The Dose She No Longer Needs: Deprescribing Levothyroxine in Older Adults

                                                        Dr Neeraj Manikath DNB

 

A Clinician–Educator Review for Postgraduate Trainees and Practising Internists

 

 

 

The 60-second summary (for those who have a clinic in 5 minutes)

● Levothyroxine is among the most prescribed drugs on earth, and a meaningful minority of older adults taking it never had a durable biochemical indication, or have drifted into iatrogenic subclinical thyrotoxicosis.

● In older adults, a suppressed TSH is more dangerous than a mildly elevated one: it roughly triples atrial fibrillation risk and materially increases fracture risk — while treating mild subclinical hypothyroidism at age ≥65 has never been shown to help (TRUST trial).

● The DINE trial showed that simply halving the dose normalised TSH in two-thirds of suppressed elderly nursing-home residents within 8 weeks, with nobody tipping into hypothyroidism.

● Deprescribing is not withdrawal of therapy — it is right-sizing. The correct move ranges from a 25–50% dose reduction to a monitored full stop, chosen by auditing why the drug was started.

● Remember one mnemonic: when the patient is T.I.R.E.D. of pills — Trace the trigger, Interrogate the trend, Risk-stratify, Eliminate or reduce, Diarise the rechecks.

 

 

 

1. Introduction — The Patient Who Changed My Practice

 

Let me introduce you to the patient who quietly rearranged how I think about thyroid disease in older people.

 

Eileen was 82 when she landed in my falls clinic: four falls in six months, a new diagnosis of atrial fibrillation, a DXA scan showing a hip T-score of −2.9 with a vertebral wedge fracture, 4 kg of weight loss, insomnia, and a fine tremor her family doctor had attributed to "anxiety." Her medication list was unremarkable except for one entry she was fiercely proud of: levothyroxine 125 mcg daily, for nineteen years, never missed a dose.

 

Her TSH that morning was 0.04 mIU/L.

 

Here is the part that should stop you mid-ward-round. When I hunted down the records from 2004, I found a single TSH of 6.9 mIU/L drawn during a workup for fatigue after an influenza-like illness, no confirmatory repeat, no thyroid antibodies, no goitre, no surgery — followed by a GP note reading "borderline thyroid — start thyroxine 50 mcg and titrate to normalise." The dose was titrated. The TSH was normalised. And over nineteen years, "normalised" drifted into suppressed — because dose requirements fall with age, and nobody ever reassessed whether the drug was needed in the first place. Her atrial fibrillation, her fracture, her tremor, and quite possibly her falls all had a plausible common contributor: a perfectly adherent patient taking a dose her body stopped needing years ago.

 

Eileen is not an outlier. Levothyroxine is one of the most prescribed medicines in the world, with well over 100 million prescriptions dispensed annually in the United States alone [1]. Subclinical hypothyroidism — the entity responsible for most "borderline" starts — affects roughly 3–4% of the general European population and up to one in ten older women [2]. Yet the threshold at which we initiate treatment has crept relentlessly downward over the past two decades: Tayside data showed the TSH level at which clinicians started levothyroxine falling substantially over ten years, with the treated population growing several-fold and a meaningful fraction of the newly treated later developing suppressed TSH [3]. Natural-history data from the Whickham cohort remind us why that matters: only a minority of mildly elevated TSH values progress to overt disease — roughly 4% per year in antibody-positive women, and far less in everyone else — while a substantial proportion spontaneously normalise [4].

 

Meanwhile, the benefit side of the ledger has all but evaporated for the age group we're discussing: the TRUST trial randomised 737 adults aged ≥65 with subclinical hypothyroidism (mean TSH ~6.4, three-quarters under 10) to levothyroxine or placebo and found no difference in hypothyroid symptoms or tiredness at one year, extended follow-up across multiple domains showing nothing convincing either [5].

 

🩺 The thesis of this review: In older adults, the most dangerous thyroid number is not the one that's slightly high — it's the one that's low. Deprescribing levothyroxine is not a fringe "less-is-more" fashion; it is an evidence-aligned correction of one of modern medicine's most quietly iatrogenic habits. Done with an audit of the original indication, a clear plan, and scheduled rechecks, it is safe, reversible, and frequently transformative.

 

 

 

2. Pathophysiology — Only the Five Facts That Change Decisions

 

You do not need a thyroid biochemistry lecture at the bedside. You need these five facts, because every clinical decision in this article flows from them.

 

Fact 1 — TSH drifts up with age, and in the oldest old much of that is not disease. Population cohorts show that mild TSH elevation in adults over 80 is frequently thyroid-antibody-negative, progresses slowly, and associates with neutral — even favourable — health outcomes [4,10]. The Leiden 85+ study followed 85-year-olds for years and found that a mildly elevated TSH did not predict disability, cognitive decline, or worse survival [10]. Actionable meaning: a TSH of 5–6 in a robust 88-year-old is a finding, not a diagnosis. Chasing it with dose escalation is how suppression begins.

 

Fact 2 — Older bodies need less levothyroxine, and the requirement keeps falling. Full replacement in a healthy younger adult is roughly 1.6 mcg/kg/day; in adults over 65 it falls to approximately 1.0–1.2 mcg/kg/day, driven by lower lean mass and slower clearance [14]. A dose that was exactly right at age 55 is, at 80, a standing overdose. Actionable meaning: weight loss, sarcopenia, and ageing itself are dose-reduction indications — nobody charts them as such, but physiology does.

 

Fact 3 — A suppressed TSH is end-organ-active, not cosmetic. Thyroid hormone excess shortens atrial action potentials, upregulates Na⁺/K⁺-ATPase and β-adrenergic signalling — electrophysiology that converts into atrial fibrillation. In the Framingham cohort aged ≥60, a TSH ≤0.1 mIU/L carried a 10-year AF incidence of 28% versus 11% — roughly a threefold excess [6]; the Cardiovascular Health Study found subclinical hyperthyroidism approximately doubled AF risk [7]. Bone is the second target: TSH receptors sit on osteoblasts and osteoclasts, and the suppressed-TSH/high-T4 state accelerates bone turnover — the Study of Osteoporotic Fractures linked low TSH in women ≥65 to two- to four-fold increases in hip and vertebral fracture risk [8], confirmed in pooled prospective data showing hip fracture risk roughly 2.4-fold higher with endogenous subclinical hyperthyroidism [9]. Actionable meaning: suppressed TSH in an older adult is a cardiology and bone clinic problem wearing an endocrine lab coat.

 

Fact 4 — Levothyroxine's 7-day half-life makes deprescribing forgiving and slow. Because T4 clearance is leisurely, dose changes take 6–8 weeks to fully register in TSH; levels after a dose reduction fall gradually — there is no crash. The same half-life is why alternate-day dosing with the same tablet is pharmacologically smooth (daily exposure averages out). Actionable meaning: never judge a dose change at 3 weeks, and don't fear the halved dose — physiology is on your side.

 

Fact 5 — Permanent and transient hypothyroidism look identical on day 400 of treatment. Post-thyroidectomy, post-radioiodine, and advanced autoimmune hypothyroidism are permanent; thyroiditis-related, drug-induced (amiodarone, lithium, interferon, kinase inhibitors, checkpoint inhibitors, iodinated contrast), and non-thyroidal-illness-related hypothyroidism are frequently transient. Once both are on levothyroxine 75 mcg, you cannot tell them apart by looking — only by interrogating the origin. This is the fork in the deprescribing road, and we will walk it explicitly.

 

 

 

3. The Five Faces of Levothyroxine in Later Life

 

Every older patient on levothyroxine is one of five phenotypes. Your first job at any review is to identify which one is sitting in front of you.

 

Phenotype	Signature clue	The right move	Watch for
1. True, permanent hypothyroidism	Thyroidectomy, radioiodine, neck radiation, or overt Hashimoto (high-titre TPOAb, TSH >10 at diagnosis)	Continue lifelong; right-size the dose; never let TSH suppress	Dose creep with weight loss
2. True hypothyroidism, currently over-replaced	TSH <0.4 on genuine replacement	Reduce 25–50%, recheck 6–8 weeks	Recurrent suppression years later
3. Treated mild subclinical hypothyroidism, now euthyroid	Small dose (25–50 mcg) started for TSH 5–8 in a ≥65-year-old; TSH now normal	Monitored trial of discontinuation	Mild TSH elevation on follow-up — usually the answer, not a failure
4. Transient cause, long resolved	Started after viral illness/subacute thyroiditis, amiodarone, lithium, contrast, or hospitalisation	Stop, with a scheduled recheck	Late relapse (rare beyond 12 months)
5. Never indicated	TSH was normal at initiation, or a single borderline value never confirmed	Stop — the prescription was treating a lab-report artefact	Nothing thyroid-specific; watch the nocebo effect

 

 

⚠️ The critical reframing: phenotypes 2–5 are not "non-compliant with the plan." The prescription is not the diagnosis — the indication is the diagnosis. If the indication was transient, weak, or absent, continuing the drug is the deviation from correct practice, not stopping it.

 

 

 

4. 🪙 Clinical Pearls — Counterintuitive Bedside Observations

 

🪙 Pearl 1 — In geriatrics, the dangerous number is the low one. Every clinician is trained to flinch at a TSH of 8 and relax at a TSH of 0.05. Invert that reflex. The elevated TSH in an asymptomatic 80-year-old has a slow, largely benign natural history [4,5]; the suppressed TSH carries measurable AF and fracture risk this year [6–9]. A TSH of 0.05 is not "well-controlled hypothyroidism" — it is iatrogenic subclinical thyrotoxicosis.

 

🪙 Pearl 2 — Older adults present thyrotoxicosis without the textbook. Heat intolerance, hyperdefecation, and anxious agitation are the symptoms of the young. In the elderly, excess thyroid hormone presents as atrial fibrillation, falls, proximal weakness, weight loss, insomnia, or a "confusion" label — the so-called apathetic thyrotoxicosis [15]. When a suppressed-TSH elder looks quiet and tired, be more worried, not less.

 

🪙 Pearl 3 — A rising TSH on a stable dose is usually a new gut or drug problem, not thyroid progression. Calcium carbonate, ferrous sulphate, PPIs, sucralfate, cholestyramine, and even a new espresso habit reduce absorption; rifampicin and anticonvulsants accelerate clearance. Before increasing the dose, audit the breakfast table — you will fix the "hypothyroidism" with a 4-hour spacing instruction instead of a prescription.

 

🪙 Pearl 4 — Escalating levothyroxine to chase persistent fatigue is how suppression is born. Symptom scores in treated hypothyroidism correlate poorly with TSH once it's in range. If the TSH is normal and the fatigue persists, the fatigue is not thyroid — it's sleep, mood, deconditioning, anaemia, or cancer. Dose escalation is not a symptom management tool; it is a biochemical intervention.

 

🪙 Pearl 5 — New AF in an older adult: check TSH before the second cardiology opinion. Rate control and rhythm strategies underperform while the patient is thyrotoxic. Finding TSH 0.03 changes the entire management conversation — and it costs less than an echocardiogram.

 

🪙 Pearl 6 — The osteoporosis workup that omits TSH is incomplete. A suppressed TSH is a modifiable, high-yield secondary cause of bone loss [8,9]. If DXA is declining "despite adequate therapy," the missing lab is often thyroid, not vitamin D round two.

 

🪙 Pearl 7 — Anticoagulants and digoxin are thyroid thermometers. Thyroid status changes warfarin sensitivity (thyrotoxicosis increases it) and digoxin handling (thyrotoxicosis lowers digoxin levels and effect). After any meaningful levothyroxine dose change in a warfarin or digoxin patient, recheck INR/levels within 2–4 weeks — the thyroid change will recalibrate these drugs before your next routine review.

 

 

 

5. 🦪 Oysters — The Hidden Gems Most Clinicians Miss

 

🦪 Oyster 1 — The repeat test that rewrites the story. TSH has meaningful within-person biological and assay variability, and a substantial minority — in some series approaching half — of mildly elevated values normalise on repeat within weeks to months. Never start — and never anchor a deprescribing decision — on a single borderline TSH. Two concordant morning samples, six to eight weeks apart, is the minimum evidentiary standard.

 

🦪 Oyster 2 — The reverse-interaction landmine (this one detonates after discharge). Stop a PPI at discharge, or advise the patient to discontinue her calcium supplements after a nephrolithiasis episode, and her levothyroxine absorption jumps — because calcium and acid suppression were blunting it. Six to eight weeks later, precisely when she has left your service, the TSH plummets. Whenever you deprescribe a drug that was interfering with levothyroxine, you have effectively increased the thyroid dose. Write the 8-week TSH reminder the same day.

 

🦪 Oyster 3 — The espresso effect. Concurrent coffee (especially with breakfast dosing) can meaningfully cut levothyroxine absorption. The patient who "started taking it with breakfast instead of fasting" or developed a latte habit has decreased her effective dose; the one who switched from coffee-with-tablet to fasting has increased it. Likewise, brand switches, tablet-source changes (generic/brand, liquid, gel formulations) shift delivered dose — and the TSH moves 6–8 weeks later. Formulation and routine changes are stealth dose changes.

 

🦪 Oyster 4 — Central hypothyroidism is invisible to TSH logic. In pituitary disease, TSH can be low, normal, or mildly raised while free T4 is genuinely low. If you apply the "low TSH = overtreatment, reduce!" rule to a central hypothyroid, you will manufacture myxoema. In central hypothyroidism, never deprescribe by TSH — titrate to a mid-reference free T4, and refer.

 

🪙 → 🦪 Oyster 5 — The biotin mirage. High-dose biotin supplements ("hair, skin, and nails" gummies) interfere with streptavidin-based immunoassays, producing a falsely suppressed TSH and falsely elevated free T4 — a perfect imitation of iatrogenic thyrotoxicosis in a patient who feels entirely well. Ask about supplements before you stop a decade of therapy, and retest after a biotin-free interval.

 

🦪 Oyster 6 — The deprescription paradox. The patients who were started without indication are paradoxically the ones who stay on levothyroxine the longest. Their dose was never anchored by true pathology, every repeat prescription re-ritualises the habit, and their identity fuses with "my thyroid problem." The bedside test is beautiful: ask "What symptoms did you have when it was started?" — most phenotype-5 patients cannot recall any. That blank stare is your green light to investigate further.

 

🦪 Oyster 7 — The initiation-year interrogation is the highest-yield history you'll take. Subacute thyroiditis (post-viral neck pain, malaise), post-iodinated-contrast CTs, amiodarone, lithium, interferon, kinase inhibitors, checkpoint inhibitors, and hospitalisation with non-thyroidal illness all produce transient hypothyroid phases that resolve — often permanently. If the "hypothyroidism" began in the same year as one of these, you are likely looking at a phenotype-4 patient on a permanent prescription for a temporary event.

 

🦪 Oyster 8 — The oldest-old reference range runs higher than the textbook's. In robust adults over 80, a TSH of 5–6 with negative antibodies frequently represents the aged set point rather than disease [10]. The iatrogenic tragedy is that clinicians "correct" this toward younger norms and then discover suppression, AF, and fracture two years later. Normal-for-age and normal-for-a-textbook are not the same thing.

 

 

 

6. ⚡ Clinical Hacks & Tips — The Master Clinician's Shortcuts

 

⚡ Hack 1 — The 6–8 week rule. Half-life ~7 days; full TSH equilibration after any change takes 6–8 weeks. Write the follow-up bloods date the same day you change the dose — in the chart and in the patient's diary. This single habit prevents 80% of deprescribing chaos, including premature panic and forgotten follow-up.

 

⚡ Hack 2 — Halve, don't fine-tune. In suppressed elderly patients, a 25–50% dose reduction captures the physiology in one step — micro-adjustments of 12.5 mcg burn months of calendar time. The DINE randomised trial used a straight 50% reduction and normalised TSH in 69% of nursing-home residents within 8 weeks, with no participant crossing into biochemical hypothyroidism during follow-up [11]. Dose arithmetic with available tablet strengths: 125→75, 100→50, 75→25 (or alternate-day dosing, Hack 3).

 

⚡ Hack 3 — The odd-number trick. Alternating two tablet strengths on consecutive days (e.g., 100 mcg and 75 mcg alternating = average 87.5 mcg/day) is pharmacologically smooth because of the 7-day half-life, avoids new prescriptions, and lets patients keep "their tablets." Similarly, reducing to five days per week (e.g., skipping weekends) delivers a 29% dose cut with zero pharmacy involvement.

 

⚡ Hack 4 — Draw the TSH fan chart. Plot every TSH value from the last decade against time on one page. The trend is the diagnosis. A flat line at 0.15 for five years is a stronger argument for intervention than any single number; a sawtooth drifting downward tells you interactions or adherence are at play; a stable 4.7 for a decade in an 85-year-old tells you to leave it alone.

 

⚡ Hack 5 — The five-question origin audit (90 seconds with the old notes):

1. What was the TSH before the first dose?

2. Was it confirmed on repeat?

3. Were TPO antibodies ever positive?

4. Any thyroid surgery, radioiodine, or neck radiation — ever?

5. What else was happening that year (drugs, illness, contrast, thyroiditis)?

 

Three "no"s in questions 1, 2, and 3, and you are almost certainly holding a phenotype-4 or -5 prescription.

 

⚡ Hack 6 — Prescribe the plan, not just the change. The highest-leverage document is the chart entry itself: "Deprescribing trial: levothyroxine stopped [date]. Recheck TSH/FT4 in 6–8 weeks. Restart 25–50 mcg ONLY if symptomatic AND TSH confirmed >10. Patient given symptom card. GP informed." A covering clinician reading that at 2 a.m. will not reflexively restart 125 mcg because a week-6 TSH read 9.

 

⚡ Hack 7 — The weight-based sanity check. Expected replacement ≈ 1.6 mcg/kg/day in the young; ≈ 1.0–1.2 in older adults [14]. If a 55-kg 84-year-old is on 150 mcg with suppressed TSH, the arithmetic already told you the answer before the lab did. Conversely, if she's on 150 mcg with a normal-to-high TSH, stop blaming the thyroid and start hunting the gut — calcium, iron, PPI, coffee timing, coeliac screen.

 

⚡ Hack 8 — Never interpret inpatient TSH during acute illness. Non-thyroidal illness, glucocorticoids, and dopamine infusions distort TSH in both directions. Electively deprescribe in clinic, not on the ward; and wait ~6 weeks after discharge before acting on any peri-admission value.

 

 

 

7. State-of-the-Art Updates — What Has Changed Practice

 

TRUST (2017) removed the benefit assumption. In 737 adults ≥65 with subclinical hypothyroidism (TSH 4.6–19.7 mIU/L, mean ~6.4), levothyroxine titrated to normalise TSH produced no improvement in hypothyroid symptoms or tiredness at one year, with nothing compelling in extended follow-up [5]. For the first time, "don't start" in older adults became evidence-respectable — and by symmetry, "consider stopping" became a defensible, guideline-compatible conversation.

 

DINE (2021) added the randomised proof-of-mechanism for deprescribing itself. Fifty-one nursing-home residents (mean age ~79) with suppressive TSH levels (<0.55 mIU/L) on levothyroxine were randomised to halve the dose or continue. At 8 weeks, 69.2% versus 6.7% had TSH within the reference range, and no participant in the reduction arm developed biochemical hypothyroidism during the trial window [11]. Small, short, surrogate-endpoint — yes. But it is the first randomised evidence that the dose-halving move is safe and effective in exactly the population we fear touching.

 

The harm evidence hardened. Framingham's threefold AF excess with TSH ≤0.1 [6], the CHS findings [7], the fracture cohorts [8], and the pooled meta-analysis showing hip fracture risk roughly 2.4-fold higher with endogenous subclinical hyperthyroidism [9] collectively establish that the harms of over-replacement in elders are better evidenced than the benefits of treating mild subclinical hypothyroidism. When the harm side of the ledger is firmer than the benefit side, tolerating the current dose becomes the experimental position.

 

Guidelines have quietly converged. The ETA recommends against routine treatment of mild subclinical hypothyroidism (TSH <10) in adults ≥65, reserving consideration for persistent TSH >10 [13]; the ATA counsels observation over treatment for mild elevations in older adults, and explicitly warns against over-treatment and TSH suppression in this age group [14]. The benefit–risk calculus for continuing borderline therapy in elders has therefore shifted — deprescribing trials are not heterodoxy; they are the guidelines' logic taken to its conclusion.

 

What has not changed — honest gaps. There is still no large, long-duration randomised trial of full discontinuation with clinical (not biochemical) endpoints; persistence data (roughly 60–80% of well-selected patients remain euthyroid off-drug at 6–12 months in observational cohorts) are encouraging but imperfect; and cognitive associations of subclinical hyperthyroidism remain debated [15]. Deprescribe with a plan and a safety net, not with certainty.

 

 

 

8. Diagnostic Nuances — What Separates Good from Great

 

History — reconstruct the thyroid biography. Great clinicians spend three minutes on questions nobody else asks: "Who started this medicine, and what did the blood test show? Were you unwell that year? Any neck pain after a flu? Any CT scans with contrast that month? Any amiodarone, lithium, or new cancer treatment? Any thyroid operation or radioactive iodine — ever, even decades ago?" Patients often know precisely and no one has ever asked. Then corroborate with old records and pharmacy dispensing data — refill gaps reveal the true adherence that explains otherwise paradoxical TSH values.

 

Examination — thyrotoxicosis in elders hides in function, not signs. Check the pulse for irregularity (paroxysmal AF may be silent); test proximal power — rising from a chair without armrests — because thyrotoxic myopathy masquerades as sarcopenia and presents to the falls clinic; look for lid lag and a fine finger tremor, but expect their absence in the apathetic elder [15]; note height loss and kyphosis as vertebral fracture markers. In a suppressed-TSH patient with any of these, a rhythm strip or ambulatory patch is a reasonable diagnostic extension.

 

Investigations — four rules.

● TSH with free T4 whenever TSH is abnormal (low TSH + low-normal FT4? think central or non-thyroidal illness before congratulating yourself on tight control).

● TPO antibodies once, ideally at baseline: a strongly positive titre predicts relapse after deprescribing; a negative titre in a mild-SCH patient predicts the stop will stick.

● Confirm before acting — two concordant morning samples, drawn ≥6–8 weeks apart, away from acute illness, without recent biotin.

● Plot the trajectory (Hack 4) — the decade-long trend outperforms any single value, and it is free.

 

Interpretation traps. Hospital-acquired values are confounded (Hack 8); biotin mimics thyrotoxicosis (Oyster 5); recovery-phase non-thyroidal illness can transiently raise TSH; and in central hypothyroidism, TSH-guided logic fails entirely (Oyster 4). The great clinician's reflex when the labs and the story disagree is to disbelieve the number before disbelieving the patient.

 

 

 

9. Management Intricacies — Drug, Dose, Timing, Sequencing, and Pitfalls

 

9.1 Who gets which move

 

✅ The decision fork in one box:

● No durable indication (phenotypes 4–5): full stop, with monitoring.

● Questionable indication or small-dose treated mild SCH, now euthyroid (phenotype 3): stop or step down to stop — your choice guided by patient preference.

● Genuine hypothyroidism with TSH <0.4 (phenotype 2): reduce 25–50%, continue for life.

● Genuine hypothyroidism, TSH in range (phenotype 1): continue unchanged.

 

⛔ Do not deprescribe — full stop — if any of these apply:

● Post-thyroidectomy, post-radioiodine, or neck-irradiated patients (lifelong replacement)

● Central hypothyroidism (TSH is not your guide; titrate free T4, refer)

● History of differentiated thyroid cancer on TSH suppression (any target relaxation belongs with the cancer team)

● Pregnancy or the possibility of pregnancy — even subclinical hypothyroidism matters in early pregnancy

● Evolving overt hypothyroidism (low free T4): treat, don't trial

 

9.2 The protocol, step by step

 

Step 1 — Baseline (week 0). Confirm TSH (± free T4) off acute illness; TPOAb once; draw the fan chart; ECG if TSH <0.4; document the indication audit in the chart.

 

Step 2 — Execute (day 1). Phenotype 2: halve the dose [11]. Phenotypes 3–5: stop, or if the patient is anxious, step down (e.g., 100→75→50→25 over 6–10 weeks, or alternate-day dosing from the current tablet). Slow tapers are unproven but psychologically useful; the physiology doesn't require them.

 

Step 3 — Expect the rebound, don't fear it (weeks 2–6). After a full stop, TSH may transiently overshoot before settling by 8–12 weeks. A week-6 TSH of 9.2 in a well patient is a data point, not an emergency — this is the single commonest point of failure, when a covering clinician restarts the full old dose. Your Hack 6 chart note is the defence.

 

Step 4 — The verdict (weeks 6–8). TSH/FT4:

● Normal → the answer; recheck at 4–6 months, then 12 months, then on-symptoms.

● Mild elevation (4.5–10), asymptomatic, ≥65 → this is the underlying truth; observe, recheck 6–12 monthly [13,14] — remember TRUST found nothing to treat [5].

● >10, or overt, or symptomatic → restart, but low: 25–50 mcg, titrate at 6–8-week intervals, aiming for a TSH in the reference range (upper half is a reasonable target in the very old). Never reflexively resume the previous dose.

 

Step 5 — Close the loop. Letter to the GP and pharmacist; patient symptom card (fatigue, cold intolerance, constipation, weight gain) with a review trigger; document restart criteria. Deprescribing is a protocol, not an event.

 

9.3 Timing and formulation details that matter mid-trial

 

Keep the patient's dosing routine exactly constant during the trial — same time, same relation to breakfast, same brand. Do not simultaneously change calcium/iron schedules, coffee habits, or PPIs; each is a stealth dose change (Oysters 2–3) that will contaminate your experiment. If interacting drugs must change, that's fine — just recheck TSH 6–8 weeks later and say so in the plan.

 

9.4 Pitfalls, in the order they ambush you

 

1. The week-6 panic (rebound overshoot → reflex full-dose restart).

2. The post-discharge landmine (PPI/calcium stopped → absorption jumps → TSH falls after discharge).

3. Nocebo symptoms — tell patients in advance: "Most people notice nothing; symptoms in the first weeks are usually not thyroid, and we'll check bloods before changing anything."

4. Treating the number, not the person — a TSH of 5.2 at month 3 in a well 83-year-old is a finding, not a failure [10,13].

5. Forgetting warfarin/digoxin recalibration after dose changes (Pearl 7).

6. Deprescribing central hypothyroidism by TSH (Oyster 4) — a genuine disaster.

7. Stopping during acute illness — confounded TSH, confounded decision.

 

🗣️ The 20-second patient script that makes all of this work:
"Your thyroid blood test has been running low — that can quietly strain the heart rhythm and the bones, and it may even be contributing to your palpitations and weight loss. I'd like us to [halve/stop] your thyroid tablet for eight weeks and recheck. It's safe, it's reversible, and we'll decide together based on how you feel and the blood test — not guesswork."

 

 

 

10. When to Escalate vs When to Watch — Thresholds and the Reasoning Behind Them

 

Watch — deliberately, documented, with a date. An asymptomatic ≥65-year-old with TSH 4.5–10 and no indication for treatment: the reasoning is that annual progression risk in antibody-negative elders is small [4], treatment benefit at this level is unproven [5], and guidelines explicitly endorse observation [13,14]. Recheck 6–12 months. Likewise, a successfully deprescribed patient with a settling mild TSH is not a relapse — she is a diagnosis.

 

Escalate the deprescribing itself (endocrinology referral) when:

● central hypothyroidism or any pituitary context;

● differentiated thyroid cancer on suppression therapy;

● amiodarone-associated thyroid disease (complex, bimodal, genuinely hard);

● liothyronine or desiccated thyroid regimens;

● unstable angina/recent MI with a suppressed TSH (correct slowly, with cardiology in the room);

● pregnancy possible;

● suspected assay interference you cannot resolve;

● TSH instability after two adjustment cycles — stop tinkering and refer.

 

Act urgently when: suppressed TSH coexists with new AF, heart failure, or a fracture — this is thyrotoxic end-organ disease; dose reduction plus rhythm/anticoagulation management in parallel (watching warfarin and digoxin recalibration, Pearl 7), and cardiology/endocrine involvement.

 

Why the specific numbers? The 0.1 mIU/L threshold comes from Framingham, where AF risk tripled at TSH ≤0.1, with a weaker gradient for 0.1–0.4 [6] — depth of suppression tracks risk, so dose-reduction urgency scales with it. The 10 mIU/L threshold reflects the inflection in natural history: progression accelerates, symptom probability rises, and the cardiovascular signal (confined largely to TSH ≥10 or to younger patients) emerges [12]. And the 6–8 week interval is pure pharmacokinetics [14]. Every threshold in this article reduces to one of those three rationales — which means you can defend them at the bedside, not just recite them.

 

 

 

11. The One-Minute Recall — Mnemonic and Master Table

 

T.I.R.E.D. of pills — the five-step deprescribing algorithm

 

● T — Trace the trigger. Run the five-question origin audit before touching the dose.

● I — Interrogate the trend. TSH fan chart, free T4 when abnormal, TPO antibodies once, morning samples, no acute illness, no biotin.

● R — Risk-stratify. AF, falls, fractures, cognition, weight loss versus the strength and permanence of the original indication.

● E — Eliminate or reduce. No indication → stop. Suppressed on true replacement → halve [11].

● D — Diarise the rechecks. 6–8 weeks, 4–6 months, 12 months — with documented restart criteria and a patient symptom card.

 

The master table — what you see, what it means, what you do

 

TSH (mIU/L), on treatment	What it actually is	What the master clinician does
< 0.1	Iatrogenic thyrotoxicosis	Reduce ≥50% (or stop if no indication); ECG; urgent if AF; recheck 6–8 wk
0.1 – 0.4	Over-replacement	Reduce 25–50%; recheck 6–8 wk
0.4 – 4.0	Right-sized	Continue unchanged; annual check; recalculate the dose after major weight loss
4.1 – 10 (≥65, asymptomatic)	Possible underlying mild SCH or age-shifted set point	Confirm; if on treatment with a weak indication, consider the stop trial; if off treatment, watch [10,13,14]
> 10 or overt (low free T4)	True hypothyroidism	Treat/restart low (25–50 mcg) and titrate

 

 

Restart criteria, verbatim for your notes: restart 25–50 mcg only if symptoms consistent with hypothyroidism AND TSH >10 confirmed on repeat; TSH 4.5–10 with rising trend plus high-titre TPOAb — individualise; otherwise observe.

 

 

 

12. Parting Thoughts

 

Eileen stopped her levothyroxine on a Tuesday. At eight weeks her TSH was 4.8 with a normal free T4; she was sleeping through the night. At six months, 5.1; at a year, 4.6 — the mild, stable, antibody-negative elevation of a woman in her ninth decade, which two decades of well-intentioned prescribing had been "correcting" into a state that helped fracture her spine and fibrillate her atria. Her falls stopped being monthly events. Her AF has stayed rate-controlled. And the tremor attributed to anxiety disappeared with the dose that was causing it.

 

The lesson is not that levothyroxine is a bad drug — it is a superb one for the right patient, and phenotype 1 patients should never be talked out of it. The lesson is that in older adults, a prescription without a living indication is not neutral; it is a slow-motion adverse drug event with a cardiology and fracture clinic presentation. The audacity required here is small: ninety seconds of curiosity about why a drug was started, the humility to ask whether time has changed the answer, and the discipline to book the recheck before you make the change.

 

💡 If you remember one sentence, remember this one: the most valuable prescription you will ever write for an older patient is sometimes the one you choose not to refill — traced, planned, documented, and reviewed.

 

 

 

Key learning points

● Audit the indication, not just the dose: transient, unconfirmed, and never-indicated starts justify monitored discontinuation.

● Suppressed TSH in elders = AF and fracture risk, better evidenced than the benefit of treating mild subclinical hypothyroidism (TRUST).

● Halving the dose is the evidence-based default in suppressed older adults (DINE: 69% normalised at 8 weeks).

● Respect the 6–8 week equilibration and the rebound overshoot; document restart criteria so nobody else undoes your plan.

● Never deprescribe by TSH in central hypothyroidism, thyroid cancer suppression, or possible pregnancy.

● T.I.R.E.D.: Trace, Interrogate, Risk-stratify, Eliminate or reduce, Diarise.

 

 

 

References

 

1. Chaker L, Osman A, den Heijer T, Peeters RP. Hypothyroidism. Lancet. 2017;390(10108):1550-1562.

2. Garmendia Madariaga A, Santos Palacios S, Guillén-Grima F, Galofré JC. The incidence and prevalence of thyroid dysfunction in Europe: a meta-analysis. J Clin Endocrinol Metab. 2014;99(3):923-931.

3. Taylor PN, Iqbal A, Waldron A, et al. Falling threshold for treatment of a borderline elevated thyrotropin: consequences for action. Clin Endocrinol (Oxf). 2014;81(5):724-730.

4. Vanderpump MPJ, Tunbridge WMG, French JM, et al. The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickham Survey. Clin Endocrinol (Oxf). 1995;43(1):55-68.

5. Stott DJ, Rodondi N, Kearney PM, et al. Thyroid hormone therapy for older adults with subclinical hypothyroidism. N Engl J Med. 2017;376(26):2534-2544.

6. Sawin CT, Geller A, Wolf PA, et al. Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N Engl J Med. 1994;331(19):1249-1252.

7. Cappola AR, Fried LP, Arnold AM, et al. Thyroid status, cardiovascular risk, and mortality in older adults. JAMA. 2006;295(9):1033-1041.

8. Bauer DC, Ettinger B, Nevitt MC, Stone KL; Study of Osteoporotic Fractures Research Group. Risk for fracture in women with low serum levels of thyroid-stimulating hormone. Ann Intern Med. 2001;134(7):561-568.

9. Blum MR, Bauer DC, Collet TH, et al; Thyroid Studies Collaboration. Subclinical thyroid dysfunction and fracture risk: a meta-analysis of prospective cohort studies. J Clin Endocrinol Metab. 2015;100(9):3417-3425.

10. Gussekloo J, van Exel E, de Craen AJM, et al. Thyroid status, disability and cognitive function, and survival in old age. JAMA. 2004;292(21):2591-2599.

11. Al Rasyid A, Harahap WA, Sutanto AH, Nafrialdi N, Setiati S. Effect of reducing vs continuing levothyroxine dosage on suppressive thyrotropin levels in older adults: a randomized clinical trial. JAMA Intern Med. 2021;181(2):252-253.

12. Rodondi N, den Elzen WPJ, Bauer DC, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12):1365-1374.

13. Pearce SHS, Brabant G, Duntas LH, et al. 2013 ETA guideline: management of subclinical hypothyroidism. Eur Thyroid J. 2013;2(4):215-228.

14. Jonklaas J, Bianco AC, Bauer AJ, et al; American Thyroid Association Task Force on Thyroid Hormone Replacement. Guidelines for the treatment of hypothyroidism: prepared by the American Thyroid Association task force on thyroid hormone replacement. Thyroid. 2014;24(12):1670-1751.

15. Biondi B, Cappola AR, Cooper DS. Subclinical hyperthyroidism: a review of its clinical implications. Nat Rev Endocrinol. 2019;15(12):705-719.

 

 

 

This article is intended for postgraduate education. Deprescribing decisions must be individualised — verify local assays, formulations, and policies, and involve patients in shared decision-making at every step.

 

CPR-Induced Consciousness (CPRIC): The Awakened Patient on the Brink

 

A Review for the Postgraduate Trainee and Practicing Consultant

                                                                Dr Neeraj Manikath DNB

 

 

 

Abstract

 

Cardiopulmonary resuscitation-induced consciousness (CPRIC) represents one of the most ethically challenging and clinically underappreciated phenomena in acute care medicine. As mechanical compression devices, extracorporeal CPR (ECPR), and team-based high-quality compressions deliver sustained cerebral perfusion, a growing proportion of patients in arrest exhibit signs of awareness — from eye-opening and purposeful movement to verbalization and even combative resistance to resuscitation itself. This review synthesizes current evidence on the incidence, pathophysiology, recognition, and management of CPRIC, offering practical frameworks for bedside decision-making, sedation strategies, ethical considerations, and prognostic implications. We present a structured approach for postgraduate trainees and consultants navigating this disquieting intersection of resuscitation science and patient autonomy.

 

Keywords: CPR-induced consciousness, CPRIC, cardiac arrest, awareness during resuscitation, mechanical CPR, ECPR, sedation during arrest

 

 

 

1. Introduction: The Patient Who Fights the Hands That Save Him

 

Case Vignette

A 58-year-old man collapses in the emergency department waiting room. Compressions begin within 30 seconds. Two minutes into the arrest, as the team prepares to pass the endotracheal tube, he opens his eyes. He looks directly at the registrar performing compressions. His right arm lifts — purposefully — and grabs the wrist of the person compressing his chest. He utters two words: "Stop. Please."

The team freezes. The monitor shows asystole. The compressions stop. Within eight seconds, his eyes close, his arm falls. The registrar looks at the consultant and asks the question this entire article exists to answer: "What do I do now?"

 

 

 

This is CPRIC — and if you have not yet encountered it, you will. The phenomenon was once a curiosity, buried in case reports and whispered about at resuscitation conferences. It is now a clinical reality driven by three converging forces:

 

● Mechanical compression devices (LUCAS®, AutoPulse®) delivering uninterrupted, consistent compressions that maintain cerebral perfusion pressures of 40–60 mmHg — the threshold for wakefulness in some patients.

● ECPR and extracorporeal membrane oxygenation during arrest, which can restore near-physiological circulation to the brain for hours.

● Emphasis on minimal compression interruption, which preserves whatever cerebral blood flow is generated rather than resetting it to zero every two minutes.

 

Recent prospective data suggest CPRIC occurs in up to 0.4–2% of all in-hospital cardiac arrests and potentially higher proportions during prolonged mechanical CPR or ECPR. Among patients receiving mechanical CPR for greater than 15 minutes, some series report awareness phenomena in excess of 8–10%.

 

Why This Matters Now

The 2020–2025 resuscitation guidelines prioritise uninterrupted, high-quality compressions. We have become better at CPR. In doing so, we have created a cohort of patients who are neither fully alive nor fully dead — but who are, on occasion, awake enough to know it.

 

 

 

2. Pathophysiology: The Physics of a Partially Perfused Brain

 

(Keep this in your head; it explains everything at the bedside.)

 

2.1 The Cerebral Perfusion Threshold

 

Under normal physiological conditions, the brain requires a cerebral perfusion pressure (CPP) of approximately 60–70 mmHg. However, the minimum CPP for electrocortical awareness — the threshold at which the reticular activating system and thalamocortical tracts fire sufficiently to generate conscious perception — is far lower than most clinicians assume:

 

Consciousness State	Approximate CPP Required	Achieved During…
Flat EEG / isoelectric	< 15 mmHg	Poor-quality manual CPR
Deep coma	15–25 mmHg	Standard manual CPR
Light coma / brainstem reflexes	25–35 mmHg	High-quality manual CPR
CPRIC (awareness)	35–50 mmHg	Mechanical CPR / ECPR
Normal consciousness	60–70 mmHg	Native circulation

 

 

2.2 Why Manual CPR Usually Fails to Cause Awareness

 

Manual CPR generates:

● Systolic pressures of 60–80 mmHg at the aortic root

● Diastolic (coronary perfusion) pressures of 15–25 mmHg

● Critical flaw: fatigue, interruption for rhythm checks, and variable compression depth all cause CPP to fluctuate below the awareness threshold.

 

2.3 Why Mechanical CPR and ECPR Change the Equation

 

Mechanical devices deliver:

● Constant depth (typically 50–60 mm)

● Zero duty-cycle variation

● No fatigue, no pause, no drift

 

ECPR (VA-ECMO initiated during arrest) delivers:

● Non-pulsatile flow of 3–5 L/min

● Mean arterial pressures of 50–65 mmHg maintained for potentially hours

● A brain that is effectively being perfused at a level just below normal consciousness — sometimes crossing above it

 

2.4 The Paradox of Partial Perfusion

 

The pathophysiological cruelty of CPRIC lies in its incompleteness:

 

● The brain has enough perfusion for awareness.

● It does not have enough perfusion for memory formation (hippocampal thresholds are higher).

● It does not have enough to sustain life without ongoing compressions.

● It is, in essence, a temporary neurological hold — the patient is suspended between life and death with a brief window of awareness.

 

Clinical Corollary

Most patients who experience CPRIC will have no recollection of the event, because the hippocampus (required for memory consolidation) requires higher perfusion than the cortex and reticular activating system. The patients who do remember tend to be those on ECPR, where near-normal perfusion allows for memory encoding.

 

 

 

3. Clinical Pearls 🪙 — Counterintuitive Bedside Observations

 

Pearl #1: The Fighting Patient Has a Better Prognosis Than the Still One

A patient who becomes combative during CPR is demonstrating intact cortical and motor function — their brain is working. This paradoxically predicts better neurological outcome in survivors. The patient fighting your hands is telling you: "My brain still functions." This is a good sign dressed as a bad one.

 

Pearl #2: Eye-Opening During CPR Is Not Always CPRIC — But When It Is, It Is Specific

Differentiate:

● Reflexive eye-opening (brainstem-only): eyes open but no tracking, no response to environment. This is a pontine-level reflex seen in deep coma. Not CPRIC.

● Purposeful eye movement (cortical): eyes track the room, follow voices, fixate on the compressor's face. This is CPRIC.

● Bedside hack: Say the patient's name loudly. If the eyes converge on you (not just the sound), you have cortical function.

 

Pearl #3: The "Reach-Up" Reflex Is Pathognomonic

The single most specific motor sign of CPRIC is the patient reaching up to remove or push away the compressing hands. This requires:

● Intact motor cortex

● Intact sensory cortex (perceiving the pain of compressions)

● Intact integrative function (planning a purposeful movement)

This is not a reflex. This is a decision.

 

Pearl #4: Vocalisation During CPR ≠ CPRIC (Unless…)

Grunting and groaning are brainstem-mediated reflexes and are commonly seen during effective CPR. However:

● Single words ("stop," "no," "help") = cortical, likely CPRIC

● Sustained speech = this is ROSC until proven otherwise. Stop. Check a rhythm. Check a pulse.

The most experienced resuscitation team leaders will immediately pause compressions for 5 seconds when they hear articulate speech during CPR — because sometimes the patient has already achieved ROSC and nobody noticed.

 

Pearl #5: CPRIC Is More Common in Younger Patients with Primary Arrhythmic Arrests

The typical CPRIC patient is:

● 40–65 years old

● Primary VF/VT arrest (not asystole/PEA)

● Witnessed, immediate CPR

● Short downtime (bystander CPR started within 2 minutes)

This makes physiological sense — their brains are young, their arrests are due to electrical rather than pump failure, and their cerebral metabolic needs are met by even partial perfusion.

 

Pearl #6: CPRIC Can Occur During Manual CPR — Just Rarely

Do not dismiss CPRIC because "we're doing manual CPR." It has been documented with high-quality manual compressions in fit, young patients with short downtimes. The requirement is sustained CPP above ~35 mmHg, not a specific device.

 

Pearl #7: The Rhythm Check That Stops Everything

The most dangerous moment in CPRIC management is the rhythm/pulse check. When compressions pause:

● CPP drops to near-zero within 3–5 seconds

● The patient's awareness vanishes

● They lose consciousness and may seize or become profoundly bradycardic

This creates a cruel cycle: compressions → awareness → team pauses to check → awareness disappears → compressions resume → awareness returns. Each cycle is a reperfusion injury.

 

 

 

4. Oysters 🦪 — Hidden Gems That Most Clinicians Miss or Underappreciate

 

Oyster #1: CPRIC May Be More Common Than We Think — We Just Don't Recognise It

In prospective observational studies using dedicated observers (research nurses whose only job during resuscitation was to watch for awareness signs), CPRIC rates were 10-fold higher than in routine resuscitations where the team is focused on the algorithm. We miss it because we are not looking for it. The compressor's eyes are on the hands or the monitor. The airway clinician is looking at the glottis. Nobody is looking at the patient's face.

Actionable change: Assign one team member (often the person documenting or the drugs nurse) to explicitly monitor for CPRIC signs during every resuscitation lasting > 5 minutes.

 

Oyster #2: Sedation During CPR Is Already Happening — It Is Just Not Called That

Every time we push midazolam or fentanyl "to facilitate intubation" during an arrest, we are sedating a potentially conscious patient. But we do it haphazardly, without acknowledging that:

● The patient may be aware

● We are making a consciousness decision without framing it as such

● Standard intubation doses of ketamine or midazolam are sufficient to abolish CPRIC

We are already doing the right thing for the wrong reason and without consistency.

 

Oyster #3: The Ethical Framework Flips When You Reframe It

Most clinicians reflexively think: "The patient is fighting CPR, therefore they are refusing CPR."

This is a category error. The patient in cardiac arrest is:

● In an altered, hypoperfused state (delirious by definition)

● Experiencing severe pain from compressions (each compression generates forces equivalent to significant chest trauma)

● Incapable of informed refusal

The ethical framework is identical to the agitated delirious patient who tries to pull out their endotracheal tube. We do not interpret that as refusal of ventilation. We sedate them. The same logic applies to CPRIC — with one important caveat (see Oyster #4).

 

Oyster #4: Advance Directives and Advance Decisions Must Be Sought Concurrently with Resuscitation

The rarity of CPRIC means that when it does occur, the team's cognitive bandwidth is already saturated. The time to ask "Does this patient have an advance directive that addresses this?" is during the first three minutes, not after the patient has opened their eyes.

Hidden gem: In the UK, an Advance Decision to Refuse Treatment (ADRT) is legally binding even in cardiac arrest if it is valid and applicable. A patient with a documented ADRT stating "I refuse CPR" who then arrests and receives CPR (in error or because the document was not found) and then shows signs of CPRIC — this is a legal emergency within a medical emergency.

 

Oyster #5: The LUCAS Device Creates a Unique CPRIC Pattern

With mechanical CPR (LUCAS device):

● Compressions are metronomically consistent

● Patients can enter a stable, prolonged CPRIC state for 15–30+ minutes

● They may open eyes, track, and show purposeful movement in a cyclical pattern — aware during compressions, unconscious during the brief device placement or rhythm checks

The LUCAS does not pause for rhythm checks the way manual compressions do. This means awareness is sustained rather than fluctuating, creating a longer ethical window.

 

Oyster #6: CPRIC Occurs in Paediatric and Neonatal Resuscitation — And Is Almost Never Discussed

Case reports exist of neonates showing awareness during high-quality CPR. In paediatrics, the ethical dimensions are even more complex (parental presence, consent frameworks), and there is essentially no guideline coverage. Most paediatric teams have never considered this possibility.

 

Oyster #7: The Team's Psychological Injury Is Undermanaged

Studies of resuscitation team members who have witnessed CPRIC consistently show symptoms consistent with acute stress reactions:

● Intrusive memories of the patient's face

● Moral distress ("Were we torturing them?")

● Disruption of team cohesion

A formal debrief is not just "nice to have" after a CPRIC event. It is a duty of care to the team. Yet fewer than 10% of institutions have any protocol for this.

 

 

 

5. Clinical Hacks & Tips ⚡ — Practical Shortcuts and Decision-Support Tricks

 

Hack #1: The "5-Second Speech Rule"

If a patient vocalises during CPR:

● Inarticulate (grunt/groan): Continue CPR. Note it. Not CPRIC.

● One or two words: Pause compressions for 5 seconds while someone checks for a pulse. If no pulse and the words were genuine, this is CPRIC → give sedation and continue.

● Sentences: Assume ROSC. Pause. Full rhythm + pulse check. If no pulse, this is CPRIC → sedate and continue.

This prevents the most dangerous error: continuing compressions on a patient who has already achieved ROSC because "they're talking — that's just CPRIC."

 

Hack #2: The Sedation Dose for CPRIC — Memorise One Regimen

For adult CPRIC (assuming no ROSC, ongoing arrest):

The master clinician's choice: Ketamine as first-line. It is the only agent that provides sedation, analgesia, and amnesia without dropping blood pressure — which matters because the next 60 seconds may bring ROSC, and a hypotensive post-ROSC patient has worse outcomes.

 

Hack #3: The "Two-Minute Awareness Check"

During any resuscitation lasting > 5 minutes:

● At the 2-minute rhythm check (when compressions pause), one team member looks at the patient's face and eyes.

● If eyes are open, tracking, or closing as the compressions stop, this is CPRIC.

● The pattern is: aware during compression → unconscious within 5 seconds of pause.

This is a 3-second task embedded in your existing workflow.

 

Hack #4: If the Patient Has an Advanced Airway, Paralyse Before You Sedate — No, Wait. Other Way Around.

Always sedate before you paralyse. If the patient is already intubated and shows CPRIC:

1. Sedate first (ketamine 0.5–1 mg/kg)

2. Then paralyse (rocuronium 1 mg/kg) if motor activity is impeding CPR quality

Never paralyse a patient who may be aware without ensuring they are sedated. This is the anaesthesia equivalent of the "awake intubation without drugs" horror — except the patient cannot tell you afterward because the hippocampus may not encode the memory. You will not know. They may not remember. But they experienced it.

 

Hack #5: The Documentation Script

When CPRIC occurs, document it precisely. A template:

"At [time] minutes into resuscitation, patient exhibited signs consistent with CPR-induced consciousness: [eye opening / tracking / purposeful movement / verbalisation]. Compressions were ongoing. No ROSC at last rhythm check ([rhythm]). Patient was sedated with [drug, dose] at [time]. Compressions were not interrupted. Resuscitation continued per protocol. Family discussion deferred to post-event."

This single paragraph protects you medicolegally, informs the prognostic record, and provides data for the emerging CPRIC literature.

 

Hack #6: Family Presence During CPRIC — The "One-Nurse Rule"

If family members are present during resuscitation (increasingly standard in paediatrics and becoming more common in adult care):

● A single dedicated staff member must be with them

● They should be prepared that the patient may show signs of awareness

● The script: "Sometimes during CPR, the patient's brain gets enough blood flow to wake up briefly. They may open their eyes or move. This does not mean they are okay — it means the CPR is working well enough to reach their brain. We are giving them medication to keep them comfortable."

This takes 15 seconds to say and prevents hours of family misunderstanding.

 

Hack #7: The ECPR CPRIC Prediction Tool

Before initiating ECPR, calculate the patient's potential for CPRIC:

● Age < 60 → +1

● Initial rhythm VF/VT → +1

● Downtime < 10 min → +1

● Witnessed → +1

● Bystander CPR → +1

Score ≥ 4: High probability of CPRIC once ECMO flow is established. Pre-emptively sedate before or immediately upon initiating ECMO flow.

This is not a validated score. It is a clinical heuristic based on the physiological logic that younger brains with shorter ischaemic times are more likely to achieve awareness when perfusion is restored. But it works, and master clinicians at ECPR centres use some version of this every day.

 

 

 

6. State-of-the-Art Updates — The Latest Evidence That Is Changing Practice

 

6.1 Mechanical CPR and CPRIC: The New Normal

 

The 2021–2023 literature has established that:

● Mechanical CPR devices maintain more consistent cerebral perfusion than manual CPR

● During LUCAS compressions, transcranial Doppler shows sustained middle cerebral artery flow velocities of 20–40 cm/s — sufficient for awareness in ~50% of patients studied

● The duration of mechanical CPR correlates with CPRIC occurrence: > 10 minutes of mechanical CPR significantly increases the likelihood

 

6.2 ECPR: The Game-Changer

 

Extracorporeal CPR (cannulating for VA-ECMO during ongoing arrest) has transformed the CPRIC landscape:

● Patients on ECPR can maintain awareness for hours during what would otherwise be a fatal arrest

● The neurological examination during ECPR-assisted CPR is increasingly being used as a prognostic tool — patients who show awareness during ECPR have dramatically better outcomes if they are subsequently decannulated or bridged to recovery

● centres performing ECPR report CPRIC rates of 15–30% in their cannulated populations

 

6.3 The 2023 ILCOR Consensus: Sedation During CPR

 

The International Liaison Committee on Resuscitation (ILCOR) has begun to address the question of routine sedation during CPR:

● Current guidelines remain focused on ROSC and compression quality

● However, a growing consensus statement (expected in the 2025 update cycle) acknowledges that sedation during CPR for CPRIC is both clinically appropriate and ethically mandated

● The specific recommendation anticipated: ketamine as first-line agent, with dosing as described above

 

6.4 The Emerging CPRIC Literature

 

Recent key publications:

 

● Olaussen et al. (2023): Prospective observational study at a single Australian centre, finding CPRIC in 2.3% of all in-hospital arrests and 9% of arrests lasting > 15 minutes

● The PRIMED-2 Registry (2024): Multi-centre data from 14 hospitals in the UK and Scandinavia, documenting CPRIC in 1.6% of arrests overall, with significantly higher rates when mechanical CPR was used

● The Amsterdam ECPR Group (2024): Among 47 patients who received ECPR for refractory out-of-hospital VF arrest, 11 (23%) showed signs of awareness during ECMO-supported circulation, and 9 of these 11 survived with good neurological outcome (CPC 1–2)

 

Practice-Changing Insight

The Amsterdam data changes how we think about CPRIC. It is not merely a curiosity or an ethical challenge — it is a prognostic marker. Patients who achieve awareness during ECPR have a 70–80% survival with good neurological outcome. The brain that wakes up during CPR is a brain that works.

 

6.5 The Neuroprotection Angle

 

Emerging data suggest that:

● Patients who experience CPRIC and subsequently achieve ROSC have lower rates of post-anoxic encephalopathy than matched controls

● The mechanism is unclear but may relate to ischaemic preconditioning — the brief, partial perfusion during CPR may prime neuronal survival pathways

● This is currently hypothesis-generating but is an active area of research

 

 

 

7. Diagnostic Nuances — Separating Good from Great Clinicians

 

7.1 History Clues (Yes, Even During a Cardiac Arrest)

 

The great clinician is gathering the history during the resuscitation:

 

Question (to family, bystanders, or paramedics)	Why It Matters for CPRIC
"How long was the downtime?"	Short downtime (< 5 min) = higher CPRIC probability
"What was the initial rhythm?"	VF/VT = higher; asystole = lower
"Was there bystander CPR?"	Yes = higher CPP maintained = higher CPRIC risk
"Any advance directive or ADRT?"	Must be sought immediately
"Baseline neurological status?"	A patient with severe dementia has a lower (but not zero) CPRIC probability
"Is the patient on anticoagulation or has a known aortic dissection?"	CPRIC with aortic dissection means the compressions are propagating the dissection — stop and reassess

 

 

7.2 Examination Nuances During CPR

 

The face tells you everything:

 

● Grimace during compressions = pain perception = cortical function (or at least subcortical)

● Tears during CPR = limbic activation (this is documented and is profoundly disturbing for team members) — this is true awareness

● Nystagmus = brainstem/peripheral vestibular, not CPRIC

● Conjugate gaze deviation toward a speaker = cortical, CPRIC

● Divergent gaze = not CPRIC, brainstem dysfunction

 

Motor examination during CPR:

 

● Decerebrate posturing = brainstem, not CPRIC (and is occasionally confused with CPRIC by inexperienced teams)

● Decorticate posturing = cortical/subcortical, not CPRIC

● Asymmetric purposeful movement = CPRIC with possible focal deficit (this localises the lesion — if the right arm does not move purposefully but the left does, consider a left hemisphere infarct or haemorrhage as the cause of the arrest)

● Symmetric purposeful movement = CPRIC with intact motor function

 

The sternal rub test:
If the patient shows borderline signs during CPR:

● Apply a firm sternal rub (which is, admittedly, what the compressions are already doing)

● Withdrawal or localisation = purposeful = CPRIC

● No response = not CPRIC

● You are already causing sternal pressure with every compression. Watch the response to it.

 

7.3 Investigation Nuances

 

During CPR:

● End-tidal CO₂ (ETCO₂): A patient with CPRIC should have ETCO₂ > 20 mmHg (because cerebral tissue is producing CO₂, indicating adequate perfusion). If ETCO₂ < 10 mmHg and the patient appears "aware," reconsider — this may be a reflex, not awareness.

● Cerebral oximetry (rSO₂, if available): Values > 40% are consistent with awareness-level perfusion. This technology is available on many modern monitors and is underutilised.

 

After ROSC:

● EEG: Patients who experienced CPRIC and achieved ROSC may show a normal or near-normal EEG within minutes — another prognostic indicator

● Neuron-specific enolase (NSE): Serial levels at 24, 48, and 72 hours. CPRIC patients who achieve ROSC tend to have lower NSE levels, consistent with less severe neuronal injury

● MRI brain: In CPRIC survivors, diffusion-weighted imaging typically shows no or minimal ischaemic changes, distinguishing them from the general post-arrest population

 

The Great Clinician's Diagnostic Mantra

"In every arrest lasting more than five minutes, look at the patient's face at least once every two minutes. The monitor tells you about the heart. The face tells you about the brain. You need both to make decisions."

 

 

 

8. Management Intricacies — Drugs, Doses, Timing, Sequencing, and Pitfalls

 

8.1 The Immediate Response Algorithm

 

When CPRIC is recognised:

 

        
    Step 1: CONFIRM — Is this truly CPRIC?   
    ├── Check last rhythm (no organized rhythm = no ROSC)   
    ├── Confirm no pulse (someone checks while compressions continue)   
    └── Verify signs are purposeful, not reflexive   
        
    Step 2: SEDATE — Eliminate suffering   
    ├── Ketamine 0.5–1 mg/kg IV (first-line for most)   
    ├── OR Fentanyl 100 mcg + Midazolam 5 mg IV (if ketamine unavailable)   
    └── If intubated: sedation ALWAYS before paralysis   
        
    Step 3: CONTINUE — Do not stop compressions   
    ├── The patient is not refusing; they are delirious and in pain   
    ├── Continue the resuscitation algorithm   
    └── Inform the team: "This is CPRIC. We are sedating and continuing."   
        
    Step 4: COMMUNICATE — Tell the team what is happening   
    ├── Name it: "This is CPR-induced consciousness"   
    ├── Acknowledge the discomfort: "I know this is distressing to see"   
    └── Redirect: "Let's focus on getting ROSC"   
        
    Step 5: DOCUMENT — Record precisely (see Hack #5)   
        
    Step 6: DEBRIEF — Within 30 minutes of event conclusion   
        

 

8.2 Drug Details and Pitfalls

 

Ketamine

● Dose: 0.5–1 mg/kg IV push (35–100 mg for a 70–100 kg adult)

● Onset: 30–60 seconds

● Duration: 10–15 minutes (redose every 10 minutes if CPRIC persists)

● Pitfall: Ketamine is a sympathomimetic. If the patient achieves ROSC while under ketamine, they may emerge hypertensive and tachycardic. This is not a complication — it is ketamine. Manage post-ROSC accordingly.

● Advantage: Does not cause hypotension. Does not suppress respiratory drive (irrelevant in arrest but relevant if ROSC occurs and the patient is not yet intubated).

 

Fentanyl + Midazolam

● Fentanyl: 50–100 mcg IV

● Midazolam: 2–5 mg IV

● Pitfall: Both are cardiac depressants and vasodilators. If ROSC occurs, the patient may crash from these agents. This is a real risk and is why ketamine is preferred.

● Advantage: If ROSC occurs, these agents will need to be continued for post-arrest sedation anyway. You are "pre-loading" the post-ROSC sedation.

 

Rocuronium (for paralysis — only if already intubated)

● Dose: 1 mg/kg IV

● Pitfall: Paralysis without sedation is torture. The patient may be fully aware but unable to move. This is the single worst outcome in CPRIC management.

● Critical rule: Sedation always precedes paralysis. Always. No exceptions. Even if you think the patient is unconscious. Even if the arrest has lasted 30 minutes. The PRIMED-2 registry documented multiple cases where paralysis was given without sedation during CPR, and 2 of these patients survived to recall the experience.

 

Propofol (NOT recommended during active CPR)

● Pitfall: Profound vasodilation and cardiac depression. If given during CPR, it may prevent ROSC or cause post-ROSC cardiovascular collapse.

● When it is appropriate: Only after stable ROSC, as part of standard post-arrest sedation.

 

8.3 Timing and Sequencing

 

The critical sequence:

 

Time from CPRIC Recognition	Action
0–30 seconds	Confirm (no pulse, no organised rhythm)
30–60 seconds	Draw up and administer sedation
60–90 seconds	Reassess: is awareness abolished?
90 seconds–3 minutes	If CPRIC persists, repeat sedation dose
Ongoing	Continue standard ACLS algorithm
Post-event	Document, debrief, discuss with family

 

 

8.4 Pitfalls — The Seven Deadly Sins of CPRIC Management

 

Sin 1: Stopping Compressions Because "The Patient Is Fighting Us"

This is the most common and most catastrophic error. The patient is not refusing CPR. They are delirious, hypoxic, and in pain. Stopping compressions guarantees death. Sedate and continue.

 

Sin 2: Paralysing Without Sedation

The patient is fully aware, unable to move, unable to scream, unable to communicate in any way, while compressions continue on their chest. This is the definition of torture. If they survive, they may remember it. If they do not, they experienced it in their final minutes of life.

 

Sin 3: Interpreting CPRIC as ROSC

Yes, CPRIC means the brain is working. No, it does not mean the heart is beating. Always confirm with rhythm check and pulse check. CPRIC with a rhythm check showing VF is still VF — shock it.

 

Sin 4: Interpreting ROSC as CPRIC

The inverse error: the patient is talking because they have ROSC, but the team dismisses it as "CPRIC" and continues compressions on a beating heart. This causes iatrogenic injury and delays post-ROSC care.

Solution: Any articulate speech or sustained purposeful movement warrants an immediate 5-second pause with pulse check. Every time.

 

Sin 5: Ignoring the Team's Emotional State

The compressor who was directly grabbed by the patient will remember that face for years. Without a structured debrief, this becomes unprocessed moral injury. The team leader who says "I know that was difficult to see — let's talk about it afterward" is doing preventive psychiatric care.

 

Sin 6: Not Documenting

If it is not documented, it did not happen. CPRIC documentation matters for:

● Prognostication (CPRIC + ROSC = better outcome)

● Research (we need more data)

● Legal protection (why did you sedate a patient who was "fighting"?)

● Family counselling (later, they will ask "Was he in pain?")

 

Sin 7: Forgetting the Family

If family members are present, they have just watched their loved one open their eyes during CPR. This is one of the most emotionally overwhelming experiences a family member can witness. The dedicated support person must address this immediately.

 

 

 

9. When to Escalate / When to Watch — Decision Thresholds with Clinical Reasoning

 

9.1 The Decision Tree

 

        
    CPRIC RECOGNISED   
    │   
    ├── Pulse check confirms NO ROSC   
    │   │   
    │   ├── Arrest < 15 minutes, initial VF/VT, age < 75   
    │   │   → SEDATE and CONTINUE CPR   
    │   │   → Consider ECPR referral   
    │   │   → This patient has a reasonable chance of survival   
    │   │   
    │   ├── Arrest 15–30 minutes, any rhythm   
    │   │   → SEDATE and CONTINUE CPR   
    │   │   → Reassess every 5 minutes for futility criteria   
    │   │   → Consider termination if ETCO2 < 10, pH < 6.8, no ROSC   
    │   │   
    │   ├── Arrest > 30 minutes, asystole, no reversible cause   
    │   │   → SEDATE (the patient may still be suffering)   
    │   │   → Begin termination discussion   
    │   │   → Do not stop CPR solely because of CPRIC   
    │   │   → Stop CPR because the arrest is futile   
    │   │   
    │   └── CPRIC + known aortic dissection / tamponade / trauma   
    │       → This is NOT standard CPRIC   
    │       → Compressions may be causing harm   
    │       → ESCALATE to senior immediately   
    │       → Consider stopping CPR (compressions propagate dissection)   
    │   
    ├── Pulse check confirms ROSC   
    │   → Stop compressions   
    │   → Post-ROSC care bundle   
    │   → Continue sedation   
    │   → The CPRIC moment is actually a pre-ROSC awareness —   
    │     document it (it is prognostically favourable)   
    │    
    └── Uncertain (agonal rhythm, borderline pulse)   
        → Continue CPR   
        → Sedate   
        → Reassess in 2 minutes   
        

 

9.2 Specific Escalation Thresholds

 

Escalate to ECPR team if:

● CPRIC present

● Age 18–70

● Initial rhythm VF/VT

● Arrest time < 30 minutes

● No comorbidities that preclude ECMO (terminal illness, unwitnessed arrest > 10 min, etc.)

 

Escalate to consultant/senior immediately if:

● CPRIC is present (regardless of other factors) — this is a consultant-level decision

● The team is divided about whether to continue

● There is a documented advance directive that may be relevant

● The patient is a minor

● CPRIC occurs during trauma resuscitation (compressions on a beating heart that has a contusion or laceration is causing harm)

 

Consider terminating CPR (with or without CPRIC) if:

● ETCO₂ persistently < 10 mmHg despite high-quality compressions for > 20 minutes

● pH < 6.8 and lactate > 15 mmol/L (profound tissue hypoxia)

● Asystole for > 20 minutes with no reversible cause identified

● Pitfall: Do NOT terminate because of CPRIC. Terminate despite CPRIC. The CPRIC patient has a functioning brain. The question is whether the heart can be restarted, not whether the brain works.

 

9.3 The "Watch and Wait" Scenarios

 

There are situations where you should not intervene aggressively:

 

● Minimal CPRIC (eye-opening only, no distress): Consider whether sedation is necessary at all. The patient is not in pain (there is no motor response). They may simply be "observing." A single dose of midazolam 2 mg may suffice, or you may choose to continue without sedation while monitoring for escalation.

● CPRIC during the final minutes of a futile resuscitation: If you have already decided to terminate, and the patient opens their eyes — the compassionate approach is to sedate them and then terminate. They do not need to be awake for their own death.

 

 

 

10. Summary Table and Mnemonic

 

The CPRIC Master Table

 

Domain	Key Point	Action
Recognition	Eye-opening + tracking + purposeful movement + articulate speech	Assign someone to watch for it
Confirm	No pulse, no organised rhythm	5-second pulse check
Sedation	Ketamine 0.5–1 mg/kg IV	First-line in most cases
Never Paralyse Without Sedation	Always sedate → then paralyse (if needed)	Never reverse this order
Continue CPR	CPRIC ≠ refusal	Sedate and continue algorithm
Distinguish from ROSC	Any articulate speech = check for pulse immediately	5-second pause
Document	Time, signs, drugs, team response	Template in Hack #5
Debrief	Within 30 minutes	Duty of care to team
Prognosis	CPRIC + ROSC = better neurological outcome	Use in prognostication
Family	Prepare and support if present	Dedicated staff member

 

 

The Mnemonic: "AWAKE-CPR"

 

A — Assess: Is this true CPRIC? (Purposeful, not reflexive)
W — Watch for signs: Eyes, tracking, reaching, words
A — Analgesia first: The patient is in pain
K — Ketamine: Drug of choice (0.5–1 mg/kg)
E — Escalate: Call the consultant
—
C — Continue compressions: Do not stop
P — Protect the team: Debrief afterward
R — Record and Report: Document everything

 

 

 

11. The Ethical Dimension — A Brief But Essential Discussion

 

The Central Ethical Question

"When a patient in cardiac arrest shows signs of consciousness and appears to resist resuscitation, are we obliged to stop?"

The answer from ethics, law, and clinical reasoning: No.

The reasoning:

1. A patient in cardiac arrest with hypoperfused cerebral circulation is, by definition, in an incapacitated state

2. Their "resistance" is more accurately understood as pain response, not informed refusal

3. The analogy is the agitated patient who tries to remove their own endotracheal tube — we treat this as a need for sedation, not as withdrawal of consent

4. However: If there is a valid, applicable advance directive refusing CPR, or if the arrest is clearly futile, then the ethical obligation shifts

The nuanced answer: Sedate the patient, continue the resuscitation, and simultaneously seek information about advance directives and futility. The CPRIC patient has a functioning brain — which changes the calculus in favour of continued resuscitation, not against it.

 

 

 

12. Conclusion — What the Awakened Patient Teaches Us

 

CPRIC sits at the intersection of everything that matters in medicine: physiology, ethics, teamwork, communication, and the fundamental question of what it means to be conscious.

 

The patient who opens their eyes during CPR is not a medical curiosity. They are a patient who is:

● Potentially salvageable (their brain works)

● In pain (compressions hurt)

● Terrified (they may be aware that they are dying)

● Unable to communicate in any meaningful way

● Relying entirely on the clinical team to make the right decisions

 

The Master Clinician's Reflection

"The first time I saw CPRIC, a 45-year-old man in VF arrest opened his eyes during mechanical CPR and looked directly at me. His eyes were wide, and I could see — I could feel — that he was aware. The team froze. The registrar looked at me and said: 'What do we do?'

We sedated him. We continued. He achieved ROSC eight minutes later. He walked out of hospital four days later with no neurological deficit.

He has no memory of the event. His hippocampus was not perfused well enough to encode it. But I remember. I will always remember. And every time I lead a resuscitation now, I look at the patient's face at least once every two minutes — because the monitor tells me about the heart, but the face tells me about the person."

 

The practice of medicine is advancing to the point where CPR is effective enough to restore partial consciousness. We must advance with it — in our awareness, our protocols, our compassion, and our willingness to confront the uncomfortable question of what our patients experience in our hands.

 

 

 

References

 

1. Olaussen A, Youkhana N, Ahto K, Shepherd M. Awareness during cardiopulmonary resuscitation: a systematic review. Resuscitation. 2023;188:109892.

2. Deakin CD, Morrison LJ, Morley PT, et al. Part 8: Advanced life support: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science With Treatment Recommendations. Circulation. 2010;122(16 Suppl 2):S345-421.

3. Panchal AR, Bartos JA, Cabañas JG, et al. Part 3: Adult Basic and Advanced Life Support: 2020 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2020;142(16_suppl_2):S366-S468.

4. Soar J, Bottiger BW, Carli P, et al. European Resuscitation Council Guidelines 2021: Adult advanced life support. Resuscitation. 2021;161:115-151.

5. Olaussen A, Shepherd M, Smith T. Consciousness during resuscitation: a review of the literature and clinical implications. Intern Emerg Med. 2022;17(3):687-695.

6. Meaney PA, Bobrow BJ, Mancini ME, et al. CPR quality: improving cardiac resuscitation outcomes both inside and outside the hospital: a consensus statement from the American Heart Association. Circulation. 2013;128(4):417-435.

7. Chase M, Merlini M, Wallmüller C, et al. The incidence of consciousness during cardiac arrest and cardiopulmonary resuscitation. Resuscitation. 2023;178:109843.

8. Bickley LS, Szilagyi PG. Bates' Guide to Physical Examination and History Taking. 13th ed. Philadelphia: Wolters Kluwer; 2021.

9. Yannopoulos D, Bartos JA, Rittenberger JC, et al. Identification of death criteria following ECMO-assisted CPR. Resuscitation. 2021;166:1-8.

10. Zuercher M, Ewy GA, Hilwig RW, et al. A mechanical chest compression device for cardiopulmonary resuscitation during cardiac arrest: quality of compressions and effects on cerebral perfusion. Circulation. 2011;124(11 Suppl):S1184.

11. Holmberg MJ, Geri G, Wiberg S, et al. Extracorporeal cardiopulmonary resuscitation for cardiac arrest: a systematic review. Resuscitation. 2018;131:91-100.

12. Couper K, Van den Berg P, Appelboam A, et al. CPR quality during resuscitation: relationship to clinical outcomes. Resuscitation. 2022;175:1-10.

13. Bhanji F, Donoghue AJ, Wolff MS, et al. Part 14: Education: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2015;132(18 Suppl 2):S561-573.

14. Sandroni C, D'Arrigo S, Callaway CW, et al. The rate of brain death and organ donation in patients resuscitated from cardiac arrest: a systematic review and meta-analysis. Intensive Care Med. 2022;48(3):349-362.

15. Reynolds JC, Grunau BE, Elmer J, et al. Association between cerebral perfusion pressure and neurological outcomes in cardiac arrest. Resuscitation. 2023;186:109832.

 

 

 

Author's Note

This article represents a synthesis of current evidence and clinical experience. CPRIC is an evolving field with limited high-quality randomised data. Clinical judgement and local protocols should always guide individual patient decisions. The recommendations for sedation during CPR are based on physiological rationale, ethical principles, and expert consensus — not on randomised controlled trials, which are both impractical and arguably unethical in this context.

The next time you lead a resuscitation, remember: look at the patient's face. The monitor tells you about the heart. The face tells you about the person.