Blood Pressure and Heart Rate Effects, Weight Loss and Maintenance During Long-Term Phentermine Pharmacotherapy for Obesity

Transcription

1 nture publishing group rticles Blood Pressure nd Hert Rte Effects, Weight Loss nd Mintennce During Long-Term Phentermine Phrmcotherpy for Obesity Ed J. Hendricks 1,2, Frnk L. Greenwy 3, Eric C. Westmn 4 nd Alok K. Gupt 3 There is perception tht phentermine phrmcotherpy for obesity increses blood pressure nd hert rte (HR), exposing treted ptients to incresed crdiovsculr risk. We collected dt from phentermine-treted (PT) nd phentermine-untreted (P0) ptients t privte weight mngement prctice, to exmine blood pressure, HR, nd weight chnges. Records of 300 sequentil returning ptients were selected who hd been treted with lowcrbohydrte ketogenic diet if their records included complete weight, blood pressure, nd HR dt from seven office exmintions during the first 12 weeks of therpy. The men time in therpy, time rnge, nd mode ws 92 (97.0), , nd 52 weeks. 14% were normotensive, 52% were prehypertensive, nd 34% were hypertensive t their first visit or hd previous dignosis of hypertension. PT subjects systolic blood pressure/distolic blood pressure (SBP/DBP) declined from bseline t ll dt points (SBP/DBP 6.9/ 5.0 mm Hg t 26, nd 7.3/ 5.4 t 52 weeks). P0 subjects declines of SBP/DBP t both 26 nd 52 weeks were 8.9/ 6.3 but the difference from the treted cohort ws not significnt. HR chnges in treted/untreted subjects t weeks 26 ( 0.9/ 3.5) nd 52 (+1.2/ 3.6) were not significnt. Weight loss ws significntly greter in the PT cohort for week 1 through 104 (P = ). These dt suggest phentermine tretment for obesity does not result in incresed SBP, DBP, or HR, nd tht weight loss ssisted with phentermine tretment is ssocited with fvorble shifts in ctegoricl blood pressure nd retrdtion of progression to hypertension in obese ptients. Obesity (2011) 19, doi: /oby Introduction Phentermine is the most widely used ntiobesity drug (1,2). Becuse it is congener of mphetmine, there is commonly held presumption tht phentermine will elevte blood pressure nd hert rte (HR). Although little clinicl dt hs been published regrding chnge in blood pressure or HR during phentermine phrmcotherpy, reviews of obesity phrmcotherpy hve mentioned increses in blood pressure nd HR s common dverse effects of phentermine phrmcotherpy (3 5). In contrst, dt from erly clinicl trils either did not mention blood pressure chnges (6), suggested low occurrence of hypertension during phentermine phrmcotherpy (7), or showed significnt reductions in blood pressure nd HR (8). A recent phentermine clinicl tril reported slight decreses in blood pressure but mde no mention of HR (9). A 1- or 2-yer long clinicl tril of phentermine monotherpy could provide the dt needed to clrify the effect of phentermine monotherpy on blood pressure nd HR but no such tril of phentermine hs been conducted. On the other hnd, phentermine hs been used cliniclly for weight mngement since the US Food nd Drug Administrtion grnted mrketing pprovl in 1959, nd hs been employed for long-term use in some clinics (2,10). Becuse dt from long-term rndomized clinicl tril of phentermine monotherpy is not vilble, crefully recorded clinicl experience is ll we hve to provide dt ddressing this inconsistency. The purpose of this retrospective study ws to exmine ptient records in britric medicine prctice tht hs employed long-term phentermine for weight mngement nd to nlyze both short-term nd long-term blood pressure nd HR effects of phentermine in the context of modern obesity tretment progrm. Methods nd Procedures Subjects Subject ptients were selected from the continuing ptient popultion of privte medicl britric prctice. Selection criteri included: (i) enrollment in prescribed rigorous weight mngement protocol, (ii) ttendnce t return visit minimum of six times within 12 weeks of the initil exm, (iii) complete vitl sign dt t seven clinic observtion 1 The Center for Weight Mngement, Roseville, Cliforni, USA; 2 The Center for Weight Mngement, Scrmento, Cliforni, USA; 3 Pennington Biomedicl Reserch Center, Louisin Stte University System, Bton Rouge, Louisin, USA; 4 Deprtment of Medicine, Duke University Medicl Center, Durhm, North Crolin, USA. Correspondence: Ed J. Hendricks (edhendricks@surewest.net) Received 23 November 2010; ccepted 11 Mrch 2011; published online 28 April doi: /oby obesity VOLUME 19 NUMBER 12 december

2 visits including weight, blood pressure, nd HR mesurements recorded t 0 time, nd t 1, 2, 3, 4, 8, nd 12 weeks, nd (iv) either phentermine monotherpy or no phrmcotherpy with ny obesity drug. Ptients who were on ntiobesity drugs other thn phentermine were excluded, s were ptients on phentermine combined with other ntiobesity drugs such s 5-hydroxytryptophn or topirmte. The selection criteri for the initil weight loss progrm excluded pregnnt women, ptients with recent myocrdil infrction, ptients with uncontrolled hypertension, ptients on drugs tht might chnge weight such s olnzpine or systemic steroids, nd ptients with cncer on chemotherpy. Ptients on ntidepressnt drugs were not excluded. Beginning on 2 Jnury 2007 the record ws reviewed of every returning ptient s they ppered for their first clinic office visit in Records were reviewed in the sequence in which the ptients ppered. The selection process continued until totl of 300 ptients were found who met the inclusion criteri. In doing so we reviewed 1,328 records, ending with ptients whose first visit in 2007 ws on 25 September Ptients who met the criteri (22.6%) were included; those who did not (77.4%) were excluded. Since ll were continuing ptients the durtion of therpy rnged from minimum of 12 weeks to mximum of 12 yers. Mny of the longterm ptients hd tken tretment holidys of vrying durtion. We included subsequent dt of ptients fter ny tretment interruption provided the tretment hitus did not exceed 1 yer. Procedures At initil exmintion the ptients hd complete medicl history tken which included n eting history, nd weight history. A focused physicl exmintion ws performed, n EKG ws performed, nd lbortory studies ordered which included chemistry profile, complete blood count, lipid studies, thyroid-stimulting hormone, nd liver function. Age, sex, birth dte, nd ethnicity were scertined nd weight, height, wist circumference, nd body ft mesured. Presence or bsence of comorbidities including dibetes, impired fsting glucose, coronry hert disese, congestive hert filure, metbolic syndrome, rthritis, etc. were noted. Pulse, weight, nd blood pressure were determined t ech visit. Redings were tken from the nondominnt rm resting t mid-sternl height fter the ptient hd been seted quietly for few minutes. Pulse rte ws determined by counting rdil pulse for 15 s. Blood pressure ws determined using mercury mnometer (Bumnometer; W.A. Bum Compny, Copige, NY). Before 1 December 2009 pulses nd blood pressures were determined once t ech visit. Blood pressure nd pulse rte dt fter 1 December 2009 ws obtined by using n utomted oscillimetric monitor (Terumo Elemno blood pressure monitor #H5503; Terumo, Tokyo, Jpn). The verge of two determintions ws recorded. If vlues differed by >10 dditionl redings were tken until two sequentil redings were within 10 mm Hg nd these were verged. Weights were determined using digitl scles (Tnit BWB 627A; Tnit, Tokyo, Jpn). Complince with the protocol nd chnges in phentermine dosges, if ny, were recorded. Adverse rections, if ny, were recorded. Ptients then hd weekly office visits t which they were interviewed nd exmined by physicin or licensed mid-level prctitioner for the first month, then every 2 weeks therefter. Mintennce ptients were seen t lest once every 3 months. This protocol utilized very lowcrbohydrte diet (11) with dily nutrient intke s follows: 1. Dily minimum protein intke set t g protein/kg idel body weight. Most ptients minimum intke ws set t between 90 nd 120 g protein/dy. 2. Crbohydrte intke ws set t 20 g or less, primrily from slds nd vegetbles. 3. Ft intke ws set t g, with emphsis on mono- nd polyunsturted fts. Once ptients were eting met fter the first week ft intke verged g dily. Ptients were instructed to use vrious selected liquid protein supplements s their sole protein source for the first week but then to begin eting solid niml protein during the second week. After the second week ptients were encourged to grdully increse their solid protein intke nd decrese their liquid protein intke but to mintin their protein intke of t lest five doses dily with totl within the initil rnge recommended. Medicl prctitioners ssessed ptient behviors t ech visit nd ny issues ddressed. No structured psychologicl testing or progrms were utilized. A prctitioner ssessed ptients exercise t ech visit. Ptients were counseled to initite dily mixture of crdiovsculr nd resistnce trining nd if their helth permitted, to grdully increse to 7 9 h of ccumultive exercise/week by the end first yer of the progrm. Phentermine dose The strting dose of phentermine rnged from 15 to 37.5 mg/dy. Phentermine-treted (PT) ptients were questioned regrding phentermine effects nd side effects t ech exmintion. When side effects occurred tht were not well tolerted, phentermine dosge ws reduced or discontinued. When phentermine-induced control of eting diminished in some ptients, dily phentermine doses for these ptients were slowly incresed using dose-to-effect titrtion until the ptient experienced improvement in eting control. Demogrphic dt Bseline demogrphic dt for the 300 study subjects re reported in Tble 1. Ninety percent of the study subjects met criteri for phrmcotherpy nd elected to include phentermine phrmcotherpy in their tretment protocol, which is typicl for new ptients in this medicl britric prctice. The percentge of mles ws higher in the phentermine-untreted (P0) cohort (26%) thn in the PT cohort (14%). Ninety-five percent of the study subjects were white. There were no significnt differences between the PT nd P0 groups in bseline ge, sex, ethnicity, weight, BMI, systolic blood pressure (SBP), or distolic blood pressure (DBP). P0 ptients hd slightly higher HR thn did PT ptients (80.4 (8.9) vs (9.9), P < 0.05). Using the criteri of the Joint Ntionl Committee on Blood Pressure (JNC 7) only 14% of the ptients hd norml blood pressure upon entry into the prctice, 54% were prehypertensive with initil blood pressures equl to or >120/80 or equl to or <139/89, nd 32% were hypertensive (12). Seven percent of the ptients hd impired fsting glucose t initil exmintion nd 4% hd type II dibetes. Sttisticl nlysis Weight, pulse, nd blood pressure over time were nlyzed using mixed model with repeted mesures. Sttisticl softwre ws mde by SAS Institute, Cry, NC. Non-normlly distributed dt like gender, ethnicity, nd percentges in the blood pressure groups ws nlyzed using the χ 2 -test. Normlly distributed dt like bseline weight, ge, pulse, BMI nd blood pressure were nlyzed using the Student t-test. Results Ptient progrm ttendnce Ner perfect progrm ttendnce during the initil 12 weeks ws requirement for inclusion. Prctitioner notes for ech visit were reviewed nd these indicted tht, t most encounters, ptients were dhering to the diet, were tking their medictions nd were following behvior modifiction dvice during the first 3 months. After 26 weeks ttendnce rte fell nd drop out rte incresed, especilly in those ptients whose weight loss ws less thn the verge. Follow-up lbortory exmintions were performed on selected ptients, generlly to follow initil bnormlities. Bicrbonte levels obtined 2352 VOLUME 19 NUMBER 12 december

3 Tble 1 Bseline demogrphic dt Totl Phen. Rx No Phen. Rx P vlue Number Sex: mle/femle, N 46/254 38/231 8/23 NS Sex: mle/femle, % 15/85 14/86 26/74 NS Ethnicity (W/B/H/O) 284/1/12/3 256/1/10/2 28/0/2/1 NS Men (s.d.) Age 49.6 (12.1) 49.6 (11.7) 49.4 (15.5) NS BMI 35.6 (8.16) 35.6 (8.38) 36.2 (5.95) NS Weight, pounds (53.1) (54.0) (43.5) NS Weight, kg 98.8 (24.1) 98.4 (24.5) (19.8) NS Systolic BP (11.3) (11.6) (8.6) NS Distolic BP 79.4 (7.6) 79.4 (7.7) 79.5 (6.7) NS Hert rte 76.5 (10.0) 76.0 (9.9) 80.4 (8.9) <0.05 Normotensive, count/% 42/14.0% 37/13.8% 5/16.1% Prehypertensive, count/% 161/53.7% 146/54.3% 17/54.8% Hypertensive, count/% 97/32.3% 86/32.0% 9/29.0% Tble 2 Weight loss, blood pressure, nd hert rte chnges for phentermine-untreted ptients (P0) Percent weight loss Weeks N Men 0 2.9% 4.2% 5.4% 6.9% 10.4% 12.8% 17.4% 19.0% 16.1% 8.4% 4.1% s.d Systolic BP Men s.d Δ s.d P b Distolic BP Men s.d Δ s.d P b Hert rte Men s.d Δ s.d P b BP, blood pressure. Chnge from bseline vlues. b Compred to bseline vlue. during the first 6 months on the progrm were lwys >22 milli-equivlents per deciliter. Lipid prmeters improved with weight loss s expected (13). Ptient retention t 26, 52, nd 104 weeks for PT ptients ws 81, 65, nd 37%. Ptient retention for P0 ptients ws significntly lower for the sme periods t 61, 48, nd 16%. These re not retention rtes clculted from initil visit becuse ptients who dropped out before 12 weeks were excluded. Blood pressure chnge Men SBP nd men DBP fell in both the P0 group (Tble 2 nd Figure 1e) nd in the PT group (Tble 3 nd Figure 1) s the ptients lost weight. In the PT group, SBP nd DBP fell rpidly during the initil 4 weeks of weight loss during which time the verge weight loss ws 7.6%. Although the verge PT subject regined some weight by yer 3, the men weight loss of PT subjects remined t bout 10% from yer 3 through yer 7 nd SBP obesity VOLUME 19 NUMBER 12 december

4 5 b % Weight loss/mm Hg/bets per minute c e d Weeks Yers % Wt loss SBP DBP HR Weeks Yers Figure 1 Weight loss, systolic blood pressure (SBP), distolic blood pressure (DBP), nd hert rte chnge by weeks/yers of therpy. () All phentermine-treted subjects (weight P < , SBP week 26, 52, 104 P , nd week P > 0.05), DBP week 26, 156 P > 0.04, nd week P > 0.05, HR P > 0.05), (b) phentermine-treted hypertensive subjects: weight P , SBP week 26, 52, 104, 156 P , week 208 P = , week 260 P = , week 312 P = , DBP week 26, 52, 104 P < , week 156 P = , week 208 P = , 260 P > 0.05, week 312 P > 0.05, HR P > 0.05, (c) phentermine-treted prehypertensive subjects (weight P < 0.01, SBP week 25 nd 52 P < , week P > 0.05, DBP P > 0.05, HR P > 0.05), (d) phentermine-treted normotensive subjects (SBP P > 0.05, DBP P > 0.05, HR P > 0.05), nd (e) phentermine-untreted subjects. (Weight week 52 P < , week 104 P > 0.05, SBP week 52 P < 0.04, week 104 P > 0.05, DBP week 52 P < 0.04, week 104 P > 0.05, HR P > 0.05). in these ptients remined significntly below bse line t week 26 (P < ), week 52 (P < ), week 104 (P = ), nd week 208 (P = ), below bseline but not significntly t week 156 (P = ) or weeks (P = ) wheres DBP in these ptients remined significntly below bseline t weeks 26 (P < ), 52 (P < ), 104 (P < ), 156 (P < ), nd 208 (P < ), below bseline but not significntly t weeks ( ). In the P0 group SBP nd DBP continued to fll until mximum weight loss occurred t 40 weeks. Although the men decrese in SBP nd DBP ws greter in the P0 group thn the PT group the SBP decrese from bseline did not persist s the ptients regined their weight reching bseline vlues t 104 weeks, even though verge weight loss ws still 8.4% (Figure 1e). Strtifiction of the PT group by initil blood pressure ctegory, reveled tht PT hypertensive ptients (Tble 4 nd Figure 1b) hd greter declines in both SBP nd DBP thn prehypertensive ptients (Tble 5 nd Figure 1c) nd tht PT ptients with initil blood pressure in the norml rnge experienced no significnt chnge in either SBP or DBP with weight loss (P > 0.05) (Tble 6 nd Figure 1d). HR chnge Men HR, decresed slightly but not significntly s ptients lost weight in the P0 cohort but HR either did not chnge or incresed slightly in the PT cohorts (P > 0.05). Ctegoricl blood pressure chnges As ptients lost weight nd blood pressure declined, some ptients who were initilly in the prehypertensive ctegory with blood pressures >119/79 but <140/90, experienced shift to the norml blood pressure ctegory. At 52 weeks 43 of 82 (52%) previously prehypertensive PT ptients nd 6 of 9 (67%) previously prehypertensive P0 ptients hd shifted to 2354 VOLUME 19 NUMBER 12 december

5 Tble 3 Weight loss, blood pressure, nd hert rte chnges for phentermine-treted ptients (PT) Percent weight loss Weeks N Men s.d Systolic BP Men Distolic BP s.d Δ s.d P b < < < Men s.d Δ s.d P b < < < Hert rte Men s.d Δ s.d P b BP, blood pressure. Chnge from bseline vlues. b Compred to bseline vlue. Tble 4 Weight loss, blood pressure, nd hert rte chnges for phentermine-treted ptients with hypertension t bseline Percent weight loss Weeks N Men s.d Systolic BP Men s.d Δ s.d P b < < < Distolic BP Men s.d Δ s.d P b < < < Hert rte Men s.d Δ s.d P b BP, blood pressure. Chnge from bseline vlues. b Compred to bseline vlue. the norml blood pressure ctegory. At 104 weeks of 27 of 55 (49%) previously prehypertensive PT ptients remined in the norml blood pressure ctegory, s did 12 of 34 (35%) t 156 weeks, 12 of 23 (52%) t 208 weeks, nd 6 of 16 (38%) t 260 weeks. The numbers of P0 ptients beyond 52 weeks were too low for meningful comprison. obesity VOLUME 19 NUMBER 12 december

6 Tble 5 Weight loss, blood pressure, nd hert rte chnges for phentermine-treted ptients with prehypertension t bseline % Weight loss Systolic BP Distolic BP Hert rte Weeks N Men s.d Men s.d Δ s.d P b < < Men s.d Δ s.d P b < < < Men s.d Δ s.d P b BP, blood pressure. Chnge from bseline vlues. b Compred to bseline vlue. Weight loss nd mintennce Men percentge weight loss for P0 nd ll PT ptients is depicted in Figure 1,e nd Tbles 2 nd 3. Percentge weight loss exmined s tretment effect by week over 2 yers in the entire dt set showed significntly greter in the PT cohort (P = ). Some ptients in the PT cohort continued to mintin >10% weight loss for s long s 8 yers. There were no significnt differences in weight loss between the ptients in the vrious blood pressure ctegory cohorts. Ctegoricl weight loss Among the 175 PT ptients with dt for 1 yer, 97% lost 5% or more of their initil weight t 1 yer. In the sme group 83% lost 10% or more, 62% lost 15% or more, 32% lost 20% or more, 17% lost 25% or more, 9% lost 30% or more, nd 3% lost 30% or more. Among the 15 P0 ptients with dt for 1 yer 80% lost t lest 5%, 73% lost 10%, 47% lost 15%, 33 lost 20%, 20% lost 25%, 13% lost 30%, nd 7% lost 35%. Figures showing individul ptient weight chnges from bseline to 1 yer re included s Supplementry Dt online. Phentermine dose Phentermine dose dt is shown in Tble 7. The initil (T0) phentermine dose rnged from 15 to 37.5 mg/dy with men dose of The modl dose ws 37.5 mg from T0 through yer 6, but s the dose ws incresed in some individuls, the men dose for the PT cohort slowly incresed, reching 57 mg/dy in yer 7. The modl dose incresed to 60 mg/dy t yer 7. The ptients in this study described experiencing two distinctly different phentermine therpeutic effects, which they believed contributed to their weight loss. The first nd most obvious ws ppetite nd hunger suppression. The second effect, usully not noticed until ptient hd been on phentermine for severl months, ws more subtle nd vriously described s improved or stronger control of eting, diminution or bsence of food crvings, or improved bility to follow their eting pln. Appetite suppression tended to fde with time in mny ptients but improved control of eting persisted nd ws still present fter ptients hd been tking phentermine for mny months or yers. Ptients experienced less hunger or even no hunger when strted on phentermine. This initil effect on hunger tended to diminish with time on ny given dose but could be re-estblished in some ptients by increses in phentermine dose. If phentermine ws discontinued, hunger nd loss of eting control quickly reppered. Ptients who took phentermine long-term to ssist with weight mintennce typiclly described noticing no effect dy-by-dy unless they skipped dy or more whereupon they noticed incresed hunger nd tht food servings grew lrger. They lso described diminution of eting self-discipline. Some ptients lso described return of crbohydrte crvings. Crbohydrte crvings typiclly disppered quickly on the low-crbohydrte ketogenic diet, but lter during mintennce s crbohydrte intke ws incresed, some ptients noted recurrence of crbohydrte crvings. For some ptients the beneficil effects of phentermine ppered to be preserved by increses in doses using dose-to-effect titrtion VOLUME 19 NUMBER 12 december

7 Tble 6 Weight loss, blood pressure, nd hert rte chnges for phentermine-treted ptients with norml blood pressure t bseline Weeks N Percent Men weight loss s.d Systolic BP Men s.d Δ s.d P b Distolic BP Men s.d Δ s.d P b Hert rte Men s.d Δ s.d P b BP. blood pressure. Chnge from bseline vlue. b Compred to bseline vlue. Tble 7 Phentermine doses in mg/dy Yer Week N Men s.d. Medin Mode Rnge To exmine for the possibility tht higher doses of phentermine might dversely ffect blood pressure or HR, the dt for ll PT subjects segregted by phentermine dose levels ws exmined t 52, 104, nd 156 weeks. There were no significnt differences in SBP, DBP, or HR compring ptient cohorts t the five dose levels 0, 18.75, , , nd mg phentermine/dy except for significnt drop in DBP in the phentermine group compred to no phentermine (P < 0.05). Although the differences in weight loss t the vrious doses were not demonstrbly sttisticlly significnt becuse of smll smple sizes, the men weight loss tended to be greter with incresed dily doses t 2 nd 3 yers. At 2 yers the men weight loss for no phentermine obesity VOLUME 19 NUMBER 12 december

8 nd for 0.5 stndrd dose ws bout 7%, for stndrd dose 11%, nd for 1.5 nd 2 stndrd dose 14%. At 3 yers weight loss for no phentermine ws 6%, for 1 nd 1.5 stndrd dose bout 10%, nd for 2 stndrd dose 12%. Adverse events To rule out the possibility of ny increses in individul ptient blood pressure vlues, the dt ws exmined crefully for occurrence of phentermine-induced blood pressure elevtions or phentermine-induced hypertension. One ptient in the study (subject #56), who initilly hd prehypertension, developed hypertension fter hving been on phentermine for 1 yer. He lost 15% of initil weight but ws still overweight. His hypertension ws esily controlled with low dose of lisinopril, nd he continued on phentermine. This ws the only instnce found where ptient developed hypertension fter being on phentermine continuously. Some of the hypertensive ptients in the study hd occsionl trnsient elevtions in either SBP or DBP. These were most often cused by omission of ntihypertensive medictions or by weight gin. In these cses phentermine ws withheld until blood pressure ws below 140/90. In the one cse where BP did not drop promptly, it did so when n ntihypertensive gent ws dded. Generlly the hypertensive study ptients on phentermine found their hypertension more esily mnged with either lower doses of medicines or no medicine. There were three PT ptients who were initilly normotensive, who did lose weight, but who subsequently developed prehypertension. One ptient (#53) developed prehypertension then ws dignosed with benign brin tumor, which ws successfully removed. She hd postopertive seizures nd is currently on topirmte nd not on phentermine. She is still prehypertensive. Another ptient (#148) developed prehypertension (BP 122/80) t week 2, nd subsequently hd intermittent SBPs in the low 120s. She dropped out, went off phentermine, then returned eighteen months lter with weight gin nd hypertension. Ptients on phentermine were routinely sked bout ny dverse phentermine effects t every encounter. The most common dverse effects were dry mouth nd insomni. Insomni occurred most often in the first few dys of phentermine therpy nd typiclly disppered within week or two. Some ptients experienced insomni even fter being on phentermine for yers if they took phentermine fter middy, but not if they took phentermine erlier. Dry mouth ws mild in most ptients nd typiclly either disppered within few weeks or becme tolerble provided the ptient drnk 64 ounces of wter dily. No other dverse effects were observed in the study group. Socil or work-relted dysfunction nd drug seeking behvior were crefully wtched for nd were not observed in the study group. Discussion The prevlence of high blood pressure nd men levels of SBP nd DBP re known to increse s BMI increses (14). Multiple studies hve shown tht weight loss induced by cloric restriction nd behvior chnge without ntiobesity phrmcotherpy is ssocited with decresed blood pressure (15). In this study, we found tht SBP nd DBP declined significntly during weight loss in ptients treted with phentermine, just s did blood pressures in the P0 ptients. Ptients with hypertension experienced the gretest declines while those with prehypertension hd lesser declines nd those with norml blood pressure hd the lest chnge. Tht weight loss induced by two-drug combintions including phentermine produced greter declines in ptients with hypertension thn in those with norml blood pressure hs been reported previously (16,17), but these reports did not strtify ptients by JNC 7 blood pressure ctegories so blood pressure declines in prehypertensive ptients were not nlyzed. In this study we found tht significnt proportion of both PT nd P0 ptients experienced ctegoricl shift from prehypertension to normotension t 12, 26, nd 52 weeks. This ctegoricl shift typiclly occurred in the first 12 weeks of therpy. Although the numbers of P0 ptients ws too low for sttisticl comprisons beyond 52 weeks, the ctegoricl shift persisted in smller numbers of PT ptients t 104 nd 156 weeks. The dt from this study suggests tht the occurrence of hypertension ssocited with phentermine phrmcotherpy is rre rther thn common. One ptient, initilly prehypertensive, of the 269 PT ptients developed hypertension fter 1 yer on phentermine. Excluding the ptients initilly hypertensive, dt for 119 ptients remined in the study t 52 weeks for n incidence rte of 0.84%/yer. If only the prehypertensive ptients re considered t 1 yer (N = 95), then the 1-yer incidence would be 1.05%. In considering these incidence rtes comprison should be mde to other dt on the incidence of new hypertension. Becuse elevtions of blood pressure in the hypertensive rnge ( 140/90) nd in the prehypertensive rnge (120/80 139/89) re very common in obese subjects, one could expect n nnul incidence rte of newly dignosed hypertension in untreted normotensive nd prehypertensive obese subjects followed for yer to be higher thn 1%. The ATTICA study (18) found n nnul incidence of new hypertension in the generl Greek popultion to be 2.86 cses per 100 in men nd 2.68 cses per 100 in women. The Strong Hert Study (19) found tht the 4-yer incidence of hypertension in group of ptients with prehypertension to be 38% nd 5-yer study in Tiwn (20) found conversion rte from prehypertension to hypertension to be 31.2%. These incidence rtes, in popultions not selected for obesity, re substntilly higher thn the rte of conversion of prehypertension to hypertension clculted for this study. These dt suggest tht weight loss ssisted by phentermine phrmcotherpy retrded progression of prehypertension to hypertension. Dt from the fixed-dose combintion of phentermine-topirmte (Qnex) clinicl trils lso suggested progression to hypertension ws retrded (17). Long-term nonphrmcologicl intervention trils hve been unsuccessful in mintining weight loss nd lower blood pressure for longer thn 3 4 yers (21). The number of ptients retined beyond 4 yers in this study is smll, but the dt suggests the possibility tht long-term phentermine therpy could 2358 VOLUME 19 NUMBER 12 december

9 be effective in mintining lower weight nd in lowering blood pressure for >4 yers in some ptients. Given the dt from this study nd the previous studies cited suggesting tht blood pressure declines rther thn increses with phentermine therpy, why does the ssumption tht phentermine dversely ffect blood pressure persist? The conception tht phentermine cn elevte blood pressure nd HR pprently derives from the often-repeted fct tht phentermine is n mphetmine congener. Tht mphetmine cn increse blood pressure nd HR nd induce or worsen hypertension is well documented (22). These dverse crdiovsculr effects re dose-relted nd dependent upon the route of dministrtion of the drug; inhltion or drug dministrtion vi intrnsl or intrvenous routes rpidly produces higher blood levels nd therefore greter toxic effect thn dose orl dministrtion. Certin mphetmine congeners such s methmphetmine (23) nd 3,4-methylenedioxymethmphetmine (24), which re typiclly bused by intrnsl, inhltion, or intrvenous routes, hve even greter toxic potentil nd re lso known to produce dverse crdiovsculr effects. Subjects who use the mphetmines of buse often self-dminister lrge doses producing very high blood levels in trying to chieve the desired stimulnt effect. On the other hnd, the route of dministrtion nd dose of mphetmines used for phrmcotherpy of ttention deficit nd ttention deficit hyperctivity disorder hve much lower toxicity potentil. Therpeutic use of these drugs is ssocited with slight increses in blood pressure nd in HR of questionble clinicl impct (25). In phentermine phrmcotherpy low (compred to the doses used of the mphetmines of buse) orl doses re used, nd phentermine blood levels chieved re comprtively lower thn blood levels chieved by subjects of the bused drugs. In ddition phentermine hs substntilly lower toxic potentil. Thus the potentil for dverse effects with phentermine phrmcotherpy is significntly lower thn for the bused mphetmines. Aside from few necdotl reports, there is no dt in the peer-reviewed medicl literture to support the perception tht phentermine increses blood pressure or HR. Phentermine clinicl tril reports, when detils of blood pressure nd HR hve been given, hve shown tht tril subjects experienced decreses in SBP nd DBP nd miniml or no chnges in HR (8,9,26). The sme is true of most trils with combintions with phentermine s one of two ntiobesity drugs. (16,17,27,28). The one exception to the ltter sttement is recent clinicl tril in which ptients given combintion of prmlintide plus phentermine showed no significnt chnge in clinic visit SBP nd DBP but n increse in HR of 5 bets/min (P < 0.01) (29). Ambultory blood pressure monitoring in the ltter study reveled bout 3 mm increse in DBP in the prmlintide plus phentermine subjects (P < 0.01) nd n increse in HR of bout 5 bets/min (P < 0.001). Becuse SBP nd HR decresed slightly in the prmlintide only group, the impliction is tht the phentermine produced the observed increses, however no subjects treted with phentermine lone were included in the study nd ptients with hypertension were excluded. We did not mesure mbultory blood pressures in this study. Our dt suggests tht phentermine phrmcotherpy does not significntly influence HR t ny time during tretment. In djustment of phentermine dose using dose-to-effect method, previously described by Rothmn (30), the desired effects re on eting behviors rther thn weight loss (31). Dose-to-effect titrtion is common prctice in medicine tht is employed with wide vriety of drugs. The most cogent comprison here would be with the methods used in djusting mphetmine dose for ttention deficit disorder nd ttention deficit hyperctivity disorder. Two pproches hve been described in treting ADD nd ADHD: (i) prescribe n initil low dose, evlute behviors, then grdully increse dose until behvior improves, or (ii) rmp up dose until undesirble side effects pper, then reduce dose to the level before they ppered (32). The first pproch ws typiclly employed in djusting phentermine doses in this study. Phentermine doseto-effect ws evluted t ech ptient encounter nd if undesirble side effects ppered the dose ws most often reduced. Men nd medin phentermine dosge trended upwrd for ptients in long-term tretment s shown in Tble 3. Ptients with dult ttention deficit disorder, which is common mong the obese (33), often derive behviorl benefit from phentermine (34) nd my tolerte higher doses. A few of the ptients on higher doses in this study were thought to hve previously undignosed dult ADD, but the mjority were simply more comfortble with their eting behvior t higher dose. Phentermine doses higher thn the US Food nd Drug Administrtion recommended mximum of 37.5 mg dily hve been previously described (2,7,30,35), but there hve been no clinicl trils tht hve exmined this issue. Phentermine is presumed to hve ddiction potentil nd is clssified s ctegory IV controlled substnce. An importnt criterion is mking dignosis of ddiction is tht drug-ddicted subjects typiclly mnifest dysfunction in one or severl spheres such s socil, work, or legl rens. Prctitioners nd stff of the clinic observed the ptients in this study crefully for signs of such dysfunction, nd no such behviors were observed. This study hs severl strengths. The study clinic is well estblished with mny long-term ptients. The physicin in chrge hs hd long experience with phentermine. As is lwys with retrospective observtionl studies, limittion could hve been bis in subject selection, which is possible even though criteri were set in plce to void bis. The methods for blood pressure nd especilly for HR determintion re nother limittion. The smll smple size in some of the cohorts is yet nother limittion. We conclude the dt from this retrospective study suggests tht the ddition of phentermine phrmcotherpy for ptients in comprehensive weight mngement progrm does not dversely ffect HR or blood pressure. Indeed, this study suggests tht in this setting phentermine phrmcotherpy, cn be especilly beneficil for the obese hypertensive nd prehypertensive ptients since persistent weight loss ssisted with such therpy my lower blood pressure long-term nd retrd the progression of prehypertension to hypertension delying obesity VOLUME 19 NUMBER 12 december

10 the nturl course of hypertension in the obese nd potentilly reducing the risk of mortlity due to congestive hert filure, stroke nd ischemic hert disese. Some of these ptients mintined weight loss of 10% or greter for s long s 8 yers suggesting long-term phentermine phrmcotherpy my be useful tretment for long-term weight mintennce. The P0 ptients who dhered to diet lone without phentermine tretment grnered similr benefits to PT ptients over the first yer of tretment, but the benefits for the P0 ptients begn to vnish s they regined weight wheres PT ptients benefits persisted. PT ptients lso hd slightly better initil weight loss. Finlly, there is suggestion tht phentermine doses >37.5 mg/ dy re sfe nd cn be useful in selected ptients. The potentil benefit to the obese popultion is huge if the findings of this study cn be confirmed nd phentermine phrmcotherpy ultimtely is used more widely. We suggest there is need for more investigtions into both the clinicl effects of phentermine nd its moleculr bsis of ction. Development of psychometric scles to ssess phentermine effect on eting behvior for guidnce of dose-to-effect titrtion could prove cliniclly useful. We suggest long-term, dose-rnging phentermine clinicl tril nd outcome study could potentilly confirm the crdiovsculr benefits implied by this study. SUPPLEMENTARY MATERIAL Supplementry mteril is linked to the online version of the pper t Disclosure The uthors declred no conflict of interest The Obesity Society REFERENCES 1. Stfford RS, Rdley DC. Ntionl trends in ntiobesity mediction use. Arch Intern Med 2003;163: Hendricks EJ, Rothmn RB, Greenwy FL. How physicin obesity specilists use drugs to tret obesity. Obesity (Silver Spring) 2009;17: Kpln LM. Phrmcologic therpies for obesity. Gstroenterol Clin North Am 2010;39: Bry GA. Lifestyle nd phrmcologicl pproches to weight loss: efficcy nd sfety. J Clin Endocrinol Metb 2008;93:S81 S Greenwy FL, Cruso MK. Sfety of obesity drugs. Expert Opin Drug Sf 2005;4: Lorber J. Obesity in childhood. A controlled tril of norectic drugs. Arch Dis Child 1966;41: Dougls A, Dougls JG, Robertson CE, Munro JF. Plsm phentermine levels, weight loss nd side-effects. Int J Obes 1983;7: Vllé-Jones JC, Brodie NH, O Hr H, O Hr J, McGhie RL. A comprtive study of phentermine nd diethylpropion in the tretment of obese ptients in generl prctice. Phrmtherpeutic 1983;3: Kim KK, Cho HJ, Kng HC, Youn BB, Lee KR. Effects on weight reduction nd sfety of short-term phentermine dministrtion in Koren obese people. Yonsei Med J 2006;47: Frnk A. The long-term mngement of obesity with continuing phrmcotherpy. Obes Res 2004;12: Foster GD, Wytt HR, Hill JO et l. Weight nd metbolic outcomes fter 2 yers on low-crbohydrte versus low-ft diet: rndomized tril. Ann Intern Med 2010;153: Chobnin AV, Bkris GL, Blck HR et l.; Joint Ntionl Committee on Prevention, Detection, Evlution, nd Tretment of High Blood Pressure. Ntionl Hert, Lung, nd Blood Institute; Ntionl High Blood Pressure Eduction Progrm Coordinting Committee. Seventh report of the Joint Ntionl Committee on Prevention, Detection, Evlution, nd Tretment of High Blood Pressure. Hypertension 2003;42: Poobln A, Aucott L, Smith WC et l. Effects of weight loss in overweight/ obese individuls nd long-term lipid outcomes systemtic review. Obes Rev 2004;5: Brown CD, Higgins M, Donto KA et l. Body mss index nd the prevlence of hypertension nd dyslipidemi. Obes Res 2000;8: Aucott L, Rothnie H, McIntyre L et l. Long-term weight loss from lifestyle intervention benefits blood pressure?: systemtic review. Hypertension 2009;54: Atkinson RL, Blnk RC, Schumcher D, Dhurndhr NV, Ritch DL. Longterm drug tretment of obesity in privte prctice setting. Obes Res 1997;5: Vivus. QNEXA FDA Advisory Committee Briefing Document (cited 8 August 2010); VIVUS, Inc. Briefing Informtion, (NDA) , proposed trdenme, QNEXA (phentermine/topirmte) Controlled Relese Cpsules, for the July 15, 2010 Joint Meeting of the Endocrinologic nd Metbolic Drugs Advisory Committee. < EndocrinologicndMetbolicDrugsAdvisoryCommittee/ucm htm>. 18. Pngiotkos DB, Chrysohoou C, Pitsvos C et l. Hierrchicl nlysis of nthropometric indices in the prediction of 5-yer incidence of hypertension in pprently helthy dults: the ATTICA study. Atherosclerosis 2009;206: De Mrco M, de Simone G, Romn MJ et l. Crdiovsculr nd metbolic predictors of progression of prehypertension into hypertension: the Strong Hert Study. Hypertension 2009;54: Liu LK, Peng LN, Chen LK, Hwng SJ, Chiou ST. Prehypertension mong middle-ged nd elderly people in Tiwn: five-yer follow-up. J Atheroscler Thromb 2010;17: Kuller LH. Weight loss nd reduction of blood pressure nd hypertension. Hypertension 2009;54: Lile JA, Bblonis S, Emurin C et l. Comprison of the Behviorl nd Crdiovsculr Effects of Intrnsl nd Orl d-amphetmine in Helthy Humn Subjects. J Clin Phrmcol 2010; e-pub hed of print 29 Jul Hrt CL, Gunderson EW, Perez A et l. Acute physiologicl nd behviorl effects of intrnsl methmphetmine in humns. Neuropsychophrmcology 2008;33: Kolbrich EA, Goodwin RS, Gorelick DA et l. Physiologicl nd subjective responses to controlled orl 3,4-methylenedioxymethmphetmine dministrtion. J Clin Psychophrmcol 2008;28: Pul H, Timothy W, Eric M et l. Crdiovsculr effects of longer-term, high-dose OROS methylphenidte in dolescents with ttention deficit hyperctivity disorder. J Peditrics 2009;155: Weintrub M, Hsdy JD, Mushlin AI, Lockwood DH. A double-blind clinicl tril in weight control. Use of fenflurmine nd phentermine lone nd in combintion. Arch Intern Med 1984;144: Weintrub M, Sundresn PR, Mdn M et l. Long-term weight control study. I (weeks 0 to 34). The enhncement of behvior modifiction, cloric restriction, nd exercise by fenflurmine plus phentermine versus plcebo. Clin Phrmcol Ther 1992;51: Colmn E, Roberts MD. FDA Clinicl Briefing Document, EMDAC, QNEXA, NDA Deprtment of Helth nd Humn Services: Wshington, D.C. < CommitteesMeetingMterils/Drugs/EndocrinologicndMetbolicDrugs AdvisoryCommittee/UCM pdf> Accessed 15 July Aronne LJ, Hlseth AE, Burns CM, Miller S, Shen LZ. Enhnced weight loss following codministrtion of prmlintide with sibutrmine or phentermine in multicenter tril. Obesity (Silver Spring) 2010;18: Rothmn RB. Tretment of obesity with combintion phrmcotherpy. Am J Ther 2010;17: Hlford JC. Lorcserin not new wepon in the bttle with ppetite. Nt Rev Endocrinol 2010;6: Mnos MJ. Phrmcologic tretment of ADHD: rod conditions in driving ptients to successful outcomes. Medscpe J Med 2008;10: Pgoto SL, Curtin C, Lemon SC et l. Assocition between dult ttention deficit/hyperctivity disorder nd obesity in the US popultion. Obesity (Silver Spring) 2009;17: Rothmn RB. Tretment of 4-yer-old boy with ADHD with the dopmine releser phentermine. J Clin Psychitry 1996;57: Hddock CK, Poston WS, Foreyt JP, DiBrtolomeo JJ, Wrner PO. Effectiveness of Medifst supplements combined with obesity phrmcotherpy: clinicl progrm evlution. Et Weight Disord 2008;13: VOLUME 19 NUMBER 12 december