Beers criteria for potentially inappropriate medication use in older adults

The Beers criteria (named after Dr. Beers and not your Friday night escapades) is a guideline to help guide safe prescribing practices in the geriatric population. Polypharmacy and medication side effects are a common, and generally under-recognized, reason for patients to present to the emergency department. Adverse events and drug interactions should almost always be on our differential diagnosis with elderly patients. Additionally, we don’t want to cause harm with our own prescriptions. Thus, the Beers criteria is a document that emergency physicians should be familiar with. 

I do have problems with the way that many guidelines are presented, and this is no exception. The underlying process is science-based, and they follow the GRADE approach, but the presentation of the results is really limited, impairing our ability to apply judgment, and impairing patients’ ability to make informed decisions. The quality of evidence is ranked from low to high, but the document is lacking clear citations that would allow practicing clinicians to check the evidence themselves. More importantly, we are presented with a blanket statement, such as “avoid”, but without any discussion of the absolute risk of these medications. Without knowing the magnitude of the risk, it is impossible for clinicians (and patients) to judge that risk against potential benefits, which severely limits the value of these guidelines, and I think explains why so many of us have ignored them for so long. For now, it is impossible to know how closely one should adhere to these guidelines, and therefore I find this document mostly useless for day to day emergency medicine practice. (This is especially true given the label “potentially inappropriate”. How am I supposed to know when it really is inappropriate?!) That being said, there is certainly a high risk of harm when prescribing in the geriatric population, and so I thought I would pick out some of the more important recommendations for emergency providers. 

The paper

2023 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2023 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023 May 4. doi: 10.1111/jgs.18372. PMID: 37139824

Potential inappropriate medications for all older patients

First generation antihistamines

• Strong recommendation to avoid, moderate evidence
• The rationale is that they are highly anticholinergic.
• I think this is probably true across all populations, as we have better antihistamines.
• They state that diphenhydramine might be reasonable in severe allergic reactions, but there are IV formulations of modern antihistamines, so we can probably just bin this class of drugs and move on.

Nitrofurantoin

• Strong recommendation with low quality of evidence to avoid long term use in patients with a creatinine clearance of less than 30 mL/min
• Has the potential for pulmonary toxicity, hepatotoxicity, and peripheral neuropathy
• We usually aren’t prescribing long term (prophylactic doses), but I see a lot of older patients on this, so keep toxicity in mind as part of your differentials. 

Warfarin

• Strong recommendation with high level evidence to use DOACs over warfarin for nonvalvular atrial fibrillation and thromboembolism
• I think we are all already doing this, but I also think the evidence is far less convincing than they make it out to be

Rivaroxaban

• They make a strong recommendation based on moderate level of evidence to use apixaban over rivaroxaban because they think rivaroxaban is associated with a higher level of bleeding.
• This is where (early in the document) I started losing a lot of faith in the evidence behind these guidelines. I have covered the observational studies comparing rivaroxaban to apixaban, and they are completely unconvincing. I will personally ignore this recommendation.

Clonidine

• Strong recommendation based on low quality evidence to avoid as routine treatment fror hypertension
• High risk of CNS adverse events, bradycardia, and orthostatic hypotension

Nifedipine

• Strong recommendation based on high quality evidence to avoid
• Risk of hypotension and precipitations of myocardial infarction.

Amiodarone

• Strong recommendation based on high quality evidence to avoid as first line therapy in atrial fibrillation unless the patient has heart failure or left ventricular hypertrophy
• Greater toxicity than other antiarrhythmics
• If you haven’t seen complications (pulmonary or thyroid) from long term amiodarone use, you probably aren’t looking hard enough

Antidepressants with strong anticholinergic activity (amitriptyline, clomipramine, nortriptyline, paroxetine, etc.)

• Strong recommendation based on high quality evidence to avoid
• Sedating, cause orthostatic hypotension

First and second generation antipsychotics (haloperidol, olanzapine, risperidone, quetiapine, etc.)

• Strong recommendation based on moderate quality of evidence
• These agents increase mortality in patients with dementia, and also increase the risk of stroke, and increase the rate of cognitive decline
• This might be the guideline I see most commonly ignored, as close to 100% of nursing home patients I see are on an antipsychotic. I don’t work in nursing homes, but the risk benefit here seems very hard to assess. Increased mortality is really bad, but these patients frequently have behavioral issues that can’t be controlled in any other way. I think every geriatrician I have ever worked with has been liberal in their prescription of these agents, so I am not sure who actually reads or follows the Beers criteria.
• They do say these can be used if nonpharmacologic options have failed, but recommend periodic deprescribing attempts, and using the lowest effective dose (which is basically true of every prescribed medicine). Deprescribing is one of the most valuable and overlooked interventions in medicine. 

Benzodiazepines

• Strong recommendation based on moderate quality evidence to avoid
• The risk of addiction and misuse applied to all patients. Older patients have more issues with metabolism, leading to accumulation, and an increased risk of cognitive impairment, delirium, falls, fractures, and motor vehicle collisions.
• Some listed exceptions are seizure disorders, withdrawal, periprocedural anesthesia, and severe generalized anxiety disorder. This is one of the areas where numbers would make this guideline so much more usable. What is the absolute risk of falls or fractures? That seems really important when trying to determine if anxiety is severe enough to warrant treatment with benzodiazepines. 

Nonbenzodiazepine benzodiazepine receptor agonist hypnotics (zopiclone, Eszopiclone, zolpidem, etc.)

• Strong recommendation based on moderate quality evidence to avoid
• Adverse effect profile is essentially the same as benzodiazepines (delirium, falls, fractures, increased emergency room visits/hospitalizations, motor vehicle crashes), with minimal improvement in sleep latency and duration.

Sulfonylureas

• Strong recommendation based on moderate to high quality evidence (depending on the outcome) to avoid
• Sulfonylureas have a higher risk of cardiovascular events, all-cause mortality, and hypoglycemia than alternative agents. Sulfonylureas may increase the risk of cardiovascular death and ischemic stroke. (This is a pretty damning statement, as the whole point of treating diabetes is to avoid cardiovascular outcomes. I think a lot of physicians have forgot that point, and focused on correcting the lab value over patient important outcomes.)
• Prolonged hypoglycemia is more likely with long acting agents, so if these agents are necessary, short acting preferred over long acting
• Again, like for all of these recommendations, there is no discussion of the absolute magnitude of the risk, nor the uncertainty around the estimate, and therefore is it essentially impossible to incorporate these guidelines into a proper evidence based practice.

PPIs

• Strong evidence based on moderate to high level evidence
• Avoid courses longer than 8 weeks unless high risk patient
• Increased risk of C difficile infection, pneumonia, GI malignancies, bones loss, and fractures
• This is, without a doubt, one of the most overused classes of drugs across all of medicine, without even incorporating the completely unnecessary use in acute GI bleeds
• Personally, I think the key is short courses, deprescribing, and transition to h2 blockers when tolerable

Metoclopramide

• Strong recommendation based on moderate evidence to avoid, unless for gastroparesis for a duration less than 12 weeks
• Can cause extrapyramidal effects, including tardive dyskinesia; the risk may be greater in frail older adults and with prolonged exposure
• This is another useful medication for some acute conditions (vertigo), where there risks and benefits of short term use need to be considered, and this guideline doesn’t help me achieve that

GI ‘antispasmodics’ (atropine, hyoscyamine, scopolamine)

• Strong recommendation based on moderate quality evidence to avoid
• Highly anticholinergic
• These agents are really popular in some circles, but I think the evidence that they do anything in any population is pretty limited

Nonselective NSAIDs

• Strong recommendation based on moderate quality evidence to avoid chronic use, and short term use combined with steroids, anticoagulants, and antiplatelet agents
• Risk of peptic ulcer disease and GI bleeds 
• Interestingly, this is the only recommendation with an absolute risk included. The risk of GI ulcers, gross bleeding, or perforation caused by NSAIDs occur in about 1% of patients treated for 3–6 months and in 2%–4% of patients treated for 1 year. In other words, the risk of a 5 day course is miniscule, especially in selected patients and when combined with a PPI.
• I think this is the one recommendation that almost everyone follows, and we have actually probably taken it too far in emergency medicine. This is not a blanket ban on NSAIDs (and even these very conservative Beers guidelines make that clear). There is a risk/benefit to consider, but short courses of NSAIDs are often incredibly valuable for quality of life, and many patients are very happy to accept a small risk to decrease pain. 
• Indomethacin is called out in its own section as having the highest adverse event rates of all NSAIDs. Personally, I just stick to naproxen across the board (although topical NSAIDs are a nice way too decrease adverse events as well).

Skeletal ‘muscle relaxants’ (cyclobenzaprine, etc)

• Strong recommendation based on moderate quality evidence to avoid
• Significant anticholinergic side effects, sedation, and increased risk of factures
• This is another class that isn’t specific to geriatrics. These ‘muscle relaxants’ do nothing to relax muscles, and just shouldn’t be used. 

Potentially inappropriate in specific conditions

The next section of the BEERs guidelines focused on specific disease states, and medications that should be avoided in those conditions. I imagine we are all pretty clear on medications to avoid in CHF (NSAIDs) and peptic ulcer disease (NSAIDS), but I think this table suggests a few common mistakes many emergency physicians make, so I will pull just a couple points.

Delirium

• It’s not the urine. It’s almost never the urine. But medication side effects are probably a cause of delirium that you are overlooking almost every shift.
• Medications that cause or increase the risk of delirium include:
  • Anticholinergics
  • Antipsychotics
  • Benzodiazepines
  • H2 receptor antagonists
  • Nonbenzodiazepine benzodiazepine receptor agonist hypnotics (“Zdrugs”)
  • Opioids

Falls

• Elderly patients falling is a daily presentation in emergency medicine. If you aren’t examining the medication lists, you are probably failing these patients.
• If you have a good geriatrics program, the list of drugs that increases falls risk is probably familiar to you:
  • Anticholinergics
  • SSRIs
  • SNRIs
  • Tricyclic antidepressants
  • Antiepileptics
  • Antipsychotics
  • Benzodiazepines
  • Nonbenzodiazepine benzodiazepine receptor agonist hypnotics (“Zdrugs”)
  • Opioids
• (Yes, the list of bad drugs is essentially the same in every category. But also, almost every elderly patient I see is on at least one of these drugs).
• Once again, it is very hard to determine the risk/benefit here without absolute numbers. Judgment and patient preference are important when looking after patients. 

SIADH / hyponatremia

• This list of medications is listed in the next section in the actual guidelines, but that formatting is weird to me. These are medications that may exacerbate or cause SIADH or hyponatremia
  • SSRIs
  • SNRIs
  • Tricyclic antidepressants
  • Mirtazapine
  • Carbamazepine
  • Antipsychotics
  • Diuretics
  • Tramadol (OK, no one should ever be prescribing tramadol)

Use with caution

This list is very similar to the first list, but rather than “potentially avoid” you get to “use with caution”. That qualitative difference is apparently important, but not really explained.

Dabigatran

• Strong recommendation with moderate evidence
• More bleeding that both warfarin and the other DOACs
• I probably didn’t need to include this, as I don’t think anyone is using this anymore

Prasugrel and Ticagrelor

• Strong recommendation with moderate evidence
• Increased risk of bleeding as compared to clopidogrel
• I think the evidence that these agents are any better than clopidogrel in any age group is incredibly questionable, but I guess that is a question for another day

Dextromethorphan

• Strong recommendation with moderate evidence
• I am not sure why this is just in the “use with caution” section, rather than the “avoid” section, as this is a useless drug that shouldn’t be used in any population, and the risks almost certainly outweigh the benefits across the board

Trimethoprim-sulfamethoxazole

• Strong recommendation based on low level evidence
• Increased risk of hyperkalemia when used concurrently with ACEI, ARB, or ARNI in presence of decreased CrCl

Sodium-glucose cotransporter-2 (SGLT2) inhibitors

• Weak recommendation based on a moderate level of evidence
• May cause increased UTIs in elderly patients, and there is also an increased risk of euglycemic DKA (which I think we have all seen by now)

Other lists

This document also contains tables looking at important drug/drug interactions, and drugs that should be adjusted or avoided in patients with renal failure. Looking through those lists, I think it is information we mostly know, and so I have not reproduced them here, but you might want to look through this document yourself at some point.

References

By the 2023 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2023 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023 May 4. doi: 10.1111/jgs.18372. Epub ahead of print. PMID: 37139824