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




There is a perception that phentermine pharmacotherapy for obesity increases blood pressure and heart rate (HR), exposing treated patients to increased cardiovascular risk. We collected data from phentermine-treated (PT) and phentermine-untreated (P0) patients at a private weight management practice, to examine blood pressure, HR, and weight changes. Records of 300 sequential returning patients were selected who had been treated with a low-carbohydrate ketogenic diet if their records included complete weight, blood pressure, and HR data from seven office examinations during the first 12 weeks of therapy. The mean time in therapy, time range, and mode was 92 (97.0), 12–624, and 52 weeks. 14% were normotensive, 52% were prehypertensive, and 34% were hypertensive at their first visit or had a previous diagnosis of hypertension. PT subjects systolic blood pressure/diastolic blood pressure (SBP/DBP) declined from baseline at all data points (SBP/DBP −6.9/−5.0 mm Hg at 26, and −7.3/−5.4 at 52 weeks). P0 subjects' declines of SBP/DBP at both 26 and 52 weeks were −8.9/−6.3 but the difference from the treated cohort was not significant. HR changes in treated/untreated subjects at weeks 26 (−0.9/−3.5) and 52 (+1.2/−3.6) were not significant. Weight loss was significantly greater in the PT cohort for week 1 through 104 (P = 0.0144). These data suggest phentermine treatment for obesity does not result in increased SBP, DBP, or HR, and that weight loss assisted with phentermine treatment is associated with favorable shifts in categorical blood pressure and retardation of progression to hypertension in obese patients.


Phentermine is the most widely used antiobesity drug (1,2). Because it is a congener of amphetamine, there is a commonly held presumption that phentermine will elevate blood pressure and heart rate (HR). Although little clinical data has been published regarding change in blood pressure or HR during phentermine pharmacotherapy, reviews of obesity pharmacotherapy have mentioned increases in blood pressure and HR as common adverse effects of phentermine pharmacotherapy (3,4,5). In contrast, data from early clinical trials either did not mention blood pressure changes (6), suggested a low occurrence of hypertension during phentermine pharmacotherapy (7), or showed significant reductions in blood pressure and HR (8). A recent phentermine clinical trial reported slight decreases in blood pressure but made no mention of HR (9). A 1- or 2-year long clinical trial of phentermine monotherapy could provide the data needed to clarify the effect of phentermine monotherapy on blood pressure and HR but no such trial of phentermine has been conducted. On the other hand, phentermine has been used clinically for weight management since the US Food and Drug Administration granted marketing approval in 1959, and has been employed for long-term use in some clinics (2,10). Because data from a long-term randomized clinical trial of phentermine monotherapy is not available, carefully recorded clinical experience is all we have to provide data addressing this inconsistency. The purpose of this retrospective study was to examine patient records in a bariatric medicine practice that has employed long-term phentermine for weight management and to analyze both short-term and long-term blood pressure and HR effects of phentermine in the context of a modern obesity treatment program.

Methods and Procedures


Subject patients were selected from the continuing patient population of a private medical bariatric practice. Selection criteria included: (i) enrollment in a prescribed rigorous weight management protocol, (ii) attendance at a return visit a minimum of six times within 12 weeks of the initial exam, (iii) complete vital sign data at seven clinic observation visits including weight, blood pressure, and HR measurements recorded at 0 time, and at 1, 2, 3, 4, 8, and 12 weeks, and (iv) either phentermine monotherapy or no pharmacotherapy with any obesity drug. Patients who were on antiobesity drugs other than phentermine were excluded, as were patients on phentermine combined with other antiobesity drugs such as 5-hydroxytryptophan or topiramate. The selection criteria for the initial weight loss program excluded pregnant women, patients with recent myocardial infarction, patients with uncontrolled hypertension, patients on drugs that might change weight such as olanzapine or systemic steroids, and patients with cancer on chemotherapy. Patients on antidepressant drugs were not excluded. Beginning on 2 January 2007 the record was reviewed of every returning patient as they appeared for their first clinic office visit in 2007. Records were reviewed in the sequence in which the patients appeared. The selection process continued until a total of 300 patients were found who met the inclusion criteria. In doing so we reviewed 1,328 records, ending with patients whose first visit in 2007 was on 25 September 2007. Patients who met the criteria (22.6%) were included; those who did not (77.4%) were excluded. Since all were continuing patients the duration of therapy ranged from a minimum of 12 weeks to a maximum of 12 years. Many of the long-term patients had taken treatment holidays of varying duration. We included subsequent data of patients after any treatment interruption provided the treatment hiatus did not exceed 1 year.


At initial examination the patients had a complete medical history taken which included an eating history, and a weight history. A focused physical examination was performed, an EKG was performed, and laboratory studies ordered which included a chemistry profile, a complete blood count, lipid studies, thyroid-stimulating hormone, and liver function. Age, sex, birth date, and ethnicity were ascertained and weight, height, waist circumference, and body fat measured. Presence or absence of comorbidities including diabetes, impaired fasting glucose, coronary heart disease, congestive heart failure, metabolic syndrome, arthritis, etc. were noted.

Pulse, weight, and blood pressure were determined at each visit. Readings were taken from the nondominant arm resting at mid-sternal height after the patient had been seated quietly for a few minutes. Pulse rate was determined by counting radial pulse for 15 s. Blood pressure was determined using a mercury manometer (Baumanometer; W.A. Baum Company, Copiage, NY). Before 1 December 2009 pulses and blood pressures were determined once at each visit. Blood pressure and pulse rate data after 1 December 2009 was obtained by using an automated oscillimetric monitor (Terumo Elemano blood pressure monitor #H5503; Terumo, Tokyo, Japan). The average of two determinations was recorded. If values differed by >10 additional readings were taken until two sequential readings were within 10 mm Hg and these were averaged. Weights were determined using digital scales (Tanita BWB 627A; Tanita, Tokyo, Japan). Compliance with the protocol and changes in phentermine dosages, if any, were recorded. Adverse reactions, if any, were recorded.

Patients then had weekly office visits at which they were interviewed and examined by a physician or licensed mid-level practitioner for the first month, then every 2 weeks thereafter. Maintenance patients were seen at least once every 3 months. This protocol utilized a very low-carbohydrate diet (11) with daily nutrient intake as follows:

  • 1. Daily minimum protein intake set at 1.6–2.0 g protein/kg ideal body weight. Most patients' minimum intake was set at between 90 and 120 g protein/day.
  • 2. Carbohydrate intake was set at 20 g or less, primarily from salads and vegetables.
  • 3. Fat intake was set at 15–25 g, with emphasis on mono- and polyunsaturated fats. Once patients were eating meat after the first week fat intake averaged 25–40 g daily.

Patients were instructed to use various selected liquid protein supplements as their sole protein source for the first week but then to begin eating solid animal protein during the second week. After the second week patients were encouraged to gradually increase their solid protein intake and decrease their liquid protein intake but to maintain their protein intake of at least five doses daily with a total within the initial range recommended.

Medical practitioners assessed patient behaviors at each visit and any issues addressed. No structured psychological testing or programs were utilized.

A practitioner assessed patients' exercise at each visit. Patients were counseled to initiate a daily mixture of cardiovascular and resistance training and if their health permitted, to gradually increase to 7–9 h of accumulative exercise/week by the end first year of the program.

Phentermine dose

The starting dose of phentermine ranged from 15 to 37.5 mg/day. Phentermine-treated (PT) patients were questioned regarding phentermine effects and side effects at each examination. When side effects occurred that were not well tolerated, phentermine dosage was reduced or discontinued. When phentermine-induced control of eating diminished in some patients, daily phentermine doses for these patients were slowly increased using dose-to-effect titration until the patient experienced improvement in eating control.

Demographic data

Baseline demographic data for the 300 study subjects are reported in Table 1. Ninety percent of the study subjects met criteria for pharmacotherapy and elected to include phentermine pharmacotherapy in their treatment protocol, which is typical for new patients in this medical bariatric practice. The percentage of males was higher in the phentermine-untreated (P0) cohort (26%) than in the PT cohort (14%). Ninety-five percent of the study subjects were white. There were no significant differences between the PT and P0 groups in baseline age, sex, ethnicity, weight, BMI, systolic blood pressure (SBP), or diastolic blood pressure (DBP). P0 patients had a slightly higher HR than did PT patients (80.4 (8.9) vs. 76.0 (9.9), P < 0.05). Using the criteria of the Joint National Committee on Blood Pressure (JNC 7) only 14% of the patients had normal blood pressure upon entry into the practice, 54% were prehypertensive with initial blood pressures equal to or >120/80 or equal to or <139/89, and 32% were hypertensive (12). Seven percent of the patients had impaired fasting glucose at initial examination and 4% had type II diabetes.

Table 1.  Baseline demographic data
inline image

Statistical analysis

Weight, pulse, and blood pressure over time were analyzed using a mixed model with repeated measures. Statistical software was made by SAS Institute, Cary, NC. Non-normally distributed data like gender, ethnicity, and percentages in the blood pressure groups was analyzed using the χ2-test. Normally distributed data like baseline weight, age, pulse, BMI and blood pressure were analyzed using the Student t-test.


Patient program attendance

Near perfect program attendance during the initial 12 weeks was a requirement for inclusion. Practitioner notes for each visit were reviewed and these indicated that, at most encounters, patients were adhering to the diet, were taking their medications and were following behavior modification advice during the first 3 months. After 26 weeks attendance rate fell and drop out rate increased, especially in those patients whose weight loss was less than the average. Follow-up laboratory examinations were performed on selected patients, generally to follow initial abnormalities. Bicarbonate levels obtained during the first 6 months on the program were always >22 milli-equivalents per deciliter. Lipid parameters improved with weight loss as expected (13). Patient retention at 26, 52, and 104 weeks for PT patients was 81, 65, and 37%. Patient retention for P0 patients was significantly lower for the same periods at 61, 48, and 16%. These are not retention rates calculated from initial visit because patients who dropped out before 12 weeks were excluded.

Blood pressure change

Mean SBP and mean DBP fell in both the P0 group (Table 2 and Figure 1e) and in the PT group (Table 3 and Figure 1a) as the patients lost weight. In the PT group, SBP and DBP fell rapidly during the initial 4 weeks of weight loss during which time the average weight loss was 7.6%. Although the average PT subject regained some weight by year 3, the mean weight loss of PT subjects remained at about 10% from year 3 through year 7 and SBP in these patients remained significantly below base line at week 26 (P < 0.0001), week 52 (P < 0.0001), week 104 (P = 0.0001), and week 208 (P = 0.0186), below baseline but not significantly at week 156 (P = 0.1126) or weeks 260–364 (P = 0.1037–0.6207) whereas DBP in these patients remained significantly below baseline at weeks 26 (P < 0.0001), 52 (P < 0.0001), 104 (P < 0.0001), 156 (P < 0.0088), and 208 (P < 0.0314), below baseline but not significantly at weeks 260–364 (0.5622–0.9576). In the P0 group SBP and DBP continued to fall until maximum weight loss occurred at 40 weeks. Although the mean decrease in SBP and DBP was greater in the P0 group than the PT group the SBP decrease from baseline did not persist as the patients regained their weight reaching baseline values at 104 weeks, even though average weight loss was still −8.4% (Figure 1e).

Table 2.  Weight loss, blood pressure, and heart rate changes for phentermine-untreated patients (P0)
inline image
Figure 1.

Weight loss, systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate change by weeks/years of therapy. (a) All phentermine-treated subjects (weight P < 0.0001, SBP week 26, 52, 104 P ≤ 0.0001, and week 156–312 P > 0.05), DBP week 26, −156 P > 0.04, and week 260–312 P > 0.05, HR P > 0.05), (b) phentermine-treated hypertensive subjects: weight P ≤ 0.0001, SBP week 26, 52, 104, 156 P ≤ 0.0001, week 208 P = 0.0217, week 260 P = 0.0074, week 312 P = 0.0047, DBP week 26, 52, 104 P < 0.0001, week 156 P = 0.0002, week 208 P = 0.0041, 260 P > 0.05, week 312 P > 0.05, HR P > 0.05, (c) phentermine-treated prehypertensive subjects (weight P < 0.01, SBP week 25 and 52 P < 0.0001, week 104–312 P > 0.05, DBP P > 0.05, HR P > 0.05), (d) phentermine-treated normotensive subjects (SBP P > 0.05, DBP P > 0.05, HR P > 0.05), and (e) phentermine-untreated subjects. (Weight week 52 P < 0.0001, 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).

Table 3.  Weight loss, blood pressure, and heart rate changes for phentermine-treated patients (PT)
inline image

Stratification of the PT group by initial blood pressure category, revealed that PT hypertensive patients (Table 4 and Figure 1b) had greater declines in both SBP and DBP than prehypertensive patients (Table 5 and Figure 1c) and that PT patients with initial blood pressure in the normal range experienced no significant change in either SBP or DBP with weight loss (P > 0.05) (Table 6 and Figure 1d).

Table 4.  Weight loss, blood pressure, and heart rate changes for phentermine-treated patients with hypertension at baseline
inline image
Table 5.  Weight loss, blood pressure, and heart rate changes for phentermine-treated patients with prehypertension at baseline
inline image
Table 6.  Weight loss, blood pressure, and heart rate changes for phentermine-treated patients with normal blood pressure at baseline
inline image

HR change

Mean HR, decreased slightly but not significantly as patients lost weight in the P0 cohort but HR either did not change or increased slightly in the PT cohorts (P > 0.05).

Categorical blood pressure changes

As patients lost weight and blood pressure declined, some patients who were initially in the prehypertensive category with blood pressures >119/79 but <140/90, experienced a shift to the normal blood pressure category. At 52 weeks 43 of 82 (52%) previously prehypertensive PT patients and 6 of 9 (67%) previously prehypertensive P0 patients had shifted to the normal blood pressure category. At 104 weeks of 27 of 55 (49%) previously prehypertensive PT patients remained in the normal blood pressure category, as did 12 of 34 (35%) at 156 weeks, 12 of 23 (52%) at 208 weeks, and 6 of 16 (38%) at 260 weeks. The numbers of P0 patients beyond 52 weeks were too low for meaningful comparison.

Weight loss and maintenance

Mean percentage weight loss for P0 and all PT patients is depicted in Figure 1a,e and Tables 2 and 3. Percentage weight loss examined as treatment effect by week over 2 years in the entire data set showed significantly greater in the PT cohort (P = 0.0144). Some patients in the PT cohort continued to maintain a >10% weight loss for as long as 8 years. There were no significant differences in weight loss between the patients in the various blood pressure category cohorts.

Categorical weight loss

Among the 175 PT patients with data for 1 year, 97% lost 5% or more of their initial weight at 1 year. In the same 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, and 3% lost 30% or more. Among the 15 P0 patients with data for 1 year 80% lost at least 5%, 73% lost 10%, 47% lost 15%, 33 lost 20%, 20% lost 25%, 13% lost 30%, and 7% lost 35%. Figures showing individual patient weight changes from baseline to 1 year are included as Supplementary Data online.

Phentermine dose

Phentermine dose data is shown in Table 7. The initial (T0) phentermine dose ranged from 15 to 37.5 mg/day with a mean dose of 33.6. The modal dose was 37.5 mg from T0 through year 6, but as the dose was increased in some individuals, the mean dose for the PT cohort slowly increased, reaching 57 mg/day in year 7. The modal dose increased to 60 mg/day at year 7.

Table 7.  Phentermine doses in mg/day
inline image

The patients in this study described experiencing two distinctly different phentermine therapeutic effects, which they believed contributed to their weight loss. The first and most obvious was appetite and hunger suppression. The second effect, usually not noticed until a patient had been on phentermine for several months, was more subtle and variously described as improved or stronger control of eating, diminution or absence of food cravings, or improved ability to follow their eating plan. Appetite suppression tended to fade with time in many patients but improved control of eating persisted and was still present after patients had been taking phentermine for many months or years. Patients experienced less hunger or even no hunger when started on phentermine. This initial effect on hunger tended to diminish with time on any given dose but could be re-established in some patients by increases in phentermine dose. If phentermine was discontinued, hunger and loss of eating control quickly reappeared. Patients who took phentermine long-term to assist with weight maintenance typically described noticing no effect day-by-day unless they skipped a day or more whereupon they noticed increased hunger and that food servings grew larger. They also described a diminution of eating self-discipline. Some patients also described a return of carbohydrate cravings. Carbohydrate cravings typically disappeared quickly on the low-carbohydrate ketogenic diet, but later during maintenance as carbohydrate intake was increased, some patients noted recurrence of carbohydrate cravings. For some patients the beneficial effects of phentermine appeared to be preserved by increases in doses using dose-to-effect titration.

To examine for the possibility that higher doses of phentermine might adversely affect blood pressure or HR, the data for all PT subjects segregated by phentermine dose levels was examined at 52, 104, and 156 weeks. There were no significant differences in SBP, DBP, or HR comparing patient cohorts at the five dose levels 0, 18.75, 30–37.5, 48.75–56.25, and 60–75 mg phentermine/day except for a significant drop in DBP in the phentermine 18.75 group compared to no phentermine (P < 0.05). Although the differences in weight loss at the various doses were not demonstrably statistically significant because of small sample sizes, the mean weight loss tended to be greater with increased daily doses at 2 and 3 years. At 2 years the mean weight loss for no phentermine and for 0.5× standard dose was about 7%, for standard dose 11%, and for 1.5× and 2× standard dose 14%. At 3 years weight loss for no phentermine was 6%, for 1 and 1.5× standard dose about 10%, and for 2× standard dose 12%.

Adverse events

To rule out the possibility of any increases in individual patient blood pressure values, the data was examined carefully for occurrence of phentermine-induced blood pressure elevations or phentermine-induced hypertension. One patient in the study (subject #56), who initially had prehypertension, developed hypertension after having been on phentermine for 1 year. He lost 15% of initial weight but was still overweight. His hypertension was easily controlled with a low dose of lisinopril, and he continued on phentermine. This was the only instance found where a patient developed hypertension after being on phentermine continuously.

Some of the hypertensive patients in the study had occasional transient elevations in either SBP or DBP. These were most often caused by omission of antihypertensive medications or by weight gain. In these cases phentermine was withheld until blood pressure was below 140/90. In the one case where BP did not drop promptly, it did so when an antihypertensive agent was added. Generally the hypertensive study patients on phentermine found their hypertension more easily managed with either lower doses of medicines or no medicine.

There were three PT patients who were initially normotensive, who did lose weight, but who subsequently developed prehypertension. One patient (#53) developed prehypertension then was diagnosed with a benign brain tumor, which was successfully removed. She had postoperative seizures and is currently on topiramate and not on phentermine. She is still prehypertensive. Another patient (#148) developed prehypertension (BP 122/80) at week 2, and subsequently had intermittent SBPs in the low 120s. She dropped out, went off phentermine, then returned eighteen months later with weight gain and hypertension.

Patients on phentermine were routinely asked about any adverse phentermine effects at every encounter. The most common adverse effects were dry mouth and insomnia. Insomnia occurred most often in the first few days of phentermine therapy and typically disappeared within a week or two. Some patients experienced insomnia even after being on phentermine for years if they took phentermine after midday, but not if they took phentermine earlier. Dry mouth was mild in most patients and typically either disappeared within a few weeks or became tolerable provided the patient drank 64 ounces of water daily. No other adverse effects were observed in the study group.

Social or work-related dysfunction and drug seeking behavior were carefully watched for and were not observed in the study group.


The prevalence of high blood pressure and mean levels of SBP and DBP are known to increase as BMI increases (14). Multiple studies have shown that weight loss induced by caloric restriction and behavior change without antiobesity pharmacotherapy is associated with decreased blood pressure (15).

In this study, we found that SBP and DBP declined significantly during weight loss in patients treated with phentermine, just as did blood pressures in the P0 patients. Patients with hypertension experienced the greatest declines while those with prehypertension had lesser declines and those with normal blood pressure had the least change. That weight loss induced by two-drug combinations including phentermine produced greater declines in patients with hypertension than in those with normal blood pressure has been reported previously (16,17), but these reports did not stratify patients by JNC 7 blood pressure categories so blood pressure declines in prehypertensive patients were not analyzed. In this study we found that a significant proportion of both PT and P0 patients experienced a categorical shift from prehypertension to normotension at 12, 26, and 52 weeks. This categorical shift typically occurred in the first 12 weeks of therapy. Although the numbers of P0 patients was too low for statistical comparisons beyond 52 weeks, the categorical shift persisted in smaller numbers of PT patients at 104 and 156 weeks.

The data from this study suggests that the occurrence of hypertension associated with phentermine pharmacotherapy is rare rather than common. One patient, initially prehypertensive, of the 269 PT patients developed hypertension after 1 year on phentermine. Excluding the patients initially hypertensive, data for 119 patients remained in the study at 52 weeks for an incidence rate of 0.84%/year. If only the prehypertensive patients are considered at 1 year (N = 95), then the 1-year incidence would be 1.05%. In considering these incidence rates comparison should be made to other data on the incidence of new hypertension. Because elevations of blood pressure in the hypertensive range (≥140/90) and in the prehypertensive range (120/80–139/89) are very common in obese subjects, one could expect an annual incidence rate of newly diagnosed hypertension in untreated normotensive and prehypertensive obese subjects followed for a year to be higher than 1%. The ATTICA study (18) found an annual incidence of new hypertension in the general Greek population to be 2.86 cases per 100 in men and 2.68 cases per 100 in women. The Strong Heart Study (19) found that the 4-year incidence of hypertension in a group of patients with prehypertension to be 38% and a 5-year study in Taiwan (20) found a conversion rate from prehypertension to hypertension to be 31.2%. These incidence rates, in populations not selected for obesity, are substantially higher than the rate of conversion of prehypertension to hypertension calculated for this study. These data suggest that weight loss assisted by phentermine pharmacotherapy retarded progression of prehypertension to hypertension. Data from the fixed-dose combination of phentermine-topiramate (Qnexa) clinical trials also suggested progression to hypertension was retarded (17).

Long-term nonpharmacological intervention trials have been unsuccessful in maintaining weight loss and lower blood pressure for longer than 3–4 years (21). The number of patients retained beyond 4 years in this study is small, but the data suggests the possibility that long-term phentermine therapy could be effective in maintaining a lower weight and in lowering blood pressure for >4 years in some patients.

Given the data from this study and the previous studies cited suggesting that blood pressure declines rather than increases with phentermine therapy, why does the assumption that phentermine adversely affect blood pressure persist? The conception that phentermine can elevate blood pressure and HR apparently derives from the often-repeated fact that phentermine is an amphetamine congener. That amphetamine can increase blood pressure and HR and induce or worsen hypertension is well documented (22). These adverse cardiovascular effects are dose-related and dependent upon the route of administration of the drug; inhalation or drug administration via intranasal or intravenous routes rapidly produces higher blood levels and therefore greater toxic effect than dose oral administration. Certain amphetamine congeners such as methamphetamine (23) and 3,4-methylenedioxymethamphetamine (24), which are typically abused by intranasal, inhalation, or intravenous routes, have even greater toxic potential and are also known to produce adverse cardiovascular effects. Subjects who use the amphetamines of abuse often self-administer large doses producing very high blood levels in trying to achieve the desired stimulant effect. On the other hand, the route of administration and dose of amphetamines used for pharmacotherapy of attention deficit and attention deficit hyperactivity disorder have a much lower toxicity potential. Therapeutic use of these drugs is associated with slight increases in blood pressure and in HR of questionable clinical impact (25). In phentermine pharmacotherapy low (compared to the doses used of the amphetamines of abuse) oral doses are used, and phentermine blood levels achieved are comparatively lower than blood levels achieved by subjects of the abused drugs. In addition phentermine has a substantially lower toxic potential. Thus the potential for adverse effects with phentermine pharmacotherapy is significantly lower than for the abused amphetamines.

Aside from a few anecdotal reports, there is no data in the peer-reviewed medical literature to support the perception that phentermine increases blood pressure or HR. Phentermine clinical trial reports, when details of blood pressure and HR have been given, have shown that trial subjects experienced decreases in SBP and DBP and minimal or no changes in HR (8,9,26). The same is true of most trials with combinations with phentermine as one of two antiobesity drugs. (16,17,27,28). The one exception to the latter statement is a recent clinical trial in which patients given a combination of pramlintide plus phentermine showed no significant change in clinic visit SBP and DBP but an increase in HR of 5 beats/min (P < 0.01) (29). Ambulatory blood pressure monitoring in the latter study revealed about a 3 mm increase in DBP in the pramlintide plus phentermine subjects (P < 0.01) and an increase in HR of about 5 beats/min (P < 0.001). Because SBP and HR decreased slightly in the pramlintide only group, the implication is that the phentermine produced the observed increases, however no subjects treated with phentermine alone were included in the study and patients with hypertension were excluded. We did not measure ambulatory blood pressures in this study.

Our data suggests that phentermine pharmacotherapy does not significantly influence HR at any time during treatment.

In adjustment of phentermine dose using a dose-to-effect method, previously described by Rothman (30), the desired effects are on eating behaviors rather than weight loss (31). Dose-to-effect titration is a common practice in medicine that is employed with a wide variety of drugs. The most cogent comparison here would be with the methods used in adjusting amphetamine dose for attention deficit disorder and attention deficit hyperactivity disorder. Two approaches have been described in treating ADD and ADHD: (i) prescribe an initial low dose, evaluate behaviors, then gradually increase dose until behavior improves, or (ii) ramp up dose until undesirable side effects appear, then reduce dose to the level before they appeared (32). The first approach was typically employed in adjusting phentermine doses in this study. Phentermine dose-to-effect was evaluated at each patient encounter and if undesirable side effects appeared the dose was most often reduced. Mean and median phentermine dosage trended upward for patients in long-term treatment as shown in Table 3. Patients with adult attention deficit disorder, which is common among the obese (33), often derive behavioral benefit from phentermine (34) and may tolerate higher doses. A few of the patients on higher doses in this study were thought to have previously undiagnosed adult ADD, but the majority were simply more comfortable with their eating behavior at a higher dose. Phentermine doses higher than the US Food and Drug Administration recommended maximum of 37.5 mg daily have been previously described (2,7,30,35), but there have been no clinical trials that have examined this issue.

Phentermine is presumed to have addiction potential and is classified as a category IV controlled substance. An important criterion is making a diagnosis of addiction is that drug-addicted subjects typically manifest dysfunction in one or several spheres such as social, work, or legal arenas. Practitioners and staff of the clinic observed the patients in this study carefully for signs of such dysfunction, and no such behaviors were observed.

This study has several strengths. The study clinic is well established with many long-term patients. The physician in charge has had long experience with phentermine. As is always with retrospective observational studies, a limitation could have been bias in subject selection, which is possible even though criteria were set in place to avoid bias. The methods for blood pressure and especially for HR determination are another limitation. The small sample size in some of the cohorts is yet another limitation.

We conclude the data from this retrospective study suggests that the addition of phentermine pharmacotherapy for patients in a comprehensive weight management program does not adversely affect HR or blood pressure. Indeed, this study suggests that in this setting phentermine pharmacotherapy, can be especially beneficial for the obese hypertensive and prehypertensive patients since persistent weight loss assisted with such therapy may lower blood pressure long-term and retard the progression of prehypertension to hypertension delaying the natural course of hypertension in the obese and potentially reducing the risk of mortality due to congestive heart failure, stroke and ischemic heart disease. Some of these patients maintained a weight loss of 10% or greater for as long as 8 years suggesting long-term phentermine pharmacotherapy may be a useful treatment for long-term weight maintenance.

The P0 patients who adhered to diet alone without phentermine treatment garnered similar benefits to PT patients over the first year of treatment, but the benefits for the P0 patients began to vanish as they regained weight whereas PT patients' benefits persisted. PT patients also had slightly better initial weight loss. Finally, there is a suggestion that phentermine doses >37.5 mg/day are safe and can be useful in selected patients.

The potential benefit to the obese population is huge if the findings of this study can be confirmed and phentermine pharmacotherapy ultimately is used more widely. We suggest there is a need for more investigations into both the clinical effects of phentermine and its molecular basis of action. Development of psychometric scales to assess phentermine effect on eating behavior for guidance of dose-to-effect titration could prove clinically useful. We suggest a long-term, dose-ranging phentermine clinical trial and outcome study could potentially confirm the cardiovascular benefits implied by this study.


Supplementary material is linked to the online version of the paper at http:www.nature.comoby


The authors declared no conflict of interest.