, it
may help with instituting preemptive weight loss strategies
when prescribing certain drugs. For example, the large and
quick effect of atypical antipsychotics on weight suggests
that some preemptive weight management measures
should be undertaken when prescribing these drugs. Several systematic reviews (40 – 43) have described such strategies. Physicians should not wait for a weight increase to
recommend lifestyle modifications when prescribing a
drug known to increase weight.
In addition to choosing a medication within a class, this
review can also help patients with multiple comorbidities
using drugs from multiple classes. Awareness of possible
additive effects on weight from various drugs provides a
rationale for developing preemptive strategies to prevent
or address undesired weight affects.
The estimates presented in this review are only one factor to consider when choosing a drug; patients and clinicians will consider other factors relating to efficacy, adverse effects, cost, treatment burden, and patients goals.
Implications for research
There is underrepresentation of ethnic minorities in
RCTs (44). The included studies in this review were no
exception. We found that Blacks, Asians, and Hispanics
were underrepresented in these datasets. Knowing the effect of various drugs on weight in these populations is
particularly important because of the increased incidence
and worse prognosis of several chronic conditions in minorities such as obesity, hypertension, chronic kidney disease, and diabetes (45– 47).
Future studies that evaluate the effect of a drug on
weight need to report the incidence of weight change in
addition to the magnitude of weight change, should have
longer follow-up, and should evaluate time trends to determine whether the effect is constant, linear, or prone to
plateau over time. Evaluating these characteristics in
meta-analyses of aggregate data is not reliable and requires individual patient data. The effect on weight after
stopping the treatment remains largely unknown.
Conclusions
Several drugs are associated with weight change of varying
magnitude. Data are provided to guide the choice of drug
when several options exist and to institute preemptive
strategies for weight management when drugs with known
weight effects are prescribed.
doi: 10.1210/jc.2014-3421
jcem.endojournals.org
Acknowledgments
Address all correspondence and requests for reprints to:
M. Hassan Murad, MD, Mayo Clinic, Knowledge and Evaluation Research Unit, 200 First Street SW, Rochester, MN 55905.
E-mail: [email protected].
This review was commissioned and funded by a contract from
The Endocrine Society.
Disclosure Summary: The authors have nothing to disclose.
17.
18.
19.
20.
21.
References
1. Barness LA, Opitz JM, Gilbert-Barness E. Obesity: genetic, molecular, and environmental aspects. Am J Med Genet A. 2007;143A:
3016 –3034.
2. World Health Organization. Fact Sheet: Obesity and Overweight.
Geneva, Switzerland: World Health Organization; 2011.
3. Vasan RS, Pencina MJ, Cobain M, Freiberg MS, D’Agostino RB.
Estimated risks for developing obesity in the Framingham Heart
Study. Ann Intern Med. 2005;143:473– 480.
4. Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific
excess deaths associated with underweight, overweight, and obesity.
JAMA. 2007;298:2028 –2037.
5. Prospective Studies Collaboration; Whitlock G, Lewington S, Sherliker P, et al. Body-mass index and cause-specific mortality in 900
000 adults: collaborative analyses of 57 prospective studies. Lancet.
2009;373:1083–1096.
6. Preston SH, Stokes A. Contribution of obesity to international differences in life expectancy. Am J Public Health. 2011;101:2137–
2143.
7. Perk J, De Backer G, Gohlke H, et al. European guidelines on cardiovascular disease prevention in clinical practice (version 2012).
The Fifth Joint Task Force of the European Society of Cardiology
and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of nine societies and by
invited experts). Eur Heart J. 2012;33:1635–1701.
8. Domecq JP, Prutsky G, Wang Z, et al. Drugs commonly associated
with weight change: umbrella systematic review and meta-analysis
(Protocol). Syst Rev. 2012;1:44.
9. Smith V, Devane D, Begley CM, Clarke M. Methodology in conducting a systematic review of systematic reviews of healthcare interventions. BMC Med Res Methodol. 2011;11:15.
10. Zhou YH, Ma XQ, Wu C, et al. Effect of anti-obesity drug on
cardiovascular risk factors: a systematic review and meta-analysis of
randomized controlled trials. PLoS One. 2012;7:e39062.
11. Smith SM, Meyer M, Trinkley KE. Phentermine/topiramate for the
treatment of obesity. Ann Pharmacother. 2013;47:340 –349.
12. Dombrowski SU, Knittle K, Avenell A, Araújo-Soares V, Sniehotta
FF. Long term maintenance of weight loss with non-surgical interventions in obese adults: systematic review and meta-analyses of
randomised controlled trials. BMJ. 2014;348:g2646.
13. Johansson K, Neovius M, Hemmingsson E. Effects of anti-obesity
drugs, diet, and exercise on weight-loss maintenance after a verylow-calorie diet or low-calorie diet: a systematic review and metaanalysis of randomized controlled trials. Am J Clin Nutr. 2014;99:
14 –23.
14. Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of
Interventions. The Cochrane Collaboration. Version 5.1.0. Updated March 2011. Hoboken, NJ: Wiley; 2011.
15. Furukawa TA, Barbui C, Cipriani A, Brambilla P, Watanabe N.
Imputing missing standard deviations in meta-analyses can provide
accurate results. J Clin Epidemiol. 2006;59:7–10.
16. Swiglo BA, Murad MH, Schünemann HJ, et al. A case for clarity,
consistency, and helpfulness: state-of-the-art clinical practice guide-
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
369
lines in endocrinology using the grading of recommendations, assessment, development, and evaluation system. J Clin Endocrinol
Metab. 2008;93:666 – 673.
Stevens J, Truesdale KP, McClain JE, Cai J. The definition of weight
maintenance. Int J Obes (Lond). 2006;30:391–399.
Montori VM, Guyatt GH. Intention-to-treat principle. CMAJ.
2001;165:1339 –1341.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control
Clin Trials. 1986;7:177–188.
Higgins JP, Thompson SG. Quantifying heterogeneity in a metaanalysis. Stat Med. 2002;21:1539 –1558.
Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088 –1101.
Lau J, Ioannidis JP, Terrin N, Schmid CH, Olkin I. The case of the
misleading funnel plot. BMJ. 2006;333:597– 600.
Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. J Clin
Epidemiol. 2009;62:e1–34.
Da Silva JA, Jacobs JW, Kirwan JR, et al. Safety of low dose glucocorticoid treatment in rheumatoid arthritis: published evidence
and prospective trial data. Ann Rheum Dis. 2006;65:285–293.
Landewé RB, Boers M, Verhoeven AC, et al. COBRA combination
therapy in patients with early rheumatoid arthritis: long-term structural benefits of a brief intervention. Arthritis Rheum. 2002;46:
347–356.
Robinson LE, Holt TA, Rees K, Randeva HS, O’Hare JP. Effects of
exenatide and liraglutide on heart rate, blood pressure and body
weight: systematic review and meta-analysis. BMJ Open. 2013;
3(1)pii:e001986.
Craddy P, Palin HJ, Johnson KI. Comparative effectiveness of dipeptidylpeptidase-4 inhibitors in type 2 diabetes: a systematic
review and mixed treatment comparison. Diabetes Ther. 2014;
5:1– 41.
Klemp M, Tvete IF, Skomedal T, Gaasemyr J, Natvig B, Aursnes I.
A review and Bayesian meta-analysis of clinical efficacy and adverse
effects of 4 atypical neuroleptic drugs compared with haloperidol
and placebo. J Clin Psychopharmacol. 2011;31:698 –704.
Gartlehner G, Hansen RA, Morgan LC, et al. Second-Generation
Antidepressants in the Pharmacologic Treatment of Adult Depression: An Update of the 2007 Comparative Effectiveness Review.
Rockville, MD: Agency for Healthcare Research and Quality; 2011.
Wang T, Gou Z, Wang F, Ma M, Zhai SD. Comparison of GLP-1
analogues versus sitagliptin in the management of type 2 diabetes:
systematic review and meta-analysis of head-to-head studies. PLoS
One. 2014;9:e103798.
Rynn M, Russell J, Erickson J, et al. Efficacy and safety of duloxetine
in the treatment of generalized anxiety disorder: a flexible-dose,
progressive-titration, placebo-controlled trial. Depress Anxiety.
2008;25:182–189.
Montgomery SA, Tobias K, Zornberg GL, Kasper S, Pande AC.
Efficacy and safety of pregabalin in the treatment of generalized
anxiety disorder: a 6-week, multicenter, randomized, double-blind,
placebo-controlled comparison of pregabalin and venlafaxine.
J Clin Psychiatry. 2006;67:771–782.
Nemeroff CB, Thase ME. A double-blind, placebo-controlled comparison of venlafaxine and fluoxetine treatment in depressed outpatients. J Psychiatr Res. 2007;41:351–359.
Moscovitch A, Blashko CA, Eagles JM, et al. A placebo-controlled
study of sertraline in the treatment of outpatients with seasonal
affective disorder. Psychopharmacology (Berl). 2004;171:390 –
397.
Wise TN, Perahia DG, Pangallo BA, Losin WG, Wiltse CG. Effects
of the antidepressant duloxetine on body weight: analyses of 10
clinical studies. Prim Care Companion J Clin Psychiatry. 2006;8:
269 –278.
Dungan KM, Povedano ST, Forst T, et al. Once-weekly dulaglutide
versus once-daily liraglutide in metformin-treated patients with type
370
37.
38.
39.
40.
41.
42.
Domecq et al
Drugs Commonly Associated With Weight Change
2 diabetes (AWARD-6): a randomised, open-label, phase 3, noninferiority trial. Lancet. 2014;384(9951):1349 –1357.
Benkhadra K, Wang Z, Murad MH. Network meta-analysis: introduction and an example that compares devices for PFO closure. Eur
Heart J. 2014;pii:ehu343.
Danielsen KK, Svendsen M, Maehlum S, Sundgot-Borgen J.
Changes in body composition, cardiovascular disease risk factors,
and eating behavior after an intensive lifestyle intervention with high
volume of physical activity in severely obese subjects: a prospective
clinical controlled trial. J Obes. 2013;2013:325464.
Montori VM, Gandhi GY, Guyatt GH. Patient-important outcomes
in diabetes–time for consensus. Lancet. 2007;370:1104 –1106.
Björkhem-Bergman L, Asplund AB, Lindh JD. Metformin for
weight reduction in non-diabetic patients on antipsychotic drugs: a
systematic review and meta-analysis. J Psychopharmacol. 2011;25:
299 –305.
Das C, Mendez G, Jagasia S, Labbate LA. Second-generation antipsychotic use in schizophrenia and associated weight gain: a critical
review and meta-analysis of behavioral and pharmacologic treatments. Ann Clin Psychiatry. 2012;24:225–239.
Ehret M, Goethe J, Lanosa M, Coleman CI. The effect of metformin
43.
44.
45.
46.
47.
J Clin Endocrinol Metab, February 2015, 100(2):363–370
on anthropometrics and insulin resistance in patients receiving atypical antipsychotic agents: a meta-analysis. J Clin Psychiatry. 2010;
71:1286 –1292.
Fiedorowicz JG, Miller DD, Bishop JR, Calarge CA, Ellingrod VL,
Haynes WG. Systematic review and meta-analysis of pharmacological interventions for weight gain from antipsychotics and mood
stabilizers. Curr Psychiatry Rev. 2012;8:25–36.
Bartlett C, Doyal L, Ebrahim S, et al. The causes and effects of
socio-demographic exclusions from clinical trials. Health Technol
Assess. 2005;9:iii–iv, ix–x, 1–152.
Sabanayagam C, Teo BW, Tai ES, Jafar TH, Wong TY. Ethnic
differences in the association between blood pressure components
and chronic kidney disease in middle aged and older Asian adults.
BMC Nephrol. 2013;14:86.
Sabanayagam C, Teo BW, Tai ES, Jafar TH, Wong TY. Ethnic
variation in the impact of metabolic syndrome components and
chronic kidney disease. Maturitas. 2013;74:369 –374.
Lanting LC, Joung IM, Mackenbach JP, Lamberts SW, Bootsma
AH. Ethnic differences in mortality, end-stage complications, and
quality of care among diabetic patients: a review. Diabetes Care.
2005;28:2280 –2288.