Pulmonary Disease · Chapter 1
The Pulmonary Decision Framework
How to think about psychotropic prescribing when the lung is the constraint — why there is no number to dose against, why sedation is the risk the blood gas won’t predict, and why the patient’s cigarettes move the pharmacokinetics more than anything you prescribe.
Bottom Line Up Front
The 30-second version
- Two independent axes, not one. Every other course in this series turns on a single organizing idea — QTc, clearance, metabolism, axis inversion. Pulmonary has two. The lung is a victim (sedating drugs depress a system with no reserve) and a wildcard (smoking induces CYP1A2 and moves your drug levels without a dose change). They fire independently. Sometimes they fire together.
- There is no Child-Pugh for the lung. FEV₁, GOLD grade, SpO₂, PaCO₂ — not one of them is a dosing instrument, and none reliably stratifies who tolerates sedation. The risk is dose-dependent, not threshold-dependent. Do not go looking for a blood gas that gives you permission.
- It’s the smoke, not the nicotine. The CYP1A2 inducers are the polycyclic aromatic hydrocarbons in combusted tobacco. Patches, gum, lozenges, and e-cigarettes do not induce CYP1A2. A patient on full nicotine replacement who has stopped smoking is, pharmacokinetically, a non-smoker — and their clozapine level is climbing. (Smokeless tobacco is less settled than it looks — see Chapter 2.)
- Two questions on every pulmonary patient. How sedating is this drug, and what else is stacked on it? (Axis 1, a drug-class question.) Is this patient smoking, and is that about to change? (Axis 2, a history question everybody forgets to ask.)
Which Number to Trust (Spoiler: None of Them)
The renal course gave you eGFR — continuous, drug-agnostic, actionable. The hepatic course took that away and handed you Child-Pugh, categorical but still label-linked. The lung gives you neither. There is no pulmonary equivalent of Child-Pugh, no FDA dose language written against GOLD grade, and no psychotropic label that stratifies by FEV₁. You will not find the number, because it does not exist.
What sits on the chart, and what each thing actually tells you:
| The number | What it measures | Why it won’t dose your drug |
|---|---|---|
| FEV₁ / GOLD grade | Severity of airflow obstruction | Describes the mechanical defect, not the ventilatory drive you are about to depress. Two patients at the same GOLD grade can have very different tolerance for a benzodiazepine. |
| SpO₂ | Oxygenation | Not ventilation. A CO₂ retainer can sit at 94% on room air while hypercapnic. Pulse oximetry is blind to the exact failure mode sedatives cause. |
| PaCO₂ | Ventilation — the intuitive candidate | The one you want to use, and the one the data don’t support as a stratifier. See below. |
| Drug level | Exposure | Nonexistent for the sedation axis — but on the smoking axis you have the best level in the whole series. That inversion is the point. |
The finding that should change how you practice
Your instinct — and everyone’s — is that the hypercapnic patient is the danger group and the normocapnic patient is safe. In Ekström’s prospective cohort of 2,249 patients on long-term oxygen for COPD, the associations between benzodiazepines and mortality were not modified by hypercapnia. The risk was dose-related, and it did not respect the blood gas. There is no PaCO₂ cutoff below which sedation is safe. Stop looking for one, and dose to the drug and the stack instead.
So the point-of-care question is never “How bad is the lung disease?” It is “How much CNS depression am I adding, to what else, in someone with no headroom — and is this patient’s CYP1A2 about to move?” Everything else in this course is those two questions applied to a drug class or a disease state.
The Two Axes
Hold these apart in your head. They are separate mechanisms with separate patients, separate drugs, and separate fixes — and the reason pulmonary is the hardest course in this series is that most clinicians only carry one of them.
| Axis 1 — The lung as victim | Axis 2 — The lung as wildcard | |
|---|---|---|
| Type | Pharmacodynamic | Pharmacokinetic |
| Mechanism | Sedating drugs depress respiratory drive in a system already at the floor. No altered metabolism required. | Polycyclic aromatic hydrocarbons in tobacco smoke induce CYP1A2, cutting levels of its substrates. Stop the smoke, induction fades, levels rise. |
| Who | Any patient with reduced reserve: COPD, OSA, neuromuscular weakness, advanced ILD, obesity hypoventilation. | Any patient on a CYP1A2 substrate whose smoking is changing — and the pulmonary patient is the one most likely to smoke, quit, relapse, or be forced to quit. |
| The drugs | Benzodiazepines, Z-drugs, opioids, gabapentinoids, sedating antipsychotics, sedating antidepressants, antihistamines. | Clozapine and olanzapine above all; also fluvoxamine, duloxetine, mirtazapine. Quetiapine is not a CYP1A2 substrate and largely sits this out. |
| Your tool | No number Clinical judgment, drug choice, and subtraction. | A real level Clozapine and olanzapine troughs are orderable and interpretable. |
| The error | Checking a gas and feeling reassured. | Never asking about cigarettes. |
Notice the inversion in the last two rows. The axis with no number is the one that kills, and the axis with a good number is the one nobody measures. That asymmetry organizes the entire course.
Axis 1: Respiratory Reserve as the Constraint
A healthy person absorbs a sedating drug because they have reserve to spend. Give the same milligram to someone whose ventilatory reserve is already committed to the work of breathing, and there is nothing left to give up: “tolerable drowsiness” becomes hypoventilation, CO₂ retention, and decompensation. The drug did not change. The margin did. No pharmacokinetic derangement is required — which is exactly why this axis is invisible to the reflexes you built in the renal and hepatic courses.
The population data are real, and more nuanced than the reflex:
| Study | What it found | How to read it |
|---|---|---|
| Vozoris 2014 177,355 Ontario adults ≥66 with COPD; new benzodiazepine users vs propensity-matched non-users |
New users were 45% more likely to need outpatient respiratory medications. In the subgroup with no exacerbation in the prior year, ED visits for COPD or pneumonia more than doubled (RR 2.46, 95% CI 1.90–3.18) and hospitalization rose (RR 1.29, 1.07–1.56). | The signal is exacerbation and acute care, not a clean mortality story — all-cause mortality in the main cohort was actually slightly lower among new users. Quote it for what it shows: benzodiazepines destabilize COPD. |
| Ekström 2014 2,249 Swedes starting long-term oxygen for COPD |
Benzodiazepines carried increased mortality (HR 1.21, 1.05–1.39) with a dose-response trend. Lower-dose opioids (≤30 mg oral morphine equivalents/day) showed no mortality signal (1.03, 0.84–1.26); higher doses did (1.21, 1.02–1.44). Neither drug increased hospital admission. | Dose-response is the headline. The low-dose opioid finding seeds the palliative inversion in Chapter 13 — but read it precisely: it shows low-dose opioids are not obviously dangerous, not that they work. Safety and efficacy are different literatures and the second one is weaker. |
The regulator has caught up on the sleeper agent in this axis. In a December 2019 Drug Safety Communication, the FDA warned that serious breathing difficulties may occur with gabapentin and pregabalin in patients with respiratory risk factors — naming COPD explicitly, alongside concurrent CNS depressants and older age — and required new respiratory-depression warnings in the prescribing information. Notably, the FDA also observed that the evidence is much weaker for healthy people taking a gabapentinoid alone. That is the shape of this entire axis: the drug is not the danger, the drug plus the stack plus the absent reserve is.
The additive-burden principle
- Nobody prescribes respiratory failure. They prescribe a benzodiazepine for anxiety, an opioid for dyspnea, a gabapentinoid for neuropathy, a Z-drug for the insomnia caused by the steroid burst, and a sedating antipsychotic for the agitation. Each was defensible. The sum was not.
- Count the CNS depressants on the list before you add the next one. The number, not the diagnosis, is your risk stratifier — and it is the only one you have. The mechanism and the arithmetic are Chapter 3.
Axis 2: The Smoking–CYP1A2 Lever
Here the ground moves under a dose you never changed. Tobacco smoke induces CYP1A2; clozapine and olanzapine are its substrates. Your patient’s smoking status is therefore a dosing variable — one that changes on their schedule, not yours, and that no lab flags for you.
The meta-analytic anchor: in Wagner’s 2020 systematic review and meta-analysis of 23 studies and more than 7,000 subjects, clozapine blood levels ran roughly a third lower in smokers than in non-smokers, with the authors recommending an approximate 30% dose reduction plus level monitoring when a patient quits. The dose-response is the part that surprises people. Haslemo put maximum induction at 7–12 cigarettes daily, but enrolled nobody below seven. Flanagan’s 231,707-sample TDM audit went lower and found the effect on plasma clozapine already maximal after 2–3 cigarettes a day in men — “perhaps fewer” — and near-maximal after some 4–5 a day in women. The practical translation: there is practically no such thing as a light smoker. Two a day is a fully induced enzyme, cutting down is not a kinetic intervention, and the patient who “barely smokes” rebounds exactly as hard as the two-pack-a-day one.
Four things determine whether you handle this well:
- Direction. Quitting raises levels toward toxicity. Starting or resuming lowers them toward relapse. Both are silent. Both are common.
- Timing. Induction does not vanish overnight — CYP1A2 activity falls over roughly the first three to four days and settles over about a week. Sedation on hospital day 3 is not a coincidence.
- It is the smoke, not the nicotine. The PAHs induce; nicotine does not. Patches, gum, and e-cigarettes leave CYP1A2 untouched. Nicotine replacement protects the cravings, not the level. (Smokeless tobacco is the one genuine uncertainty — Chapter 2.)
- Which drugs. Clozapine and olanzapine are the headline. Fluvoxamine, duloxetine, and mirtazapine are also affected. Quetiapine, risperidone, and aripiprazole largely are not — and knowing that is often the cleanest fix available.
Olanzapine deserves a caveat this chapter will honor and Chapter 5 will settle: the published magnitudes scatter badly — from a modest clearance difference in some pharmacokinetic work to several-fold differences in small concentration-to-dose studies. The direction is not in doubt; the multiplier is. Treat the widely used 1.5× dose-correction factor as a starting heuristic to be checked against a level, not as a number to trust.
Which Psychotropics Need Caution (Orientation)
The map, not the agent-by-agent dosing — that lives in the member chapters. Note that a drug can be dangerous on one axis and irrelevant on the other: lorazepam is an Axis 1 problem and an Axis 2 non-event; clozapine is an Axis 2 nightmare that is also sedating enough to matter on Axis 1. Sort every agent on both.
| Tier | Agents (orientation only) |
|---|---|
| Avoid / highest concern Axis 1 — direct respiratory risk |
Benzodiazepines — the defining pulmonary psychotropic risk, with population-level exacerbation and mortality signals, and the class most reflexively prescribed for the dyspnea–anxiety overlap that drives these patients to you. Opioid co-prescription — additive, and the combination carries FDA boxed warnings in both classes. Z-drugs — not a benzodiazepine in name only; the receptor and the respiratory consequence are close cousins. Gabapentinoids — low risk alone, FDA-warned in COPD and with CNS depressants. |
| Reduce / watch sedation load or a moving level |
Clozapine and olanzapine — both axes at once: sedating, and hostage to the smoking lever. Mirtazapine — sedation plus weight gain, which is an OSA problem you are creating. Sedating antipsychotics generally (quetiapine included — kinetically clean, pharmacodynamically not). TCAs — sedation plus anticholinergic drying of secretions. Fluvoxamine — a potent CYP1A2 inhibitor, which sets up the theophylline trap in Chapter 4. |
| Relatively preferred non-sedating, kinetically stable |
The non-sedating SSRIs and SNRIs (sertraline, escitalopram, venlafaxine) — no meaningful respiratory depression, and mostly outside the CYP1A2 story. Buspirone for the anxiety you would otherwise reach for a benzodiazepine to treat. Bupropion — non-sedating, and it doubles as a cessation agent, though that is its own decision with its own chapter. “Preferred” is not “free”: every agent still gets counted in the sedation stack, and dyspnea that is really air hunger will not respond to an anxiolytic strategy at all. |
Two moves carry most of the benefit in this course, and both are subtraction rather than addition: treat the dyspnea–anxiety overlap without a benzodiazepine wherever you can, and ask about cigarettes before you touch a CYP1A2 substrate.
Stable Disease vs Exacerbation: A Fixed Floor vs a Falling One
In stable pulmonary disease, reserve is reduced but predictable. You can build a regimen against it: choose the least-sedating effective agent, keep the stack short, know the smoking status, and monitor. The patient is not fragile so much as unforgiving — there is simply no slack for an error you would get away with in anyone else.
An exacerbation changes every input at once. Reserve falls; the sedation ceiling drops with it. A corticosteroid burst arrives, bringing insomnia, agitation, and sometimes frank steroid-induced mania or psychosis — the subject of Chapter 8, and the reason a psychiatrist gets called in the first place. The cigarettes usually stop, whether the patient chose that or not. And the psychiatric symptoms that prompt the consult — anxiety, panic, air hunger, insomnia — are precisely the ones that invite the drugs this chapter is warning you about.
So dose to the trajectory, not the snapshot. Two habits follow directly: reassess sedation load at every transition of care, because the level of respiratory reserve that made a regimen safe last month may not exist this week; and treat new confusion or drowsiness in a pulmonary patient on a sedating agent as drug-induced hypoventilation until proven otherwise — because at the bedside it is indistinguishable from the exacerbation itself, and the two managements diverge completely.
The Admission Trap: When Both Axes Fire at Once
This is the scenario that justifies the whole course, and it is not rare — it is routine, it is iatrogenic, and it is almost always misread as the disease progressing. A patient on clozapine is admitted with a COPD exacerbation.
| Timeline | Axis 2 (silent, kinetic) | Axis 1 (visible, pharmacodynamic) |
|---|---|---|
| Day 0 Admission | Cigarettes stop — hospital policy, not a clinical decision. Nobody documents it as a medication event. Nicotine replacement is offered, which helps the cravings and does nothing to the enzyme. | Baseline sedation load. Home regimen continues. |
| Days 1–3 | CYP1A2 induction begins to fade. The clozapine level starts climbing toward its non-smoking steady state — roughly a third higher — on an unchanged dose. | Prednisone causes insomnia → a Z-drug. Dyspnea causes anxiety → lorazepam. Air hunger → low-dose opioid. Three CNS depressants, each individually reasonable. |
| Days 3–5 | Level near or above the non-smoking steady state. Sedation, hypersalivation, unsteadiness. | Stacked depression of a drive with no reserve. Hypoventilation. Rising CO₂. |
| The read | Somnolent, confused, hypercapnic. The chart says “delirium, likely hypercapnic, COPD not improving.” Two iatrogenic mechanisms just summed, and neither is named. The clozapine dose was never changed — which is exactly the problem. | |
And then the mirror image, four weeks later, which almost nobody connects: the patient goes home, resumes smoking, CYP1A2 re-induces, the clozapine level falls by a third, and they relapse. That gets written up as non-adherence or loss of response. It is a pharmacokinetic event with a lighter, and it was predictable on the day of discharge. Chapter 2 takes this apart in full — it is the single most under-appreciated interaction in this course.
The Action Ladder
Any pulmonary diagnosis, stable
Choose the least-sedating effective agent and count the CNS depressants already on the list. Ask about cigarettes — number per day, and any plan to change. If the patient is on a CYP1A2 substrate, document smoking status as a medication-relevant variable, not social history.
Advanced disease: oxygen-dependent, hypercapnic, or severe OSA
Benzodiazepines and Z-drugs come off the table absent a compelling, documented reason. Treat the dyspnea–anxiety overlap by other means. Do not let a normal PaCO₂ reassure you — the risk is dose-related and does not track the gas.
Acute exacerbation or admission
Both axes are live. Recount the sedation stack from zero — the safe regimen from last month may not be safe today. Flag forced smoking cessation on day 0 as a pharmacokinetic event: for clozapine and olanzapine, anticipate the rise, check a level, and pre-plan the dose reduction rather than waiting for sedation to announce it.
Advanced, end-stage, or palliative
The calculus deliberately inverts — because the goal changes and the low-dose risk is acceptable, not because the drug is proven to work. The best-designed trial to date found no benefit of morphine or fentanyl over placebo for persistent dyspnea in COPD. Different goal, different arithmetic, honest uncertainty — handled in its own chapter.
Clinical Pearls
Pearls
- It’s the smoke, not the nicotine. The single most transferable fact in this course. A patient on a patch who has stopped smoking is kinetically a non-smoker for CYP1A2 purposes. If you remember one line from the free preview, make it this one.
- There is practically no such thing as a light smoker. Induction looks maximal by 2–3 cigarettes a day in men and 4–5 in women. “I cut down to half a pack” is not a kinetic intervention and will not soften the rebound — and the sex difference means the threshold is not even the same for the patient sitting next to them.
- There is no PaCO₂ that gives you permission. Hypercapnia did not modify the benzodiazepine–mortality association in oxygen-dependent COPD. Dose to the drug and the stack, not the gas.
- Count the depressants before you add one. Nobody prescribes respiratory failure; they prescribe the fifth reasonable sedative. The number on the list is your only real risk stratifier.
- Smoking status is a medication variable, not social history. If it lives only in the social history, it will not be reviewed when the dose matters — which is at admission and at discharge.
- Discharge is a dosing event. Resumed smoking re-induces CYP1A2 and drops clozapine and olanzapine levels. A relapse three to four weeks after discharge is a kinetic event until proven otherwise, not non-adherence.
- Quetiapine is kinetically clean and pharmacodynamically not. It sidesteps the smoking lever entirely but still counts on the sedation stack. Escaping one axis is not escaping both.
- Sedation in a pulmonary patient is hypoventilation until proven otherwise. New drowsiness or confusion on a sedating agent looks exactly like the exacerbation, and the two treatments diverge.
Red Flags — Stop and Reassess
Hold the agent and look closer
- New drowsiness, confusion, morning headache, or a falling respiratory rate in a pulmonary patient on any sedating psychotropic — assume drug-induced hypoventilation; hold, reassess, and check a gas for the CO₂, not the saturation.
- A patient on clozapine or olanzapine whose smoking has stopped for any reason — admission, a smoke-free facility, illness, or a genuine quit attempt. Levels are rising on an unchanged dose. This is an action item on day 0, not a wait-and-see.
- Three or more CNS depressants on the list of a patient with COPD, OSA, or advanced lung disease — regardless of how defensible each one was on its own.
- A benzodiazepine and an opioid running together in a patient with reduced respiratory reserve — the combination carries FDA boxed warnings, and this is the population in which it matters most.
- Relapse of psychosis three to four weeks after a hospital discharge in a smoker on clozapine or olanzapine — check a level before you conclude non-adherence or loss of response.
Patient Counseling Script
Plain-language script
“Two things about this medication and your lungs. First, some medicines for anxiety and sleep can slow your breathing down, and because your lungs are already working hard, you don’t have as much room to spare as someone else would. So if you or your family notice you getting unusually drowsy, hard to wake, confused, or waking with a headache, call us — that’s the medication, not your lungs getting worse, and we can fix it. Second, and this surprises everybody: cigarette smoke changes how fast your body clears this medication. Not the nicotine — the smoke. So if you stop smoking, or you’re in the hospital and can’t smoke, this medication builds up even though the dose hasn’t changed. Please don’t stop — quitting is one of the best things you can do — but tell me before you do, so I can lower the dose and check a level. Same the other way: if you start again, tell me, because the medicine may stop working.”
EMR / Documentation Template
References
- Vozoris NT, Fischer HD, Wang X, et al. Benzodiazepine drug use and adverse respiratory outcomes among older adults with COPD. Eur Respir J. 2014;44(2):332–340. (Increased outpatient exacerbations and emergency department visits with new benzodiazepine use; hospitalization elevated in the exacerbation-naive subgroup.)
- Ekström MP, Bornefalk-Hermansson A, Abernethy AP, Currow DC. Safety of benzodiazepines and opioids in very severe respiratory disease: national prospective study. BMJ. 2014;348:g445. (Benzodiazepine–mortality association with dose-response; associations not modified by hypercapnia; lower-dose opioids not associated with increased admissions or deaths.)
- Wagner E, McMahon L, Falkai P, Hasan A, Siskind D. Impact of smoking behaviour on clozapine blood levels — a systematic review and meta-analysis. Acta Psychiatr Scand. 2020;142(6):456–466. (23 studies, >7,000 subjects; clozapine levels approximately one-third lower in smokers; ~30% dose reduction with monitoring recommended on cessation.)
- Haslemo T, Eikeseth PH, Tanum L, Molden E, Refsum H. The effect of variable cigarette consumption on the interaction with clozapine and olanzapine. Eur J Clin Pharmacol. 2006;62(12):1049–1053. (7–12 cigarettes daily probably sufficient for maximum induction; 50% lower starting dose in non-smokers. No subjects enrolled below 7/day.)
- Flanagan RJ, Hunter S, Obee SJ. Effect of cigarette smoking on clozapine dose and on plasma clozapine and N-desmethylclozapine (norclozapine) concentrations in clinical practice. J Clin Psychopharmacol. 2023;43(6):514–519. (231,707 TDM samples; effect appeared maximal after 2–3 cigarettes daily in males and near-maximal after ~4–5 in females — filling the band Haslemo lacked.)
- Faber MS, Fuhr U. Time response of cytochrome P450 1A2 activity on cessation of heavy smoking. Clin Pharmacol Ther. 2004;76(2):178–184. (Time course of de-induction after cessation.)
- U.S. Food & Drug Administration. FDA warns about serious breathing problems with seizure and nerve pain medicines gabapentin (Neurontin, Gralise, Horizant) and pregabalin (Lyrica, Lyrica CR) — when used with CNS depressants or in patients with lung problems. Drug Safety Communication, December 19, 2019. (COPD named as a respiratory risk factor; new respiratory-depression warnings required in labeling.)
- Zhang H, Zhang Y, Sheng S, et al. Global prevalence and risk factors of depression in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis from 2000 to 2022. J Psychosom Res. 2023. (Pooled prevalence of variably defined depression 34.5%, 95% CI 30.9–38.1; odds 3.53-fold higher than non-COPD comparators. Estimates across the literature range widely by instrument and population.)
- Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global Strategy for the Diagnosis, Management, and Prevention of COPD. (Severity classification — a disease-staging tool, not a psychotropic dosing instrument.)
Last reviewed July 2026. Part of the Psychiatry Education Forum Academy; for clinician education — it supports, and does not replace, individual clinical judgment and current local protocols.
You have the two axes. Now learn the one that hides.
This chapter gave you the framework: no number to dose against, sedation as the risk the blood gas won’t predict, and cigarettes as a dosing variable. The next chapter is the free preview that most psychiatrists tell us they wish they’d read years earlier — the smoking–CYP1A2 lever, and the forced-quit toxicity trap that sends patients to the ICU on an unchanged dose.
Educational use only. Refer to the sources cited above and current prescribing information for clinical decisions. Psychiatry Education Forum and authors assume no liability for use of this material.
Responses