--- topic: "Pharmaceutical Side Effects" title: "Safely Reducing or Stopping a Medication" order: 4 description: "Which medicines can be tapered with a prescriber, which are dangerous to stop, what withdrawal and rebound look like, and how to plan a safe review." question: "Can you safely reduce or stop a medication, and how?" answer: "Some medicines can likely be reduced safely, but only with your prescriber; stopping suddenly can cause withdrawal, rebound or relapse." actions: - "Never stop or cut a prescribed medicine alone: book a review with your list, supplements included, and ask what each is for, if still needed and the taper plan [6]." - "Ask for a written, gradual plan with follow-up dates: guidelines advise step-down or on-demand acid reducers [2] and slow benzodiazepine tapers [3]; experts propose months-long SSRI tapers [4]." - "Add non-drug support where it exists, such as behavioral sleep advice during a sleeping-pill taper; people do better when they understand why and help build the plan [3]." cautions: - "Some medicines are dangerous to stop unsupervised: after heart attack, stopping aspirin, beta-blocker and statin together carried 3.8 times one-year death risk [1]; stopping steroids can cause adrenal insufficiency [9]." - "Antidepressants: in the ANTLER trial, 56 percent who stopped relapsed within a year versus 39 percent who stayed on, so plan tapers with your prescriber and watch for relapse [10]." - "Pregnant or planning pregnancy? Some medicines are riskier to continue and others riskier to stop, so decide with your prescriber rather than stopping on your own." confidence: 3 confidenceNote: "Guidelines and trials support supervised tapering of acid reducers, sleeping pills and some blood-pressure drugs and show dangers of stopping alone, but taper schedules, especially antidepressants', rest on limited evidence." evidenceSummary: rctCount: 3 observationalCount: 1 metaAnalysisCount: 4 totalParticipants: "34,143 older adults in the deprescribing review; 21,002 in the antidepressant analysis" replicated: true newestStudyYear: 2024 references: - id: "1" pmid: "17000940" title: "Impact of medication therapy discontinuation on mortality after myocardial infarction" authors: "Ho PM, Spertus JA, Masoudi FA, et al." year: 2006 journal: "Arch Intern Med" - id: "2" pmid: "28500192" title: "Deprescribing proton pump inhibitors: Evidence-based clinical practice guideline" authors: "Farrell B, Pottie K, Thompson W, et al." year: 2017 journal: "Can Fam Physician" - id: "3" pmid: "29760253" title: "Deprescribing benzodiazepine receptor agonists: Evidence-based clinical practice guideline" authors: "Pottie K, Thompson W, Davies S, et al." year: 2018 journal: "Can Fam Physician" - id: "4" pmid: "30850328" title: "Tapering of SSRI treatment to mitigate withdrawal symptoms" authors: "Horowitz MA, Taylor D" year: 2019 journal: "Lancet Psychiatry" - id: "5" pmid: "27077231" title: "The feasibility and effect of deprescribing in older adults on mortality and health: a systematic review and meta-analysis" authors: "Page AT, Clifford RM, Potter K, et al." year: 2016 journal: "Br J Clin Pharmacol" - id: "6" pmid: "25798731" title: "Reducing inappropriate polypharmacy: the process of deprescribing" authors: "Scott IA, Hilmer SN, Reeve E, et al." year: 2015 journal: "JAMA Intern Med" - id: "7" pmid: "19362552" title: "Proton-pump inhibitor therapy induces acid-related symptoms in healthy volunteers after withdrawal of therapy" authors: "Reimer C, Søndergaard B, Hilsted L, et al." year: 2009 journal: "Gastroenterology" - id: "8" pmid: "32453368" title: "Effect of Antihypertensive Medication Reduction vs Usual Care on Short-term Blood Pressure Control in Patients With Hypertension Aged 80 Years and Older: The OPTIMISE Randomized Clinical Trial" authors: "Sheppard JP, Burt J, Lown M, et al." year: 2020 journal: "JAMA" - id: "9" pmid: "25844620" title: "Adrenal Insufficiency in Corticosteroids Use: Systematic Review and Meta-Analysis" authors: "Broersen LH, Pereira AM, Jørgensen JO, et al." year: 2015 journal: "J Clin Endocrinol Metab" - id: "10" pmid: "34587384" title: "Maintenance or Discontinuation of Antidepressants in Primary Care" authors: "Lewis G, Marston L, Duffy L, et al." year: 2021 journal: "N Engl J Med" - id: "11" pmid: "38851198" title: "Incidence of antidepressant discontinuation symptoms: a systematic review and meta-analysis" authors: "Henssler J, Schmidt Y, Schmidt U, et al." year: 2024 journal: "Lancet Psychiatry" - id: "12" pmid: "30292574" title: "A systematic review into the incidence, severity and duration of antidepressant withdrawal effects: Are guidelines evidence-based?" authors: "Davies J, Read J" year: 2019 journal: "Addict Behav" --- Some medicines can be reduced or stopped safely, but only with your prescriber, because stopping the wrong one, or stopping suddenly, can be dangerous. After a heart attack, people who stopped aspirin, a beta-blocker and a statin within a month had about 3.8 times the one-year risk of death of those who stayed on at least one [1]. Acid reducers, sleeping pills and antidepressants each call for a planned, usually gradual, change [2][3][4]. Trials support tailored, prescriber-led reviews [5][6]. ## How it works ### Deprescribing is a studied practice Deprescribing means tapering or stopping drugs that no longer help, to cut the harm of taking too many. A five-step process lists every drug and why it is taken, weighs benefit against harm, targets drugs with the lowest benefit-harm ratio and least risk of withdrawal, then monitors the change [6]. A review of 132 studies (34,143 older adults) found no clear fall in deaths across randomized trials (odds ratio 0.82) but a fall when deprescribing was tailored to the patient (0.62) [5]. See [informed consent and being your own advocate](/pharmaceutical-side-effects/informed-consent). ### Where reduction has worked - **Acid reducers (PPIs).** Guidelines advise reducing the dose, stopping or using them on demand once heartburn has settled, but not after Barrett esophagus or severe esophagitis [2]. Rebound heartburn can occur: after 8 weeks of esomeprazole, 44 percent of healthy volunteers had acid symptoms after stopping, against 15 percent on placebo [7]. See [acid reflux](/acid-reflux/natural-management). - **Sleeping pills (benzodiazepines and "Z-drugs").** Slow tapering is advised for adults 65 and older, and for adults 18 to 64 who have used them for over 4 weeks [3]. - **Blood-pressure pills in the very old.** In OPTIMISE, 569 adults aged 80 or older, with systolic pressure under 150 mm Hg on two or more drugs, had one drug removed or kept their regimen. At 12 weeks, 86.4 versus 87.7 percent still had pressure under 150 mm Hg [8]. ### Where stopping is risky - **Heart medicines.** Stopping all three of aspirin, a beta-blocker and a statin early after a heart attack went with a higher death rate [1]. - **Steroid tablets.** Adrenal insufficiency after stopping glucocorticoids is common, and no dose, duration or route excludes it with certainty [9]. - **Antidepressants.** Two risks overlap: relapse of depression (56 percent after stopping versus 39 percent on staying in ANTLER) [10] and withdrawal symptoms. Estimates of withdrawal differ: about 31 percent after stopping an antidepressant versus 17 percent after stopping a placebo in a 2024 meta-analysis [11], against an average of 56 percent in a 2019 review of varied studies [12]. Short two-to-four-week tapers showed little advantage over stopping abruptly; slow tapers over months to very low doses have been proposed [4]. ### How to do it safely Book a review, bring every medicine and supplement, and agree a written taper with dose steps, timing and a follow-up date. Treat new symptoms as information: withdrawal, rebound and relapse can feel alike, and only your prescriber can sort them out. See also [which medications deplete nutrients](/medication-nutrient-depletion/which-medications-deplete-nutrients). ## The evidence in detail ### Does deprescribing help? Page and colleagues pooled 132 papers (34,143 participants, mean age 73.8). Non-randomized studies showed lower mortality with deprescribing (odds ratio 0.32, 95% CI 0.17 to 0.60), but randomized studies did not (0.82, 0.61 to 1.11). Patient-specific interventions did lower mortality (0.62, 0.43 to 0.88), while generalized education programs did not (1.21, 0.86 to 1.69) [5]. Scott and colleagues note that the number of prescribed drugs is the single strongest predictor of inappropriate prescribing and adverse drug events in older patients [6]. In OPTIMISE, 569 patients (mean age 84.8) were randomized to remove one blood-pressure drug or continue usual care. Pressure stayed under 150 mm Hg in 86.4 versus 87.7 percent (adjusted relative risk 0.98), meeting the non-inferiority margin of 0.90; mean systolic pressure was 3.4 mm Hg higher after reduction, and serious adverse events were 4.3 versus 2.4 percent (relative risk 1.72, 0.7 to 4.3). The trial lasted only 12 weeks, so long-term outcomes are unknown [8]. ### Acid reducers and sleeping pills Reimer's trial gave 120 healthy volunteers 12 weeks of placebo, or 8 weeks of esomeprazole 40 mg a day followed by 4 weeks of placebo. In weeks 9 to 12, 44 percent (26 of 59) of the esomeprazole group reported a relevant acid-related symptom, against 15 percent (9 of 59) on placebo, a sign of rebound acid hypersecretion [7]. The PPI guideline recommends deprescribing in adults who have completed at least 4 weeks of treatment for heartburn or mild-to-moderate reflux disease or esophagitis and whose symptoms have resolved; it does not apply to Barrett esophagus, severe grade C or D esophagitis, or a history of bleeding ulcers [2]. The benzodiazepine-receptor-agonist guideline recommends offering a slow taper and found that tapering improved cessation rates compared with usual care without serious harms; it does not apply to untreated anxiety, depression or other sleep disorders [3]. ### When stopping is dangerous In a registry of 1,521 people discharged after a heart attack on aspirin, a beta-blocker and a statin, 184 stopped all three by one month, 56 stopped two, 272 stopped one and 1,009 continued all three. One-year survival was 88.5 versus 97.7 percent for those who stopped all three versus those who kept at least one (hazard ratio 3.81, 1.88 to 7.72), and results were consistent drug by drug. It was observational, so people who stop may differ in other ways [1]. A meta-analysis of 74 articles (3,753 participants) found adrenal insufficiency after glucocorticoids in 4.2 percent with nasal use to 52.2 percent with intra-articular use, rising with higher dose and longer duration; testing should be considered in people with non-specific symptoms after stopping [9]. ### Antidepressants: relapse and withdrawal In ANTLER, 478 UK primary-care patients who felt well enough to consider stopping (a history of at least two depressive episodes, or two or more years on an antidepressant) were randomized to maintain citalopram, fluoxetine, sertraline or mirtazapine or to taper to matching placebo. By 52 weeks, relapse occurred in 92 of 238 (39 percent) who stayed on versus 135 of 240 (56 percent) who stopped (hazard ratio 2.06, 1.56 to 2.70), and the stopping group had more depressive, anxiety and withdrawal symptoms. Adherence to the assigned plan was 70 versus 52 percent [10]. A 2024 meta-analysis of 79 studies (21,002 patients) found at least one discontinuation symptom in 31 percent (27 to 35) after stopping an antidepressant and 17 percent (14 to 21) after stopping placebo, a difference of 8 points in randomized trials; severe symptoms occurred in 2.8 versus 0.6 percent. It put the excess incidence at about 15 percent, one in six or seven, with substantial heterogeneity [11]. The 2019 review of 24 studies of varied design reported incidence from 27 to 86 percent (average 56 percent), with many people affected for over two weeks and some for months [12]. A review proposes tapering SSRIs hyperbolically over months, based on imaging of serotonin-transporter occupancy and limited trial data, rather than a randomized comparison of schedules [4]. ### Where the evidence conflicts The two withdrawal estimates differ in method: the 2024 analysis allows for symptoms that also appear after stopping placebo, while the 2019 review pooled figures from studies of diverse design [11][12]. Mortality benefits of deprescribing appear in non-randomized work but are not confirmed in randomized trials overall [5]. Taper schedules for antidepressants rest on pharmacology and clinical experience more than on head-to-head trials [4]. ### Bottom line Confidence is moderate (3 of 5). Supervised dose reduction is supported for acid reducers, sleeping pills and some blood-pressure regimens in selected people [2][3][8]. Stopping heart medicines or steroids abruptly, or stopping antidepressants too soon, carries real risks [1][9][10]. The safest path is a plan made with your prescriber.