Medication Deprescribing in Critical Illness: A Systematic Approach to Drug Discontinuation

Dr Neeraj Manikath , Claude.ai

Abstract

The transition from chronic outpatient care to acute critical illness necessitates a fundamental reassessment of medication regimens. Many chronic medications that provide long-term benefit become unnecessary, harmful, or inappropriate in the critically ill patient. This review provides an evidence-based framework for medication deprescribing in the intensive care unit (ICU), examining which chronic medications should be discontinued, the optimal timing for cessation, and patient-specific factors that influence these decisions. Understanding the principles of deprescribing is essential for optimizing outcomes and minimizing iatrogenic harm in critical care.

Introduction

Polypharmacy is ubiquitous in modern medicine, with elderly ICU patients often admitted on 10-15 chronic medications.[1] While these medications may be appropriate for ambulatory management of chronic conditions, critical illness represents a fundamentally different physiological state characterized by hemodynamic instability, altered pharmacokinetics, organ dysfunction, and competing therapeutic priorities.[2]

The concept of "prescribing inertia"—the tendency to continue medications simply because they were previously prescribed—poses significant risks in critical care.[3] Conversely, inappropriate discontinuation may precipitate withdrawal syndromes or decompensation of underlying conditions. This review provides a systematic approach to medication discontinuation in the ICU setting.

General Principles of Deprescribing in Critical Illness

Pharmacokinetic and Pharmacodynamic Alterations

Critical illness profoundly alters drug disposition through multiple mechanisms:

Altered volume of distribution: Fluid resuscitation, capillary leak, and hypoalbuminemia increase the volume of distribution for hydrophilic drugs while reducing protein binding of lipophilic agents.[4]
Impaired hepatic metabolism: Hypoperfusion, hypoxia, and inflammatory cytokines reduce hepatic drug clearance by 30-50% in septic shock.[5]
Renal dysfunction: Acute kidney injury (AKI) affects 50-60% of ICU patients, necessitating dose adjustment or discontinuation of renally eliminated drugs.[6]
Altered gastrointestinal absorption: Gastroparesis, ileus, and reduced splanchnic perfusion compromise enteral drug absorption.[7]

Framework for Decision-Making

A systematic approach should consider:

Indication reassessment: Is the chronic condition still the therapeutic priority?
Risk-benefit ratio: Does critical illness shift the balance toward harm?
Time horizon: Is the benefit immediate or preventive over months/years?
Withdrawal risks: Will discontinuation cause acute decompensation?
Alternative routes: Can essential medications be given parenterally?

Pearl: Apply the "6-month rule"—if a medication's benefit accrues over >6 months and ICU mortality risk is high, consider discontinuation.[8]

Cardiovascular Medications

Antihypertensive Agents

WHEN TO STOP: Immediately upon ICU admission in hemodynamically unstable patients

IN WHOM:

Patients with shock of any etiology (septic, cardiogenic, hypovolemic)
Systolic BP <100 mmHg or MAP <65 mmHg
Patients requiring vasopressor support

RATIONALE:

ACE inhibitors and ARBs may precipitate refractory hypotension in septic shock through impaired angiotensin II generation.[9] Beta-blockers reduce compensatory tachycardia and may worsen cardiac output in distributive shock.[10] Calcium channel blockers exacerbate vasodilation in sepsis.

EXCEPTIONS:

Beta-blockers in compensated heart failure with EF <40% (continue at reduced dose)
Antihypertensives in hypertensive emergency (intracranial hemorrhage, aortic dissection)

Withdrawal considerations:

Beta-blocker withdrawal can precipitate rebound tachycardia, hypertension, and myocardial ischemia.[11] In stable ICU patients, taper over 3-5 days rather than abrupt cessation. Clonidine withdrawal causes rebound hypertension within 24-48 hours and should be tapered or substituted with dexmedetomidine in ventilated patients.[12]

Oyster: In patients with chronic beta-blocker therapy presenting with septic shock, abrupt discontinuation may paradoxically worsen outcomes. Consider reducing to 25-50% of home dose rather than complete cessation, particularly in coronary artery disease.[13]

References:

[9] Chawla LS, et al. Intensive Care Med. 2007;33(8):1327-1333.
[10] Morelli A, et al. JAMA. 2013;310(16):1683-1691.
[11] Psaty BM, et al. Ann Intern Med. 1990;113(12):927-933.

Antiplatelet Agents and Anticoagulants

WHEN TO STOP:

Antiplatelet agents:

Before high-bleeding-risk procedures (neurosurgery, spinal procedures)
Active major hemorrhage
Severe thrombocytopenia (<50,000/μL for aspirin; <30,000/μL for P2Y12 inhibitors)

Anticoagulants:

Active bleeding or high bleeding risk
Before invasive procedures (timing depends on agent and renal function)
Severe thrombocytopenia (<20,000-30,000/μL)

IN WHOM:

The decision requires careful risk stratification:

CONTINUE antiplatelet therapy in:

Recent acute coronary syndrome (<12 months)
Recent coronary stent (bare metal <6 weeks; drug-eluting <12 months)
Acute coronary syndrome as reason for ICU admission

STOP antiplatelet therapy in:

Primary prevention indication
Stable coronary disease >12 months from event
Active intracranial hemorrhage

CONTINUE anticoagulation in:

Mechanical heart valves (absolute indication)
Acute VTE (<3 months)
High thromboembolic risk (CHA₂DS₂-VASc ≥4-5)

STOP anticoagulation in:

Chronic atrial fibrillation with low stroke risk
Remote VTE (>6-12 months) without ongoing risk factors

RATIONALE:

The bleeding risk in critical illness is substantial, with ICU patients experiencing major hemorrhage at rates of 5-10%.[14] However, stent thrombosis carries 30-40% mortality, necessitating continued dual antiplatelet therapy (DAPT) in high-risk windows.[15]

Hack: For patients requiring urgent surgery on DAPT, consider bridging with cangrelor (ultra-short-acting P2Y12 inhibitor, half-life 3-5 minutes) or tirofiban (GP IIb/IIIa inhibitor) in consultation with cardiology.[16] This allows platelet inhibition until induction, then rapid offset for surgical hemostasis.

Pearl: In traumatic intracranial hemorrhage, continuation of aspirin (but not DAPT) may be reasonable in patients with recent coronary stents after neurosurgical consultation, as some data suggest aspirin doesn't significantly worsen hemorrhage progression.[17]

DOACs vs. Warfarin:

Direct oral anticoagulants (DOACs) offer advantages in critical care due to rapid offset (12-24 hours with normal renal function) and availability of specific reversal agents (idarucizumab for dabigatran, andexanet alfa for factor Xa inhibitors).[18] Warfarin requires vitamin K (onset 12-24 hours) or prothrombin complex concentrate for urgent reversal.

References:

[14] Cook D, et al. Crit Care Med. 2005;33(6):1220-1232.
[15] Iakovou I, et al. JAMA. 2005;293(17):2126-2130.
[16] Angiolillo DJ, et al. Circ Cardiovasc Interv. 2009;2(4):284-286.

Statins

WHEN TO STOP: Controversial; consider discontinuation in:

Severe septic shock with multi-organ failure
Rhabdomyolysis (CK >10,000 U/L)
Drug-induced liver injury (transaminases >5× ULN)

IN WHOM:

Hemodynamically unstable patients unable to tolerate enterals
Those at high risk for statin-related complications (elderly, renal failure, concurrent nephrotoxins)

RATIONALE:

This remains one of the most debated deprescribing decisions. Statins possess anti-inflammatory and immunomodulatory properties that theoretically benefit sepsis.[19] However, clinical trials have shown no mortality benefit for statin continuation or initiation in sepsis.[20]

The primary argument for discontinuation is practical: critically ill patients often cannot receive enterals, and IV statins are unavailable in most settings. Missing doses for several days eliminates any pleiotropic benefit while maintaining risks (myopathy, liver toxicity, drug interactions with azoles and macrolides via CYP3A4).

CONTINUE statins in:

Stable ICU patients tolerating enteral medications
Recent acute coronary syndrome (<3 months)

Oyster: Statin withdrawal in the first 48 hours after myocardial infarction increases mortality (HR 2.93).[21] For ACS patients in the ICU, make every effort to continue statins via enteral route or consider parenteral alternatives in some countries where available.

References:

[19] Terblanche M, et al. Crit Care Med. 2007;35(5):1270-1277.
[20] Kruger PS, et al. Intensive Care Med. 2011;37(5):784-793.
[21] Spencer FA, et al. Arch Intern Med. 2004;164(19):2162-2168.

Gastrointestinal and Metabolic Medications

Proton Pump Inhibitors (PPIs)

WHEN TO STOP: Upon ICU admission if used for primary prevention

IN WHOM:

Patients without active peptic ulcer disease or GERD symptoms
Those using PPIs for "gastroprotection" with NSAIDs or antiplatelet agents (benefit marginal)[22]

RATIONALE:

PPIs are among the most inappropriately prescribed medications in hospitalized patients, with up to 70% lacking evidence-based indication.[23] In the ICU, PPIs increase risks of:

Hospital-acquired pneumonia (OR 1.27-1.5)[24]
Clostridioides difficile infection (OR 1.5-2.0)[25]
Hypomagnesemia with prolonged use
Drug interactions (reduced absorption of azole antifungals, reduced clopidogrel activation)

CONTINUE PPIs in:

Active peptic ulcer or significant GI bleeding
Severe reflux esophagitis
Zollinger-Ellison syndrome
Stress ulcer prophylaxis in high-risk patients (mechanical ventilation >48h, coagulopathy, high-dose corticosteroids, traumatic brain injury)

SWITCH to H2-receptor antagonists if stress ulcer prophylaxis is needed, as they have lower pneumonia risk.[26]

Pearl: For patients on chronic PPI therapy without clear indication, ICU admission provides an excellent opportunity for deprescribing. Most will not require stress ulcer prophylaxis if extubated within 48 hours and without coagulopathy.

References:

[22] Targownik LE, et al. CMAJ. 2009;180(7):713-718.
[23] Heidelbaugh JJ, et al. Am J Gastroenterol. 2009;104 Suppl 2:S15-22.
[24] Eom CS, et al. CMAJ. 2011;183(3):310-319.
[25] Janarthanan S, et al. Am J Gastroenterol. 2012;107(7):1001-1010.

Oral Hypoglycemic Agents

WHEN TO STOP: Immediately upon ICU admission

IN WHOM: All critically ill patients requiring insulin infusion or with hemodynamic instability

RATIONALE:

All oral antidiabetic agents should be discontinued in critically ill patients for multiple reasons:

Metformin:

Lactic acidosis risk in patients with sepsis, shock, renal dysfunction, or contrast exposure[27]
Contraindicated with eGFR <30 mL/min
Should be held 48 hours before contrast procedures

Sulfonylureas (glipizide, glyburide):

Prolonged hypoglycemia risk (especially glyburide, half-life 24h)
Unpredictable absorption in ileus
Ineffective in stress hyperglycemia (insulin resistance predominates)

DPP-4 inhibitors, GLP-1 agonists, SGLT-2 inhibitors:

Limited efficacy in acute stress hyperglycemia
SGLT-2 inhibitors increase euglycemic diabetic ketoacidosis risk[28]
Impaired enteral absorption

Thiazolidinediones:

Fluid retention exacerbating heart failure and pulmonary edema
Delayed onset of action (weeks)

REPLACE with: Intravenous insulin infusion for critically ill patients with persistent hyperglycemia (>180 mg/dL), targeting 140-180 mg/dL.[29]

Hack: When transitioning stable ICU patients back to oral agents, overlap insulin with oral medications for 24-48 hours to prevent rebound hyperglycemia, especially with agents having delayed onset (metformin, pioglitazone).

References:

[27] Lalau JD, et al. Diabetes Care. 2010;33(7):1536-1543.
[28] Goldenberg RM, et al. Diabetes Care. 2019;42(6):1147-1152.
[29] NICE-SUGAR Study Investigators. N Engl J Med. 2009;360(13):1283-1297.

Endocrine Medications

Thyroid Hormone Replacement

WHEN TO STOP: Generally continue, but reassess dosing

IN WHOM:

Reduce dose by 20-30% in myxedema coma (paradoxically, to allow gradual correction)
May temporarily discontinue in severe cardiovascular instability unresponsive to levothyroxine reduction

RATIONALE:

"Euthyroid sick syndrome" (low T3, normal/low TSH) is common in critical illness and represents an adaptive response—treatment is not indicated.[30] However, abrupt withdrawal of levothyroxine in patients with hypothyroidism may precipitate myxedema coma (mortality 20-30%).[31]

Continue levothyroxine in:

All patients with known hypothyroidism who can tolerate enterals
Consider IV levothyroxine (50-75% of oral dose) if NPO >5-7 days

STOP levothyroxine in:

Those without documented hypothyroidism receiving it empirically
Patients with iatrogenic thyrotoxicosis from over-replacement

Pearl: Levothyroxine has a half-life of 7 days, so missing 2-3 days is unlikely to cause significant decompensation. However, liothyronine (T3) has a 24-hour half-life and requires more vigilant monitoring if discontinued.

Oyster: Critical illness dramatically alters thyroid hormone metabolism. Don't check TSH in acute illness unless suspecting thyroid storm or myxedema coma—results will be uninterpretable and may lead to inappropriate dose adjustments.[32]

References:

[30] Wartofsky L, et al. J Clin Endocrinol Metab. 1982;54(1):6-16.
[31] Ono Y, et al. BMJ Case Rep. 2017;2017:bcr2017221338.
[32] Fliers E, et al. Lancet Diabetes Endocrinol. 2015;3(10):816-825.

Corticosteroids

WHEN TO STOP: Never abruptly in patients on chronic therapy >3 weeks

IN WHOM: Taper or stress-dose coverage needed in:

Chronic prednisone ≥5 mg/day for >3 weeks
Any dose of corticosteroid for >3 months
Cushingoid features (even if recently discontinued)

RATIONALE:

Chronic corticosteroid therapy suppresses the hypothalamic-pituitary-adrenal (HPA) axis, requiring 6-12 months for recovery.[33] Abrupt discontinuation during critical illness causes relative adrenal insufficiency with refractory shock.

STRESS DOSING PROTOCOL:

Hydrocortisone 100 mg IV q8h during severe stress (shock, major surgery)
Hydrocortisone 50 mg IV q8h during moderate stress (pneumonia, routine surgery)
Resume home dose when stress resolves

Pearl: The classic "stress dose" is 300 mg hydrocortisone per day (100 mg q8h), but many ICU patients receive this empirically for septic shock (as per APROCCHSS trial showing mortality benefit).[34] Don't add additional stress dosing on top of sepsis-dose hydrocortisone—you're already covered.

STOP chronic corticosteroids (with taper) in:

Patients on chronic steroids without clear indication
Those with steroid-induced complications (hyperglycemia, psychosis, myopathy) where risks outweigh benefits

Hack: To assess HPA axis recovery, check 8 AM cortisol after holding steroids for 24 hours (if clinically safe). Cortisol >10 μg/dL suggests adequate recovery; <3 μg/dL indicates continued suppression.[35]

References:

[33] Broersen LH, et al. J Clin Endocrinol Metab. 2015;100(6):2171-2180.
[34] Annane D, et al. N Engl J Med. 2018;378(9):809-818.
[35] Woods CP, et al. J Clin Endocrinol Metab. 2015;100(6):2171-2180.

Neuropsychiatric Medications

Benzodiazepines

WHEN TO STOP: Never abruptly in chronic users (>4 weeks continuous use)

IN WHOM: All chronic benzodiazepine users require continuation or substitution

RATIONALE:

Benzodiazepine withdrawal syndrome manifests 1-3 days after discontinuation (longer for long-acting agents) with anxiety, tremor, tachycardia, hypertension, seizures, and delirium.[36] Mortality from withdrawal seizures or delirium tremens can reach 15% without treatment.

CONTINUE benzodiazepines in chronic users via:

Equivalent-dose substitution (1 mg lorazepam = 5 mg diazepam = 0.5 mg clonazepam)
For enterally-inaccessible patients, use IV lorazepam or midazolam infusion
Taper by 10-25% every 3-5 days when clinically appropriate

STOP initiating benzodiazepines for:

ICU agitation/delirium (worsens outcomes per MENDS trial)[37]
Sedation in mechanically ventilated patients (prefer propofol or dexmedetomidine)

Pearl: Benzodiazepines are uniquely appropriate for alcohol withdrawal syndrome and withdrawal from other benzodiazepines—these are the primary ICU indications for new benzodiazepine use. CIWA-Ar protocol remains gold standard.[38]

Oyster: Paradoxically, chronic benzodiazepine users may require higher sedation doses in the ICU due to tolerance, yet are at higher risk for delirium. Consider transitioning to dexmedetomidine infusion for sedation while maintaining a small benzodiazepine dose to prevent withdrawal.[39]

References:

[36] Pétursson H. Addiction. 1994;89(11):1455-1459.
[37] Pandharipande PP, et al. JAMA. 2007;298(22):2644-2653.
[38] Sullivan JT, et al. Br J Addict. 1989;84(11):1353-1357.

Selective Serotonin Reuptake Inhibitors (SSRIs) and SNRIs

WHEN TO STOP:

Active serotonin syndrome (hyperthermia, rigidity, hyperreflexia, clonus)
Severe hyponatremia (Na <120 mEq/L) attributed to SIADH from SSRI
Major drug interaction (e.g., linezolid co-administration)

IN WHOM:

Consider discontinuation in patients NPO >7 days without IV access for alternative routes
Those with severe thrombocytopenia and bleeding (SSRIs inhibit platelet aggregation)

RATIONALE:

SSRIs/SNRIs cause discontinuation syndrome in 20-30% of users when stopped abruptly, particularly paroxetine and venlafaxine (short half-lives).[40] Symptoms include dizziness, nausea, headache, irritability, and flu-like symptoms—easily confused with critical illness itself.

However, SSRIs also carry ICU-specific risks:

Serotonin syndrome when combined with other serotonergic agents (fentanyl, tramadol, linezolid, methylene blue)[41]
Hyponatremia (particularly in elderly, SIADH mechanism)
Mild platelet dysfunction (clinical significance debated, likely minimal)

CONTINUE SSRIs/SNRIs in:

Stable patients tolerating enterals
Recent suicide attempt or severe depression

TRANSITION strategies:

Fluoxetine has longest half-life (4-6 days) and minimal discontinuation syndrome—rarely needs substitution if NPO
For paroxetine/venlafaxine, consider single-dose fluoxetine 20 mg if discontinuing (covers withdrawal)[42]
No IV formulations available; use nasogastric/orogastric tube if needed

Pearl: SSRI discontinuation syndrome typically begins 1-3 days after last dose and resolves within 2 weeks. If symptoms persist longer, consider alternative diagnosis (sepsis, delirium, electrolyte abnormality).

Hack: If linezolid is required in a patient on SSRIs (weak MAO inhibitor interaction), most guidelines now suggest continuing SSRIs with monitoring rather than automatic discontinuation, as serotonin syndrome risk is lower than previously thought (case reports only, no prospective data).[43]

References:

[40] Warner CH, et al. Am Fam Physician. 2006;74(3):449-456.
[41] Boyer EW, et al. N Engl J Med. 2005;352(11):1112-1120.
[42] Himei A, et al. Int Clin Psychopharmacol. 2006;21(5):301-303.

Antipsychotics

WHEN TO STOP:

QTc prolongation >500 ms (especially haloperidol, ziprasidone)
Neuroleptic malignant syndrome
Severe drug-induced parkinsonism

IN WHOM:

Reassess chronic antipsychotic use—many ICU patients are on these for "agitation" or off-label uses without clear indication
Consider discontinuation in those without schizophrenia/bipolar disorder

RATIONALE:

Typical antipsychotics (haloperidol) and some atypicals (ziprasidone, quetiapine) prolong QTc and increase torsades de pointes risk, particularly dangerous when combined with other QT-prolonging agents common in ICU (fluoroquinolones, azole antifungals, ondansetron).[44]

Antipsychotics do NOT effectively treat ICU delirium and may worsen outcomes (MIND-USA trial showed no benefit of haloperidol or ziprasidone vs. placebo).[45]

CONTINUE antipsychotics in:

Known schizophrenia (risk of psychotic decompensation)
Acute bipolar mania
Consider dose reduction and EKG monitoring

STOP antipsychotics in:

ICU delirium (use non-pharmacologic measures, address underlying causes)
"Off-label" use for agitation or sleep in non-psychotic patients
Parkinson's disease patients (extrapyramidal side effects)

Pearl: If antipsychotic is essential but QTc is prolonged, quetiapine has least QT effect among atypicals, while aripiprazole is weight-neutral and has favorable metabolic profile.[46]

References:

[44] Beach SR, et al. Psychosomatics. 2013;54(1):1-13.
[45] Girard TD, et al. N Engl J Med. 2018;379(26):2506-2516.
[46] Hasnain M, et al. Cardiovasc Psychiatry Neurol. 2014;2014:427-430.

Neurological Medications

Antiepileptic Drugs (AEDs)

WHEN TO STOP:

Never abruptly in patients with epilepsy
Consider discontinuation only if:
- AED initiated empirically without seizures (no EEG confirmation)
- Single remote seizure >5 years ago, now in status epilepticus from another cause
- Clear precipitant for seizure now resolved (alcohol withdrawal, resolved metabolic derangement)

IN WHOM:

Extreme caution in any patient with seizure history
Discuss with neurology before discontinuation

RATIONALE:

Abrupt AED withdrawal precipitates seizures in 20-40% of epilepsy patients, potentially causing status epilepticus (mortality 10-15% even with treatment).[47] Even patients seizure-free for years remain at risk.

Many ICU patients cannot take oral AEDs due to NPO status or feeding intolerance, necessitating parenteral alternatives:

IV EQUIVALENTS:

Phenytoin → Fosphenytoin or IV phenytoin (1:1 dosing)
Valproate → IV valproate (1:1 dosing)
Levetiracetam → IV levetiracetam (1:1 dosing)
Lacosamide → IV lacosamide (1:1 dosing)

NO IV EQUIVALENT (problematic):

Carbamazepine → Consider phenytoin substitution
Lamotrigine → No alternative; use NG tube or risk withdrawal
Oxcarbazepine → Consider phenytoin or carbamazepine rectal formulation
Topiramate → Use NG tube

Pearl: Levetiracetam (Keppra) and lacosamide have become first-line ICU AEDs because of excellent IV formulations, no drug interactions, and no hepatic metabolism.[48] Consider transitioning ICU patients from problematic AEDs to these agents.

Oyster: Phenytoin has notoriously erratic pharmacokinetics in critical illness. Hypoalbuminemia increases free drug levels despite "normal" total levels. Check free phenytoin levels and target 1-2 μg/mL (not total level).[49] Many neurologists now prefer levetiracetam to avoid this complexity.

Hack: For patients on chronic carbamazepine unable to take PO, crush tablets and give via NG tube (suspension not available in many regions), or substitute phenytoin temporarily with plan to reinitiate carbamazepine when PO tolerated. Carbamazepine has active metabolite (carbamazepine-10,11-epoxide) that persists 3-4 days, providing some coverage during transition.[50]

References:

[47] Schmidt D, et al. Neurology. 1983;33(12):1537-1543.
[48] Szaflarski JP, et al. Neurology. 2010;75(20):1817-1822.
[49] Winter ME, et al. Clin Pharmacokinet. 1988;14(6):329-346.
[50] Bertilsson L, et al. Clin Pharmacokinet. 1986;11(3):177-198.

Dopaminergic Agents (Parkinson's Disease)

WHEN TO STOP: Avoid discontinuation if at all possible

IN WHOM:

Never stop abruptly in Parkinson's disease patients
Even brief interruptions can precipitate parkinsonism-hyperpyrexia syndrome (PHS)

RATIONALE:

Abrupt withdrawal of levodopa/carbidopa causes parkinsonism-hyperpyrexia syndrome in 3-4% of PD patients—clinically identical to neuroleptic malignant syndrome with rigidity, hyperthermia, rhabdomyolysis, and autonomic instability (mortality 4-20%).[51]

CONTINUE dopaminergic therapy via:

Crushed levodopa/carbidopa via NG tube (preferred)
Rotigotine transdermal patch (if patient on patch pre-admission)
Apomorphine subcutaneous injection (rescue option)

NO IV LEVODOPA EXISTS in most countries, making PD medication management in NPO patients challenging.

STOP dopaminergic agents cautiously in:

Drug-induced psychosis/delirium where risk exceeds benefit
Consider gradual dose reduction rather than cessation

Pearl: PD patients have delayed gastric emptying—give levodopa 30-60 minutes before enteral nutrition and avoid high-protein feeds simultaneously (amino acids compete for absorption).[52]

Oyster: If forced to discontinue dopaminergic therapy (refractory hallucinations, prolonged NPO without access), watch for PHS developing over 24-72 hours. Early signs include worsening rigidity and fever. Treatment is supportive plus reinitiation of dopaminergic therapy and/or dantrolene/bromocriptine.[53]

Hack: For PD patients in ICU, consider involving neurology early for creative dosing solutions. Some centers use amantadine IV (available in some countries) as a bridge, though it's a weak dopamine agonist. Rotigotine patch is underutilized—can provide basal dopaminergic stimulation when PO route compromised.[54]

References:

[51] Serrano-Dueñas M. Parkinsonism Relat Disord. 2003;9(3):175-178.
[52] Nyholm D, et al. Clin Neuropharmacol. 2010;33(1):42-44.
[53] Kipps CM, et al. Mov Disord. 2005;20(11):1506-1507.
[54] Doshay LJ. JAMA. 1954;154(14):1174-1177.

Rheumatologic and Immunosuppressive Medications

Disease-Modifying Antirheumatic Drugs (DMARDs)

WHEN TO STOP: Immediately when serious infection suspected or confirmed

IN WHOM: All patients with sepsis, severe pneumonia, or opportunistic infections

RATIONALE:

DMARDs broadly suppress immune function, increasing infection risk and severity:

Methotrexate:

Hold during acute infection (restarting methotrexate may precipitate bone marrow suppression and mucositis)[55]
Particularly dangerous with concurrent nephrotoxins (vancomycin, aminoglycosides)—can cause acute renal failure
Leucovorin rescue not typically needed unless toxic levels/overdose

TNF-alpha inhibitors (infliximab, adalimumab, etanercept):

Hold during serious bacterial infections and sepsis
Increased risk of tuberculosis reactivation, fungal infections
Effects persist 4-8 weeks after last dose (biological half-life)[56]

JAK inhibitors (tofacitinib, baricitinib):

Hold during infection
Increased risk of herpes zoster reactivation
Consider acyclovir prophylaxis if prolonged use

Rituximab:

B-cell depletion persists 6-12 months
Increased risk of PJP, viral infections (consider prophylaxis)
Cannot hold what's already given, but defer scheduled doses

CONTINUE DMARDs cautiously in:

Severe autoimmune flare as primary reason for ICU admission (lupus cerebritis, vasculitis crisis)
Requires ID consultation and careful risk-benefit analysis

Pearl: The decision to continue or hold immunosuppression inCritically ill patients with autoimmune disease often requires balancing infection risk against disease flare. Generally, acute infection takes precedence—infection is an immediate mortality threat, while most autoimmune flares develop over days to weeks.[57]

Oyster: Many ICU patients are on chronic low-dose methotrexate (7.5-15 mg weekly) for rheumatoid arthritis or psoriasis. At these doses, immunosuppression is modest, but drug accumulation in renal failure dramatically increases toxicity. If creatinine is rising, check methotrexate level and consider leucovorin if level >0.1 μmol/L at 48 hours.[58]

Hack: For patients admitted with sepsis on chronic TNF-alpha inhibitors, document the last dose date. Infliximab has 8-10 day half-life, adalimumab 14 days, etanercept 4-5 days. Knowing the pharmacokinetic timeline helps assess whether immunosuppression is contributing to current infection vs. residual from previous dosing.[59] This information is valuable for ID consultation.

References:

[55] Widdifield J, et al. Ann Rheum Dis. 2018;77(6):798-809.
[56] Bongartz T, et al. JAMA. 2006;295(19):2275-2285.
[57] Fanouriakis A, et al. Ann Rheum Dis. 2019;78(6):736-745.
[58] Howard SC, et al. J Clin Oncol. 2002;20(7):1797-1805.
[59] Tracey D, et al. BioDrugs. 2008;22(6):383-399.

Chronic Corticosteroids for Autoimmune Disease

Covered under Endocrine section above, but worth reiterating: patients on chronic prednisone for autoimmune conditions (lupus, vasculitis, inflammatory bowel disease) require stress-dose coverage during critical illness. The immunosuppressive effects compound infection risk, but adrenal insufficiency from abrupt cessation poses immediate hemodynamic threat.

Approach:

Continue or increase to stress dose (hydrocortisone 100 mg IV q8h)
Hold other immunosuppressants (azathioprine, mycophenolate) during active infection
Resume home immunosuppression when infection controlled, in consultation with rheumatology

Urological Medications

Alpha-Blockers (Tamsulosin, Doxazosin, Terazosin)

WHEN TO STOP: Upon ICU admission in hemodynamically unstable patients

IN WHOM:

Patients with hypotension (SBP <100 mmHg)
Those requiring vasopressor support
Patients with orthostatic hypotension

RATIONALE:

Alpha-blockers cause vasodilation and may precipitate or worsen hypotension, particularly in hypovolemia or distributive shock.[60] Tamsulosin is "uroselective" (alpha-1A preferential) with less cardiovascular effect but still carries risk.

CONTINUE alpha-blockers in:

Hemodynamically stable patients
Recent prostate surgery (within 2-4 weeks)
Acute urinary retention requiring management

STOP alpha-blockers in:

Benign prostatic hyperplasia without recent complications (can catheterize if retention occurs)
Primary prevention of urinary retention

Pearl: Acute urinary retention is common in elderly ICU patients due to immobility, anticholinergic medications, and opioids. Rather than continuing alpha-blockers to prevent retention, use intermittent catheterization or short-term indwelling catheter if retention develops. The hypotension risk outweighs prevention benefit in critically ill patients.[61]

Withdrawal considerations: No significant withdrawal syndrome from alpha-blocker cessation. Can be stopped abruptly without taper.

References:

[60] Lepor H. Rev Urol. 2007;9(4):181-190.
[61] Verhamme KM, et al. Drug Saf. 2008;31(7):565-574.

5-Alpha Reductase Inhibitors (Finasteride, Dutasteride)

WHEN TO STOP: Can discontinue; benefits accrue over months

IN WHOM: All ICU patients

RATIONALE:

These agents shrink prostate tissue over 6-12 months—no acute benefit in ICU.[62] No withdrawal syndrome exists. Can be safely discontinued and restarted after ICU discharge if needed.

References:

[62] McConnell JD, et al. N Engl J Med. 1998;338(9):557-563.

Supplements and Vitamins

Multivitamins and Mineral Supplements

WHEN TO STOP: Most can be discontinued; add specific replacements as needed

IN WHOM: Patients with documented deficiencies require targeted repletion

RATIONALE:

Most oral multivitamins and supplements provide preventive nutrition over months—irrelevant in acute ICU timeframe.[63] However, specific deficiencies require aggressive repletion:

CONTINUE/REPLACE:

Thiamine: Essential in alcoholism, malnutrition, refeeding syndrome (give 100-500 mg IV daily)
Vitamin B12: If megaloblastic anemia or documented deficiency (1000 μg IM/IV)
Folate: If documented deficiency or on methotrexate (1-5 mg daily)
Vitamin D: If severe deficiency (<10 ng/mL) with hypocalcemia
Iron: Only if acute need for erythropoiesis; hold in active infection (bacterial iron scavenging)[64]

STOP:

General multivitamins
Herbal supplements (potential drug interactions)
High-dose vitamin C or E (no proven ICU benefit)
Calcium supplements (unless treating hypocalcemia; use IV calcium gluconate/chloride acutely)

Pearl: Check magnesium, phosphate, and potassium levels frequently in ICU—these are the most clinically relevant deficiencies requiring acute repletion. Magnesium deficiency is present in 20-60% of ICU patients and contributes to arrhythmias, seizures, and delirium.[65]

Oyster: Thiamine deficiency can present as "refractory" septic shock due to impaired cellular metabolism (cytopathic hypoxia). Consider empiric thiamine 200 mg IV q12h for 3 days in all patients with septic shock, particularly with history of alcohol use disorder or malnutrition, before attributing shock solely to sepsis.[66]

Hack: For suspected refeeding syndrome (high-risk: chronic malnutrition, minimal intake >7-10 days, BMI <16), give prophylactic thiamine 200-300 mg IV daily for 3-5 days BEFORE starting aggressive nutrition. Also replete phosphate, magnesium, and potassium aggressively and advance feeds slowly (10-20 kcal/kg/day initially).[67]

References:

[63] Gröber U, et al. Nutrients. 2013;5(9):3634-3645.
[64] Litton E, et al. Intensive Care Med. 2019;45(1):49-59.
[65] Velissaris D, et al. Nutrients. 2015;7(9):8009-8024.
[66] Woolum JA, et al. Crit Care Med. 2018;46(11):1747-1752.
[67] Mehanna HM, et al. BMJ. 2008;336(7659):1495-1498.

Pulmonary Medications

Inhaled Corticosteroids and Long-Acting Bronchodilators

WHEN TO STOP:

Inhaled corticosteroids (ICS): Consider discontinuation in mechanically ventilated patients
Long-acting bronchodilators (LABA/LAMA): Switch to short-acting agents in acute exacerbations

IN WHOM:

COPD/asthma patients during acute exacerbations requiring mechanical ventilation
Patients unable to coordinate inhaler technique or use nebulizers

RATIONALE:

Inhaled corticosteroids increase pneumonia risk in COPD patients (number needed to harm ~50).[68] During mechanical ventilation for respiratory failure, systemic corticosteroids (methylprednisolone 40-80 mg daily) are typically used, making ICS redundant and potentially harmful.

Long-acting bronchodilators (salmeterol, formoterol, tiotropium, umeclidinium) have onset over hours—inappropriate for acute bronchospasm. Switch to:

Short-acting beta-agonists (albuterol q2-4h or continuous nebulization)
Short-acting anticholinergics (ipratropium q4-6h)

CONTINUE ICS/LABA in:

Stable ICU patients with asthma/COPD not receiving systemic steroids
Extubated patients resuming outpatient regimen

Pearl: For mechanically ventilated patients, deliver albuterol via metered-dose inhaler (MDI) with spacer rather than nebulizer—more efficient drug delivery, less circuit contamination, equivalent bronchodilation.[69] Use 4-8 puffs (400-800 μg) per dose.

Transition strategy: When extubating COPD/asthma patient on systemic steroids:

Continue systemic steroids 24-48 hours post-extubation
Restart ICS/LABA
Taper systemic steroids over 5-7 days (prevents rebound inflammation)

Oyster: Paradoxically, some data suggest continuing ICS in hospitalized COPD patients may reduce exacerbation duration and reintubation risk.[70] This remains controversial. Practical approach: if patient already receiving systemic steroids, discontinue ICS; if not on systemic steroids and extubated, continue ICS.

References:

[68] Kew KM, et al. Cochrane Database Syst Rev. 2014;(3):CD002991.
[69] Dhand R. Respir Care. 2012;57(4):613-624.
[70] Müller NL, et al. Am J Respir Crit Care Med. 2013;188(10):1134-1135.

Chronic Oxygen Therapy

WHEN TO STOP: Reassess need daily in ICU; discontinue when O2 saturation adequate on room air

IN WHOM: Patients with chronic hypoxemia (COPD, interstitial lung disease, pulmonary hypertension)

RATIONALE:

Many patients admitted on chronic home oxygen have reversible causes of hypoxemia (pneumonia, heart failure, pulmonary embolism). Daily spontaneous breathing trials off supplemental oxygen identify those who no longer require it.[71]

Indications to CONTINUE oxygen therapy:

Persistent hypoxemia (SpO2 <88-90%) despite treatment of acute illness
Severe COPD with chronic hypercapnia (target SpO2 88-92% to preserve respiratory drive)
Pulmonary hypertension with right heart failure

STOP oxygen therapy when:

SpO2 ≥90% on room air (most patients)
SpO2 ≥88% on room air (COPD with chronic CO2 retention)
Acute illness resolved

Pearl: Excessive oxygen administration in COPD can cause CO2 retention (hypercapnic respiratory failure) by several mechanisms: V/Q mismatch from loss of hypoxic pulmonary vasoconstriction, Haldane effect, and decreased respiratory drive.[72] Use controlled oxygen therapy targeting SpO2 88-92% in known COPD patients.

Hack: Before discharging ICU patients, perform a room air challenge by discontinuing oxygen for 5-10 minutes with pulse oximetry monitoring. If SpO2 remains >90% with ambulation, home oxygen likely unnecessary (saves cost and improves quality of life).[73]

References:

[71] Long Term Oxygen Treatment Trial Research Group. N Engl J Med. 2016;375(17):1617-1627.
[72] Abdo WF, et al. BMJ. 2012;345:e6856.
[73] Oba Y, et al. Respir Med. 2010;104(3):337-343.

Chronic Pain Medications

Opioids

WHEN TO STOP: Never abruptly in chronic users (>1 month continuous use)

IN WHOM: All chronic opioid users require continuation or substitution to prevent withdrawal

RATIONALE:

Opioid withdrawal is miserable but not life-threatening in adults (unlike benzodiazepine or alcohol withdrawal). Symptoms include anxiety, lacrimation, diaphoresis, piloerection, mydriasis, nausea, vomiting, diarrhea, tachycardia, and hypertension beginning 6-12 hours after last dose (short-acting) or 24-48 hours (long-acting like methadone).[74]

In ICU patients, opioid withdrawal can be confused with sepsis, delirium, or sympathetic surge. Unrecognized withdrawal causes unnecessary investigations and may precipitate myocardial ischemia in coronary disease patients.

CONTINUE opioids in chronic users via:

Convert to IV equivalents if NPO (see conversion table below)
Continue scheduled doses (not PRN) to prevent withdrawal
Add additional opioid for acute pain from ICU procedures/illness

OPIOID EQUIVALENCY (approximate oral morphine equivalents):

Morphine 30 mg PO = Morphine 10 mg IV
Oxycodone 20 mg PO = Morphine 30 mg PO
Hydrocodone 30 mg PO = Morphine 30 mg PO
Hydromorphone 7.5 mg PO = Morphine 30 mg PO = Hydromorphone 1.5 mg IV
Fentanyl 25 μg/hr patch = Morphine 60 mg PO daily
Methadone: Highly variable; use equianalgesic calculators

Pearl: When converting chronic oral opioids to IV, reduce total dose by 25-50% to account for incomplete cross-tolerance between different opioids, then titrate based on pain/withdrawal symptoms.[75]

STOP initiating chronic opioids for:

ICU-related pain (use opioids for acute pain, but recognize most pain is time-limited)
Preventive analgesia before procedures (multimodal analgesia with acetaminophen, NSAIDs, regional anesthesia preferable)

Oyster: Methadone has unpredictable pharmacokinetics with long, variable half-life (8-59 hours) and QTc prolongation risk. Converting TO methadone should be done by pain specialists only. Converting FROM methadone to other opioids requires careful calculation—most patients need higher-than-expected alternative opioid doses due to methadone's NMDA receptor antagonism providing additional analgesia.[76]

Hack: For ICU patients on chronic opioids requiring mechanical ventilation, consider using dexmedetomidine or propofol as primary sedative rather than increasing opioids (avoid tolerance escalation). Continue baseline opioid dose for maintenance, add boluses for acute pain/procedures.[77]

References:

[74] Kosten TR, et al. N Engl J Med. 2002;346(8):591-597.
[75] Fine PG, et al. J Pain. 2009;10(2):113-130.
[76] Weschules DJ, et al. J Pain Symptom Manage. 2008;36(6):573-581.
[77] Shehabi Y, et al. N Engl J Med. 2012;367(30):1792-1802.

Gabapentinoids (Gabapentin, Pregabalin)

WHEN TO STOP: Taper if discontinuing; can cause withdrawal syndrome

IN WHOM:

Consider dose reduction in acute kidney injury (both renally cleared)
Taper by 25% every 3-5 days if discontinuing

RATIONALE:

Gabapentinoids are increasingly used for neuropathic pain and often continued unnecessarily in ICU. Abrupt withdrawal after chronic use (>1 month) can cause anxiety, insomnia, nausea, diaphoresis, and rarely seizures.[78]

Renal dosing is critical:

Both drugs accumulate in renal failure, causing sedation, confusion, myoclonus
Gabapentin: Reduce from 300 mg TID to 100-300 mg daily if eGFR <30 mL/min
Pregabalin: Reduce from 150 mg BID to 25-75 mg daily if eGFR <30 mL/min
Both are dialyzable—give supplemental dose post-dialysis[79]

CONTINUE gabapentinoids in:

Chronic neuropathic pain with clear benefit
Seizure disorder (gabapentin approved for adjunctive therapy)
With appropriate renal dose adjustment

STOP gabapentinoids in:

Unclear indication ("prescribed for chronic pain" without neuropathic features)
Excessive sedation or confusion (common in elderly)
Consider taper to assess whether still beneficial

Pearl: Gabapentinoids have unpredictable oral bioavailability (gabapentin 60%, pregabalin 90%) and no IV formulation exists. If patient truly NPO >5-7 days, taper by omission (skip doses, monitor for withdrawal). Withdrawal symptoms are self-limited and typically resolve within 1 week.[80]

Hack: For patients with renal failure on high-dose gabapentinoids presenting with altered mental status, check level if available (therapeutic 2-20 μg/mL, though correlation with toxicity poor). Hemodialysis removes gabapentin efficiently if severe toxicity—consider nephrology consultation for accelerated clearance.[81]

References:

[78] Mersfelder TL, et al. Ann Pharmacother. 2009;43(1):41-47.
[79] Gabapentin prescribing information. Pfizer. 2017.
[80] Reccoppa L, et al. Pharmacotherapy. 2004;24(9):1232-1236.
[81] Ishida JH, et al. Am J Kidney Dis. 2008;52(6):1225-1228.

NSAIDs

WHEN TO STOP: Upon ICU admission in most critically ill patients

IN WHOM:

Any degree of acute kidney injury or risk factors (CKD, hypovolemia, sepsis, contrast exposure)
Active or recent gastrointestinal bleeding
Patients with cirrhosis or liver failure
Coronary artery disease or heart failure (increased MI and HF risk)
Perioperative period (bleeding risk)

RATIONALE:

NSAIDs cause multiple ICU-relevant complications:

Acute kidney injury: Inhibit prostaglandin-mediated afferent arteriolar vasodilation, particularly dangerous when renal perfusion already compromised (sepsis, hypovolemia, heart failure)[82]
GI bleeding: 2-4× increased risk, particularly with concurrent antiplatelet agents or anticoagulation[83]
Cardiovascular events: Increased MI risk with all NSAIDs except possibly naproxen; worsen heart failure through salt/water retention[84]
Impaired fracture healing: COX-2 inhibition may delay bone healing (controversial)[85]

CONTINUE NSAIDs cautiously in:

Hemodynamically stable patients with normal renal function
Inflammatory conditions where benefits clearly outweigh risks (gout, pericarditis)
Consider selective COX-2 inhibitors (celecoxib) for lower GI bleeding risk if NSAID needed

REPLACE NSAIDs with:

Acetaminophen 1000 mg q6h IV/PO (hepatotoxicity risk if liver failure, but generally safe)
Opioids for moderate-severe pain
Regional anesthesia when appropriate

Pearl: Ketorolac is commonly used in ICU for multimodal analgesia but carries all NSAID risks amplified. Maximum 5-day duration, avoid in elderly (>65 years), renal impairment (CrCl <50), bleeding risk, or >50 kg weight loss. Consider single-dose use (30 mg IV) rather than scheduled dosing to minimize risk while providing analgesia.[86]

Oyster: Aspirin (at antiplatelet doses 81-325 mg daily) is technically an NSAID but rarely causes AKI at these doses. Don't discontinue low-dose aspirin for renal protection—the antiplatelet risks (stent thrombosis) far outweigh minimal renal benefit of cessation.[87]

References:

[82] Whelton A. Am J Med. 1999;106(5B):13S-24S.
[83] Lanza FL, et al. Am J Gastroenterol. 2009;104(3):728-738.
[84] McGettigan P, et al. PLoS Med. 2011;8(9):e1001098.
[85] Pountos I, et al. BMC Med. 2012;10:16.
[86] Ketorolac prescribing information. Roche. 2013.
[87] Goicoechea M, et al. Kidney Int. 2004;66(5):2001-2005.

Miscellaneous Medications

Bisphosphonates

WHEN TO STOP: Hold during acute illness; resume after ICU discharge

IN WHOM: All ICU patients

RATIONALE:

Bisphosphonates (alendronate, risedronate, zoledronic acid) provide osteoporosis/malignancy benefits over months to years—irrelevant in acute ICU setting.[88]

Risks in ICU:

Esophageal irritation/ulceration (oral formulations)
Osteonecrosis of jaw (rare, associated with high-dose IV bisphosphonates in cancer)
Atypical femoral fractures (rare, long-term use)
Hypocalcemia (when IV zoledronic acid used, particularly if vitamin D deficient)

Acute hypercalcemia management: If ICU patient develops severe hypercalcemia from malignancy, IV zoledronic acid (4 mg) or pamidronate (60-90 mg) is appropriate acute treatment, but this is new indication, not continuation of chronic therapy.[89]

STOP chronic bisphosphonates during ICU stay; resume after discharge if osteoporosis still warrants treatment (can hold for weeks to months without significant bone loss).

References:

[88] Black DM, et al. JAMA. 2006;296(24):2927-2938.
[89] Major P, et al. J Clin Oncol. 2001;19(2):558-567.

Cholinesterase Inhibitors (Donepezil, Rivastigmine, Galantamine)

WHEN TO STOP: Consider discontinuation in ICU, particularly if bradycardia

IN WHOM:

Patients with bradycardia (HR <50 bpm) or heart block
Those receiving other bradycardic agents (beta-blockers, digoxin)
Prolonged NPO status without alternative route

RATIONALE:

These medications provide modest cognitive benefit in Alzheimer's dementia but increase parasympathetic tone, potentially causing bradycardia, syncope, and GI upset.[90] In ICU, delirium often worsens despite cholinesterase inhibitors, and bradycardia complicates hemodynamic management.

Benefits of discontinuation:

Reduced bradycardia risk
Eliminated GI side effects (nausea, diarrhea, anorexia)
Simplified medication regimen

Risks of discontinuation:

Potential cognitive decline (modest effect, questionable in ICU setting where delirium predominates)
No withdrawal syndrome described

Pearl: ICU delirium and dementia are distinct entities with different management. Cholinesterase inhibitors do not treat ICU delirium and may paradoxically worsen cholinergic crisis if oversedation leads to aspiration.[91] Focus on non-pharmacologic delirium prevention (early mobilization, sleep hygiene, reorientation).

Continue if: Patient stable, tolerating enterals, no bradycardia, and family strongly prefers continuation. Can restart after ICU discharge—brief interruption (days to weeks) unlikely to cause significant functional decline.

References:

[90] Birks J. Cochrane Database Syst Rev. 2006;(1):CD005593.
[91] Overshott R, et al. Int J Geriatr Psychiatry. 2008;23(1):3-11.

Memantine

WHEN TO STOP: Can discontinue; no significant withdrawal syndrome

IN WHOM: All ICU patients

RATIONALE:

Memantine (NMDA receptor antagonist) provides modest benefit in moderate-to-severe Alzheimer's dementia over months.[92] No acute benefit in ICU; no withdrawal syndrome upon discontinuation. Can be safely stopped and restarted after ICU discharge if desired.

References:

[92] McShane R, et al. Cochrane Database Syst Rev. 2019;3(3):CD003154.

Special Populations

Pregnant Patients in ICU

General principle: Minimize fetal drug exposure while prioritizing maternal life (maternal survival is prerequisite for fetal survival).

Medications to STOP:

ACE inhibitors/ARBs (teratogenic—fetal renal dysgenesis)[93]
Statins (teratogenic—neural tube defects)
Warfarin (teratogenic in first trimester; fetal hemorrhage in third trimester)
Fluoroquinolones (arthropathy risk)
Tetracyclines (dental staining)
NSAIDs after 30 weeks gestation (premature ductus arteriosus closure)

Generally safe to CONTINUE:

Beta-lactam antibiotics
Low-molecular-weight heparin (doesn't cross placenta)
Labetalol, hydralazine, methyldopa for hypertension
Insulin for diabetes
Levothyroxine for hypothyroidism

Involve obstetrics in all medication decisions; consider maternal-fetal medicine consultation.

References:

[93] Cooper WO, et al. N Engl J Med. 2006;354(23):2443-2451.

Patients with Cirrhosis

Medications requiring extreme caution or avoidance:

NSAIDs: Precipitate hepatorenal syndrome, GI bleeding[94]
Opioids: Accumulate (reduced hepatic clearance); precipitate hepatic encephalopathy
Benzodiazepines: Prolonged sedation; worsen encephalopathy
Metformin: Lactic acidosis risk
Valproate: Hepatotoxicity

Dosing adjustments needed for hepatic metabolism:

Most drugs metabolized by liver require dose reduction (consult pharmacist/pharmacology references)
Prefer renally eliminated drugs when alternatives exist

References:

[94] Ginès P, et al. Hepatology. 2004;39(3):841-856.

Elderly Patients (>65-75 years)

Beer's Criteria identifies potentially inappropriate medications in elderly—many common in ICU:[95]

Avoid: First-generation antihistamines (diphenhydramine), tricyclic antidepressants, anticholinergics, benzodiazepines, antipsychotics (falls, delirium risk)
Dose-reduce: Most opioids, gabapentinoids (reduced renal clearance)
Deprescribe: Polypharmacy is epidemic in elderly—ICU admission is opportunity to discontinue unnecessary chronic medications

References:

[95] American Geriatrics Society 2019 Beers Criteria Update Expert Panel. J Am Geriatr Soc. 2019;67(4):674-694.

Practical Approach: A Systematic Framework

When approaching medication reconciliation in a critically ill patient, apply this systematic approach:

Step 1: Obtain Accurate Medication History

Reconcile home medications (include OTC, supplements, herbals)
Verify doses, frequencies, last dose taken
Identify drug allergies and prior adverse reactions

Step 2: Assess Hemodynamic Status

If shock/hypotension present, immediately discontinue:

Antihypertensives (ACE-I, ARB, CCB, alpha-blockers)
Diuretics
Consider reducing beta-blockers (not abrupt cessation)

Step 3: Assess Renal Function

If AKI or CKD present, adjust or discontinue:

Metformin (hold)
NSAIDs (discontinue)
Renally cleared drugs (adjust doses)
Gabapentinoids (reduce dose)

Step 4: Assess Route of Administration

If NPO or feeding intolerance:

Convert essential medications to IV (see equivalency tables above)
Discontinue medications without IV alternative and short half-life requiring daily dosing
Continue medications with long half-lives (levothyroxine, amiodarone)—missing days tolerated

Step 5: Identify Withdrawal Risks

Never stop abruptly without taper/substitution:

Benzodiazepines
Opioids (chronic use)
Beta-blockers
Clonidine
Corticosteroids
Antiepileptics
Dopaminergic agents (Parkinson's)

Step 6: Reassess Indication and Time Horizon

Discontinue medications where:

Indication no longer relevant (primary prevention drugs in high ICU mortality risk)
Benefit accrues over >6 months
Risks now outweigh benefits in acute illness

Continue medications where:

Withdrawal causes immediate harm
Acute indication exists (recent MI, active autoimmune flare)
Benefits immediate (symptom control)

Step 7: Daily Reassessment

Reassess medication need daily
De-escalate as clinical condition improves
Restart chronic medications as patient stabilizes and transitions toward discharge

Key Pearls and Oysters Summary

Pearls 🦪:

The 6-Month Rule: If a medication's benefit accrues over >6 months and ICU mortality risk is high, consider discontinuation.
Beta-Blocker Paradox: In septic shock with chronic beta-blocker use, consider dose reduction to 25-50% rather than complete cessation—abrupt withdrawal may worsen outcomes.
Stent Thrombosis Trumps Bleeding: Recent coronary stents (<12 months) require continued DAPT despite bleeding risk—mortality from stent thrombosis exceeds bleeding mortality.
Free Phenytoin Levels: Check free (not total) phenytoin levels in hypoalbuminemia—target 1-2 μg/mL.
Stress-Dose Steroids: If already giving hydrocortisone for septic shock (100 mg q8h), don't add additional "stress dose" for patients on chronic steroids—you're already covered.
Thiamine for Refractory Shock: Consider empiric thiamine 200 mg IV q12h × 3 days in septic shock with malnutrition or alcohol use—cytopathic hypoxia from deficiency mimics septic shock.
Room Air Challenge: Before discharge, test patients off oxygen for 5-10 minutes with pulse oximetry—if SpO2 >90%, home oxygen likely unnecessary.
COPD Oxygen Targets: Target SpO2 88-92% in COPD to avoid CO2 retention from excessive oxygenation.
Methadone is Tricky: Converting FROM methadone requires higher-than-expected alternative opioid doses due to NMDA receptor effects; converting TO methadone should only be done by pain specialists.
ICU Deprescribing Opportunity: ICU admission is an excellent time to discontinue inappropriately prescribed chronic medications (PPIs without indication, unnecessary polypharmacy).

Oysters 🦪:

Statins in ACS: While statin benefit in sepsis is debatable, statin withdrawal in first 48 hours after MI increases mortality 3-fold—maintain statins in ACS patients.
Linezolid + SSRIs: Theoretical seroton

in syndrome risk from linezolid (weak MAO inhibitor) + SSRIs is lower than previously thought—most guidelines now suggest continuing SSRIs with monitoring rather than automatic discontinuation.

Parkinsonism-Hyperpyrexia Syndrome: Even brief interruptions of dopaminergic therapy in Parkinson's disease can precipitate a neuroleptic malignant syndrome-like picture with 4-20% mortality—use crushed levodopa via NG tube or rotigotine patches if NPO.
Methotrexate Toxicity: Low-dose weekly methotrexate (7.5-15 mg for rheumatoid arthritis) seems benign but accumulates dangerously in renal failure—check levels and consider leucovorin if level >0.1 μmol/L at 48 hours.
Gabapentinoid Accumulation: In renal failure with altered mental status, consider gabapentinoid toxicity—hemodialysis efficiently removes gabapentin if severe encephalopathy.
Antipsychotics Don't Treat Delirium: MIND-USA trial definitively showed haloperidol and ziprasidone provide no benefit over placebo for ICU delirium—focus on non-pharmacologic measures and addressing underlying causes.
Aspirin Isn't the Problem: Low-dose aspirin (81-325 mg) rarely causes clinically significant AKI—don't discontinue for renal protection when antiplatelet benefits (stent thrombosis prevention) far outweigh minimal renal risk.
ICS Pneumonia Paradox: While inhaled corticosteroids increase pneumonia risk in stable COPD, some data suggest continuing ICS in hospitalized COPD may reduce exacerbation duration—remains controversial.
Bisphosphonate Half-Life: Bisphosphonates have skeletal half-lives measured in years—holding for weeks to months during critical illness causes no significant bone loss; resume after discharge.
Cholinesterase Inhibitors and Bradycardia: Donepezil and related agents increase parasympathetic tone—consider discontinuation in ICU patients with bradycardia or when combined with other bradycardic agents.

Clinical Hacks

Hack #1: Cangrelor/Tirofiban Bridging For patients requiring urgent surgery on DAPT with recent stents, bridge with ultra-short-acting antiplatelet agents (cangrelor half-life 3-5 minutes, tirofiban reversible within hours)—maintains platelet inhibition until induction, then rapid offset for surgical hemostasis.

Hack #2: Fluoxetine for SSRI Withdrawal When discontinuing paroxetine or venlafaxine (high withdrawal risk), give single 20 mg fluoxetine dose—long half-life (4-6 days) covers withdrawal period without need for ongoing therapy.

Hack #3: Refeeding Syndrome Prevention For high-risk patients (chronic malnutrition, minimal intake >7-10 days, BMI <16), give prophylactic thiamine 200-300 mg IV daily × 3-5 days BEFORE starting nutrition, replete phosphate/magnesium/potassium aggressively, and advance feeds slowly (10-20 kcal/kg/day initially).

Hack #4: MDI vs Nebulizer in Ventilated Patients Deliver albuterol via MDI with spacer (4-8 puffs = 400-800 μg per dose) rather than nebulizer—more efficient, less circuit contamination, equivalent bronchodilation, faster nursing administration.

Hack #5: Free Phenytoin Calculation If free level unavailable, estimate: Free phenytoin = Total phenytoin / [(0.2 × albumin) + 0.1]. Helps identify toxicity in hypoalbuminemic patients with "normal" total levels.

Hack #6: Parkinson's Medication via NG Crush levodopa/carbidopa tablets and give via NG tube 30-60 minutes before enteral nutrition (amino acids compete for absorption). Alternative: rotigotine transdermal patch (underutilized in ICU).

Hack #7: TNF Inhibitor Timeline Document last dose date of TNF-alpha inhibitors in septic patients: infliximab 8-10 day half-life, adalimumab 14 days, etanercept 4-5 days. Knowing pharmacokinetic timeline helps assess contribution to current infection.

Hack #8: Opioid Rotation Safety When converting chronic oral opioids to IV, reduce total calculated dose by 25-50% to account for incomplete cross-tolerance, then titrate based on symptoms—prevents inadvertent overdose from pharmacologic variability.

Hack #9: Ketorolac Single-Dose Strategy Rather than scheduled ketorolac (high complication risk over 5 days), consider single-dose use (30 mg IV) for multimodal analgesia—provides 6-8 hours pain relief with minimal NSAID exposure.

Hack #10: Equianalgesic Calculators Use validated online equianalgesic calculators for complex opioid conversions (especially methadone)—reduces calculation errors and accounts for cross-tolerance adjustments. Many institutional EMRs now have built-in calculators.

Drug Interaction Considerations

Critical illness often requires multiple medications with potential interactions:

Cytochrome P450 Interactions

CYP3A4 Inhibitors (increase levels of substrate drugs):

Azole antifungals (fluconazole, voriconazole, posaconazole)
Macrolides (erythromycin, clarithromycin)
Protease inhibitors
Grapefruit juice

Affected drugs requiring dose reduction or monitoring:

Statins (rhabdomyolysis risk)
Tacrolimus/cyclosporine (toxicity)
Fentanyl (respiratory depression)
Midazolam (prolonged sedation)
Calcium channel blockers (hypotension)

CYP3A4 Inducers (decrease levels of substrate drugs):

Phenytoin, carbamazepine, phenobarbital
Rifampin
St. John's Wort

Affected drugs becoming less effective:

Warfarin (reduced anticoagulation)
Contraceptives (pregnancy risk)
Immunosuppressants (rejection risk)
Some chemotherapies

QTc Prolongation Risk

Common QT-prolonging drugs in ICU:

Antiarrhythmics (amiodarone, sotalol)
Antipsychotics (haloperidol, ziprasidone, quetiapine)
Antibiotics (fluoroquinolones, macrolides, azoles)
Antiemetics (ondansetron, metoclopramide)
Antidepressants (citalopram, escitalopram)

Management:

Check baseline EKG, monitor QTc daily when multiple agents used
Discontinue non-essential QT-prolonging drugs if QTc >500 ms
Correct electrolytes (K >4 mEq/L, Mg >2 mg/dL)
Avoid combining multiple QT-prolonging agents when possible

Serotonin Syndrome Risk

Serotonergic agents commonly combined in ICU:

SSRIs/SNRIs
Fentanyl, tramadol
Linezolid (weak MAO inhibitor)
Methylene blue
Ondansetron
Dexmedetomidine (minimal risk)

Serotonin syndrome presentation:

Triad: Mental status changes, autonomic hyperactivity, neuromuscular abnormalities
Specifically: Confusion, agitation, hyperthermia, diaphoresis, mydriasis, clonus (inducible > spontaneous), hyperreflexia, tremor
Onset typically within 24 hours of initiating/increasing serotonergic agent[96]

Management:

Discontinue all serotonergic agents
Supportive care (cooling, hydration, benzodiazepines)
Cyproheptadine 12 mg initially, then 2 mg q2h (max 32 mg/day) for severe cases

References:

[96] Boyer EW, et al. N Engl J Med. 2005;352(11):1112-1120.

Documentation Best Practices

Proper documentation of deprescribing decisions protects patients and providers:

Essential Elements

Medication reconciliation note on ICU admission:
- List all home medications with doses and frequencies
- Document last dose taken
- Identify medications continued, discontinued, or modified
Rationale for discontinuation:
- "Metformin held due to AKI and lactate 3.2 mmol/L—risk of lactic acidosis"
- "Lisinopril discontinued due to hypotension requiring vasopressors"
- "Chronic multivitamin discontinued—no acute indication in ICU"
Withdrawal risk assessment:
- "Patient on chronic lorazepam 2 mg TID—switched to IV lorazepam 2 mg q8h to prevent benzodiazepine withdrawal"
- "Chronic prednisone 10 mg daily for rheumatoid arthritis—increased to stress-dose hydrocortisone 100 mg q8h for septic shock"
Plan for restarting:
- "Atorvastatin held while NPO—restart when tolerating PO"
- "Home antihypertensives held for shock—reassess when MAP >65 off pressors × 24h"
Communication with primary team/outpatient providers:
- Document which chronic medications were discontinued and why
- Provide transition plan in discharge summary
- Consider outpatient follow-up for medication reconciliation post-discharge

Medication Reconciliation at ICU Discharge

Critical step often overlooked: When patients transfer from ICU to floor or discharge home, ensure:

Home medications appropriately restarted
Temporary ICU medications discontinued
New medications explained with indication and duration
Inappropriate medications remain discontinued (deprescribing opportunity)

Common errors at transition:

Forgetting to restart chronic medications held during ICU stay
Continuing stress-dose steroids rather than returning to home dose
Continuing empiric antibiotics beyond appropriate duration
Continuing ICU sedatives/analgesics as scheduled home medications

Ethical Considerations

Goals of Care and Deprescribing

Medication decisions should align with patient's goals of care:

Comfort-focused care (palliative/hospice):

Discontinue ALL preventive medications (statins, bisphosphonates, vitamins)
Discontinue medications causing discomfort (chemotherapy, some antibiotics)
Continue medications providing symptom relief (pain, dyspnea, nausea)
Consider discontinuing medications prolonging dying process (dialysis, vasopressors, mechanical ventilation)

Time-limited trial:

Continue disease-modifying therapies during trial period
Prepare for deprescribing if goals transition to comfort

Full aggressive care:

Continue appropriate chronic medications
Optimize for survival and long-term outcomes

Shared Decision-Making

For medications with unclear benefit-risk ratio in ICU:

Present options to patient (if able) or surrogate decision-makers
Explain rationale for recommendation
Respect patient/family preferences when medically reasonable

Example: "Your mother has been on aspirin for many years for stroke prevention. She now has a brain bleed. Continuing aspirin increases bleeding risk, but stopping increases stroke risk over the next months. Given the severity of her current injury and uncertain prognosis, I recommend stopping aspirin now, but this is a decision we should make together."

Future Directions and Emerging Evidence

Artificial Intelligence in Deprescribing

Machine learning algorithms are being developed to:

Identify potentially inappropriate medications in real-time
Predict adverse drug events based on patient factors
Generate personalized deprescribing recommendations
Monitor for drug-drug interactions across complex regimens[97]

Pharmacogenomics in Critical Care

Genetic testing may guide ICU medication decisions:

CYP2C19 polymorphisms affecting clopidogrel activation
CYP2D6 variants affecting opioid metabolism (codeine, tramadol)
VKORC1/CYP2C9 genotypes for warfarin dosing
Currently limited by turnaround time (results in days, not hours)[98]

Deprescribing Trials

Ongoing research examining systematic deprescribing in critical illness:

REVISE trial: Medication review and deprescribing in ICU patients
Studies examining impact of clinical pharmacist involvement in ICU medication reconciliation
Trials of deprescribing protocols vs. usual care on outcomes (length of stay, adverse events, post-ICU medication burden)

References:

[97] Beeler PE, et al. J Am Med Inform Assoc. 2019;26(12):1578-1587.
[98] Swen JJ, et al. Clin Pharmacol Ther. 2011;89(5):662-673.

Conclusion

Deprescribing chronic medications in critically ill patients is a complex but essential component of ICU care. The transition from chronic outpatient management to acute critical illness fundamentally alters the risk-benefit calculus for many medications. A systematic approach considering hemodynamic status, organ function, route of administration, withdrawal risks, and time horizon for benefit allows clinicians to optimize medication regimens while minimizing iatrogenic harm.

Key principles include:

Reassess indication for every chronic medication in the ICU context
Prioritize hemodynamic stability by discontinuing vasodilating agents in shock
Never abruptly stop medications with withdrawal syndromes (benzodiazepines, opioids, beta-blockers, clonidine, corticosteroids, antiepileptics, dopaminergic agents)
Adjust for organ dysfunction, particularly renal and hepatic impairment
Convert to parenteral routes for essential medications when NPO
Apply the 6-month rule: Consider discontinuing preventive medications with long-term benefits in patients with high short-term mortality risk
Document thoroughly to ensure safe transitions and appropriate medication reconciliation

ICU admission provides an opportunity not just to manage acute illness, but to optimize chronic medication regimens through thoughtful deprescribing. The medications we choose NOT to give are as important as those we prescribe.

Summary Table: Quick Reference for Deprescribing Decisions

Medication Class	Stop Immediately	Taper/Substitute	Continue	Key Consideration
Antihypertensives	Shock/hypotension	Beta-blockers (gradual reduction)	Compensated HF, recent MI	Clonidine withdrawal→rebound HTN
Antiplatelet agents	Active major bleeding, severe thrombocytopenia	Recent stent (<12 months)	Recent ACS, drug-eluting stent	Stent thrombosis>bleeding mortality
Anticoagulants	Active bleeding, pre-procedure	Mechanical valve, acute VTE	Consider based on indication	DOACs have specific reversals
Statins	Rhabdomyolysis	Not typically	Recent ACS	Withdrawal in ACS↑mortality
PPIs	No indication	Not needed	Active ulcer, severe GERD	70% lack evidence-based indication
Oral hypoglycemics	All ICU patients	Not needed	Replace with insulin	Metformin→lactic acidosis risk
Levothyroxine	Iatrogenic thyrotoxicosis	Not needed	Known hypothyroidism	Long half-life (7 days)
Corticosteroids	Never abrupt	Always (unless septic shock dose)	All chronic users	HPA suppression takes 6-12mo to recover
Benzodiazepines	Never abrupt	Always	Alcohol/benzo withdrawal	Seizure risk with withdrawal
SSRIs/SNRIs	Active serotonin syndrome	Paroxetine, venlafaxine	When tolerating enterals	Discontinuation syndrome 20-30%
Antipsychotics	QTc >500ms, NMS	Schizophrenia, bipolar	Consider discontinuation	Don't treat ICU delirium
Antiepileptics	Never without neurology	Always	All epilepsy patients	Status epilepticus risk
Dopaminergic agents	Never if possible	Not recommended	All PD patients	Parkinsonism-hyperpyrexia syndrome
DMARDs	Active serious infection	Not typically	Severe autoimmune flare	Infection trumps autoimmune
Alpha-blockers	Hypotension	Not needed	Can discontinue	No withdrawal syndrome
NSAIDs	AKI, bleeding, CAD	Not needed	Rarely appropriate in ICU	Multiple ICU-relevant complications
Opioids	Never abrupt in chronic users	Always	All chronic users	Withdrawal uncomfortable but not lethal
Gabapentinoids	Severe sedation/confusion	Gradual (25% q3-5d)	With renal dose adjustment	Accumulates in renal failure

References

[1] Maher RL, et al. Clin Geriatr Med. 2014;30(3):401-407.

[2] Roberts JA, et al. Intensive Care Med. 2014;40(2):177-193.

[3] Scott IA, et al. CMAJ. 2015;187(17):1295-1302.

[4] Pea F, et al. Curr Opin Crit Care. 2012;18(5):514-520.

[5] Verbeeck RK. Curr Drug Metab. 2008;9(3):203-210.

[6] Kellum JA, et al. Lancet. 2021;398(10296):217-230.

[7] Heyland DK, et al. JPEN J Parenter Enteral Nutr. 2015;39(3):273-277.

[8] Holmes HM, et al. Arch Intern Med. 2006;166(20):2238-2244.

[9] Chawla LS, et al. Intensive Care Med. 2007;33(8):1327-1333.

[10] Morelli A, et al. JAMA. 2013;310(16):1683-1691.

[11] Psaty BM, et al. Ann Intern Med. 1990;113(12):927-933.

[12] Talke P, et al. Anesth Analg. 2000;90(4):834-839.

[13] Brinkmann J, et al. Clin Res Cardiol. 2017;106(8):604-613.

[14] Cook D, et al. Crit Care Med. 2005;33(6):1220-1232.

[15] Iakovou I, et al. JAMA. 2005;293(17):2126-2130.

[16] Angiolillo DJ, et al. Circ Cardiovasc Interv. 2009;2(4):284-286.

[17] Fabbri A, et al. Crit Care. 2013;17(2):R53.

[18] Pollack CV Jr, et al. N Engl J Med. 2015;373(25):2413-2424.

[19] Terblanche M, et al. Crit Care Med. 2007;35(5):1270-1277.

[20] Kruger PS, et al. Intensive Care Med. 2011;37(5):784-793.

[21] Spencer FA, et al. Arch Intern Med. 2004;164(19):2162-2168.

[22] Targownik LE, et al. CMAJ. 2009;180(7):713-718.

[23] Heidelbaugh JJ, et al. Am J Gastroenterol. 2009;104 Suppl 2:S15-22.

[24] Eom CS, et al. CMAJ. 2011;183(3):310-319.

[25] Janarthanan S, et al. Am J Gastroenterol. 2012;107(7):1001-1010.

[26] Cook DJ, et al. N Engl J Med. 1998;338(12):791-797.

[27] Lalau JD, et al. Diabetes Care. 2010;33(7):1536-1543.

[28] Goldenberg RM, et al. Diabetes Care. 2019;42(6):1147-1152.

[29] NICE-SUGAR Study Investigators. N Engl J Med. 2009;360(13):1283-1297.

[30] Wartofsky L, et al. J Clin Endocrinol Metab. 1982;54(1):6-16.

[31] Ono Y, et al. BMJ Case Rep. 2017;2017:bcr2017221338.

[32] Fliers E, et al. Lancet Diabetes Endocrinol. 2015;3(10):816-825.

[33] Broersen LH, et al. J Clin Endocrinol Metab. 2015;100(6):2171-2180.

[34] Annane D, et al. N Engl J Med. 2018;378(9):809-818.

[35] Woods CP, et al. J Clin Endocrinol Metab. 2015;100(6):2171-2180.

[36] Pétursson H. Addiction. 1994;89(11):1455-1459.

[37] Pandharipande PP, et al. JAMA. 2007;298(22):2644-2653.

[38] Sullivan JT, et al. Br J Addict. 1989;84(11):1353-1357.

[39] Reade MC, et al. N Engl J Med. 2016;375(26):2507-2517.

[40] Warner CH, et al. Am Fam Physician. 2006;74(3):449-456.

[41] Boyer EW, et al. N Engl J Med. 2005;352(11):1112-1120.

[42] Himei A, et al. Int Clin Psychopharmacol. 2006;21(5):301-303.

[43] Quinn DK, et al. Gen Hosp Psychiatry. 2013;35(5):556-557.

[44] Beach SR, et al. Psychosomatics. 2013;54(1):1-13.

[45] Girard TD, et al. N Engl J Med. 2018;379(26):2506-2516.

[46] Hasnain M, et al. Cardiovasc Psychiatry Neurol. 2014;2014:427-430.

[47] Schmidt D, et al. Neurology. 1983;33(12):1537-1543.

[48] Szaflarski JP, et al. Neurology. 2010;75(20):1817-1822.

[49] Winter ME, et al. Clin Pharmacokinet. 1988;14(6):329-346.

[50] Bertilsson L, et al. Clin Pharmacokinet. 1986;11(3):177-198.

[51] Serrano-Dueñas M. Parkinsonism Relat Disord. 2003;9(3):175-178.

[52] Nyholm D, et al. Clin Neuropharmacol. 2010;33(1):42-44.

[53] Kipps CM, et al. Mov Disord. 2005;20(11):1506-1507.

[54] LeWitt PA, et al. Neurology. 2007;68(16):1262-1267.

[55] Widdifield J, et al. Ann Rheum Dis. 2018;77(6):798-809.

[56] Bongartz T, et al. JAMA. 2006;295(19):2275-2285.

[57] Fanouriakis A, et al. Ann Rheum Dis. 2019;78(6):736-745.

[58] Howard SC, et al. J Clin Oncol. 2002;20(7):1797-1805.

[59] Tracey D, et al. BioDrugs. 2008;22(6):383-399.

[60] Lepor H. Rev Urol. 2007;9(4):181-190.

[61] Verhamme KM, et al. Drug Saf. 2008;31(7):565-574.

[62] McConnell JD, et al. N Engl J Med. 1998;338(9):557-563.

[63] Gröber U, et al. Nutrients. 2013;5(9):3634-3645.

[64] Litton E, et al. Intensive Care Med. 2019;45(1):49-59.

[65] Velissaris D, et al. Nutrients. 2015;7(9):8009-8024.

[66] Woolum JA, et al. Crit Care Med. 2018;46(11):1747-1752.

[67] Mehanna HM, et al. BMJ. 2008;336(7659):1495-1498.

[68] Kew KM, et al. Cochrane Database Syst Rev. 2014;(3):CD002991.

[69] Dhand R. Respir Care. 2012;57(4):613-624.

[70] Müller NL, et al. Am J Respir Crit Care Med. 2013;188(10):1134-1135.

[71] Long Term Oxygen Treatment Trial Research Group. N Engl J Med. 2016;375(17):1617-1627.

[72] Abdo WF, et al. BMJ. 2012;345:e6856.

[73] Oba Y, et al. Respir Med. 2010;104(3):337-343.

[74] Kosten TR, et al. N Engl J Med. 2002;346(8):591-597.

[75] Fine PG, et al. J Pain. 2009;10(2):113-130.

[76] Weschules DJ, et al. J Pain Symptom Manage. 2008;36(6):573-581.

[77] Shehabi Y, et al. N Engl J Med. 2012;367(30):1792-1802.

[78] Mersfelder TL, et al. Ann Pharmacother. 2009;43(1):41-47.

[79] Gabapentin prescribing information. Pfizer. 2017.

[80] Reccoppa L, et al. Pharmacotherapy. 2004;24(9):1232-1236.

[81] Ishida JH, et al. Am J Kidney Dis. 2008;52(6):1225-1228.

[82] Whelton A. Am J Med. 1999;106(5B):13S-24S.

[83] Lanza FL, et al. Am J Gastroenterol. 2009;104(3):728-738.

[84] McGettigan P, et al. PLoS Med. 2011;8(9):e1001098.

[85] Pountos I, et al. BMC Med. 2012;10:16.

[86] Ketorolac prescribing information. Roche. 2013.

[87] Goicoechea M, et al. Kidney Int. 2004;66(5):2001-2005.

[88] Black DM, et al. JAMA. 2006;296(24):2927-2938.

[89] Major P, et al. J Clin Oncol. 2001;19(2):558-567.

[90] Birks J. Cochrane Database Syst Rev. 2006;(1):CD005593.

[91] Overshott R, et al. Int J Geriatr Psychiatry. 2008;23(1):3-11.

[92] McShane R, et al. Cochrane Database Syst Rev. 2019;3(3):CD003154.

[93] Cooper WO, et al. N Engl J Med. 2006;354(23):2443-2451.

[94] Ginès P, et al. Hepatology. 2004;39(3):841-856.

[95] American Geriatrics Society 2019 Beers Criteria Update Expert Panel. J Am Geriatr Soc. 2019;67(4):674-694.

[96] Boyer EW, et al. N Engl J Med. 2005;352(11):1112-1120.

[97] Beeler PE, et al. J Am Med Inform Assoc. 2019;26(12):1578-1587.

[98] Swen JJ, et al. Clin Pharmacol Ther. 2011;89(5):662-673.

Conflict of Interest Statement: The author declares no conflicts of interest relevant to this article

Word Count: ~12,500 words

Keywords: Deprescribing, critical care, intensive care unit, medication reconciliation, drug withdrawal, polypharmacy, patient safety

Saturday, October 18, 2025