Bedside Monitoring of Respiratory Mechanics

Educational reference — a way to understand what the numbers behind the waveforms mean and how they are obtained. It is not a bedside directive, a substitute for a device’s operator’s manual, or a replacement for clinical judgment. Several of the measurements below are still research-grade or held at abstract depth in this corpus; where that is true, this chapter says so.

Why monitor mechanics at the bedside

A ventilator’s pressure, volume, and flow tracings tell you a great deal about the patient — but only up to a point. Reading a scalar waveform lets you infer the elastic load, the resistive load, and the presence of patient effort by watching how the curve deviates from its expected shape [Mireles-Cabodevila 2022]. Inference, however, has a hard limit: the ventilator sees only what happens at the airway opening. It cannot separate the pressure the patient’s muscles generate (Pmus) from the pressure it delivers itself (Pvent), and while the muscles are active you generally cannot reliably separate the resistive load from the elastic load either [Mireles-Cabodevila 2022]. To go past inference — to actually measure patient effort, the pressure across the lung, the division of work, or the energy a mode deposits in the tissue — you need instruments that reach beyond the airway sensor.

This chapter walks the three that the corpus uses. Esophageal pressure turns the ventilator’s blind spot for Pmus into a measurable signal. Work of breathing puts a number on how the total work of a breath is split between patient and machine — the magnitude axis of patient–ventilator interaction. Mechanical power asks a different question entirely: not who does the work, but how much energy the ventilator pours into the lung per minute, and whether that energy injures it. The first two are native to the Mireles-Cabodevila / Chatburn school. The third is not, and this chapter flags that boundary prominently wherever it appears.

Esophageal pressure: reaching past the airway sensor

flowchart LR
  Pmus["Muscle effort<br/>Pmus"] -.->|"reflected in"| Pes["Esophageal pressure<br/>Pes ≈ pleural"]
  Paw["Airway pressure<br/>Paw"] --> Ptp["Transpulmonary pressure<br/>Ptp = Paw − Pes"]
  Pes --> Ptp

Figure 1 — esophageal pressure (a pleural-pressure surrogate) yields transpulmonary pressure and a bedside estimate of muscle effort. Adapted from [Mireles-Cabodevila 2023 · Pes].

The single most useful thing a bedside clinician can do to escape the inference-only trap is to place an esophageal balloon catheter. The esophagus runs through the thorax, so the pressure inside a correctly placed balloon (Pes) tracks pleural pressure closely enough to serve as a window on what the chest wall and the respiratory muscles are doing. Over the last decade newer ventilators and bedside monitors have made this measurement feasible in routine critical care rather than only in the research lab [Mireles-Cabodevila 2023 · Pes] [established]. The corpus uses Pes for two distinct jobs.

First, estimating patient effort. Reading the magnitude and timing of the Pes swings lets you evaluate respiratory-muscle activity directly [Mireles-Cabodevila 2023 · Pes] [established]. This is exactly what waveform inference can only approximate. In the patient–ventilator interaction (PVI) framework, the “true” reference signal against which every interaction is judged is the patient’s muscle-pressure waveform, Pmus — and Pmus cannot be measured directly. It is estimated through surrogates: the electrical activity of the diaphragm (EAdi), thoracic–abdominal motion belts, the ventilator’s own flow and pressure waveforms, and esophageal pressure [Mireles-Cabodevila 2026 · PVI] [established]. Each surrogate carries technical limits and a time lag relative to the true muscle signal [Mireles-Cabodevila 2026 · PVI]. Of the practical options, Pes (with EAdi) is named as the reference standard for the patient signal — the thing you turn to when reading effort off the waveform is not enough [Mireles-Cabodevila 2022] [established].

Second, computing transpulmonary pressure. The pressure the ventilator displays is airway pressure, which is spent partly on the lung and partly on the chest wall. The pressure that actually distends the lung is the difference between the airway and the pleural space — and Pes supplies the pleural term. In the corpus’s careful vocabulary of what “pressure” can mean, transpulmonary pressure is a difference between two points in space:

Ptp = Paw − Pes

with the lung being, by definition, the structure that lies between the two points [Chatburn 2026 · Waveforms] [established]. Knowing Ptp lets a clinician evaluate whether the lung itself — not the whole respiratory system — is being over-stretched. The Pes primer lists respiratory-system driving pressure and transpulmonary driving pressure among its keywords [Mireles-Cabodevila 2023 · Pes]; driving pressure is classified separately as a pressure difference between two points in time [Chatburn 2026 · Waveforms] [established]. Together, transpulmonary and driving pressure are what tie esophageal manometry to the safety goal of ventilation — protecting the lung from its own treatment.

What we can and cannot say about the technique

Here the chapter must be honest about depth. The primer this section rests on is human-reviewed, but it is held at abstract depth in the corpus — the article body is not machine-extractable, so the actual how-to is not in our frozen artifact. The primer states plainly that, as with any measurement, technique, fidelity, and accuracy are paramount, and it flags named areas of uncertainty and ongoing development [Mireles-Cabodevila 2023 · Pes] [established]. But the specifics that make the measurement trustworthy — balloon placement, filling volume, the occlusion and validation tests that confirm the balloon is reading pleural pressure, artifact handling, and any numeric targets — live in the article body and are held [open], full text needed. This is a place to resist the temptation to fill in what “everyone knows”; the honest statement is that the corpus asserts a rigorous technique exists without, at this depth, teaching it. Do not treat Pes as plug-and-play from the abstract alone.

Work of breathing: the magnitude axis

Once you can measure Pmus, a second question opens up: of the total work needed to move a breath, how much is the patient doing and how much is the ventilator doing? This is work of breathing (WOB), and in the PVI framework it is one of the two axes along which every interaction is judged. The 2022 taxonomy paper names the two axes as synchrony — the timing of Pvent relative to Pmus — and work of breathing — the distribution of work between ventilator and patient [Mireles-Cabodevila 2022] [established]. The 2026 maxims paper reframes the same pair as timing and magnitude, with WOB as the magnitude axis and its signature phenomenon called work shifting [Mireles-Cabodevila 2026 · PVI] [established]. Timing and magnitude are independent: a breath can be perfectly in phase and still be badly mis-shared, and vice versa.

The physics is inherited straight from the equation of motion. During an active breath the patient and ventilator push together against the elastic and resistive loads:

Pvent + Pmus = E·V + R·V̇

so whoever supplies the pressure is supplying the work — and work is pressure integrated over volume, W = ∫P dV. That single identity is what makes work-shifting measurable in principle: if you know the pressure each party contributes and the volume moved, you know each party’s share of the work.

Measuring the split: the Campbell diagram and the Work Shifting Index

Turning “in principle” into a number is where esophageal pressure and WOB meet. The reference standard for partitioning work between patient and ventilator is the esophageal-pressure–based Campbell diagram, which the maxims paper flags as technically demanding and rarely used clinically [Mireles-Cabodevila 2026 · PVI]. It is the gold standard precisely because it is built on a real Pmus surrogate rather than on inference, which is why this section depends on the esophageal-pressure section that precedes it.

The 2026 maxims paper makes measurement its distinguishing contribution and offers a summary metric for the magnitude axis, the Work Shifting Index (WSI):

WSI = (Wpt / Wtot) × 100

the percent of the total passive work that has been shifted onto the patient. The scale is intuitive: 0% means the ventilator does all the work; 100% means the patient does all of it; and greater than 100% means Pmus is actually overriding Pvent — the patient is doing work on the ventilator, the domain of loaded, unproductive breaths [Mireles-Cabodevila 2026 · PVI] [established]. The WSI is noted as most tractable in simulations, and the paper is candid about the gap: at present no method gives a simple, continuous bedside index of work shifting [Mireles-Cabodevila 2026 · PVI] [open]. The Campbell diagram works but is technically demanding and rarely used clinically; a bedside-friendly continuous index remains an explicitly named research need.

Work the equipment imposes: the SBT example

WOB is not only what the patient’s disease demands — the equipment itself imposes work, and that imposed work can quietly corrupt a clinical decision. A bench study makes the point concretely in the context of the spontaneous breathing trial (SBT). Observational data had suggested that both a T-piece and a zero-pressure-support / zero-PEEP setting impose roughly the work a patient will face after extubation, so the two are often treated as interchangeable SBT modalities. On a breathing simulator across three lung models (normal, moderate ARDS, COPD) and three ventilators, the imposed WOB differed significantly between the T-piece and zero-PSV/zero-PEEP — and, crucially, the difference was ventilator-dependent and unpredictable in direction: on the Carescape R860 the zero/zero setting raised WOB by 5–6%, while on the Servo-u it lowered WOB by 15–21% [Sameed 2023] [established for the bench setting]. Because the same nominal setting behaves so differently across machines, zero-PSV/zero-PEEP is an imprecise SBT modality for judging extubation readiness — the number you are basing a liberation decision on partly reflects the ventilator, not the patient [Sameed 2023]. (The per-model WOB values and the exact computation are in the article body and held [open]; the percentage ranges above are quoted directly from the abstract and are safe to cite.)

Which goal WOB serves

Getting the work split right is the operational content of comfort. The 2013 mode-selection framework defines comfort as optimizing patient–ventilator synchrony, including coordinating the ventilator’s work output with the patient’s demand, analyzed through muscle pressure [Mireles-Cabodevila 2013] [established] — which is to say, getting work-shifting into an appropriate range. Too much unloading risks disuse; too little risks fatigue and injury. The SBT result above shows the same axis reaching into liberation: imposed work during weaning is part of what an SBT is supposed to measure, and it must be measured cleanly.

Mechanical power and VILI — an adjacent, out-of-school topic

Scope flag — read this first. Everything in this section sits

outside the single-school corpus. Mechanical power and ventilator-induced lung injury (VILI) are the framework of a different research lineage — the Gattinoni / Marini VILI-energetics school — and they enter this wiki only through Chatburn’s collaborations on two co-authored modeling papers. Treat what follows as an adjacent extension, not native taxonomy work. And note the second caveat, which the authors themselves insist on: both anchor papers are simulations — a mathematical model, in one case checked against a physical lung simulator — and their authors explicitly flag the VILI inferences as unvalidated in vivo. None of the numbers below is a bedside rule.

If work of breathing asks who does the work, mechanical power (MP) asks how much energy per unit time the ventilator delivers to the respiratory system. Formally it is the energy delivered per breathing cycle multiplied by the breathing frequency — energy per breath × frequency [Marini 2021] — and it has emerged as a construct that can meaningfully affect outcomes from mechanical ventilation [El-Khatib 2024] [established as a concept in the cited work]. Its principal components are the ones a clinician already sets: tidal volume, breathing frequency, and PEEP [El-Khatib 2024]. The appeal of MP is that it rolls several separately-titrated settings into one number that stands as a proxy for the injurious load a ventilator imposes.

What the settings do to power (El-Khatib 2024)

The first anchor is an interactive mathematical model of ventilator output, run across three inspiratory patterns — volume control with constant flow (VCV-CF), volume control with a descending-ramp flow, and pressure control (PCV) — for simulated mild, moderate, and severe ARDS. Whenever modeled MP exceeded the study’s chosen safety target of 17 J/min, one setting was manipulated to bring it back down [El-Khatib 2024]. The findings:

• VCV with constant flow always produced the lowest MP; pressure control produced the highest for matched conditions [El-Khatib 2024].
• Reducing tidal volume was the single most efficient lever for keeping MP at a safe/protective level [El-Khatib 2024].
• Safe-MP strategies trended toward lower-than-normal tidal volume plus higher-than-normal frequency, and the optimal volume–frequency pairing moved progressively lower and faster as ARDS worsened (from roughly 250–350 mL at 32–35 breaths/min in mild disease toward 200–300 mL at 37–45 breaths/min in severe) [El-Khatib 2024].

Two honesty notes travel with these results. The 17 J/min figure is a study-chosen target, a modeling assumption, not a validated universal cutoff. And the full per-severity tables and the model equations are in the article body and held [open] — the abstract-level results above are safe to state; the detailed grid is not held.

Intracycle power — when the energy arrives (Marini 2021)

Per-minute MP averages energy over a whole minute, which can hide when within a breath the energy is delivered. The second anchor sharpens this to intracycle power (ICP) — the instantaneous product of pressure and flow at each moment of a single inflation [Marini 2021]. Because a breath’s flow contour determines how the pressure–flow product is distributed over time, the shape of the flow waveform, not just its amplitude, becomes a variable of interest. The modeling — built on the same equation-of-motion single-compartment frame as the rest of the corpus, and verified against an ASL-5000 physical lung simulator before being run on 5,000 “virtual ARDS patients” — found:

• The common flow waveforms (constant, decelerating, exponentially decelerating = pressure control, and sinusoidal) deliver similar total alveolar energy per breath, but distribute it very differently over time [Marini 2021].
• Pressure control, with its abruptly rising flow, produced the highest maximal intracycle power; constant flow produced the lowest [Marini 2021].
• The elastic energy area did not vary much with flow contour; it was the power peak that discriminated sharply between flow patterns [Marini 2021].
• The authors therefore propose flow amplitude and waveform as relatively neglected, modifiable determinants of VILI risk in ARDS [Marini 2021] [proposed; model-based, not validated in vivo].

Notice how neatly the two anchors converge with the taxonomy’s first axis: the same control-variable and flow-waveform choice that the mode taxonomy uses merely to describe a breath turns out to govern how much energy is delivered (El-Khatib: VCV-CF lowest, PCV highest MP) and how peaky its delivery is (Marini: constant flow lowest, PC highest intracycle power). The description axis and the energetics axis point the same way. But — to repeat the scope flag — this is an inference from two simulations whose own authors call it a starting point awaiting experimental and clinical validation. It is not a licence to choose a mode “for VILI protection” at the bedside.

How this chapter connects

• The physics underneath all three tools is the same force balance that runs the rest of the wiki. Esophageal pressure supplies the Pmus term, work of breathing integrates pressure over volume, and mechanical power multiplies the pressure–flow product across the breath — every one of them is the equation of motion read for energy rather than for load. See Respiratory Mechanics and the Equation of Motion.
• Esophageal pressure and work of breathing are the measurement backbone of the magnitude axis of PVI. Pes is the practical Pmus surrogate; WOB and the Work Shifting Index quantify the split it makes visible. See Patient-Ventilator Interaction and Discordance.
• Each tool answers to a goal. Transpulmonary pressure serves safety; work-shifting serves comfort; imposed work in the SBT serves liberation; and mechanical power — with all its out-of-school caveats — is a safety construct. Choosing among modes and settings in light of these measurements is the subject of Goals of Ventilation and Choosing a Mode.

Open items and scope caveats

• Esophageal technique is [open]. The measurement procedure, occlusion / validation tests, and any numeric targets are held at abstract depth (PMID 37433629) — a full-text upgrade would resolve them. Bedside technique should not be presented as settled from the abstract alone [Mireles-Cabodevila 2023 · Pes].
• WOB and MP detail is [open]. The Sameed per-model WOB values and the El-Khatib per-severity MP tables live in article bodies not held in the corpus; only the abstract-level figures quoted here are citable. No simple continuous bedside index of work shifting yet exists [Mireles-Cabodevila 2026 · PVI] [open].
• Mechanical power / VILI is out-of-school and simulation-based. No paper native to the Mireles-Cabodevila / Chatburn taxonomy school synthesizes MP/VILI on its own terms; the corpus touches it only through Chatburn’s co-authorships with the El-Khatib (American University of Beirut) and Marini / Gattinoni (VILI) groups, and both anchors are simulations flagged by their authors as unvalidated in vivo [El-Khatib 2024; Marini 2021]. Widening scope to the primary VILI literature would require the human’s explicit go-ahead; until then this section is deliberately an adjacent extension, not a native chapter.

Sources

Every claim above is drawn from the frozen primary literature in the machine layer (raw/literature/, one artifact per paper). The short keys used inline resolve to:

• [Mireles-Cabodevila 2013] — Mireles-Cabodevila E, Hatipoğlu U, Chatburn RL. A rational framework for selecting modes of ventilation. Respir Care 2013. PMID 22710796.
• [Marini 2021] — Marini JJ, Crooke PS, Tawfik P, Chatburn RL, Dries DJ, Gattinoni L. Intracycle power and ventilation mode as potential contributors to ventilator-induced lung injury. Intensive Care Med Exp 2021;9(1):55. PMID 34719749. (Out-of-school collaborator paper; simulation.)
• [Mireles-Cabodevila 2022] — Mireles-Cabodevila E, Siuba MT, Chatburn RL. A Taxonomy for Patient-Ventilator Interactions and a Method to Read Ventilator Waveforms. Respir Care 2022;67(1):129–148. PMID 34470804.
• [Mireles-Cabodevila 2023 · Pes] — Mireles-Cabodevila E, Fischer M, Wiles S, Chatburn RL. Esophageal Pressure Measurement: A Primer. Respir Care 2023;68(9):1281–1294. PMID 37433629. (Held at abstract depth — technique [open].)
• [Sameed 2023] — Sameed M, Chatburn RL, Hatipoğlu U. Bench Assessment of Work of Breathing During a Spontaneous Breathing Trial on Zero Pressure Support and Zero PEEP Compared to T-Piece. Respir Care 2023;68(6):767–772. PMID 37225650. (Bench / simulator study.)
• [El-Khatib 2024] — El-Khatib MF, Zeinelddine SM, HajAli TH, Rizk M, van der Staay M, Chatburn RL. Effect of Ventilator Settings on Mechanical Power During Simulated Mechanical Ventilation of Patients With ARDS. Respir Care 2024;69(4):449–462. PMID 38538014. (Out-of-school collaborator paper; mathematical-model simulation.)
• [Chatburn 2026 · Waveforms] — Chatburn RL. How to interpret ventilator waveforms using the taxonomy for modes of mechanical ventilation. Respir Care 2026;71(6):566–587. PMID 41631602.
• [Mireles-Cabodevila 2026 · PVI] — Mireles-Cabodevila E, Vaporidi K, Blanch L, Chatburn RL. Defining and Measuring Patient-Ventilator Interactions: 10 Fundamental Maxims. Respir Care 2026;71(6):601–629. PMID 41913371.

This chapter is the human-layer synthesis of three one-concept-per-page records in the machine layer — Esophageal Pressure Monitoring, Work of Breathing, and Mechanical Power — and the Lit — … source summaries they cite (see the derived_from field). The mechanical-power / VILI material is retained as an explicitly labeled out-of-school, collaborator, simulation-based extension per the corpus’s single-school scope rule.