Content uploaded by Ed J. Hendricks
Author content
All content in this area was uploaded by Ed J. Hendricks on Jun 30, 2016
Content may be subject to copyright.
ORIGINAL ARTICLE
Addiction potential of phentermine prescribed during long-term
treatment of obesity
EJ Hendricks
1
, M Srisurapanont
2
, SL Schmidt
3
, M Haggard
1
, S Souter
1
, CL Mitchell
1
, DG De Marco
1
, MJ Hendricks
1
,
Y Istratiy
1
and FL Greenway
4
OBJECTIVE: To investigate if phentermine treatment induces phentermine abuse, psychological dependence (addiction) or
phentermine drug craving in overweight, obese and weight loss maintenance patients. To investigate whether amphetamine-like
withdrawal occurs after abrupt cessation of long-term phentermine treatment.
DESIGN: Clinical intervention trial with interruption of phentermine treatment in long-term patients.
SUBJECTS: 269 obese, overweight or formerly obese subjects (age: 20–88 years, BMI: 21–74 kg m
2
) treated with phentermine
long-term (LTP, N¼117), 1.1–21.1 years, or short-term (ATP, N¼152), 4–22 days, with phentermine doses of 18.75–112.5 (LTP) and
15–93.75 (ATP) mg per day.
MEASUREMENTS: Module K of the Mini International Neuropsychiatric Interview modified for phentermine (MINI-SUD), Severity of
Dependence Scale (SDS), 45-item Cocaine Craving Questionnaire-NOW (CCQ-NOW) modified for phentermine (PCQ-NOW), and
Amphetamine Withdrawal Questionnaire (AWQ) modified for phentermine (PWQ).
RESULTS: MINI-SUD interviews were negative for phentermine abuse or psychological dependence in all LTP patients. SDS
examination scores were low for all LTP and ATP patients, indicating they were not psychologically dependent upon phentermine.
PCQ-NOW scores were low for all LTP and ATP patients, indicating neither short-term nor long-term phentermine treatment had
induced phentermine craving. Other than an increase in hunger or eating, amphetamine-like withdrawal symptoms did not occur
upon abrupt phentermine cessation as measured by sequential PWQ scores.
CONCLUSIONS: Phentermine abuse or psychological dependence (addiction) does not occur in patients treated with phentermine
for obesity. Phentermine treatment does not induce phentermine drug craving, a hallmark sign of addiction. Amphetamine-like
withdrawal does not occur upon abrupt treatment cessation even at doses much higher than commonly recommended and after
treatment durations of up to 21 years.
International Journal of Obesity (2014) 38, 292–298; doi:10.1038/ijo.2013.74
Keywords: phentermine; weight loss; addiction; abuse; dependence; withdrawal
INTRODUCTION
As the worldwide obesity epidemic continues unabated,
pharmacotherapy remains an underutilized modality in treating
obesity.
1
Of course a major reason for this is that there is a
paucity of approved effective drugs, but another reason is that
use of certain drugs has been marginalized.
2
Among the older
drugs, approved before the US Food and Drug Administration
(FDA) began to require long-term trials for obesity drugs,
phentermine is known to be effective, both for short-term
weight loss and for long-term weight maintenance.
3–9
Although
long-term, FDA-supervised, clinical trials have never been
conducted for phentermine, 97% of US obesity treatment
specialists use phentermine as their agent-of-choice drug
for treating obesity.
10
The use of this drug for obesity
treatment by other physicians throughout the world has long
been curtailed because of what we believe to be misapprehensions
regarding phentermine safety.
2
Concerns of phentermine-induced
addiction and of adverse cardiovascular reactions are two fears
that have had a profound negative impact on phentermine
prescribing.
To determine whether concerns of addiction resulting from
phentermine treatment are reasonable, we carried out this
addiction potential of phentermine prescribed during long-term
treatment of obesity (APPLO) trial, which aimed to examine if
psychological dependence, abuse, craving and withdrawal actu-
ally did occur in obese patients receiving long-term treatment
with phentermine.
PATIENTS AND METHODS
Patient recruitment
Patients were recruited in two cohorts from a private fee-for-service
obesity medicine specialty practice. Overweight, obese and weight loss
maintenance patients in this practice are typically treated with antiobesity
drugs over a long term.
8,11
Patients were recruited for the long-term
phentermine-treated (LTP) cohort if they had been treated with
phentermine for a minimum of 1 year, while patients for the acute-
term phentermine (ATP) cohort were recruited after they had been on
phentermine for 7–14 days. All patients were 18 years of age or older; there
were no other age restrictions. Patients who had taken cumulative
phentermine drug holidays exceeding 60 days in the previous 12 months
1
Center for Weight Management, Roseville, CA, USA;
2
Department of Psychiatry, Chiang Mai University, Chiang Mai, Thailand;
3
Department of Endocrinology, Metabolism &
Diabetes, University of Colorado, Anschutz Medical Campus, Aurora, CO, USA and
4
Pennington Biomedical Research Center of the Louisiana State University System, Baton
Rouge, LA, USA. Correspondence: Dr EJ Hendricks, Center for Weight Management, 2510 Douglas Boulevard, Suite 200, Roseville, CA 95661, USA.
E-mail: edhendricks@surewest.net
Received 11 February 2013; revised 7 April 2013; accepted 10 April 2013; accepted article preview online 17 May 2013; advance online publication, 11 June 2013
International Journal of Obesity (2014) 38, 292–298
&
2014 Macmillan Publishers Limited All rights reserved 0307-0565/14
www.nature.com/ijo
or who had taken any drug holiday in the previous 30 days were excluded
from the LTP cohort. Patients with current Axis I psychiatric diagnoses were
included provided these were stable and under treatment. Patients with
dependence on drugs other than nicotine were excluded, as phentermine
was not prescribed for such patients. Ethical and Independent Review
Services, Independence, Missouri, a commercial Independent Review
Board, approved the study procedures and informed consent document.
After the risks of the study had been fully explained, all participating
patients provided informed consent. Participants were given no remunera-
tion for participation in the study; each paid the clinic’s customary fees for
their care without discount. Beginning in August 2011, patients newly
started on phentermine were recruited for the ATP cohort as they
appeared for follow-up examinations after 7–14 days of treatment. At the
same time eligibility of returning patients for inclusion in the LTP cohort
was determined as the patients appeared at the clinic. Those eligible were
invited to participate. Few ATP and LTP eligible candidates declined
participation. Patient demographics are detailed in Table 1. This clinical
trial is registered at clinicaltrials.gov as NCT01402674.
Examination procedures
The Mini International Neuropsychiatric Interview (MINI)
12
is a structured
interview guide for making a DSM-IV-TR diagnosis of mental
disorders, including substance dependence and abuse. Module K, for
diagnosing non-alcohol psychoactive substance use disorder (MINI-
SUD), was used and modified by restricting questioning to phentermine.
Module K is for assessing addictive behavior over the 12 months prior
to assessment. Module K, version 5.0.0 of the MINI was modified for
this study.
The Severity of Dependence Scale (SDS),
13
a four-point, Likert-type
(score 0–3), five-question psychometric scale used for assessing the
severity of substance dependence was modified by replacing the word
‘drug’ with ‘phentermine.’ The SDS used for the LTP patients is shown in
Table 2. The SDS used for the ATP patients was identical except the time-
frame was weeks rather than years.
The Cocaine Craving Questionnaire NOW (CCQ-NOW), a seven-point
Likert-type (score 1–7) 45-question psychometric scale
14
originally
developed for assessment of cocaine drug craving, but also used for
methamphetamine craving in methamphetamine-dependent subjects,
15
was modified for phentermine by replacing the words ‘cocaine’ and
‘coke’ with ‘phentermine,’ to create a PCQ-NOW.
The Amphetamine Withdrawal Questionnaire (AWQ),
16
a five-point
Likert-type (score 0–4), ten-question psychometric scale for assessing
the severity of amphetamine withdrawal in addicted subjects was
modified to a Phentermine Withdrawal Questionnaire (PWQ) by
changing the word ‘amphetamine’ to ‘phentermine.’
Examination schedule
ATP patients were tested once in the clinic with SDS and PCQ-NOW upon
recruitment.
LTP patients were tested in the clinic with MINI-SUD, SDS, PCQ-NOW and
PWQ (Day 0 or D0) upon recruitment. LTP patients were instructed to
continue to take their usual dose of phentermine until 1 day prior to their
next office visit when they were to skip their phentermine, and self-
examine with the PWQ (D1) at B24 h after their last phentermine dose at
home. They were instructed to skip their phentermine again the following
Table 1. Demographic and clinical characteristics of patients
APT (N ¼152) LPT (N ¼117) Significant difference
a
Asked, consented, completed, (LTP) withdrawal completed 158, 152, 152 124, 117, 117, 76
Mean (s.d.) Mean (s.d.)
Age 44.18 (12.39) 51.18 (11.40) t¼4.757, df ¼267, Po0.001
BMI 34.88 (7.31) 33.79 (7.59) t¼1.198, df ¼267, P¼0.232
Days of phentermine treatment 9.30 (3.35)
Years of phentermine treatment 8.35 (5.16)
Rx Duration, Range (days/years) 4–22 1.1–21.5
Phentermine-HCl (mg) 34.46 (9.23) 53.43 (19.46) t¼9.732, df ¼267, Po0.001
Phentermine dose range (mg/d) 15–93.75 18.75–112.5
Number of office examinations 72.1 (51.7)
Patients who had taken treatment hiatuso1 month 117 (100%)
Patients who had taken treatment hiatus41 month 70 (60%)
Average treatment hiatus (mo) 19.5 (19.9)
N(%) N(%)
Sex (female) 134 (88.2) 108 (92.3) w
2
¼1.261, df ¼1, P¼0.261
Race
White 138 (90.8) 103 (88.0) w
2
¼1.812, df ¼3, P¼0.612
Hispanic 10 (6.6) (7.7)
Black 1 (0.7) (2.6)
Asia 3 (2.0) 2 (1.7)
Concomitant antidepressant 2 (1.3) 24 (20.5)
Bupropion 3
Citalopram 5
Desvenlafaxine 1
Duloxetine 1
Escitalopram 5
Fluoxetine 4
Paroxetine 1
Sertraline 2
Trazodone 3
Venlafaxine 1
Other antiobesity medications 0 6 (5.1)
Topiramate 4
Diethylpropion 2
LPT patients were older and received higher doses of phentermine.
a
Only the differences of age, BMI, phentermine dose, sex and race between groups were
compared.
Investigation for phentermine dependence
EJ Hendricks et al
293
&2014 Macmillan Publishers Limited International Journal of Obesity (2014) 292 – 298
day, the day of their scheduled office visit. Upon arrival at the clinic they
were again tested with the PWQ (D2) at B48 h after their last dose of
phentermine, after which they resumed phentermine treatment.
Data analysis
Mean and standard deviation were calculated for patient age, BMI,
phentermine dose, duration of phentermine treatment and number of
office visits. Scores for SDS, PCQ-NOW and PWQ are ordinal data. Ordinal
data statistical methods, including Mann–Whitney, Kruskal–Wallis, Wil-
coxen signed rank, and Friedman tests calculated by SPSS V17.0 ( SPSS Inc.,
Chicago, IL, USA) for Windows, were employed for statistical analysis for
the psychometric scale scores.
RESULTS
Phentermine dose and duration distribution
Phentermine doses and durations of treatment are listed in
Table 1. Patients were treated variously with 15 or 30 mg capsules
of phentermine hydrochloride or 37.5 mg tablets of phentermine
hydrochloride. The latter tablets are scored enabling patients to
take one-half tablet of 18.73 milligrams of phentermine hydro-
chloride. As has been previously reported,
8
phentermine doses for
clinic patients may be adjusted over time using dose-to-effect
titration where the effect monitored is control of eating behavior.
Because of use of this method of ‘dose-to-effect’ titration of
phentermine dosage, we find average phentermine doses of clinic
patients gradually increase with duration of treatment. Among the
ATP cohort phentermine doses were: 19.1% on 15–18.75, 78.9%
on 30–37.5 and 2% on greater than 37.5 mg per day. The
distribution of doses in the LTP cohort was: 41% at 18.75–37,
24.8% at 45–56.25, 30.8% at 60–75 and 3.4% at 475 mg per day.
Duration of treatment among the ATP cohort varied slightly
with a median value of 8.0 days. More striking were the variations
in duration of treatment in the LTP cohort ranging from 1.1 years
to 21.1 years with a median value of 7.2 years.
Mini International Neuropsychiatric Interview
MINI-SUD interviews for each of the 117 LTP patients examined
were negative for phentermine dependence or abuse. These long-
term patients were well known at the clinic as indicated by the
mean number (±s.d.) of clinical examination visits for the LTP
patients of 72.1 (51.7). In reviewing the LTP patient records, we
found no suggestion that any of the LTP patients had ever
exhibited maladaptive drug use symptoms listed as the criteria
for drug abuse or dependence in the DSM-IV TR. Nor did any of
these patients exhibit drug-seeking behaviors, manifest signs
of compulsive phentermine use, or have apparent clinically
significant impairment or distress as a result of their phentermine
use. The finding that 100% of the LTP MINI-SUD examinations
were negative for either phentermine abuse or psychological
dependence is consistent with the clinic’s 24 years experience in
treating obesity with phentermine.
Severity of Dependence Scale
SDS data collected from the 152 ATP patients were compared with
that collected from the 117 LTP patients using the Mann–Whitney
Utest. LTP patients’ mean SDS scores were slightly higher (0.50
(0.91)) than the ATP patients’ mean scores (0.42 (0.75)), but the
difference was not significant (P¼0.528). The distribution of
patients’ individual question and total scores is shown in Table 3.
Note that the majority of both ATP and LTP had total scores of 0 or
1 corresponding to answering never or sometimes, respectively.
The range of total scores among the ATP patients was 0–3. The
range of total scores among the LPT patients was 0–5, with one
patient scoring 5 and one patient scoring 4. Examination of the
individual question scoring for these two subjects revealed that
Table 2. Severity of Dependence Scale—SDS.
Never/
almost
never
Sometimes Often Always/
nearly
always
1. Do you think your use of phentermine was out of control? 0 1 2 3
2. Did the prospect of missing a dose make you anxious or worried? 0 1 2 3
3. Did you worry about your use of phentermine? 0 1 2 3
4. Did you wish you could stop? 0 1 2 3
Not
difficult
Quite
difficult
Very
difficult
Impossible
5. How difficult did (or would) you find it to stop or go without
phentermine?
0123
The above questions are about your phentermine during the last year. Please indicate your
response to each question by drawing a circle around the number in the box with your
answer.
The SDS is in the public domain and may be downloaded from the Alcohol and Drug Abuse Institute Library of the University of Washington: http://
adai.washington.edu/instruments/pdf/Severity_of_Dependence_Scale_397.pdf.
Table 3. Distribution of patient scores for each SDS question and SDS total for ATP and LTP patients
Scores ATP ATP ATP ATP ATP ATP LTP LTP LTP LTP LTP LTP
Q1 Q2 Q3 Q4 Q5 Total Q1 Q2 Q3 Q4 Q5 Total
0 149 140 128 150 134 107 115 96 104 106 108 78
1 3 12 20 2 18 31 2 21 13 11 7 28
2 00300 9 00 0015
3 00100 5 00 0014
4 00000 0 00 0001
5 00000 0 00 0001
Values tabulated are number of patients. ATP, N¼152; LTP, N¼117. The sum of each column is 152 for ATP and 117 for LTP.
Investigation for phentermine dependence
EJ Hendricks et al
294
International Journal of Obesity (2014) 292 – 298 &2014 Macmillan Publishers Limited
the patient with a total of 5 answered sometimes for a score of
1 on each of the five SDS questions, while the patient with a total
of 4 answered sometimes, scoring 1 for questions 2, 3, 4 and 5.
Twenty-four of the 117 LTP or 20.5% of patients had a diagnosis
of depression, and were taking an antidepressant medication
when recruited as indicated in Table 1. There was no significant
difference in mean total SDS scores between the 24 patients
on antidepressant medicines, 0.54 (1.18) and the 93 LTP patients
who were not, 0.49 (0.83) (Mann–Whitney Utest: Z¼0.419,
P¼0.676).
Phentermine Craving Questionnaire—NOW
In scoring the PCQ-NOW we followed the scoring conventions
suggested by Heinz et al.
17
in which total a score for 41 of the 45
questions is calculated and scores subtotaled for four domains.
The four domains as suggested by Heinz et al.
17
are: factor 1,
desire, factor 2, lack of self-efficacy, factor 3, compulsivity, and
factor 4, relief. PCQ-NOW scoring results are presented in Table 4.
The total and each of the four domain scores for the LTP patients
are significantly lower than the corresponding scores for the ATP
patients. There was no significant difference in PCQ-NOW mean
total scores between the 24 patients on antidepressant medicines
(2.06 (0.69)) and the 93 LTP patients who were not (1.90 (0.63))
(Mann–Whitney Utest, Z¼0.915, P¼0.360).
Phentermine Withdrawal Questionnaire
LTP patients were tested with the PWQ three times. The first test
was done on a day they had taken phentermine (D0). They were
also tested at B24 h after cessation (D1) and again at 48 h after
cessation (D2). PWQ scoring is shown in Table 5. Ten-item PWQ
total scores for D0, D1 and D2 differed significantly (Friedman test,
P¼0.046) in which the peak and bottom-most of withdrawal
states were found on D1 and D2, respectively (Z¼2.50,
P¼0.013). When the three PWQ component syndromes were
compared, scores for the hyperarousal and anxiety syndromes
were not significantly different, but the reversed vegetative
syndrome score was different (Friedman test, P¼0.011). Again,
the peak of withdrawal was found on D1 (Z¼2.478, P¼0.013).
The reversed vegetative syndrome score is a combination of three
symptoms: fatigue, hyperphagia and sleepiness. Individual ques-
tion scores for fatigue and sleepiness did not differ significantly
among D0, D1 and D2 scores. Thus the hyperphagia score was the
only PWQ question significantly different within the reversed
vegetative syndrome score among the D0, D1 and D2 scores
(P¼0.027). Nine-item total PWQ scores, omitting scores for the
hyperphagia question, were not significantly different for D0, D1
and D2.
DISCUSSION
The results of the MINI-SUD structured interviews were unambig-
uous with no hint of either phentermine abuse or psychological
dependence in any LTP subject.
The SDS is considered a measure of compulsive drug use
13
that
has been used as a screening tool to identify drug users at risk
for addiction.
18
Total SDS scores 44 are thought to indicate
problematic amphetamine use in known amphetamine users
19
requiring further investigation. Suggested SDS cutoff values for
other drugs are 43 for cocaine,
20
and 43 for cannabis.
21
Gossop
et al.,
13
in a study of the SDS in drug users, found that the mean
SDS scores in two samples from heroin abusers were 5.2 (5.0),
N¼222, and 8.7 (4.0), N¼408. The range of values in each sample
Table 4. Scores from PCQ-NOW, a 45-item, Likert-type, psychometric scale adapted from the Cocaine Craving Questionnaire NOW of Tiffany by
replacing the words ‘cocaine’ and ‘coke’ with ‘phentermine’
APT (N ¼152) LPT (N ¼117) Significant difference
a
PCQ-NOW desire 1.30 (0.53) 1.17 (0.42) Z ¼3.155, P¼0.002
PCQ-NOW lack of self- efficacy 3.29 (1.29) 2.74 (1.31) Z¼3.549, Po0.001
PCQ-NOW relief 1.74 (0.67) 1.58 (0.66) Z¼2.391, P¼0.017
PCQ-NOW compulsivity 2.55 (0.97) 2.13 (0.89) Z¼3.716, Po0.001
PCQ-NOW total 2.25 (0.71) 1.93 (0.64) Z¼3.668, Po0.001
Comparison of the PCQ-NOW total scores and 4 PCQ-NOW domain scores obtained from the APT (N¼152) and LPT (N¼117) patients.
a
Mann–Whitney Utests
(PCQ-NOW scores treated as ordinal data).
Table 5. Phentermine Withdrawal Questionnaire (PWQ) scores obtained while on phentermine (D0), one day after the last dose of phentermine (D1)
and two days after the last dose of phentermine (D2)
Measure Mean PWQ score (s.d.) Significant difference
a
D0 D1 D2
Total score (10 items) 3.39 (3.48) 3.55 (3.05) 2.99 (2.65) w
2
¼6.146, df ¼2, P¼0.046
b
Hyperarousal syndrome score 0.51 (0.86) 0.43 (0.825) 0.37 (0.85) w
2
¼1.852, df ¼2, P¼0.396
Anxiety syndrome score 0.57 (0.96) 0.49 (1.01) 0.43 (0.82) w
2
¼1.776, df ¼2, P¼0.412
Reversed vegetative syndrome score 1.85 (2.0) 2.35 (1.72) 2.09 (1.79) w
2
¼8.941, df ¼2, P¼0.011
c
Fatigue symptom score 0.86 (0.88) 0.91 (0.91) 0.75 (0.80) w
2
¼2.849, df ¼2, P¼0.241
Hyperphagia symptom score 0.77 (0.97) 1.16 (1.02) 1.03 (1.09) w
2
¼7.260, df ¼2, P¼0.027
d
Sleepiness symptom score 0.25 (0.59) 0.32 (0.62) 0.30 (0.67) w
2
¼1.000, df ¼2, P¼0.607
nine-item PWQ total score
e
2.65 (3.01) 2.35 (2.60) 1.95 (2.38) w
2
¼4.962, df ¼2, P¼0.084
a
Friedman test, if significantly difference found, followed by Wilcoxon Signed Rank test.
b
Mean D2 score significantly lower than mean D1 score (Z¼2.50,
P¼0.013).
c
Mean D1 score significantly higher than mean D0 score (Z¼2.478, P¼0.013).
d
Mean D1 significantly higher than mean D0 score (Z¼2.829,
P¼0.005).
e
PWQ total score—PWQ hyperphagia symptom score.
Investigation for phentermine dependence
EJ Hendricks et al
295
&2014 Macmillan Publishers Limited International Journal of Obesity (2014) 292 – 298
was 0–15, the latter being the maximum possible SDS score.
13
In the same study the mean SDS scores in two cohorts of
amphetamine abusers were 3.7 (4.0), N¼231, and 4.3 (3.2),
N¼301. The range of values in the two amphetamine user
samples was 1–15 and 0–14, respectively. The mean SDS score in a
cohort of cocaine-using patients (N¼15) was 4.2 (3.3) with a range
of 0–13. The mean SDS scores and the range of SDS scores in our
study patients were substantially lower than those reported for
amphetamine-addicted subjects. One LTP patient did have a SDS
score of 5, scoring 1 in each of the five questions, and one LTP
patient had a SDS score of 4, scoring 1 in questions 2, 3, 4 and 5.
However, these two subjects had been patients for 10.4 and
13.2 years, and had negative MINI-SUD interviews; neither had
shown drug-seeking behavior nor any sign of compulsive use of
phentermine. The SDS individual and total question scoring in this
study is evidence that the addiction potential of phentermine is
vanishingly low in a clinical setting.
The PCQ-NOW was included as a measure of phentermine
craving. Drug craving is considered a hallmark symptom of
addiction and relapse.
22,23
The CCQ-NOW has been used as a
measure of cocaine and other drug craving intensity in known
drug users.
17
Addictive drugs and substances vary in their
addictive potential and the number of exposures required to
induce addiction. Variation in onset of addiction and drug craving
also depends on how rapidly peak drug levels are reached, and
this varies by route of administration. Thus, addiction occurs
quickly with few exposures to inhaled crystal methamphetamine,
while multiple exposures are required to induce addiction to
hydrocodone via the oral route. Our rationale in comparing
phentermine-craving intensity between the ATP and LTP cohorts
was that if orally-administered phentermine has any addiction
potential it is of a very low order, and that one would not expect
phentermine craving to appear after 1–2 weeks of phentermine
treatment. We found that both total and domain subtotal
PCQ-NOW scores were significantly higher in the ATP patients
than in the LTP patients. This is the opposite result from the one
expected if long-term phentermine treatment is indeed addicting
and induces phentermine craving.
The AWQ is a measure of severity of the ten major symptoms
found to occur in amphetamine-addicted subjects after abrupt
cessation of amphetamine.
16
The ten questions assess drug
craving, dysphoria, anhedonia, anxiety, motor retardation,
agitation, fatigue, increased appetite or eating, vivid or
unpleasant dreams, and craving for sleep or hypersomnia. Each
question has five possible answers ranging from ‘‘not at all’’ to
‘very much’ with a range of scores of 0–4. AWQ total score is the
total of all ten questions with a possible range of 0–40. One can
also analyze AWQ by calculating three syndrome scores:
hyperarousal syndrome, a combination of scores for drug craving,
agitation, and vivid or unpleasant dreams; reversed vegetative
syndrome, combined scores for fatigue, increased appetite, and
craving for sleep or hypersomnia; and anxiety syndrome,
combined scores for anhedonia, anxiety and motor retardation.
In one study of the AWQ in hospitalized amphetamine abusers
the total scores were highest at D0, then fell slightly on D1 and
again on D2.
24
As seen in panel 1, page 1323, the total AWQ scores
for the inpatients appear to be approximately D0E17, D1E15
and D2E13. Also depicted are AWQ scores for a comparison
group of healthy nondependent individuals, which followed that
same pattern but were substantially lower: D0E3.0, D1E1.5 and
D2E1.7. PWQ scores in LTP patients were also low and were
comparable to the scores of the comparison group.
As seen in Table 5, different PWQ scores among those obtained
on D0, D1 and D2 found on the total, reversed vegetative
syndrome and hyperphagia scores suggested the possibility of
phentermine withdrawal. However, in examining the syndrome
and symptom related to the withdrawal effect, hyperphagia was
the only withdrawal symptom found. If hyperphagia was not
taken into account, the nine-item PWQ total scores obtained from
those three time points were not significantly different.
In addition, the mean nine-item PWQ total scores on D1 (2.35)
and on D2 (1.95) were actually lower than that on D0 (2.65), which
were contrast to the concept of stimulant withdrawal.
Thus, the only amphetamine-like withdrawal symptom we
found after abrupt phentermine cessation was an increase in
appetite or an increase in eating. This represents a loss of
phentermine therapeutic effect rather than an amphetamine-like
withdrawal symptom. As seen in Table 1, 100% of LTP subjects
had discontinued phentermine and stayed off it for less than
1 month. Most of these episodes were patient initiated in an effort
to see if they could continue to lose weight or could maintain
a weight loss without continuing phentermine. Occasionally
the clinician initiated these episodes in an attempt to ascertain
efficacy or to determine if phentermine was the cause of an
unfavorable reaction. These patients invariably reported return of
appetite, increased hunger, increased craving or increased eating,
either singly or in combination. Typically weight gain slowly
ensued within a month of phentermine cessation, but occasional
patients reported successful weight maintenance for longer
intervals. 70% of LTP patients gave a history of having
discontinued phentermine for 41 month. The majority of these
returned after having regained weight during the hiatus. Hence,
reluctance to enroll as an LTP subject was rare because these
patients had all tried abrupt cessation of phentermine many times,
knew well the symptoms that would ensue, and were unafraid of
abrupt cessation because they were confident they could avoid
weight gain during a brief treatment hiatus.
Evidence supporting the concept that phentermine is an
addictive substance is missing in the medical literature. The
idea that phentermine has addiction potential appears to stem
from early work in rats
25
and from the fact that phentermine
is a substituted phenethylamine, a large group of compounds
with widely varying effects that includes addicting substances
such as methamphetamine, cathinone, MDMA (3,4-methylene-
dioxymethamphetamine or ‘Ecstasy’) and DOM (2,5-dimethoxy-4-
methylamphetamine or ‘STP’), but also compounds without addiction
potential such as ephedrine (N-methyl-b-hydroxyamphetamine)
and bupropion (3-chloro-teri-butyl-b-ketoamphetamine).
There are no published reports that orally administered
phentermine treatment for obesity has been associated with
abuse or psychological dependence (addiction) as defined by the
DSM-IV TR. Although methamphetamine abuse, addiction and
withdrawal syndromes are now well described and characterized
in the peer-reviewed medical literature,
24,26–28
no comparable
syndrome for phentermine has been described.
Prior to our addiction medicine study of abrupt phentermine
cessation in patients who had taken it long-term as treatment for
obesity,
11
there were only a few reports of human studies of
phentermine abuse liability.
29–31
None of the latter studies were
conducted in subjects being phentermine-treated for obesity.
Subjects tested in a laboratory environment were found to ‘like’
phentermine to varying degrees or to have subjective effects
similar in some ways to those the same subjects experienced with
amphetamine.
Two reports indicated that phentermine had been found
in urine drug screening samples from truck drivers
32
and in
hair samples from Korean drug suspects.
33
In 1979, Jain et al.
34
had reported that 3.4% of over 10 000 urine specimens from
the Los Angeles County probation department and methadone
maintenance programs tested positive for amphetamine. and that
1.7% of the 10 000 specimens tested positive for phentermine
using a gas-liquid chromatography technique.
34
These reports
suggest that although apparently phentermine abuse does occur,
it may be limited to subjects who abuse other stimulants.
In a recently conducted telephone survey of 23 drug
rehabilitation centers in Kentucky these centers reported that
Investigation for phentermine dependence
EJ Hendricks et al
296
International Journal of Obesity (2014) 292 – 298 &2014 Macmillan Publishers Limited
during 2010 and 2011 they had treated B24 000 patients but that
not one had been admitted or treated for phentermine use or
addiction (personal communication, Julie Swindler, M.D., Lexington,
Kentucky). In another recently conducted telephone survey of
medical directors of 50 drug treatment centers and 50 major
hospital emergency rooms in each of 50 US states, no physician
could recall ever having diagnosed or treated phentermine abuse,
addiction or withdrawal (personal communication, David Bryman,
D.O., Scottsdale, Arizona).
The drug has now been in use for more than 50 years and
although a few anecdotes suggesting phentermine is abused have
appeared,
32–34
there is still no clinical evidence in the peer-
reviewed medical literature to support the hypothesis that
phentermine is addicting or has any significant human addiction
potential.
35
Research in addiction medicine has undergone
noteworthy development since 1959. Concepts of addiction
have shifted from an early focus on tolerance and withdrawal to
a current emphasis on psychological components of dependence.
13
Standardized diagnostic criteria for drug abuse, addiction and
withdrawal were agreed upon among the American Psychiatric
Association, and adopted for the Diagnostic and Statistical Manual
of Mental Disorders—IV.
36
Although there was agreement on
definition of addiction, the label for this condition was changed to
dependence, a change which has created considerable confusion
as there is a clinical difference between psychological dependence,
meaning a compulsive drug-taking condition or addiction, and
physical dependence, which can occur in anyone taking a medicine
affecting the central nervous system.
37
In this paper we use the
terms ‘psychological dependence’ and ‘addiction’ synonymously.
Psychometric testing methods have been developed, validated and
applied clinically for measurements of psychological dependence,
drug craving and withdrawal for a wide variety of substances of
abuse including cocaine, heroin, amphetamine and others.
16,17,24,38,39
Currently drug craving is considered a characteristic symptom of
both drug abuse, and psychological dependence or addiction.
22,23
We are aware of two studies that compared the addiction
potential of phentermine with the addiction potential of other
drugs using drug liking scales.
30,31
Both studies found that when
experimental subjects were given phentermine some liked the
drug. Neither of these studies was conducted in subjects given
phentermine for obesity treatment. The only study we are aware
of that has investigated phentermine addiction potential in
phentermine-treated patients using modern addiction medicine
metrics is one we recently conducted looking for amphetamine-
like withdrawal symptoms in patients treated with long-term
phentermine in a weight management program who ceased
taking phentermine abruptly.
11
We used a psychometric scale
originally developed for use in evaluating withdrawal symptoms
in cocaine abusers occurring after abrupt cocaine cessation.
38
The scale has been used to characterize cocaine withdrawal,
40,41
and has also been modified for use in characterizing amphetamine
withdrawal.
15,24
We found patients on long-term phentermine who
ceased phentermine abruptly did not develop an amphetamine-like
withdrawal symptom complex. Significantly we also found no
evidence that long-term phentermine use induced phentermine
craving.
This report is the first clinical trial conducted in which patients
treated with phentermine for obesity have been examined with
validated, currently used addiction medicine metrics. The study
was conducted among patients in a fee-for-service obesity
medicine specialty practice that has existed over two decades.
Our data strongly suggest that long-term phentermine pharma-
cotherapy for obesity for up to 21.5 years and at doses up to 112.5
mg per day does not induce abuse or psychological dependence
(addiction), that long-term phentermine pharmacotherapy does
not induce phentermine drug craving, and that abrupt treatment
cessation does not induce amphetamine-like withdrawal. These
trial data suggest fears of causing addiction with long-term
phentermine are exaggerated and present a needless barrier to
better care for overweight and obese patients worldwide.
CONFLICT OF INTEREST
Dr Hendricks has received honoraria from Akramax Pharmaceuticals, Eurodrug
Laboratories, Citius Pharmaceuticals and Vivus pharmaceuticals. Dr Greenway has
received honoraria from Baronova, Basic Research, Citius, Diabetic Living, Eisai, GNC,
Jenny Craig, Lithera, Merck, Naturalpha, Nume Health, Orexigen, Plensat, Takeda,
Thetis, Unigene and Zafgen. The remaining authors declare no conflict of interest.
REFERENCES
1 Samaranayake NR, Ong KL, Leung RY, Cheung BM. Management of obesity in the
National Health and Nutrition Examination Survey (NHANES), 2007-2008. Ann
Epidemiol 2012; 22: 349–353.
2 Colman E. Anorectics on trial: a half century of federal regulation of prescription
appetite suppressants. Ann Intern Med 2005; 143: 380–385.
3 Glazer G. Long-term pharmacotherapy of obesity 2000: a review of efficacy and
safety. Arch Intern Med 2001; 161: 1814–1824.
4 Ioannides-Demos LL, Piccenna L, McNeil JJ. Pharmacotherapies for obesity: past,
current, and future therapies. J Obes 2011; 2011: 179674.
5 Kang JG, Park CY, Kang JH, Park YW, Park SW. Randomized controlled trial to
investigate the effects of a newly developed formulation of phentermine diffuse-
controlled release for obesity. Diabetes Obes Metab 2010; 12: 876–882.
6 Kim KK, Cho HJ, Kang HC, Youn BB, Lee KR. Effects on weight reduction and safety
of short-term phentermine administration in korean obese people. Yonsei Med J
2006; 47: 614–625.
7 Frank A. The long-term management of obesity with continuing pharmacother-
apy. Obes Res 2004; 12: 1821–1827.
8 Hendricks EJ, Greenway FL, Westman EC, Gupta AK. Blood pressure and heart rate
effects, weight loss and maintenance during long-term phentermine pharma-
cotherapy for obesity. Obesity (Silver Spring) 2011; 19: 2351–2360.
9 Rothman RB. Treatment of obesity with ‘combination’ pharmacotherapy. Am J
Ther 2010; 17: 596–603.
10 Hendricks EJ, Rothman RB, Greenway FL. How physician obesity specialists use
drugs to treat obesity. Obesity (Silver Spring) 2009; 17: 1730–1735.
11 Hendricks EJ, Greenway FL. A study of abrupt phentermine cessation in patients
in a weight management program. Am J Ther 2011; 18: 292–299.
12 Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E et al. The
Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and
validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10.
J Clin Psychiatry 1998; 59(Suppl 20): 22–33quiz 34-57.
13 Gossop M, Darke S, Griffiths P, Hando J, Powis B, Hall W et al. The Severity of
Dependence Scale (SDS): psychometric properties of the SDS in English and
Australian samples of heroin, cocaine and amphetamine users. Addiction 1995; 90:
607–614.
14 Tiffany ST, Singleton E, Haertzen CA, Henningfield JE. The development of a
cocaine craving questionnaire. Drug Alcohol Depend 1993; 34: 19–28.
15 McGregor C. Amphetamine withdrawal:nature, time course, and treatment. PhD
Thesis. Department of Clinical and Experimental Pharmacology, University of
Adelaide, Adelaide, 2005, 224pp.
16 Srisurapanont M, Jarusuraisin N, Jittiwutikan J. Amphetamine withdrawal: I.
Reliability, validity and factor structure of a measure. Aust N Z J Psychiatry 1999;
33: 89–93.
17 Heinz AJ, Epstein DH, Schroeder JR, Singleton EG, Heishman SJ, Preston KL. Heroin
and cocaine craving and use during treatment: measurement validation and
potential relationships. J Subst Abuse Treat 2006; 31: 355–364.
18 Bruno R, Matthews AJ, Topp L, Degenhardt L, Gomez R, Dunn M. Can the severity
of dependence scale be usefully applied to ‘ecstasy’? Neuropsychobiology 2009;
60: 137–147.
19 Topp L, Mattick RP. Choosing a cut-off on the Severity of Dependence Scale (SDS)
for amphetamine users. Addiction 1997; 92: 839–845.
20 Kaye S, Darke S. Determining a diagnostic cut-off on the Severity of Dependence
Scale (SDS) for cocaine dependence. Addiction 2002; 97: 727–731.
21 Martin G, Copeland J, Gates P, Gilmour S. The Severity of Dependence Scale (SDS)
in an adolescent population of cannabis users: reliability, validity and diagnostic
cut-off. Drug Alcohol Depend 2006; 83: 90–93.
22 Tiffany ST, Wray JM. The clinical significance of drug craving. Ann N Y Acad Sci
2012; 1248: 1–17.
23 Newton TF, De La Garza 2nd R, Kalechstein AD, Tziortzis D, Jacobsen CA. Theories
of addiction: methamphetamine users’ explanations for continuing drug use and
relapse. Am J Addict 2009; 18: 294–300.
Investigation for phentermine dependence
EJ Hendricks et al
297
&2014 Macmillan Publishers Limited International Journal of Obesity (2014) 292 – 298
24 McGregor C, Srisurapanont M, Jittiwutikarn J, Laobhripatr S, Wongtan T, White JM.
The nature, time course and severity of methamphetamine withdrawal. Addiction
2005; 100: 1320–1329.
25 van Rossum JM, Simons F. Locomotor activity and anorexogenic action.
Psychopharmacologia 1969; 14: 248–254.
26 Srisurapanont M, Arunpongpaisal S, Wada K, Marsden J, Ali R, Kongsakon R.
Comparisons of methamphetamine psychotic and schizophrenic symptoms: a
differential item functioning analysis. Prog Neuropsychopharmacol Biol Psychiatry
2011; 35: 959–964.
27 Kish SJ. Pharmacologic mechanisms of crystal meth. CMAJ 2008; 178: 1679–1682.
28 Zorick T, Nestor L, Miotto K, Sugar C, Hellemann G, Scanlon G et al. Withdrawal
symptoms in abstinent methamphetamine-dependent subjects. Addiction 2010;
105: 1809–1818.
29 Brauer LH, Johanson CE, Schuster CR, Rothman RB, de Wit H. Evaluation of
phentermine and fenfluramine, alone and in combination, in normal, healthy
volunteers. Neuropsychopharmacology 1996; 14: 233–241.
30 Jasinski DR, Faries DE, Moore RJ, Schuh LM, Allen AJ. Abuse liability assessment of
atomoxetine in a drug-abusing population. Drug Alcohol Depend 2008; 95: 140–146.
31 Schoedel KA, Addy C, Chakraborty B, Rosko K, Dunbar S, Maes A et al. Human abuse
potential and cognitive effects of taranabant, a cannabinoid 1 receptor inverse
agonist: a randomized, double-blind, placebo- and active-controlled, crossover
study in recreational polydrug users. J Clin Psychopharmacol 2012; 32: 492–502.
32 Couper FJ, Pemberton M, Jarvis A, Hughes M, Logan BK. Prevalence of drug use in
commercial tractor-trailer drivers. J Forensic Sci 2002; 47: 562–567.
33 Lee S, Kim J, In S, Choi H, Chung H, Chung KH. Detection of phentermine in hair
samples from drug suspects. Forensic Sci Int 2011; 207: e5–e7.
34 Jain NC, Budd RD, Sneath TC. Frequency of use or abuse of amphetamine-related
drugs. Am J Drug Alcohol Abuse 1979; 6: 53–57.
35 Greene WM, Sylvester M, Abraham J. Addiction liability of pharmacotherapeutic
interventions in obesity. Curr Pharm Des 2011; 17: 1188–1192.
36 American Psychiatric Association. Diagnostic and Statistical Manual of Mental
Disorders. 4th edn (American Psychiatric Association: Washington, DC, USA, 2000.
37 O’Brien CP, Volkow N, Li TK. What’s in a word? Addiction versus dependence in
DSM-V. Am J Psychiatry 2006; 163: 764–765.
38 Kampman KM, Volpicelli JR, McGinnis DE, Alterman AI, Weinrieb RM, D’Angelo L
et al. Reliability and validity of the Cocaine Selective Severity Assessment. Addict
Behav 1998; 23: 449–461.
39 Shoptaw SJ, Kao U, Heinzerling K, Ling W. Treatment for amphetamine
withdrawal. Cochrane Database Syst Rev 2009: CD003021: doi: 10.1002/14651858.
40 Ahmadi J, Kampman K, Dackis C. Outcome predictors in cocaine dependence
treatment trials. Am J Addict 2006; 15: 434–439.
41 Ahmadi J, Kampman KM, Oslin DM, Pettinati HM, Dackis C, Sparkman T. Predictors
of treatment outcome in outpatient cocaine and alcohol dependence treatment.
Am J Addict 2009; 18:81–86.
Investigation for phentermine dependence
EJ Hendricks et al
298
International Journal of Obesity (2014) 292 – 298 &2014 Macmillan Publishers Limited
