Phentermine is the oldest weight drug still in wide use, it costs a fraction of a weekly injection, and it is the one a primary care office is most likely to reach for first. The fixed-dose combination of phentermine with topiramate, sold as Qsymia, is a different product with a different label and a much larger effect. Both get compared to the incretin class constantly, and the comparison has a shape that a single quoted number cannot hold — as with the metformin comparison, the answer depends on when the question was asked.
Two different drugs, and only one of them may be taken indefinitely
Phentermine hydrochloride is a Schedule IV sympathomimetic amine. Its label indicates it as a short-term (a few weeks) adjunct to a weight-reduction regimen, and describes “the limited usefulness of agents of this class.” It is contraindicated in people with a history of cardiovascular disease — coronary artery disease, stroke, arrhythmias, congestive heart failure, uncontrolled hypertension — and in hyperthyroidism, glaucoma, agitated states, a history of drug abuse and pregnancy. The usual adult dose is 37.5 mg once daily before or shortly after breakfast.
Qsymia’s label says the opposite about duration. It is indicated to reduce excess body weight and maintain weight reduction long term, in adults and in patients aged 12 and over with obesity, or adults with overweight plus a weight-related condition. The phrase “duration of use” appears nowhere in it. In exchange, the label carries a Risk Evaluation and Mitigation Strategy restricting distribution to certified pharmacies, a recommendation for a negative pregnancy test before starting and monthly during treatment, a required blood chemistry panel before and during use, and an instruction to taper the top dose to every other day for at least a week before stopping, because abrupt topiramate withdrawal has caused seizures in people with no seizure history.
Its stopping rule is two rules. After 12 weeks at 7.5/46, an adult who has not lost at least 3% of baseline body weight escalates to 11.25/69 and then 15/92. After a further 12 weeks at 15/92, an adult who has not lost at least 5% discontinues. That is roughly 26 weeks to a stop decision, which is a longer runway than the equivalent checkpoints described in the titration article.
What the combination actually produced
CONQUER randomized 2,487 adults with a body-mass index of 27 to 45 and two or more weight-related conditions to placebo, 7.5/46 or 15/92 for 56 weeks. Least-squares mean weight change was −9.8% (95% CI, −10.4 to −9.3) at the top dose and −7.8% (95% CI, −8.5 to −7.1) at the middle dose, against −1.2% (95% CI, −1.8 to −0.7) on placebo. A loss of 5% or more was reached by 70% at 15/92 (OR 9.0; 95% CI, 7.3 to 11.1) and 21% on placebo.[1]
EQUIP tested the same combination in severe obesity, at a body-mass index of 35 or higher, and reported −10.9% at 15/92 against −1.6% on placebo over 56 weeks.[2] SEQUEL extended CONQUER by another 52 weeks: at 108 weeks, weight change was −10.5% at 15/92, −9.3% at 7.5/46 and −1.8% on placebo.[3] That is two full years of randomized, placebo-controlled data, which is longer than most of the incretin obesity trials have run.
The attrition belongs next to those figures. In EQUIP, the study dropout rate was 47.1% on placebo, 39.0% at 3.75/23 and 33.6% at 15/92 by 56 weeks.[2] SEQUEL enrolled only 676 of the 866 eligible participants into its extension, so the 108-week result describes roughly a quarter of the people CONQUER randomized.[3] Adverse events at the top dose ran to paresthesia in 21%, dry mouth in 21%, dysgeusia in 10%, and depression-related events in 7% against 4% on placebo, with anxiety-related events in 8% against 3%.[1]
The rank that inverted
A network meta-analysis republished in 2024 pooled 132 randomized trials and 48,209 participants and put phentermine-topiramate at the top of the field: mean difference in percentage body weight of −7.98% against lifestyle modification alone (95% CI, −9.27 to −6.69), with an odds ratio of 8.02 (95% CI, 5.24 to 12.27) for reaching a 5% loss. The GLP-1 receptor agonists as a class came second at −5.79% (95% CI, −6.34 to −5.25). A post-hoc split put semaglutide alone at −11.40% (95% CI, −12.51 to −10.29).[4]
A 2026 analysis asked the same question over 262 trials and 99,791 participants, searched through November 2025. Among the agents carrying moderate-to-high certainty at one year, the order is tirzepatide −14.9% (95% CI, −16.0 to −13.9), cagrilintide-semaglutide −14.8%, oral semaglutide −10.9%, orforglipron −9.9%, subcutaneous semaglutide −9.8%, and phentermine-topiramate −8.1% (95% CI, −9.7 to −6.5) — last of the six.[5]
Read the two estimates for the drug itself: −7.98% and −8.1%. Nothing happened to phentermine-topiramate. What happened is that five agents that did not have mature evidence in 2021 now do, and the combination that was once the most effective option available is now the weakest of the well-evidenced ones. A page still quoting the 2024 sentence is quoting a correctly reported result about a field that has been replaced. The numbers those newer agents earned in their own trials are set out in the semaglutide article and the tirzepatide article.
The guideline changes its answer at a body-mass index of 30
The American College of Physicians living guideline of April 2026 makes two recommendations. For nonpregnant adults with obesity, meaning a body-mass index of 30 or higher, semaglutide and tirzepatide are first-line on moderate-certainty evidence, and phentermine-topiramate is second-line on low-certainty evidence, ahead of liraglutide and naltrexone-bupropion.[6]
For adults with overweight — a body-mass index of 27 to under 30 — plus type 2 diabetes, dyslipidemia, hypertension, obstructive sleep apnea or cardiovascular disease, the recommended list is semaglutide and tirzepatide first-line and liraglutide second-line. Phentermine-topiramate is not on that list at all. Phentermine by itself appears at no line in either recommendation.[6] The drug ranked second in one population is absent from the next population down, which is a sharper fact than “second-line” on its own conveys.
One clause in that guideline needs separating from the labels. Its counseling text names a contraindication to phentermine-topiramate in people with cardiovascular disease.[6] The current Qsymia prescribing information lists five contraindications — pregnancy, glaucoma, hyperthyroidism, recent monoamine oxidase inhibitor use, and hypersensitivity — and cardiovascular disease is not among them. Phentermine’s own label does contraindicate a cardiovascular history. The caution is real for the single agent; for the combination it is a warning about heart rate rather than a contraindication, and the label quantifies it: 24-hour average heart rate rose 3.6 beats per minute (95% CI, 2.1 to 5.2) against placebo in an eight-week ambulatory monitoring study.
Nobody has run the head-to-head
No randomized trial has assigned adults to phentermine, or to phentermine-topiramate, or to a GLP-1 receptor agonist, with weight as the endpoint. The living systematic review underpinning the 2026 guideline covered 69 studies and 112,511 participants across twelve drugs and states its own limitation plainly: direct head-to-head comparisons of different treatments were limited.[7] Every figure that puts the two side by side above comes from separate trials with separate populations, pooled indirectly.
The evidence base is also thinner on one side than it looks. A 2025 systematic review of 56 obesity pharmacotherapy trials counted two phentermine-topiramate trials against fourteen for semaglutide, eleven for liraglutide and six for tirzepatide.[8] Those two trials are CONQUER and EQUIP. That is why every guideline rates the combination at low certainty while rating the incretins at moderate, and it is a fact about the quantity of evidence rather than about the drug.
What people actually do with the prescription
A claims analysis of 26,522 adults newly prescribed a long-term weight medication found 6-month persistence of 27.3% on phentermine-topiramate against 41.8% on liraglutide 3.0 mg, with an adjusted hazard ratio for discontinuation of 0.64 on liraglutide (p < 0.0001).[9] A 2026 chart review of psychiatric outpatients across four weight drugs found early discontinuation at twelve weeks differed by drug (p = 0.017), running highest on phentermine-topiramate at 72.2% and lowest on metformin at 39.1%.[10]
Neither study is a trial, and the persistence analysis was authored by employees of the company that markets liraglutide. But the direction is the same in both, and it matters more than the effect sizes do: a drug stopped at week twelve produces the week-twelve result, not the 56-week one. The same arithmetic applies across this class, which is why the side-effect profile decides more outcomes than potency does.
Where phentermine wins, and it is not on weight
A 40-year Markov model comparing five weight medications from a United States payer perspective found naltrexone-bupropion and phentermine plus topiramate the least costly, with all five agents separated by no more than 0.5 quality-adjusted life-years. Tirzepatide’s incremental cost-effectiveness ratio came to $355,616 per quality-adjusted life-year, and liraglutide and semaglutide were dominated by phentermine plus topiramate — more expensive without a compensating gain in the model’s own currency.[11] That is a modeled result rather than an observed one, and it is the strongest published case for the older drug. What an actual year on a cash plan costs is worked through in the annual cost article, with a version you can run in the cost calculator.
The trial phentermine has never had
Phentermine alone reached the market before long-term randomized evidence was required, and it has never been tested that way. A 24-month, double-blind, placebo-controlled trial of phentermine 24 mg daily against placebo, enrolling 870 participants with a body-mass index of 27 to 44.9 across six sites, is underway with 24-month percent weight loss, change in systolic blood pressure and adverse event rates as its three primary outcomes. Its own rationale states that phentermine’s long-term efficacy and safety have never been evaluated in a randomized trial because its market approval predates such requirements.[12] It has not reported.
The nearest existing evidence is observational. An electronic health record cohort of 13,972 adults starting phentermine between 2010 and 2015 found that people using it continuously for more than twelve months lost 7.4% more weight at 24 months than those using it three months or less (p < 0.001), with the composite of cardiovascular disease or death occurring in 0.3% — 41 events — and no significant difference in hazard between duration groups.[13] People who stay on a drug for a year are not the same people as those who stop at three months, so that comparison describes a selected group rather than an effect of duration. Whether anyone should be taking it for a year is a prescriber’s question, and the general version of it is covered in the contraindications article.
What this comparison settles
On magnitude, the incretins win and the gap is roughly two to one at the top doses. On price, phentermine wins and the gap is larger than that. On duration, the two products called “phentermine” part company completely: one is labeled for a few weeks, the other for indefinite use under a pregnancy-testing program. On evidence, the combination has two trials and the incretins have dozens. And on the question anyone actually wants answered — which of these works better for a particular person — no randomized comparison exists to answer it.
One limit applies to every incretin figure above. They came from trials of FDA-approved products at labeled doses. Compounded semaglutide and tirzepatide are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before dispensing, so a compounded vial priced closer to a phentermine prescription is not thereby carrying the trial result with it. How the figures on this site are established before publication is set out in the methodology.