Polypharmacy and Deprescribing Viva

Long Case Viva Defence

Candidate's opening statement (model answer)

"Mrs Eleanor Davies is an 84-year-old retired schoolteacher who has been admitted after a fall at home in which she sustained a minor head laceration but no fracture. She has atrial fibrillation, ischaemic heart disease, heart failure with reduced ejection fraction, hypertension, type 2 diabetes, osteoporosis, overactive bladder, insomnia, and depression. Her Clinical Frailty Scale is 5, indicating mild frailty. [1]

She is taking 14 regular medications: warfarin, bisoprolol, frusemide, ramipril, spironolactone, metformin, glibenclamide, atorvastatin, alendronate, oxybutynin, diazepam, amitriptyline, omeprazole, and aspirin. This is polypharmacy, with at least 7 of these drugs flagged as potentially inappropriate by the STOPP/START or Beers criteria. [1]

Her admission observations show orthostatic hypotension (104/58 lying, 88/52 standing), bradycardia at 54 in atrial fibrillation, a low body weight of 46 kilograms, an eGFR of 29, an HbA1c of 6.3 percent, and an INR of 3.8. She is on both aspirin and warfarin. [1]

Her main problems are:

1. A multifactorial fall with the medications as the most modifiable contributors — orthostatic hypotension from her antihypertensives and diuretics, bradycardia from the bisoprolol, sedation and ataxia from the diazepam and the amitriptyline, and the over-anticoagulated INR with dual antiplatelet-anticoagulant therapy increasing her bleeding risk.
2. Polypharmacy with multiple potentially inappropriate medications — the glibenclamide (prolonged hypoglycaemia in renal impairment, on the Beers list), the diazepam (long-acting benzodiazepine, falls risk), the oxybutynin (high anticholinergic burden), the amitriptyline (anticholinergic and sedating), the aspirin combined with warfarin (bleeding), the omeprazole (long-term without a current specialist indication), and the spironolactone in renal impairment with an ACE inhibitor (hyperkalaemia risk).
3. Renal impairment with a falsely reassuring creatinine-based eGFR, given her low weight and muscle mass — her drugs need renal dose adjustment and I would request a cystatin C-based eGFR.
4. Possible prescribing omissions — I would assess whether she is missing guideline-directed heart failure therapy such as an SGLT2 inhibitor, and ensure her vaccinations and osteoporosis management are optimised. [1]

My integrated plan is a structured medication review using the five-step deprescribing process of Scott and colleagues. I will immediately stop the glibenclamide and the aspirin, reduce the warfarin, hold or reduce the spironolactone, and reduce the frusemide and the antihypertensives to correct the orthostatic hypotension and the bradycardia. I will plan a supervised taper of the diazepam combined with cognitive behavioural therapy for insomnia, a step-down of the omeprazole, a switch from the oxybutynin to mirabegron, and a review of the amitriptyline. I will request a cystatin C-based eGFR and dose-adjust her renally-cleared drugs. I will apply the START criteria to identify any missing indicated therapy. I will anchor every decision to her goals of care, her frailty, and her renal function, and I will communicate the reconciled and deprescribed medication list to her, her family, and her GP at discharge, with a follow-up review in 4 to 6 weeks." [1]

Examiner probing questions and model answers

Q1: "You plan to stop the glibenclamide. Walk me through the pharmacological reasoning and what you will use instead." [1]

"Glibenclamide, also known as glyburide, is a long-acting sulfonylurea that is on the 2023 AGS Beers Criteria (PMID 37140026) as potentially inappropriate in older adults because of its prolonged and unpredictable hypoglycaemia, which is mediated by its active metabolites that accumulate in renal impairment. Mrs Davies has an eGFR of 29, so the metabolites will accumulate, and her HbA1c of 6.3 percent indicates that her diabetes is over-treated — the risk of a severe hypoglycaemic event, which can cause falls, confusion, arrhythmia, and death in an older adult, far exceeds any microvascular benefit of pushing the HbA1c lower. The STOPP criteria also flag long-acting sulfonylureas in this context. So stopping the glibenclamide is both evidence-based and patient-specific. For her ongoing diabetes management, in a frail 84-year-old with an HbA1c already at 6.3 percent, I would relax the glycaemic target — a target HbA1c of 7.5 to 8.5 percent is appropriate in frailty, prioritising the avoidance of hypoglycaemia over tight control. I would continue the metformin at a renal-appropriate dose (the eGFR of 29 is at the threshold where metformin is still permitted with caution and dose reduction, though I would consider holding it acutely while her renal function is assessed and during the admission). If a second agent is needed, I would add a DPP-4 inhibitor such as linagliptin or vildagliptin (linagliptin is hepatically cleared and does not need renal dose adjustment), or consider an SGLT2 inhibitor if her heart failure and renal function make it appropriate. I would explicitly avoid a thiazolidinedione (fluid retention, heart failure) and avoid initiating insulin unless there is a clear symptomatic hyperglycaemia. The principle is that the glycaemic target must match the patient's frailty, life expectancy, and goals — and in this patient, the current regimen is causing harm." [1]

Q2: "Her INR is 3.8 and she is on both aspirin and warfarin. How do you reconcile her antithrombotic therapy?" [1]

"The combination of aspirin and warfarin is flagged by the STOPP criteria unless there is a clear indication for dual therapy — typically a recent coronary stent, a mechanical heart valve, or a recent acute coronary syndrome. Mrs Davies has atrial fibrillation and ischaemic heart disease, but without a recent coronary event or stent, the aspirin adds bleeding risk without a meaningful reduction in ischaemic events on top of the anticoagulation. Her INR of 3.8 is above the therapeutic range for atrial fibrillation (2.0 to 3.0), which reflects her age (older adults are more sensitive to warfarin, pharmacodynamically), her low weight (46 kilograms, reducing the volume of distribution), and her reduced warfarin clearance. The combination of an above-target INR with aspirin is a high-risk situation for a major bleed, and it is likely a significant contributor to her fall (an intracranial bleed must be excluded given her head laceration — I would arrange a CT brain on admission). My plan is to stop the aspirin immediately (the anticoagulant alone provides adequate stroke prevention for the atrial fibrillation and reasonable ischaemic protection for the coronary disease), to withhold the warfarin until the INR is in range, and to reduce her maintenance warfarin dose (a frail, low-weight older adult typically needs a lower maintenance dose — perhaps 1.5 to 2.5 mg daily). I would monitor the INR daily until it is stable, and I would review her for interactions (amiodarone, antibiotics, and the recently stopped drugs all interact with warfarin). I would also address whether she should be on a DOAC instead of warfarin — a DOAC has a lower intracranial bleeding risk than warfarin and does not require INR monitoring, which is an advantage in a patient with falls and labile INRs, though the dose must be carefully chosen for her renal function and weight. This is a decision I would make with her and her GP after the acute admission." [1]

Q3: "How would you assess her renal function, given her weight and frailty?" [1]

"Her creatinine-based eGFR of 29 by the 2021 CKD-EPI equation (Inker et al, NEJM 2021, PMID 34554658) is a starting point, but I am concerned that it may overestimate her true renal function because the equation derives the GFR from the serum creatinine, and creatinine is generated by muscle — a frail, 46-kilogram, 84-year-old woman has very low muscle mass, so she produces less creatinine for any given GFR, and the creatinine-based estimate can be falsely reassuring. For drugs with a narrow therapeutic index that are renally cleared — and several of her drugs fall into this category (metformin, the DOACs if she switches, gabapentin or pregabalin if she were on them) — I would refine the estimate with a cystatin C-based eGFR. Cystatin C is produced by all nucleated cells at a relatively constant rate and is much less dependent on muscle mass, so it provides a more accurate estimate in the low-muscle patient. Alternatively, a measured creatinine clearance from a 24-hour urine collection, though logistically harder in a frail patient, would provide a direct measurement. For drug dosing, many references still use the Cockcroft-Gault creatinine clearance in mL per minute, which I would calculate using her actual (or an estimated) weight. The clinical principle is that in a frail, low-muscle older patient, you must not trust the serum creatinine or the creatinine-based eGFR alone for a high-stakes dosing decision — you verify with cystatin C or a measured clearance, and you dose conservatively with close monitoring of the drug effect and the renal function." [1]

Q4: "She is on oxybutynin and amitriptyline. Why are these a particular concern in this patient, and what would you do?" [1]

"Both oxybutynin and amitriptyline have high anticholinergic (antimuscarinic) activity, and together they create a significant anticholinergic burden — the cumulative central antimuscarinic effect that is associated with cognitive impairment, delirium, falls, dry mouth, constipation, and urinary retention. Oxybutynin is the highest-burden agent on the Anticholinergic Burden Scale and is on the Beers list; amitriptyline, a tricyclic antidepressant, is also on the Beers list for its anticholinergic, sedating, and orthostatic-hypotension effects. In a patient who has had a fall and who is on a benzodiazepine as well, the combined CNS depression and the anticholinergic burden are a major, modifiable contributor to her presentation. My plan is to stop the oxybutynin — she is on it for an overactive bladder, and I would offer mirabegron, a beta-3 adrenergic agonist that relaxes the detrusor without anticholinergic activity, though I would monitor her blood pressure as mirabegron can raise it. If pharmacological treatment of the bladder is not essential, I would prioritise non-pharmacological measures — timed voiding, prompted voiding by the family or the carers, fluid management, and ensuring easy access to a toilet or a commode. For the amitriptyline, she is taking it for sleep and possibly for depression — I would stop it, and if there is a genuine current depressive episode, I would switch to sertraline, an SSRI with a lower anticholinergic burden, while being alert to the hyponatraemia risk (she is on diuretics, which compounds this). For the sleep, I would use sleep hygiene and CBT-I principles. The overarching principle is to calculate and actively reduce the anticholinergic burden in every older patient with cognitive symptoms or falls — it is one of the highest-yield deprescribing targets." [1]

Q5: "How would you explain the deprescribing plan to Mrs Davies and her family?" [1]

"I would frame the conversation around her goals and her safety, not around a list of drug names. I would start by asking her and her family what matters most to her — and in a woman who has just been admitted with a fall, the answer is likely to be staying independent, avoiding another fall, and not coming back to hospital. I would then explain that several of her medications, which were all started with good intentions at different times, are now working against those goals — that the sleeping tablet is making her unsteady, that the bladder tablet and the antidepressant are clouding her thinking and contributing to the fall, that the diabetes tablet is lowering her blood sugar too much, and that the combination of the aspirin and the warfarin is making her prone to bleeding. I would explain that stopping or reducing these is an active, positive step to make her safer and clearer-headed, not a withdrawal of care — and that we will do it carefully, one change at a time, with a plan and a follow-up, so that she is in control. I would acknowledge that some of these medications have been part of her life for years and that she may have strong feelings about them — she may believe the sleeping tablet is essential, or that stopping a heart medication is dangerous — and I would address each concern with evidence and patience. I would provide a written, reconciled medication list and a deprescribing plan with the timeline, and I would involve her GP and her pharmacist in the follow-up so that the changes are sustained after discharge. The conversation is the intervention — a medication list changed without the patient's understanding and agreement will simply be reversed at the next GP visit." [1]

DCE Short-Case Viva — The Bedside Medication Review

Examiner instruction: "This 82-year-old man is in hospital with a fall. Please review his medications at the bedside." [1]

Candidate's examination routine (narrated)

"I would begin by confirming the patient's identity, introducing myself, and explaining that I am going to review his medications to make sure each one is still right for him. [1]

First, I obtain the best possible medication history — the verified list of everything he is actually taking. I ask the patient, but I do not rely on his recall alone, because many older patients cannot name all their drugs. I ask his family or his carer, I check the drug chart, and I ask to see the bag of medications he brought in, or I contact his GP and his pharmacist for the pre-admission list. I explicitly ask about over-the-counter medications (ibuprofen, aspirin, antihistamines, which patients often do not think of as 'medications'), complementary and herbal medicines, 'as needed' medications, topical agents (gels, creams), inhalers, eye drops, and recently stopped medications. I record the dose, the frequency, the route, the indication, the duration, and the adherence for each drug. [1]

Next, I apply a structured tool to identify potentially inappropriate medications. I use the STOPP and START criteria version 3 as my primary tool because they are systems-based and they detect both over-prescribing (STOPP) and under-prescribing (START), and I cross-check with the 2023 AGS Beers Criteria for the high-yield drugs — benzodiazepines, anticholinergics, long-acting sulfonylureas, NSAIDs, and first-generation antihistamines. [1]

I then check the renal function and dose-adjust every renally-cleared drug — and if the patient is frail or low-muscle, I consider a cystatin C-based estimate to refine the creatinine-based eGFR. [1]

I assess the falls risk from medications — I look specifically for benzodiazepines, antipsychotics, antidepressants, opioids, and antihypertensives, all of which the Woolcott meta-analysis (PMID 19933962) identified as independently associated with falls in older adults. [1]

I assess the anticholinergic burden — I identify every drug with antimuscarinic activity (oxybutynin, solifenacin, tricyclics, first-generation antihistamines, quetiapine) and score the cumulative burden, flagging a score of 3 or more for review. [1]

I check for prescribing omissions using the START criteria — is there an indicated drug that is missing, such as an anticoagulant in atrial fibrillation at high stroke risk, or disease-modifying heart failure therapy, or osteoporosis treatment after a fracture? [1]

Finally, I stand back and present the plan — the prioritised list of changes, the drugs to stop, the drugs to taper with the taper schedule, the drugs to switch, the drugs to start, the dose adjustments for renal function, and the review dates — and I communicate the reconciled list to the patient, the family, the GP, and the pharmacist." [1]

Presentation template (model answer)

"I performed a bedside medication review on this 82-year-old man admitted with a fall. His verified medication list comprises 12 regular medications and 2 over-the-counter preparations that he had not previously reported. Using the STOPP/START version 3 criteria and the 2023 AGS Beers Criteria, I identified the following: first, three potentially inappropriate medications to stop — the diazepam (a long-acting benzodiazepine on the Beers list, an independent falls-risk factor, which I will taper by 10 to 25 percent every 1 to 2 weeks with cognitive behavioural therapy for insomnia), the oxybutynin (the highest anticholinergic burden agent, contributing to his cognitive impairment, which I will stop and replace with mirabegron or non-pharmacological bladder management), and the diclofenac he has been buying over the counter for arthritis (an NSAID, on the Beers list and STOPP, contributing to his elevated blood pressure and his falling eGFR, which I will stop, managing his pain with paracetamol and a topical agent). Second, one prescribing omission to address — he has atrial fibrillation at high stroke risk (CHA2DS2-VASc elevated) and is not anticoagulated; after discussing the bleeding risk and his falls, I will start an anticoagulant, dose-adjusted for his renal function. Third, his renal function (eGFR 32 by creatinine, but he is frail and low-muscle, so I have requested a cystatin C-based estimate) requires a dose reduction of his gabapentin and his metformin. His anticholinergic burden score is currently 3, driven by the oxybutynin; stopping it reduces the score to 1. I have communicated the plan to him and his family, and I will reconcile the medication list at discharge and arrange a GP and pharmacist review within 2 weeks." [1]

Examiner discussion

Q: "How do the STOPP and Beers criteria differ, and when would you use each?" [1]

"The Beers Criteria, published by the American Geriatrics Society and updated in 2023 (PMID 37140026), are a US-origin explicit list of potentially inappropriate medications organised into five categories: PIMs in older adults generally, PIMs in specific diseases or syndromes, medications to use with caution, clinically important drug-drug interactions, and medications requiring dose adjustment for renal function. They are widely used in the United States and are the standard reference in ABIM-style questions. The STOPP/START criteria, now in version 3 (O'Mahony et al, 2023, PMID 37231111), are a European-origin tool organised by physiological system, with two halves: STOPP (133 criteria flagging potentially inappropriate medications to stop) and START (57 criteria flagging potentially prescribing omissions to start). The key difference is that STOPP/START detects both over-prescribing and under-prescribing, whereas Beers is primarily an over-prescribing tool. In my practice, I use STOPP/START as my primary bedside tool because it is systems-based, comprehensive, and catches omissions, and I cross-reference Beers for the highest-yield drug classes (the benzodiazepines, anticholinergics, and long-acting sulfonylureas that appear most often in exams and in practice). The two tools overlap substantially but are not identical — each catches some drugs the other does not — so using both is more sensitive than using either alone. The final prescribing decision, however, always rests on clinical judgement of the individual patient, not on a criterion alone." [1]

Q: "What is the single most important principle in deprescribing?" [1]

"The single most important principle is that deprescribing is a patient-centred, supervised process, not a drug-stopping reflex. Every decision must be anchored to the patient's current goals of care, their life expectancy, their frailty, and the current balance of benefit and harm of each drug — and it must be done with the patient's understanding and agreement, with a taper where withdrawal or rebound is possible, and with a monitoring plan. The five-step framework of Scott and colleagues (JAMA Intern Med 2015, PMID 25798731) operationalises this: ascertain all drugs, identify the PIMs, assess each against the patient, plan the taper or cessation, and monitor for withdrawal, rebound, and return of the original condition. The evidence consistently shows that most deprescribing attempts succeed — the fear of relapse is greater than the reality — provided the process is supervised and the patient is engaged. A medication list changed unilaterally, without the patient's agreement and without a follow-up plan, will simply be reversed." [1]

Q: "What is the prescribing cascade, and why is it so important in geriatric medicine?" [1]

"The prescribing cascade, first formally described by Rochon and Gurwitz, is the sequence in which an adverse effect of one drug is misinterpreted as a new medical condition, and a second drug is prescribed to treat it — adding to the polypharmacy and the harm, and potentially triggering a third prescription. The classic examples are metoclopramide causing drug-induced parkinsonism that is treated with levodopa; a diuretic causing gout that is treated with allopurinol; an NSAID causing hypertension that is treated with an additional antihypertensive; and an anticholinergic for overactive bladder opposing a cholinesterase inhibitor for dementia. It is important in geriatric medicine because it is the commonest engine of polypharmacy growth in older adults, because the symptoms it generates (falls, confusion, incontinence, parkinsonism) are common and non-specific and are easily misattributed to ageing or a new disease, and because preventing and reversing it is a high-yield clinical intervention. The clinical rule is simple: whenever you are about to start a new drug for a new symptom in an older patient, ask whether the symptom could be an adverse effect of a drug they are already taking, and address the first drug before adding the second." [1]