This document discusses the lack of agreement on what constitutes a valid breath and the importance of clearly defining breath criteria for clinical algorithms. It notes that breath definitions depend on clinical context and environment. Valid breaths are those that clear the deadspace volume and represent a patient effort, though small breaths may still indicate hypoventilation. The criteria should be optimized for each environment and therapeutic procedure. Without clear disclosure of breath detection methods, algorithms cannot be properly judged against clinical standards.

1. What is a “valid” breath? - Methodological Issues
Michael B. Jaffe PhD
Philips-Respironics, Wallingford, CT
 One dictionary defines a breath as “the air taken in and
expelled by the expansion and contraction of the thorax” while
the Webster’s Revised Unabridged Dictionary‘s definition
associates a breath with gas exchange and defines it as:
The air inhaled and exhaled in respiration; air which, in the process of
respiration, has parted with oxygen and has received carbonic acid,
aqueous vapor, warmth, etc.
 As in different dictionaries, there is in the medical community
generally a lack of agreement about what constitutes a valid
breath. A clear definition is particularly important to
developers of computer based algorithms which estimate
clinically important measures such as a respiratory rate, tidal
volume and end-tidal gas measurements that are considered
critical in the management of patients in clinical environments
ranging from pre-hospital to the OR and ICU.
1. Govindarajan N, Prakash O. Breath detection algorithm in digital
computers. Int J Clin Monit Comput. 1990 Jan;7(1):59-64.
2. Folke M, Cernerud L, Ekström M, Hök B. Critical review of non-invasive
respiratory monitoring in medical care. Med Biol Eng Comput. 2003
Jul;41(4):377-83
3. Karlen, W., Turner, M., Cooke, E., Dumont, G. Ansermino, J. M.
CapnoBase: Signal database and tools to collect, share and annotate
respiratory signals. 2010 Society for Technology in Anesthesia Annual
Meeting.
4. Orr JA, Brewer LM, Jaffe MB. Evaluation of Adequacy of Tidal Volumes
Using a Volumetric Capnography Reference Data Set. AARC 2010
5. Idris AH, Banner MJ, Wenzel V, Fuerst RS, V, Becker LB, Melker RJ.
Ventilation caused by external chest compression is unable to sustain
effective gas exchange during CPR: a comparison with mechanical
ventilation. Resuscitation 1994;28:143-150.
 The criteria for what constitutes a valid breath needs to clearly
defined, context specific and clinically relevant. Algorithms
need to better disclose their breath detection criteria and to be
judged against relevant bench and clinical standards.
Conclusions
Table 3 – Clinical Environments, Context and Breath Criteria
Table 1– Respiratory monitoring methods and variables that
can be estimated (Adapted from Folke et al., 2003 )
DiscussionDiscussionIntroductionIntroduction
 The output of computer based algorithms is dependent upon
the proper detection and clear definition of respiratory events
such as the start of breath (SOB), end of breath (EOB), and the
transition between inspiratory and expiratory phases. These
boundaries can be inferred in a number of ways – using
pressure, flow, a constituent component of the breath (e.g. gas
such CO2), combinations of these measurements (1) or through
more indirect measurements such as chest wall movement, or
acoustic measurements. (Table 1) (2)
 The definitions of SOB, EOB and what constitutes sufficient
volume to be considered a breath are dependent upon clinical
environment, context and technology. The criteria for what
constitutes a patient effort or breath may vary between the pre-
hospital and hospital environments (Figure 1) and context
(Table 2). Also what constitutes a useful gold standard needs
further clarification. In support of this waveform databases
with annotations and/or reference waveforms are being
developed. (3) It is suggested that the criteria for breath
detection and measurement should be optimized for the
environment of use, clinical expectations and therapeutic
procedure (e.g. procedural sedation, CPR, general anesthesia,
and invasive and non-invasive ventilation).(Table 3)
 The relevant clinical and physiological questions asked in
determining what is a breath also vary in a similar manner (e.g.
is the breath “effective”, does it clear the deadspace, and does
the breath represent a patient effort?). Folke (1) notes as a
caveat that devices providing only respiratory rate and lacking
information about actual gas exchange may have limited
clinical value.
 The issue of breath size and rate is readily apparent with some
algorithms where the reported breath rate can vary widely in
presence of artifacts and small patient efforts. A recent study
(4) using a large OR and ICU dataset found that the fraction of
breaths for which the tidal volume was too small to clear the
serial dead volume can be significant (Table 4), and that
algorithms which do not indicate the presence of very small
breaths may fail to indicate hypoventilation. Similarly,
algorithms must be careful to distinguish between chest
compressions which fail to clear the deadspace and
mechanically delivered breaths which likely will (5).
ReferencesReferences
Table 4 - Frequency of Inadequate Breaths* (4)
Patient Type
Percent Breaths
Too Small
N (Breaths) N (patients)
ICU, Adult 3.57 % 229,187 28
OR, Adult 0.6 % 32,331 38
Adult, Non-
intubated
17.3 % 28,078 55
Pediatric, OR 0.44 % 50,398 13
Volume (ml)
Gas movement > 0
Patient effort 10-20
Breath attempt (fails to clear deadspace but
sufficient to trigger)
< 150
Breath (clears deadspace and provides trigger) 150-750
Table 2 – Possible volume criteria for breaths (adult)
Respiratory
rate
estimation
Tidal
volume
estimation
Estimation
of CO2
elimination
Estimation
of O2
saturation
Notes
Airway sensing
Flow/pressure sensing ● ● ●* With CO2
Temperature sensing ●
Humidity sensing ●
Acoustic ●
Gas sensing (e.g.CO2) ● ●* With flow
Movement, volume and
tissue composition detection
Transthoracic impedance/
Inductance/ Fiber-optic
plethysmography
● ●
Strain-gauge transducers ● ●
Mutual inductance ● ●
Magnetometer ● ●
Capacitance displacement ● ●
Microwave radiation ●
Sensors in mattress ●
Photoplethysmography ● ● * Using POX
Muscle activity ●
Clinical
Environment
Context Representative Clinical Problems and
Breath Criteria
OR General
anesthesia
Small patient efforts (e.g. inadequate
anesthesia) may be obscured with sidestream
gas monitoring
OR Procedural
sedation
Small patient efforts may be labeled as breaths
indicating normal breath rate with
hypoventilation
Pre-hospital CPR Small fluctuations in volume and gas may or
may not indicate gas exchange
Pre-hospital Respiratory
distress
Nasal cannula placement problematic
ICU Invasive
ventilation
Patient asynchrony may complicate definition of
SOB/EOB
ICU Non-invasive
ventilation
Patient efforts/breath may be obscured in the
presence of mask leak
*inadequate breaths defined as fraction of breaths for which the tidal volume
was too small to clear the serial deadspace volume of the patient.
Figure 1- Different Criteria for different clinical environments .
With the capnogram as an example – CO2 waveforms in (a)
mechanically ventilated ICU patient with significant rebreathing
(with flow shown); (b) patient receiving procedural sedation with
small breath efforts; and (c) patient during CPR with compression
oscillations which fail to clear the deadspace (5)
a.
b.
c.
12/15/10 MBJ final