The Taxonomy of Ventilator Modes

Educational reference — a way to understand and name what a ventilator is doing. It is not a bedside directive or a substitute for a device’s operator’s manual or for clinical judgment.

Why a taxonomy exists

Open any modern ICU ventilator and you meet a wall of proprietary names — “pressure-regulated volume control,” “AutoFlow,” “SIMV,” “Spontaneous/Timed,” “IntelliVent-ASV.” A recent respiratory-care equipment textbook counted 495 unique mode brand names across 55 ventilators; the free Ventilator Mode Map catalog counts 569 names on 78 ventilators [Chatburn 2026 · Waveforms]. The names multiply, and — like drug brand names — they hide the fact that many different labels often describe the same underlying behavior. Worse, a brand name often does not describe the behavior at all: a mode called “volume control” may not control volume the way you would guess.

The response, originated by Robert Chatburn (with Frank Primiano) in 2001 and developed with Eduardo Mireles-Cabodevila from 2013 onward, is a taxonomy: a small set of definitions that let you translate any mode name on any ventilator into a short, unambiguous classification. Those 495 brand names collapse to 74 distinct classifications; the 569-name catalog collapses to 94 [Chatburn 2026 · Waveforms]. The taxonomy is a proposed standard — its authors advocate it, many current textbooks now reproduce it “virtually verbatim,” and it appears in a growing body of peer-reviewed work, but it is not universally adopted nomenclature, and manufacturers have not agreed on a common language [Chatburn 2026 · Waveforms; Chatburn 2007]. Keep that distinction in mind throughout: the physics the taxonomy sits on is settled; the vocabulary is a well-argued proposal.

The one idea to hold onto: a mode is nothing more than the predetermined pattern of interaction between the ventilator and the patient [Chatburn 2026 · Waveforms], and that pattern is fully specified by answering just three questions:

1. Control variable — does the ventilator control pressure or volume?
2. Breath sequence — how do mandatory and spontaneous breaths coexist (CMV, IMV, or CSV)?
3. Targeting scheme — what feedback logic does the ventilator use to hit its targets?

Answer those three and you have the mode’s tag — its generic name. The rest of this chapter is those three questions, how to combine them into a tag, and the worked examples that make it click.

Where it comes from: the equation of motion

The taxonomy is not arbitrary; it is read straight off the physics. The Equation of Motion for the Respiratory System describes every assisted breath with three variables — pressure, volume, and flow — tied together as

Pvent + Pmus = E·V + R·V̇ + PEEPauto

where the ventilator’s pressure (Pvent) plus the patient’s muscular pressure (Pmus) must overcome the elastic load (E·V) and the resistive load (R·V̇) [Chatburn 2026 · Waveforms; Mireles-Cabodevila 2022]. The crucial consequence: during inspiration, if you predetermine any one of pressure, volume, or flow, the other two are no longer free — they fall out of the mechanics. That single fact is what makes the first axis of the taxonomy possible.

Axis 1 — The control variable: pressure or volume

The control variable is the function the ventilator predetermines during inspiration [Chatburn 2014]. Because the equation of motion leaves only one degree of freedom, there are exactly two clinically meaningful choices:

• Volume control (VC): the ventilator presets both volume and flow before the breath. Because volume and flow are fixed, any change in the patient’s mechanics or effort has nowhere to go but the pressure waveform — so in VC you read the patient off the pressure curve [Chatburn 2014; Mireles-Cabodevila 2022].
• Pressure control (PC): the ventilator presets the inspiratory pressure–time waveform. That waveform can be a fixed shape (e.g., constant pressure) or pressure made proportional to the patient’s own effort (as in NAVA, where pressure tracks the diaphragm signal, or PAV). Because pressure is fixed, changes in mechanics and effort show up in the flow and volume waveforms [Chatburn 2014; Mireles-Cabodevila 2022].

Two subtleties matter for classifying real modes correctly:

• Setting a tidal-volume target is not volume control. VC requires that both volume and flow be preset. Several “volume” modes let the operator request a target VT while the ventilator decides the pressure and lets flow vary — those are pressure control with a volume target, not VC [Chatburn 2014]. This is the classic hidden misclassification the taxonomy exists to catch (PRVC is the textbook case) [Mireles-Cabodevila 2022] — a point a 2025 letter makes its very title: Volume Control Is Not Always Volume Control [Majumdar 2025].
• A residual “time control” category exists for the rare case where pressure, volume, and flow all depend on mechanics and nothing about the waveform is preset — only the timing is controlled. High-frequency oscillatory ventilation is the classic example [Chatburn 2014].

Identifying it at the bedside. Under simple (set-point) control, change the mechanical load and watch the waveforms: if peak inspiratory pressure stays constant as load changes, the control variable is pressure; if peak pressure moves but tidal volume stays constant, it is volume [Chatburn 2007]. (Careful: the control variable is what the ventilator controls, not what it manipulates — a device may manipulate flow moment to moment in order to control pressure.) When a mode switches between VC and PC within a single breath (dual targeting), the control variable is defined as whichever is in effect at the start of inspiration, because the second target may never be reached [Chatburn 2014].

Axis 2 — The breath sequence: CMV, IMV, CSV

The second axis describes how breaths are strung together. It is built on just two breath types, defined by who controls timing — specifically, who triggers (starts) and who cycles (ends) the breath:

• A spontaneous breath is patient-triggered AND patient-cycled — the patient controls both the start and the end, independent of any ventilator timing settings.
• A mandatory breath is machine-triggered AND/OR machine-cycled — the moment the ventilator is involved in either the start or the end, the breath is mandatory [Chatburn 2014; Chatburn 2026 · Five IMV Types].

To apply this you need to know how each trigger/cycle event is attributed: time-triggering and time- or **volume-**cycling count as the machine acting; pressure-, flow-, or diaphragm-signal-triggering and pressure- or **flow-**cycling count as the patient acting (in flow cycling, the patient’s own mechanics and effort set the inspiratory time) [Chatburn 2007]. A useful sharpening: in this taxonomy “assist” means the ventilator does some of the work of breathing — a ventilator assists even a paralyzed patient — so the old manufacturer usage of “assist” to mean “patient-triggered” is treated as obsolete [Chatburn 2026 · Waveforms].

With two breath types in hand, there are exactly three breath sequences:

Sequence	Definition	Set rate is a…
CMV — continuous mandatory ventilation	Spontaneous breaths are not possible between mandatory breaths (every valid patient trigger yields a machine-cycled breath). This is classic “assist/control.”	minimum — actual rate can be higher, never lower
IMV — intermittent mandatory ventilation	Spontaneous breaths are permitted between mandatory breaths.	maximum — patient efforts between mandatory breaths start spontaneous breaths
CSV — continuous spontaneous ventilation	All breaths are spontaneous.	(no mandatory breaths)

[Chatburn 2014; Mireles-Cabodevila 2022]

The “set rate is a floor vs. a ceiling” distinction is the practical heart of the difference between CMV and IMV, and it is exactly what the synchronization window governs. Historically, “SIMV” added an S for synchronized: near the end of each expiratory period a short window (≈5 s) lets a patient effort trigger and align the scheduled mandatory breath, and the ventilator then stretches expiratory time to hold the mandatory rate steady; the rest of expiration is the trigger window, where an effort instead starts a patient-triggered, patient-cycled spontaneous breath [Chatburn 2026 · Five IMV Types; Chatburn & Liu 2022]. The taxonomy drops the “S”: nearly every ventilator can be patient-triggered, so synchronization adds no classificatory information (though it is clinically desirable — it is associated with lower airway pressures and better interaction). The 2007 paper already called the “S” an anachronism [Chatburn 2026 · Five IMV Types; Chatburn 2007].

IMV is deep enough — and has evolved enough (from four “varieties” in 2022 to five types, IMV(1)–IMV(5), in 2026) — to get its own treatment. See Intermittent Mandatory Ventilation for the five mechanisms, their bedside trade-offs, and example ventilators.

Axis 3 — The targeting scheme: how the ventilator hits its targets

The first two axes tell you what is controlled and how breaths are arranged. The third tells you how smart the feedback is. A targeting scheme is the model relating operator inputs to ventilator outputs — usually a feedback control system — used to achieve the intended breath pattern; a target is a predetermined goal of ventilator output, which may be reached within a breath or adjusted between breaths [Chatburn 2014]. There are seven, each with a one-letter code used in the tag:

Code	Scheme	What it does	Example
s	set-point	Operator sets all waveform parameters; simple, but fixed settings can become inappropriate as the patient changes.	conventional VC and PC
d	dual	Automatically switches between VC and PC within one breath.	VC that can convert to PC on large effort
b	bio-variable	Randomly varies inspiratory pressure/VT to mimic natural breath-to-breath variability.	”variable” pressure support
r	servo	Output follows a varying input — pressure proportional to inspiratory effort.	PAV, NAVA, automatic tube compensation
a	adaptive	Automatically sets one target between breaths (e.g., pressure) to achieve another (e.g., average VT).	PRVC
o	optimal	Automatically adjusts targets to minimize/maximize an overall performance measure (e.g., work rate).	Adaptive Support Ventilation
i	intelligent	Adjusts targets using AI methods (fuzzy logic, rule-based expert systems, neural networks).	SmartCare/PS, IntelliVent-ASV

[Chatburn 2014; Mireles-Cabodevila 2013]

Primary vs. secondary, and compounds. A scheme is specified per breath type. CMV and CSV have a single breath type, so they carry one (primary) scheme. IMV has two — a mandatory (primary) and a spontaneous (secondary) — so it carries two, written with a comma between them (e.g., PC-IMVa,s = adaptive mandatory breaths, set-point spontaneous breaths). If two schemes act together on the same breath, their letters are run together with no comma (e.g., PC-CMVar = adaptive + servo) [Chatburn 2014; Mireles-Cabodevila 2013].

Spotting a scheme on the screen [Chatburn 2026 · Waveforms]:

• Set-point — the preset inspiratory pressure is independent of R and C; it does not move as mechanics change.
• Adaptive — inspiratory pressure decreases as inspiratory effort increases (the ventilator lowers its pressure target when the patient’s effort pushes delivered volume above the set target).
• Servo — inspiratory pressure is proportional to inspiratory effort.

Putting it together: the mode tag

flowchart TD
  Start(["Any mode"]) --> CV{"Control variable?"}
  CV -->|"pressure preset"| PC["PC"]
  CV -->|"volume and flow preset"| VC["VC"]
  PC --> BS{"Breath sequence?"}
  VC --> BS
  BS -->|"no spontaneous between mandatory"| CMV["CMV"]
  BS -->|"spontaneous permitted between"| IMV["IMV"]
  BS -->|"all breaths spontaneous"| CSV["CSV"]
  CMV --> TS["+ targeting scheme<br/>s / d / b / r / a / o / i"]
  IMV --> TS
  CSV --> TS
  TS --> TAG["Mode tag<br/>e.g. PC-IMVa,s"]

Figure 1 — classifying any mode in three steps: control variable → breath sequence → targeting scheme → the mode tag. Adapted from [Chatburn 2014].

Stack the three answers and you get the tag (a Taxonomic Attribute Grouping, or TAG): control variable – breath sequence – targeting letter(s). Combining the 2 control variables with the 3 breath sequences yields five basic ventilatory patterns:

VC-CMV · VC-IMV · PC-CMV · PC-IMV · PC-CSV

There is no sixth. VC-CSV is impossible: volume control requires the ventilator to cycle the breath (to guarantee the set volume), and machine cycling makes every breath mandatory — so no breath in a volume-controlled mode can be spontaneous [Chatburn 2014; Chatburn 2007].

Append the targeting letters and you have the full generic name. Four worked translations — brand name in, tag out [Chatburn 2014]:

Brand name (ventilator)	Reasoning	Tag
A/C Volume Control (PB840)	volume + flow set → VC; every breath volume-cycled → mandatory → CMV; operator sets all parameters → set-point	VC-CMVs
Pressure Support (PSV)	inspiratory pressure set → PC; all breaths patient-triggered and flow-cycled (= patient-cycled) → CSV; no auto-adjustment → set-point	PC-CSVs
PRVC (pressure-regulated volume control)	pressure adjusted between breaths to hit an average VT → PC, CMV, adaptive	PC-CMVa — not the VC mode its name implies
SIMV Volume Control Plus (PB840)	VT set but not flow → control variable is pressure; spontaneous allowed between mandatory → IMV; pressure adapted to an average VT (mandatory) + operator-set pressure support (spontaneous)	PC-IMVa,s

Notice what the tags reveal: “PRVC” and “SIMV Volume Control Plus” both sound like volume control but are pressure control — the kind of hidden truth the taxonomy exists to surface.

A mental model: the biological analogy

flowchart LR
  O["Order<br/>control variable"] --> F["Family<br/>breath sequence"] --> G["Genus<br/>primary targeting"] --> Sp["Species<br/>secondary targeting"] --> Var["Variety<br/>unique algorithm"]

Figure 2 — the taxonomy mapped onto the five Linnaean ranks. Adapted from [Mireles-Cabodevila 2013].

Mireles-Cabodevila’s 2013 framework maps the same structure onto the Linnaean ranks of biology, which many readers find the easiest way to remember the hierarchy [Mireles-Cabodevila 2013]:

• Order = control variable (VC or PC) — the broadest split.
• Family = breath sequence (CMV, IMV, CSV).
• Genus = the primary-breath targeting scheme.
• Species = the secondary-breath targeting scheme — present only in IMV, which alone has two breath types.
• Variety = a fifth level for modes that share order/family/genus/species but run a unique algorithm. For instance, the servo form of PC-CSV has three varieties: automatic tube compensation, NAVA, and proportional assist ventilation.

The taxonomy proper has four hierarchical levels (control variable → breath sequence → primary targeting → secondary targeting), with variety as an optional fifth for fine distinctions [Chatburn 2014; Mireles-Cabodevila 2013].

The formal backbone: ten maxims

The 2014 paper states the taxonomy as ten fundamental maxims — its axioms. You have already met all of them in this chapter; grouped, they are the scaffold that keeps the whole system internally consistent [Chatburn 2014]:

• Maxims 1–3 — the control variable. (1) A breath is one cycle of positive then negative flow. (2) A breath is assisted if the ventilator does some or all of its work (airway pressure rises during inspiration). (3) A ventilator assists using either pressure control or volume control, read from the equation of motion.
• Maxims 4–8 — the breath sequence. (4) Breaths are classified by their trigger and cycle criteria. (5) Trigger and cycle events are patient- or machine-initiated (with the trigger-window vs. synchronization-window distinction to keep the bookkeeping consistent — APRV is the worked case). (6) Breaths are spontaneous or mandatory based on both trigger and cycle. (7) There are three breath sequences (CMV, IMV, CSV). (8) Two control variables × three sequences give five ventilatory patterns, because VC-CSV is impossible.
• Maxim 9 — the targeting scheme. Within a pattern, modes differ by their targeting scheme; there are seven.
• Maxim 10 — assembly. A mode is classified by control variable + breath sequence + targeting scheme(s), across the four hierarchical levels.

To turn a mode name into a tag, the paper gives a three-step procedure: (1) identify the primary-breath control variable; (2) identify the breath sequence from the trigger/cycle analysis; (3) identify the targeting scheme(s) for the primary and — in IMV — the secondary breath [Chatburn 2014]. That is exactly the reasoning in the worked-examples table above.

Don’t confuse this with the PVI maxims. A separate set of “ten fundamental maxims” concerns Patient-Ventilator Interaction, not modes. The 2026 PVI paper deliberately mirrors this 2014 method and cites it as its model, but the two cover distinct subjects and neither supersedes the other [Mireles-Cabodevila 2026 · PVI]. See Ten Fundamental Maxims of Patient-Ventilator Interaction.

Why the payoff is large

Human working memory holds only a handful of variables at once, which is why raw mode names are unmanageable and why the shift from names to tags is not cosmetic [Mireles-Cabodevila 2013]. The collapse is dramatic and has been demonstrated repeatedly as the corpus grew:

• 2013: across four common ICU ventilators, 52 mode names (47 unique) reduce to 17 tags (≈22 modes once varieties are counted); one equipment text listed 174 names [Mireles-Cabodevila 2013].
• 2014: a catalog of 290 names across 30 ventilators → 45 tags [Chatburn 2014].
• 2026: 495 names / 55 ventilators → 74 tags; the Ventilator Mode Map app, 569 names / 78 ventilators → 94 tags [Chatburn 2026 · Waveforms].

The three most common adult modes worldwide remain Volume A/C, Pressure A/C, and Pressure Support — and the framework points out, pointedly, that these popular modes are technologically unsophisticated and not obviously the safest or most comfortable. Volume A/C’s staying power is nonetheless defensible: the one capability with hard outcome evidence is that lower tidal volume reduces mortality [Mireles-Cabodevila 2013].

The 2025 update: sharper labels, same structure

The most recent statement of the taxonomy (2025, full text) confirms the 2014 system without altering it — the ten principles, the three-component tag, the five basic patterns, and the seven targeting schemes are all restated verbatim. What it adds is notation, plus a Spanish-language glossary [Fajardo-Campoverdi 2025]:

• The IMV type goes inside the tag, in parentheses — PC-IMV(1)s,s, PC-IMV(2)s,s, VC-IMV(4)d,d, PC-IMV(4)ar,sr — folding the IMV(1)–IMV(5) distinction into the written name.
• The “variety” goes in square brackets — PC-CSVr[Flow] (ATC), PC-CSVr[Pmus] (PAV), PC-CSVr[Edi] (NAVA), PC-CSVs[CPAP] vs PC-CSVs[PS], VC-CMVd[PLimit] — the explicit written form of the fifth Linnaean level.

This is a labeling refinement, not a structural revision. (Earlier there was an open question about whether the 2025 paper might overturn the canon; the full-text PDF settled it — see [Fajardo-Campoverdi 2025] in the Sources.)

Beyond the ICU

The same three-part scheme has been applied to noninvasive modes on contemporary portable ventilators, showing it generalizes past invasive ICU machines and helps locate modes of comparable function across brands [Hatipoğlu 2024].

How this chapter connects

• The taxonomy is read off the physics — the equation of motion for the respiratory system.
• It is the first move in reading a tracing: identify the mode, then read the patient off the waveform opposite the control variable — see Reading Ventilator Waveforms.
• A mode is a pattern of interaction, which is why the same framework underlies Patient-Ventilator Interaction.
• Choosing among modes is driven by the Goals of Mechanical Ventilation (safety, comfort, liberation).
• Teaching all of this to mastery is the mission of SEVA (Standardized Education for Ventilatory Assistance) and its Ventilator Mode Map app.

Loose ends and honest caveats

• A real terminology shift in the corpus. The 2007 paper used the term “control type” with a different seven-item list (set-point, auto-set-point, servo, adaptive, optimal, knowledge-based, neural-network) organized into tactical/strategic/intelligent classes. The 2013/2014 papers reframed this as the “targeting scheme” axis with today’s seven (set-point, dual, bio-variable, servo, adaptive, optimal, intelligent). The lists do not map cleanly 1:1, and the exact correspondence is itself an unresolved corpus-evolution question the authors flag rather than settle: 2007’s “knowledge-based” and “neural-network” types are subsumed by the single 2013/2014 “intelligent” scheme (which covers rule-based expert systems and neural networks), while whether 2007’s “auto-set-point” (the within-breath VC↔PC switch) is simply the 2014 “dual” scheme is left open. This is a genuine evolution across the corpus, recorded rather than silently reconciled [Chatburn 2007; Chatburn 2014; Mireles-Cabodevila 2013].
• What “adopted” means. The taxonomy appears in textbooks and papers “virtually verbatim,” but adoption is not universality; manufacturers still name modes as they please [Chatburn 2026 · Waveforms].

Sources

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

• [Chatburn & Primiano 2001] — Chatburn RL, Primiano FP Jr. A new system for understanding modes of mechanical ventilation. Respir Care 2001. PMID 11353550.
• [Chatburn 2007] — Chatburn RL. Classification of ventilator modes: update and proposal for implementation. Respir Care 2007. PMID 17328828.
• [Mireles-Cabodevila 2013] — Mireles-Cabodevila E, Hatipoğlu U, Chatburn RL. A rational framework for selecting modes of ventilation. Respir Care 2013. PMID 22710796.
• [Chatburn 2014] — Chatburn RL, El-Khatib M, Mireles-Cabodevila E. A taxonomy for mechanical ventilation: 10 fundamental maxims. Respir Care 2014;59(11):1747–1763. PMID 25118309.
• [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. PMID 34470804.
• [Chatburn & Liu 2022] — Chatburn RL, Liu PH. The Evolution of Intermittent Mandatory Ventilation. Respir Care 2023;68(3). PMID 36195349.
• [Hatipoğlu 2024] — Hatipoğlu U, Lewarski J, Chatburn RL. A taxonomy for the modes on contemporary portable ventilators used for noninvasive ventilation. Respir Care 2024. PMID 38290749.
• [Fajardo-Campoverdi 2025] — Fajardo-Campoverdi A, Mireles-Cabodevila E, Chatburn RL, et al. Update of the taxonomy of mechanical ventilation modes. Med Intensiva 2025. PMID 40480865.
• [Chatburn 2026 · Waveforms] — Chatburn RL. How to interpret ventilator waveforms using the taxonomy for modes of mechanical ventilation. 2026. PMID 41631602.
• [Chatburn 2026 · Five IMV Types] — Chatburn RL, Liu PH, Mireles-Cabodevila E. Characteristics of the five different types of intermittent mandatory ventilation. 2026. PMID 41830556.
• [Mireles-Cabodevila 2026 · PVI] — Mireles-Cabodevila E, Vaporidi K, Blanch L, Chatburn RL. Defining and Measuring Patient-Ventilator Interactions: 10 Fundamental Maxims. 2026. PMID 41913371.

Recent addition — metadata-only:

• [Majumdar 2025] — Majumdar U, Chatburn RL, Mireles-Cabodevila E. Volume Control Is Not Always Volume Control. Chest 2025;168(4):e129–e130. PMID 41073049. (A brief letter, cited only for its title-level thesis; specifics remain [open] pending a full-text PDF.)

This chapter is the human-layer synthesis of six one-concept-per-page records in the machine layer — Control Variable, Breath Sequence, Targeting Scheme, Taxonomy for Modes of Mechanical Ventilation, Ten Fundamental Maxims (Mode Taxonomy), and Modes of Mechanical Ventilation — and the Lit — … source summaries they cite (see the derived_from field).