British Journal of Anaesthesia 91 (2): 249±64 (2003)

DOI: 10.1093/bja/aeg154

REVIEW ARTICLES
Hidden hazards and dangers associated with the use of
HME/®lters in breathing circuits. Their effect on toxic metabolite
production, pulse oximetry and airway resistance
E. G. Lawes

Shackleton Department of Anaesthesia, Southampton University Teaching Hospitals, Tremona Road,

SO16 6YD, UK
E-mail: riclawes@ntlworld.com

Br J Anaesth 2003; 91: 249±64

Keywords: complications; equipment, ®lters

Filter technology industrial applications such as microelectronic clean rooms,

Current airway dust ®lter technology emanates from the but have too high a resistance for medical breathing
USA mining industry during the 1930s. It was further apparatus.
developed and improved by the USA military during World All HEPA ®lters are manufactured from glass ®bre
War II to prevent workers from inhaling very ®ne materials supported on a rigid frame. In order to reduce
radioactive particles in the nuclear industry. These later resistance to air ¯ow and increase ef®ciency, the surface
®lters we now recognize as high ef®ciency particulate air area is increased by pleating. Filtration is achieved for larger
®lters (HEPA). particles (>0.3 m) by inertial impaction and interception;
The ef®ciency of ®lters varies according to the number smaller particles are captured by Brownian diffusion. The
and size of particles they capture. Filters may be classi®ed size of particles used to test ®lters is measured in microns.
commonly as 95, 99.95, and 99.97% ef®cient (95, 99 and Microns are units used for particles that can be seen with a
100, respectively). The higher the ef®ciency the higher the light microscope. A micron is a thousandth of a millimeter,
classi®cation number used. The ef®ciency of ®lters may which is in turn a thousandth of a meter (1 m=1000 nm or
also be affected by volatile chemicals which may be inhaled 0.001 mm). The variation in ®ltration ef®ciency is tested by
with the particles. Industrial ®lters which are categorized by the British BS3928 Sodium Flame method and the USA Hot
the above system may also have a pre®x of N, R, or P; N for DOP method. The most dif®cult particle size to capture by
Not resistant to oil, R for Resistant to oil, and P for oil Proof; ®ltration is a particle of 0.3 m as at this size the effects of
thus, a ®lter would be classi®ed as for example a P100. inertial impaction, interception and Brownian motion are
Requirements for military and environmental biohazard least effective. Particles of 0.3 m are also most likely to
®lters are based on these classi®cations and surprisingly be deposited in the lungs if inhaled. The di-octyl-phthalate
may only achieve 95% ef®ciency. (DOP) test, used for testing the ef®ciency of ®lters, takes
A generally perceived increased threat from biological advantage of the properties of DOP. In particulate form,
hazards that may be inhaled, including bioterrorism threats, DOP has a constant mean diameter of 0.3 m. By passing
especially in hospital environments, has increased interest in DOP through the ®lter, the capturing ef®ciency can be
respiratory ®lters.94 Protecting personnel from inhaled classi®ed. Bacterial size generally equates to 0.3 m; virus
biological hazards requires a ®lter ef®ciency of several sizes are considerably smaller (Table 1). Wilkes has
orders of magnitude greater than the industrial dust ®lters recently used the sodium chloride test for measuring ®lter
used in respirators alluded to above. As few as 10 inhaled ef®ciency on 33 breathing system ®lters (nine pleated
smallpox viruses may be suf®cient to infect patients with the hydrophobic and 24 electrostatic ®lters). The particles in
disease.94 This has resulted in the sub-classi®cation of this test are considerably smaller than those of the DOP
HEPA ®lters into true HEPA ®lters and HEPA type ®lters. A test and have a size distribution with a median diameter of
further sub group has also developed. ULPA ®lters (Ultra 0.07 m and a geometric standard deviation not exceeding
Low Penetration Air) and `absolute' ®lters are designed for 1.83.110

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2003
 Lawes

Table 1 Target organisms which ®lters are designed to capture can vary in size. A comparative table illustrates this point. Light microscopes resolve images
down to about 200 nm = 0.2 m. 0.2 m = 200 nm, m = 0.001 mm = 1000 nm

Bacteria Organism Size (m) Condition

Staphylococcus aureus 1.0 m Pneumonia, sepsis, renal failure

Pseudomonas diminuta 0.62 m Septicaemia
Pseudomonas aeruginosa 0.5 m Septicaemia, endocarditis, osteomyelitis, pneumonia
Mycobacterium tuberculosis 0.3 3 1.0 m Tuberculosis
Viruses Smallpox 0.3 3 0.2 m Smallpox
Orthomyxovirus 0.12 m In¯uenza
Cytomegalovirus 0.1 m Intrauterine birth defects, hepatitis
HIV 0.08 m Acquired Immunode®ciency Syndrome
Adenovirus 0.07 m Respiratory infections
Hepatitis C virus 0.03 m Hepatitis, cirrhosis, liver failure, hepatocellular carcinoma
Hepatitis A virus 0.02 m Hepatitis
T1-Coliphage 0.017 m Test virus for ®ltration ef®ciency testing
Cells Red blood cell 5m
Lymphocyte 5 to 8 m
Particles Smog (moisture droplet) 0.5 mm to <2 m Asthma

Table 2 The colour coding of medical ®lters to differentiate their uses

cerning the construction of the ®lter housing, connections
suitable for anaesthetic circuits, labelling, and packaging.
Heat and moisture exchanging ®lters (HMEF) Green The most important section of this standard is Section 6,
Heat and moisture exchanging (HME) Blue
Breathing ®lters Yellow `Test Methods'. This section ensures all devices are tested
by the same methods, under the same conditions of tidal
volume, breathing frequency, and duration of test (24 h),
thus facilitating objective appraisal of similar devices. The
A ®ltration ef®ciency of 99.999% indicates that only one
ef®ciency of a HME is tested by adding a precise level of
particle of 100 000 challenging the ®lter has the potential for
moisture at a precise temperature setting (34°C and 37.6 mg
penetration beyond the ®lter. Even if a ®lter package states
H2O litre±1 of inspired air) and is maintained during the 24 h
that it is 99.97% ef®cient it is not necessarily a true HEPA
test period. A detailed review of the structure of HMEs in
®lter unless this ef®ciency was achieved using particles of
relation to their function is given by Wilkes.109
0.3 m. HEPA type ®lters are made of the same materials as
true HEPA ®lters but they may achieve as little as 25%
ef®ciency. A third type of ®lter known as an electrostatic Ef®cacy of HME/®lters is not proven
®lter was, until recently, unable to achieve this level of The idea that the `wrong ®lter' can lead to disaster will come
®ltration, but some are now capable of achieving ef®cien- as no surprise to observers of modern history or investiga-
cies of 99.999%. tors of civil aviation disasters. Spectacular examples are the
Aside from presenting a barrier to inhalation of organisms abortive Iranian hostage rescue in the Middle East in the
and latex particles, ®lters may be modi®ed to perform early 1980s, during which the USA marines' helicopters
additional functions. These are to conserve patient's body were grounded by sand in their engines, and the failure of
heat and to ensure an adequate moisture content of inhaled Britain's main battle tank in the Gulf and more recently in
anaesthetic and respiratory gases in order to protect airways Oman. All were because of the inappropriate use of the
from drying out. With these modi®cations, ®lters function as engine air intake ®lters that were unsuited to the prevailing
heat and moisture exchanger ®lters (HMEF). The ®lters in environment and, in the latter cases, as the result of cost
industrial HEPA respirators are usually colour coded, for cutting. Similar problems have repeated themselves in the
example, magenta (reddish-purple), in order to quickly current Afghan crisis.
identify the purpose for which the ®lter is designed, that is to Recently, anaesthetic departments have been asking
protect against dust, volatile solvents, spray paints, etc. The whether they should carry the cost of single use, disposable
different functions of medical ®lters are also identi®ed by at anaesthetic circuits, discarded between cases, in order to
least one manufacturer (Intersurgicalâ) by colour coding avoid cross contamination between patients. Alternatively,
their products in a way similar to industrial ®lters thus should we rely on the properties of single use disposable
enabling the most appropriate use of the device (Table 2). ®lters, whilst retaining the circuits? Disposing of ®lters is
The manufactures of all heat and moisture exchangers cheaper than disposing of circuits, but which ®lters are the
(HME) are expected to adhere to voluntary standard ISO right ones and which are the wrong ones? Are we changing
9360 `Anaesthetic and Respiratory EquipmentÐHeat and one set of problems for another? The debate is neither new
Moisture Exchangers for Use in Humidifying Respired nor resolved but it would be advisable to examine these
Gases in Humans'.59 This standard provides advice con- options closely.

250
 Hazards and dangers associated with HME/®lters

Christopher and colleagues, and Wille highlight the risks persisted. Pelosi and colleagues warned that as HME
of cross infection that are presumed to exist following long- substantially increase minute ventilation, ventilatory drive,
term mechanical ventilation and short-term anaesthe- and work of breathing, these devices should be used
sia.29 115 Numerous articles and reviews attest to the carefully in patients with acute respiratory failure during
advantages of breathing system ®lters, circuit ®lters, pressure-support ventilation.83 However, they pointed out
HME, and HMEF, with or without bactericidal properties, that an increase in pressure-support ventilation (5±10 cm
introduced to avoid the risk of cross contamination. The H2O) might compensate for the increased work of breathing.
properties and performance characteristics of these devices Iotti and colleagues concluded that HMEs cause unfavour-
have been investigated by Hedley and Allt-Graham in the able mechanical effects by increasing inspiratory resistance,
laboratory under simulated clinical conditions.52 As a result ventilation requirements, and dynamic intrinsic PEEP.58
of this, and similar investigations that demonstrate the They warn clinicians to consider these effects when setting
bene®cial properties of ®lters in the laboratory, there is a mechanical ventilation and when assessing patients' ability
danger of implementing poorly thought out guidelines to breathe spontaneously.
requiring the use of ®lters in clinical settings without In contrast, Subayi and colleagues reviewed the literature
considering all of the consequences. Filters are being used in following a Medline search. They concluded that despite
the interests of economy and legislation, aside from clinical reservations such as those above and others like them, that
indications, and without regard to the properties of these the evidence in the literature not only supported using an
devices. The potential detrimental consequences of making HMEF but that they were highly recommended in intensive
such a simple change to clinical practice as to include one of care and are essential during anaesthesia.99
these devices in an anaesthetic circuit are rarely highlighted.

Filters in ITU breathing circuits

Filters in the ITU There is no doubt that circuit colonization occurs in
There is some overlap between the requirements for heat un®ltered long-term ventilated patients in ITU. No one
and moisture conservation and biological ®ltration for long- would risk cross contamination between patients by reusing
term mechanical ventilation in intensive care, and short- such circuits on consecutive patients. The signi®cance of
term ventilation using circle absorbers in operating theatres, circuit colonization in a patient's own breathing system is,
but there are also marked differences. It is instructive to however, largely unknown. Circuit colonization is not
consider the evolution of ®lters introduced in intensive care necessarily associated with an increased incidence of
units (ITU) following laboratory investigations. Their pneumonia. Dreyfuss and colleagues were able to demon-
introduction was not without mishap. strate that ventilator tubing contamination was considerably
reduced with the use of a heat and moisture exchanger.37 In
contrast, bacterial colonization of the pharynx and trachea
Airways resistance was identical to that found in the control group without a
Numerous studies highlight the increased resistance to HME. The authors suggest that provided the usual hygiene
breathing associated with ®lters. HME with bacterial and maintenance precautions are applied, circuit coloniz-
®ltering capabilities were investigated in an intensive care ation plays little or no role in the occurrence of ventilator-
environment by Cohen and colleagues.30 They concluded associated pneumonia.
that not only did this early model of a HME not provide The way in which HMEs are used in ITU is changing
suf®cient airway humidi®cation but also their use was subtly with doubts being cast on the necessity to change
associated with an increase in tracheal tube occlusion, ®lters frequently. The need to replace more recent designs of
atelectasis, and an increased incidence of pneumonia. HMEs on a daily basis during mechanical ventilation has
Not all technical problems had been resolved 6 yr later been investigated by Davis and colleagues.36 The authors
and it was noted that increased resistance could interfere investigated recently introduced hydrophobic or hygro-
with patient monitoring. Manthous and Schmidt were one of scopic HMEs and found that using the device for 3 days
the ®rst to criticize the indiscriminate use of humidi®ers in between changes does not diminish their ef®ciency, increase
ITU.70 They concluded that the humidi®er added a signi®- resistance, or alter bacterial colonization. They noted that
cant resistance to the ventilator circuit that may lead to with this regime there was also no increase in the frequency
incorrect assessment of respiratory system mechanics; to of nosocomial pneumonia. They concluded that it is safe and
inappropriate therapy (e.g. bronchodilators); or to dif®culty cost effective to use HMEs for >24 h, and certainly up to
in weaning from mechanical ventilation. Le Bourdelles and 72 h. Ricard and colleagues extended the change interval
colleagues also reported dif®culties weaning patients from and concluded that mechanical ventilation can be safely
mechanical ventilation when using HMEs in an ITU.63 conducted in non-COPD patients using an HME changed
These authors' ®ndings and their concerns were similar to only once a week, and that this is safe, ef®cient, and cost-
those of others dealing with weaning from mechanical effective.88 This interval between the need to change used
ventilators. Problems primarily related to airways resistance ®lters for fresh ones has been con®rmed by Han and

251
 Lawes

colleagues who investigated the possibility of retaining the provide additional guidance;112 however, Stevens warns of
HME for 7 days and also found the devices to be equally inconsistencies in performance between ®lter manufac-
effective after this interval.48 Littlewood and Durbin take turers, the consequence of which is that not all ®lters are
the argument a stage further and contend that there is little equally effective.98 This difference in ®lter ef®ciency
evidence to support using HME at all, and that there is between manufacturers may explain the contradictory
substantial evidence to support the changing of ventilator nature of articles investigating the bene®cial effects on
circuits no more frequently than once every 7 days.67 contamination of ®ltration of anaesthetic circuits.
However, until the change interval has been investigated The original report by Chant and colleagues that
further, it is important that anaesthetists understand the reawakened us to the possible role that liquid from an
signi®cance and responsibilities of using devices outside the un®ltered breathing circuit played in transmission of
limits of their product licences. Hepatitis C virus came from Australia.28 In this report,
consecutive use of the anaesthetic circuit was thought to be
the link between the patients. Hepatitis transmission through
Filters in anaesthetic circuits a reusable part of an anaesthetic system was thought to be a
Atkinson and colleagues found that the terms for ®lters, possible mode of transmission in a further incident reported
HMEFs and HMEs are commonly used interchangeably by Heinsen and colleagues.53 Although the virus involved
without reference to their ef®ciency in respect of either was clearly genetically identical in the two patients, this
bacteriological function or heat and moisture conservation.6 speci®c route of transmission could not be established
A consequence of the interchangeable use of the terms for conclusively.
®lters and HMEs is that anaesthetist's attitudes towards If cross contamination with hepatitis did occur in the
®lters are varied and inconsistent as to both their cost manner suggested, what would have been the effect of using
effectiveness and their ability to prevent cross contamin- a ®lter? If a ®lter is used, which one should it be? It is by no
ation between patients. Wilkes and colleagues remind us means clear that the presence of a ®lter would have
that the use of HMEs may be inappropriate for some patient prevented such an unusual means of transmission of
groups and the signi®cance of their placement in anaesthetic Hepatitis C, or even whether we can trust all ®lters to
circuits is poorly understood.106 This lack of differentiation protect our patients against this or any other organism if
between different devices may have an unexpected impact ®lters are used. The barrier to organism transmission varies
on the more complex circumstances of an anaesthetic circuit in effectiveness between different ®lter manufacturers'
compared with an intensive care ventilator, especially in products. Lloyd and colleagues investigated this problem
conjunction with circle absorbers. Elucidating the problems by examining the transfer of viruses through ®lters and
associated with ®lters in these circumstances depends more concluded that not all ®lters are equally suited to preventing
on laboratory investigations, case reports and correspond- the passage of viruses.68
ence rather than prospective clinical trials. Leijten and colleagues contested the old idea that
Current product licences require that single use anaes- halothane and soda lime had bactericidal properties and
thetic circuits be used once and replaced between patients to demonstrated, in a laboratory experiment, that without a
eliminate the potential, but unclear risk, of infection HMEF the whole interior of the anaesthetic circuits
between consecutive patients. The ®nancial implications becomes contaminated with bacteria thus supporting the
are enormous. The strategy of incorporating a fresh use of ®lters.64 The protection of the anaesthetic circuits by
disposable ®lter in the circuit and changing this, whilst a ®lter is also tentatively supported by Vezina and
retaining the circuit between cases, has been suggested by colleagues.103 These authors demonstrated that using a
Berry and Nolte as a cheaper alternative to discarding the sterile DAR Barrierbac S breathing ®lter for every patient
circuit.15 The consequence of this recommendation is that was highly effective at preventing, but not eliminating
®lters are now commonly placed on every anaesthetic contamination of the anaesthesia breathing circuit. They
circuit without reference to pertinent arguments regarding calculated that the use of this ®lter would result in a cross
the wisdom of such action. contamination rate of the breathing circuit less than once in
Laboratory studies point to the potential of bacterial cross every 250 cases.
contamination between patients using unprotected anaes- du Moulin and Sauberman investigated this problem
thetic circuits. The magnitude of potential cross contamin- many years ago and concluded that anaesthetic machines
ation between patients during routine and short procedures were unlikely sources of contamination and that basic
in ®t, well patients, whose immune system is not com- hygienic management of anaesthesia machines ensured
promised is, however, largely unknown. Contamination of patient safety.38 Garibaldi and colleagues, in a blinded
circuits is not the same as contamination occurring between prospective study of 520 patients undergoing anaesthesia,
two consecutive patients on an operating list. The investigated the use of ®lters during anaesthesia and
Association of Anaesthetists (GBI) have published recom- concluded that circuit ®lters played no part in the incidence
mendations regarding the use of bacteriological ®lters and of postoperative chest infection.43 Hogarth goes one step
their use in preventing cross infection.5 Wilkes and Stevens further, and points out that although ®lters can be shown to

252
 Hazards and dangers associated with HME/®lters

decrease circuit colonization in laboratory studies, they out-perform electrostatic ®lters, but were associated with
have not been proven to decrease ventilator-pneumonia increases in airway resistance. The variation in performance
infection rates.57 He contends that the use of ®lters is based again raises the question, which ®lter in which circum-
less on science than on defensive medical practice. In a stances? They remind us that ®lters are contraindicated in
clinical study, Rathgeber and colleagues found that only the presence of copious secretions (either sputum or
13% of anaesthetic circuits become contaminated with pulmonary oedema).
bacteria if used without ®lters.87 In this study the author In an attempt to untangle the results, Wilkes and
concedes that circuit contamination occurs, but despite this colleagues investigated the bacterial and viral ®ltration
he was unable to con®rm either an increase or a decrease in qualities of ®lters using the test methods speci®ed in the
the incidence of patient infections. More recently, Body and draft European standards BS EN 13328-1.106 Whilst vari-
Philip came to the conclusion that the use of ®lters may not ations in performance were found, the essential value of this
in¯uence the incidence of cross contamination and paper was to enable comparisons between ®lters to be made
expressed their reservations in using circuit ®lters.21 as a result of a common testing method. More recently, the
Following their investigations of water trap contamination performance characteristics of ®lters were found by Wilkes
on their anaesthetic machines despite using ®lters, they have to vary depending not only on their mode of construction
elected to use disposable circle anaesthesia circuits and a (pleated hydrophobic or electrostatic), but also the orienta-
regular (daily) cleaning programme without the use of tion of the ®lter layer, adding a further uncertainty to their
®lters. universal ef®cacy.110
The inconsistency between the results of these studies lies Some authors contend that even if a ®lter is used, it is not
in the failure of manufacturers, until recently, to state the suf®cient just to change ®lters between patients. Neft and
ef®ciency of their ®lters. Many manufacturers claim colleagues undertook a study on anaesthetized patients in
99.95% ef®ciency for their ®lters, whilst the early the operating room, comparing bacteriostatic vs non-
recommendations from Lumley and colleagues were that bacteriostatic HMEs (BHMEs/NHMEs) to assess their role
®lters should be at least 99.9977% ef®cient (i.e. only 23 per in preventing bacterial transmission to the anaesthesia
13106 organisms to pass through). 99.99999% ef®cient is breathing circuit. Neither HME prevented contamination of
10 times more ef®cient than 99.9999%, 100 times more the machine side of the circuit.81 Their results support
ef®cient than 99.999% and 1000 times more ef®cient than discarding breathing circuits even if a ®lter is used. The idea
99.99%.69 of throwing away a ®lter and retaining the circuit between
Even if ®lter ef®ciency is published by a manufacturer, a each patient is attractive on economic grounds, although
confounding factor is the microenvironment of the circuit Hess points out that this raises several ethical issues,
and the ®lter that develops in a clinical setting. Not all ®lters especially when some authors state that the ®lters they
perform in the same manner when presented with either wet investigated failed to keep the circuits clean, as alluded to
or dry challenges. Hedley and Allt-Graham published two above.55 Snowdon points out in his editorial that few
papers revealing their ®ndings that some ®lters are better at anaesthetists would suggest or know which medical breath-
®ltering dry particles, and some at wet particles or ¯uid.51 52 ing circuit ®lters are suf®ciently ineffective to risk cross
Mebius investigated, in the laboratory, the physical charac- contamination with hepatitis viruses or tuberculosis.95 Even
teristics, humidi®cation ef®ciency, ®ltration capability, and with a ®lter, most anaesthetists would prefer to discard both
resistance to ¯ow of six commercially available HMEs with ®lter and circuit in the above circumstances as the ef®ciency
built-in bacterial ®ltration.75 He also reported small but of any given ®lter is hard to identify at present.
signi®cant differences between average ®ltration capability Some attempt at standardization of biohazard ®lter
and humidi®cation ef®ciency between different manufac- ef®ciency has been made in the USA. Nardell looked at
turers. The clinical signi®cance of this range of ®lter biohazard ®lters for protecting workers against airborne
ef®ciency is very important when making the judgement to pathogenic organisms to which health workers are fre-
either use or omit the ®lter, and which other circuit quently exposed.80 Two organisms feature in this standard-
components should also be discarded between patients, ization, tuberculosis and hantavirus; infection with the latter
even if a ®lter has been used. Differences in ef®ciency are has a mortality rate of approximately 70%. Existing
not easily available to users; information on product inserts recommendations for protecting workers from biological
is lacking. hazards require the use of half-mask or full face-piece air-
Wilkes and colleagues investigated the ability of a mixed purifying respirators with particulate ®lter ef®ciencies
selection of electrostatic and pleated hydrophobic mem- ranging surprisingly from N95 (95% ef®cient for hazards
brane ®lters to prevent the passage of ¯uid across ®lter such as pulmonary tuberculosis) to P100 (99.97% ef®cient,
membranes.107 113 Their model was designed to accurately referred to as 100% for hazards such as hantavirus) as a
re¯ect the behaviour of contaminated ¯uid present in used minimum level of protection.
anaesthetic circuits. Again, inconsistencies in performance Many commercial circuit ®lters now exceed this mini-
were demonstrated, and in some instances the ®lters failed mum level of protection. In the UK, safety documents
to form an effective barrier. Hydrophobic ®lters tended to relating to exposure to biohazards are published by the

253
 Lawes

Health and Safety Executive and COSHH, both accessible ®lters. Furthermore, patient studies fail to demonstrate
online at www.ukonline.gov.uk. HEPA ®lter terminology is differences in the incidence of infection in these two
still in use in the UK but has largely been discarded in the circumstances. Despite all this contradictory evidence, the
USA. P100 ®lters equate to HEPA ®lters, but this does not temptation is to assume that placing ®lters on every
distinguish true HEPA ®lters from HEPA-like ®lters. anaesthetic circuit is a prudent intervention, devoid of
Clearly terminology remains confusing. adverse consequences. This represents defensive medicine
In response to this dilemma of how effective ®lters are, and has little bearing on the known facts and, as so often
and in order to contain the costs of discarding circuits, happens because of hidden hazards, this policy may actually
initially one, then other manufacturers, developed highly do more harm than it seeks to avoid. No general article has
ef®cient products with the objective of enabling circuits to sought to warn of some of the numerous disadvantages
be retained whilst discarding ®lters between cases. They incurred by introducing such devices into anaesthetic
have obtained product licences indemnifying users to use a circuits or of the potential consequences of mixing circle
circuit for intervals up to a week, provided they use a fresh absorbers and HMEs.
®lter (that they also manufacture) for every case. This
product licence may not re¯ect the ef®ciency of this option
so much as it re¯ects the lack of evidence that there is a real Disadvantages of the use of HME ®lters
problem in routine anaesthesia, as has been noted above.
The product licence has various stipulations. Publications of Filter obstruction
this nature are of course helpful in demonstrating manufac-
The routine use of ®lters is not without hazard. Filters are
turers' willingness to respond to clinical problems. The data
designed to obstruct the passage of one component or other
are, however, inadmissible for evidence-based practice
of a gas or liquid presented to them. That this ability to
without independent veri®cation, as there is a clear con¯ict
obstruct should cross the cusp of selective obstruction to
of interest. Independent clinical veri®cation of the ef®cacy
total obstruction is, in certain circumstances, inevitable.
of this strategy is now beginning to emerge with demon-
Filters can obstruct completely, partially, suddenly or over a
strable cost savings. Daggan and colleagues demonstrated,
longer time span. Detection and interpretation of the
in a clinical study, that it is possible to reuse an anaesthetic
changes to resistance that occur is dif®cult. Extreme caution
breathing circuit at least twice if the circuit is ®ltered by one
should be exercised when the ®lter is out of sight, dif®cult to
of these highly ef®cient ®lters, but suggest that further access and concealed by sterile drapes. Following a near
studies are needed.35 fatal accident with an obstructed ®lter, McEwan and
Adopting a policy based on these product licences may colleagues recommended that the routine use of ®lters be
not be as expedient as it ®rst appears. From the viewpoint of discouraged.72 His recommendations some of which are
a company lawyer, without recording equipment batch listed in Table 3 are as valid now as when they were made in
numbers, implementing and auditing rigid hygiene training 1993.
protocols and monitoring all patients on an operating list for These recommendations cover most but not all of the
pathological organisms both before and after anaesthesia, problems associated with ®lter use. Other authors have
the indemnity provided by this product licence is moot. reported similar near fatal complications that have resulted
Proving a failure of ®ltration will be almost impossible. The as a consequence of partial or complete airways obstruction
laboratory and labour costs of implementing these changes during anaesthesia when using a ®lter.10 24 27 50 84 85
would seem to be in excess of potential savings. Similar problems of ®lter obstruction have been reported
Numerous publications attest to the bene®cial effect of by Barnes and Normoyle (Table 4) who, when attempting to
HMEs, HMEFs and ®lters variously on the economics of ensure adequate airway humidi®cation of an infant's airway
disposable devices, heat conservation, preservation of found that the ®lter obstructed.9 Active humidi®cation is a
tracheobronchial mucosal integrity, and the potential for contraindication to the use of the HME. Stacey and
reduction in cross contamination. Barbara and colleagues colleagues and others had similar experiences of obstruction
have even shown them to be effective in the ®ltration of following the use of nebulized medications.97 Walton and
latex particles.7 Laboratory studies, however, as noted colleagues reported two cases of ®lter obstruction during the
previously, show circuit colonization both with and without management of bronchospasm with the nebulized beta-2-

Table 3 Recommendations for the use of ®lters, as detailed by McEwan and colleagures

Filters, if used, should be placed in the inspiratory limb or at the tracheal tube connector
Routine use of ®lters should be discontinued
Expiratory volume meters should be installed on all anaesthesia machines
Anaesthesia machine pressure alarms should be located on the patient side of the expiratory valve
Each anaesthesia machine should have a self-in¯ating bag. If a problem with the anaesthesia circuit should occur, the self-in¯ating bag should be readily available

254
 Hazards and dangers associated with HME/®lters

Table 4 Complications reported during use of a ®lter with some of the relevant references

Complication Authors

Pneumothorax McEwan,72 Smith93

Disconnection Bengtsson,14 Caplan27
Disable Disconnect Alarm Milligan,77 Pryne,86 Campbell26
Increased Airway Aarhus,1 Barnes,9 Buckley,24 Stacey,97 Villafaine,104 Cohen30
Resistance
Obstruction Barton,10 Australian Therapeutic Device Bulletin,50 Prados,84 Prasad,85 Barnes and Normoyle,9 Walton,105 Williams and Stacey116
Delayed Induction Goddard,45 Grodin,46 Grodin,47 Janshon and Dudziak,60 Marcus,71 Da Fonseca34

agonist salbutamol.105 They comment that death or injury gradual obstruction of the ®lter or of adjacent devices is not
was avoided only by the timely diagnosis of breathing the only hazard, and HMEs may affect adjacent downstream
circuit obstruction. Diagnosis of breathing circuit obstruc- apparatus also.
tion originating in the ®lter can clearly be dif®cult. Both
Smith and colleagues,93 and Aarhus and co-workers warn
that this can be confused with bilateral pneumothorax and Filters and HMEs, and carbon monoxide
bronchospasm.1 production
Williams and Stacey also described a case of rapid airway In certain circumstances, carbon dioxide absorbers degrade
obstruction which was dif®cult to differentiate from either all currently available volatile anaesthetics to compounds
pneumothorax or bronchospasm following the onset of that are toxic, the most important of which is carbon
pulmonary oedema.116 These authors warn that even small monoxide. Various authors have reported this effect with
amounts of proteinaceous ¯uid in the ®lter may precipitate des¯urane, en¯urane, and iso¯urane, all of which may be
this type of critical incident. The risks of tracheal tube degraded to carbon monoxide.11 13 16 40 41 60
obstruction secondary to the use of HMEs for long-term Long anaesthetic maintenance times are not required for
ventilation were highlighted many years ago by Cohen and signi®cant carbon monoxide poisoning to occur. Short
colleagues.30 periods of exposure to des¯urane in the presence of
The risk of a sudden increase in airway resistance or of desiccated baralyme are potentially lethal. Berry and
airway obstruction, partial or complete, is particularly acute colleagues reported that after as little as 15 min exposure
during anaesthesia when ®lters are concealed beneath to this combination, the arterial blood gases of a ®t young
drapes or during anaesthesia in the prone position.10 ASA 1 experimental subject demonstrated the following
Given the uncertain ef®cacy of ®lters and the known results: oxygenated haemoglobin (HbO2), 63%; carboxy-
problems of ®lter obstruction, it would seem wiser to haemoglobin (COHb), 36%; and methaemoglobin (MetHb),
exclude the ®lter in circumstances where both visual and 1%.16 High levels of COHb are not always required to
physical access is dif®cult. In these situations, the circuit increase patient morbidity or mortality. Allred and col-
should be discarded between patients. leagues reported that during graded exercise in patients with
The performance characteristics of different HMEF types coronary artery disease, COHb levels as low as 2.9±4.5%
are clearly important in terms of breathing circuit patency. can exacerbate myocardial ischaemia in a dose related
In addition to variations in airway resistance intrinsic to the manner.3 Unexpected ST segment changes or a turbulent
®lter as it becomes wetted, HMEs may alter the function and anaesthetic on a Monday morning list may be the only clue
behaviour of attached anaesthetic or ventilator equipment. to carbon monoxide poisoning.
These changes in performance characteristics induced in Tibbles and Perrotta discuss the management of carbon
adjacent upstream equipment may evolve so subtlety that monoxide poisoning and emphasize that a 1 h exposure to
they are dif®cult to detect. The differences between 1000 p.p.m. carbon monoxide will produce approximately
hydrophobic and hydroscopic HMEs were highlighted in 25% COHb, suf®cient to cause severe neuropsychiatric
relation to long-term ventilation by Villafane and colleagues impairment, whereas 67% COHb causes death.101 Myers
who reported increased resistance in tracheal tubes and Britten, on the other hand, found a poor correlation
during mechanical ventilation when different humidi®ca- between the levels of COHb and the damage incurred to the
tion devices were used.104 patient. Some neurological damage can only be demon-
If a HME/®lter is not to be used at the patient end of the strated following cognitive testing and neurobehavioural
circuit, it is tempting to protect circle absorber systems from studies.79 Signs and symptoms of carbon monoxide toxicity
bacterial contamination by incorporating these devices are masked during and after anaesthesia. Although COHb
immediately before the absorber. The purpose of this is to cannot be detected by pulse oximetry, it is possible to detect
utilize the biological ®ltering function of the device. This COHb with blood gas analysers if they also have carbon
tactic does not take into consideration the consequences of monoxide-oximetry. Where elevated levels of carbon
the heat and moisture function of the device. Sudden or monoxide have been demonstrated, it may be reasonable

255
 Lawes

for patients to be followed up for evidence of postoperative An assumption is made that water production from the
cognitive dysfunction. Some would argue that carbon reaction between soda lime and carbon dioxide will be
monoxide production in circle absorber systems is a non- suf®cient to prevent desiccation during subsequent use.
problem or is only relevant if prolonged procedures occur. Fang and colleagues showed that drying of absorbent will
Woehlck warns that this apathetic position may generate occur at fresh gas ¯ow rates greater than 3 litre min±1.40
signi®cant harm.120 These authors investigated carbon monoxide production in
Janshon and colleagues reported two similar incidents of relation to anaesthetic agent used, fresh gas ¯ow, and the
carbon monoxide production that occurred during induction water content of the absorber and made some recommend-
of anaesthesia with both en¯urane and sevo¯urane. Delay in ations. Their results suggest that carbon monoxide gener-
induction, over-hot absorber canisters and visible conden- ation can be avoided for all anaesthetics by using soda lime
sate in the tubing were characteristic ®ndings in these cases. with 4.8% (or more) water or Baralyme with 9.7% (or more)
The machines had not been in use for 2 weeks.60 A common water, and by using in¯ow rates of less than 2±3 litre min±1.
factor with all of these case reports was that the anaesthetic They contend that such in¯ow rates are low enough to
machines were not in regular use, thus enabling desiccation ensure that the absorbent does not dry out. Baxter reported
of absorbent to occur with time. All the authors who that even lower (minimal) ¯ow rates are required to prevent
reported these cases have since changed their clinical drying.12 Baxter suggested ¯ow rates of <1 litre min±1 are
practice to ensure fresh absorbent is used at the beginning of required to eliminate this risk. Baum and colleagues take the
each day. As only dry absorbent has been implicated in this fresh gas ¯ow even lower and suggest that rates as low as
reaction, of especial concern is the observation that 0.5 litre min±1 are required to avoid drying.11
condensing water was visible in the anaesthetic tubing of Even with these low ¯ows, the effect of an HME placed
the above cases. This indicates that visible condensation is between the circuit and the absorber may be to tip the
not proof of adequate circuit hydration, contrary to the balance in favour of drying. In the absence of knowing how
®ndings of Beydon and colleagues17 who hold the view that much moisture is present in the absorbent. Woehlck and
if you can see condensation in the circuit tubing then colleagues suggest that in order to recognize and end
humidi®cation is adequate. patient exposures that occur, despite preventative measures,
The source of carbon monoxide in circle systems indirect mass spectrometer monitoring for carbon monoxide
remained a puzzle for some time until drying of the should be used during general anaesthesia.119 These
absorbent was implicated following investigations by workers also recommend that anaesthesia technicians and
various authors including Middleton and co-workers,76 housekeeping personnel should be instructed to turn off all
who made the original observation, Lentz,65 Moon,78 and anaesthesia machines after the last case of the day in each
Frink and colleagues.41 room.118 This has implications for anaesthetic machines
Brown and his co-workers have demonstrated that the whose oxygen failure alarm is battery operated with a life of
function of the absorbent is directly related to the degree of 40 h. Alternative strategies may involve the measurement of
hydration.23 The water content of circle systems is partly circuit humidity or inclusion of moisture reservoirs or
derived from the granules themselves either from residual moisture dye indicators within the circle absorber. A messy
moisture included at the time of manufacture or through prophylactic solution suggested by Woehlck and Kharasch
chemical production (1 mol of carbon dioxide produces is to dip the absorbent in water before the start of
1 mol (18 g) of water); or from exhaled moisture from the anaesthesia.117 However, even under these circumstances,
patient and from re-circulated moisture. carbon monoxide production may occur unless the water is
HME/®lters are designed to retain exhaled moisture on evenly distributed throughout the canister.
the patient side of the HMEF. They can only contribute to HMEs only contribute to patient humidi®cation if placed
patient humidi®cation if they are placed in the circuit where on the Y connector between the circuit and the patient, as
to and fro movement of gas occurs, that is at the Y piece. If a they depend on to and fro air movement. The relationship
HMEF is placed immediately before the absorber in an between circle absorber systems and HMEs has been
attempt to reduce bacterial contamination, not only will investigated to a limited degree. Henriksson and colleagues
moisture exhaled from the patient be largely prevented from investigated the effect of inserting HMEs into the Y joint of
rehydrating the soda lime, but also the re-circulated circuits on circle absorber systems.54 They wished to
moisture from the soda lime itself will be prevented from determine if the inherent inadequacy of humidi®cation of
rehydrating the granules. If the water content of soda lime or these systems at high fresh gas ¯ows (>5 litre min±1) could
Baralyme in particular, is reduced to such an extent that be compensated. They recommended that at high ¯ow rates,
desiccation occurs, carbon monoxide production occurs HMEs should be included in the circuit. The ¯ow rates
secondary to degradation of the volatile agents. Degradation investigated were beyond those commonly used for circle
is inversely related to absorbent water content. Baxter and absorber systems in clinical practice. The last measurements
Kharasch remind us that the greatest carbon monoxide were made 60 min after inclusion of the HME, a time when
concentrations occur with des¯urane and fully desiccated other authors have stated that HMEs are just reaching their
Baralyme.13 steady state. Bissonette and colleagues had demonstrated

256
 Hazards and dangers associated with HME/®lters

previously that in-line HMEs become more ef®cient at time on drying did not feature in their conclusion. Even if
retaining expired moisture with increasing duration of the water is added to soda lime at the beginning of a long
procedure, reaching maximum ef®ciency after 1 h.20 procedure, Soro and co-workers found, amongst other
Bisinotto and colleagues investigated the relationship factors, that drying occurs exponentially with time.96 The
between humidity and tracheobronchial damage. They time constant is directly proportional to the volume of soda
demonstrated that the intrinsic humidifying action of circle lime contained within the canister and is inversely propor-
systems at low ¯ow (<2 litre min±1) was adequate at 60 min, tional to the continuous fresh gas ¯ow rate.
but that at ¯ow rates of 5 litre min±1 (very high ¯ow rates) The effect of interposing an HME between the circuit and
extra humidi®cation was required if a recommended lower the absorbent is unknown but Wilkes and Mecklenburgh
limit of absolute humidity greater than 20 mg litre±1 is to be demonstrated that the moisture conserving capabilities of
achieved.19 The authors did not measure carbon monoxide ®lters and HMEs can be calculated from a formula for
production. different fresh gas ¯ows.111 Not all HMEFs are equally
The measurement of humidity and the time to a steady ef®cient. The effect of adding the most ef®cient HMEFs to
state within anaesthetic circuits can be achieved by a circle systems is likely to exacerbate the risk of soda lime
number of means including gravimetric hygrometry, dew desiccation by decreasing the natural replenishment of
point hygrometry, wet±dry bulb psychrometry, mass moisture from the patient's exhaled gases and the re-
spectrometry, spectroscopic hygrometry, and electrical circulated intrinsic water content of the absorber.
hygrometry. Kleemann investigated circle absorber humid- Moon investigated the relationship between drying and
ity by designing a custom-made apparatus suitable for carbon monoxide production.78 This author described 31
prolonged use under clinical conditions.62 He was able to cases of intraoperative carbon monoxide poisoning during
continuously measure humidity and temperature in the en¯urane or iso¯urane anaesthesia using circle systems that
inspired and expired gas mixtures of a breathing circuit. He became desiccated. Some carbon monoxide concentrations
validated his results against the standard psychrometer exceeded 1000 p.p.m. and COHb levels reached 30% or
technique. He demonstrated that at ¯ow rates below 1.5 litre more. It is clear that at high fresh gas ¯ows, loss of moisture
min±1, an absolute humidity of 21.3 mg H2O could be through the expiratory waste gas valve can exceed water
achieved after 120 min. Flow rates of greater than 3.0 litre production from the reaction between soda lime and carbon
min±1 were found to be too high for maintaining humidity dioxide.
during prolonged anaesthesia. Although his study sought the Carbon monoxide monitoring is not part of the standard
relationship between fresh gas ¯ow, humidity, and tracheo- gas analysis on anaesthetic machines but would identify if
bronchial damage, the results con®rm that ¯ow rates and when carbon monoxide production occurred. Although
through circles should be below 3 litre min±1 to ensure both industrial and domestic carbon monoxide monitors are
adequate humidi®cation of the respiratory tract. The now commonplace and cheap, some are affected by
implication is that ¯ow rates in excess of this cause halogens, ammonia, and nitrogen gases. Adding an appro-
tracheobronchial drying as a result of drying of the absorber priate device for measuring carbon monoxide to standard
granules. The possible detrimental effects on these results of anaesthesia monitoring may be an inexpensive option that
an HME placed between the circuit and the absorbent were does not require mass spectrometry.119
not investigated, but the results clearly demonstrate that
drying of anaesthetic circuits does occur and that intrinsic
water production through the chemical reaction of soda lime
and carbon dioxide is inadequate to maintain humidity at Filters and HMEs, and compound A
higher ¯ow rates. Compound A (2-(Fluoromethoxy)-1,1,3,3,3-penta¯uoro-1-
Flow rates of less than 3 litre min±1 are required to protect propene) is variously known as a halo-alkene or a vinyl
the respiratory tract from drying, but what ¯ow rates are ether, or a vinyl halide. It is one of several products formed
required to protect the absorbents from drying out and are when sevo¯urane dehydro¯uorinates in the presence of
they the same? Fang and colleagues have investigated the strong bases in soda lime and barium hydroxide lime,
minimum per cent water in both baralyme and soda lime especially at higher temperatures. Compound A is the most
required to avoid carbon monoxide production.40 These common sevo¯urane degradation product in anaesthesia
authors state that soda lime containing 4.8% or more water machines. Although these compounds have been demon-
(standard soda lime contains 15% water) generates no strated to be nephrotoxic in mice, their signi®cance in
carbon monoxide and that the same applies to baralyme with humans remains unknown.
9.7% (or more) water. They used ¯ow rates of less than The production of compound A, like carbon monoxide, is
2±3 litre min±1 during this study. Baxter and Kharasch found related to the dryness (and temperature) of the carbon
a minimum safe level of humidity required to avoid dioxide absorber granules. Cunningham and colleagues
desiccation and carbon monoxide production, and estab- demonstrated that drying Sodasorb from 16 to 11% moisture
lished that this could be achieved by adding the full had little effect on the concentrations of compound A, but
complement (13%) of water to Baralyme.13 The effect of drying to 2±6% moisture produced greater concentrations.33

257
 Lawes

Ruzicka and co-workers found that the release of volatile carbon monoxide has occurred, leading to carbon monoxide
sevo¯urane degradation products in an anaesthesia circuit is poisoning, errors in estimation of SpO2 occur.25 Oxygen
highly dependent on soda lime temperatures and water saturation, as measured by pulse oximetry, failed to
content.90 Higuchi and colleagues have been working on the decrease to less than 96% despite COHb levels as high as
production of new carbon dioxide absorber formulations to 44%. Pulse oximetry is unreliable in estimating HbO
address the production of these toxic compounds.56 In the saturation in patients massively exposed to carbon mon-
United States, the Food and Drug Administration recom- oxide and should also be interpreted with caution when used
mends the use of sevo¯urane with fresh gas ¯ow rates of at to estimate oxygen saturation in smokers. Barker and
least 1 litre min±1 for exposures up to 1 h and at least 2 litre Tremper warn that in the presence of COHb, the SpO2 is
min±1 for exposures greater than 1 h in an attempt to reduce approximately the sum of COHb and HbO, and may
exposure by washing the substance from the circuit. These seriously overestimate the latter.8
recommendations are empirical and do not relate to some These results contrast with the ®ndings of Berry and co-
experimental evidence. Interestingly, the fresh gas ¯ow workers who noted a small decrease in SpO2 in association
requirements found by some authors for avoiding substance with a COHb of 30%, although the decrease in saturation
A accumulation (>3 litre min±1) are in con¯ict with the did not re¯ect the entire increase in COHb.16 Presumably,
requirements for reducing carbon monoxide production these warnings also apply in patients exposed to CO from
(<1 litre min±1). In addition, Egar and colleagues found that desiccated soda lime and volatile anaesthetics, where we
baralyme dehydration increases, and soda lime dehydration know we can generate clinically signi®cant levels of carbon
decreases, degradation of the inhaled anaesthetic sevo- monoxide? Until we measure it, we will not know. Unless
¯urane to compound A.39 COHb can be speci®cally excluded, we do not even know
It would seem prudent to investigate the need for the signi®cance of the SpO2. Woehlck roughly estimated that
inclusion of a carbon monoxide analyser and a compound in the USA the possible extent of the number of patients
A analyser amongst gas monitoring equipment if an HME/ exposed to carbon monoxide as a result of anaesthetic
®lter is used in a circle system. In the absence of direct breakdown each year to be approximately 1000Ð10 000.120
carbon monoxide monitoring, it is possible to identify if At present, without the ability or willingness to measure
carbon monoxide production is occurring during anaesthe- toxic substance production or humidity, it is important to
sia by measuring the percent of COHb in the patient's blood. understand contributory factors to the risk in order to reduce
Using current blood gas analysers that incorporate carbon it. Greatest risks are where absorbent drying most com-
monoxide-oximetry enables this solution but this is not a monly occurs such as Monday mornings, following the use
satisfactory order of events. As we know the humidity of high fresh gas ¯ows (>3 litre min±1) especially in remote
required to avoid carbon monoxide production, it may be sites, and in outlying hospitals or theatres where general
easier to avoid the potential risk of poisoning from both anaesthesia is used occasionally (e.g. opthalmic threatres). It
toxins by measuring and/or controlling humidity. is not suf®cient to just look in the theatre log book for the
last period of use of an anaesthetic machine. New absorbent
should be started at the beginning of the list, each day, and
fresh gas ¯ow rates of less than 1 litre min±1 used for
Filters and HMEs and false elevation of pulse maintenance of anaesthesia, provided there is minimal
oximetry reading circuit leakage. All gas supplies to anaesthetic machines
Aside from the direct damage incurred by carbon monoxide, should be turned off at the end of lists, but not the electricity
it has an impact on the interpretation of pulse oximetry. supply, even at weekends, if oxygen failure alarms are
Many anaesthetists may consider that carbon monoxide dependant on re-chargeable batteries to function. It would
exposure and intoxication is only a matter of concern when seem prudent not to place a ®lter with HME characteristics
dealing with heavy smokers with ischaemic heart disease, or between the circuit and the absorber, but to use a pure
patients exposed to ®res in enclosed spaces. The impact of biological ®lter. The impact of incorporating a HME device
carbon monoxide on oxygen saturation would not normally between the circuit and the absorber has not been investi-
enter into consideration in other circumstances. We are gated in relation to these recommendations. It would seem
lulled into a false sense of security believing that our sensible to clearly differentiate HMEs, HMEFs, and ®lters,
oximeters truly re¯ect patients' SpO2 levels. If our patients thus enabling the insertion of a ®lter alone, when required,
have been exposed to desiccated absorbers, this con®dence immediately before the absorber without restricting the
may be inappropriate. circulation of airway humidity.
Vegfors and Lennmarken investigated the ability of pulse
oximeters to measure COHb as HbO and compared results
using a co-oximeter.102 They concluded that pulse oxi- Delay or failure of gas induction
meters measure COHb as HbO and warn of the dangers of The work of McGinty and co-workers, Lerman and
relying on oximetry when COHb is present. Buckley and co- colleagues, and Paris and colleagues indicates that some
workers found that when signi®cant patient exposure to patients, especially young adults and children who have

258
 Hazards and dangers associated with HME/®lters

never received chair dental anaesthetics or ether, are opting anaesthetic agent reaches the patient, irrespective of the
for a gas induction now that sevo¯urane has permitted the circuit, at fresh gas ¯ows of less than 3 litre min±1.
reintroduction of this not unpleasant option.73 66 82 Seropian When attempting to hyperoxygenate the patient and using
and co-workers remind us of complications related to low fresh gas ¯ow rates, gas diversion away from the patient
prolonged induction which include, but are not limited to, also occurs if the facemask is merely suspended over the
coughing, vomiting, and/or laryngospasm, all of which have patient's face before induction of anaesthesia. This further
an impact on patient morbidity and mortality.91 Delays in reduces the ef®cacy of this already inef®cient manoeuvre.
anaesthetic induction can occur directly or indirectly as a McGowan and Skinner emphasize the importance of a good
result of using an HME if it facilitates the drying of carbon ®t between mask and face during accelerated induction
dioxide absorbers (Table 4). Grodin and co-workers in two techniques, and this is especially important if a ®lter or
separate papers46 47 found that not only does dry soda lime HME is incorporated in the circuit.74 In comparative trials
absorb halothane, iso¯urane, and en¯urane akin to a sponge of inhalation induction by Lerman and co-workers,66
but also, dry absorber degrades the volatile agent, thus Sigston and co-workers,92 and Paris and co-workers82
reducing the amount available to the patient, as observed by sevo¯urane was compared favourably against halothane.
Janshon and Dudziak60 and Funk.42 In these paediatric studies HMEFs are conspicuously absent
Thwaites and co-workers recommend sevo¯urane gas from the detailed descriptions of the induction techniques
induction as a technique of choice in patients with dif®cult used.
airways.100 Reduction in the inspired concentration of
sevo¯urane under these circumstances increases the risk
of a turbulent induction, as there is prolongation of the Filters and HMEs as a disconnection hazard
second stage. All the consequences of laryngospasm,
Adams, in a symposium on mishap or negligence,
desaturation, and loss of airway may ensue.
emphasized that disconnections occur most frequently at
Resistance to ¯ow induced by ®lters and HMEs is not just
the site of connection of the tracheal tube to the breathing
a problem of long-term mechanical ventilation, it is also of
system. Adams cited a survey of members of the Canadian
concern during induction and maintenance of anaesthesia,
Anaesthetists' Society, in which 63% of the respondents
as is dead space, especially in paediatrics. Wilkinson and
stated that such disconnections occurred at least `occasion-
co-workers investigated the resistance to ¯ow for 12
ally' at the tracheal tube connection, and that such
hygroscopic HME in order to determine the possible effects
disconnections occur so often as to be considered as
on the work of breathing during steady state anaesthesia.114
routine.2 In an analysis by Caplan and co-workers of closed
They found the dead space to be equal to standard catheter
claims that had resulted in death or permanent disability,
mounts (12 ml), and they comment that resistance to disconnection at the junction between the distal end of the
breathing was `low, compared to other anaesthetic equip- breathing circuit and the tracheal tube accounted for 16% of
ment'. The use of some HMEFs in paediatrics has been cases (Table 4).27 For head and neck and similar surgery,
restricted by the dead space of the devices. Booker reported adding ®lters to a circuit constitutes an additional discon-
that the deadspace of the smallest HMEF (45 ml) used in nection hazard by introducing an extra connection. This is
his study had restricted their use to children of greater than despite the requirement for manufacturers to comply with
15 kg.22 the International Standard for HMEs (ISO 5356-1: 1987),
Even if the dead space permits the use of a ®lter or HME associated with tapered 15 and 22 mm connections.108
during induction, the resistance to ¯ow may not. The The disconnection hazard is exacerbated when the ®lter is
pressure gradient across a new ®lter or HME is in the order bulky, angular, within the surgeons operating environment,
of 1.0±1.5 cm H2O at fresh gas ¯ows of 50 litre min±1. This and concealed by drapes. An increased risk is evident for
gradient is suf®cient to allow diversion of fresh gas down both spontaneously breathing patients and patients receiving
the expiratory limb, resulting in failure or delay in gas IPPV. Russell and co-workers reviewed the ®rst 2000
induction. Goddard reported dif®culty in performing gas incidents reported to the Australian Incident Monitoring
inductions when using an Ayres T piece with a ®lter, and Study and found 317 incidents which involved problems
this technique34 was the cause of concern for Marcus and with ventilation.89 Eighty-nine per cent of problems were
colleagues.45 71 Da Fonseca and co-workers recommend related to circuit disconnects or misconnects, of which 48%
that the HME should be removed from the Mapleson F were disconnections. During IPPV, detection of patient
breathing system until inhalation induction is complete, or disconnection from the circuit is not necessarily guaranteed
that the reservoir bag is completely occluded unless an by the use of low pressure disconnect alarms. Monitor
effective seal is obtained with the mask. This problem is the detection was by a low circuit pressure alarm in 37%, but
subject of a hazard warning within our own department, and this alarm failed to warn of non-ventilation in 12 incidents.
the use of a ®lter or HME during induction with a T piece In one-third of the cases, disconnection was associated with
has been discontinued. When using your cupped hand at the interference to the anaesthetic circuit by a third party, and in
beginning of a gas induction, no measurable volatile nearly half with surgery on the head and neck.

259
 Lawes

Milligan warns of the dangers of disconnect alarms that tubing could not be incorporated in both non-circle
are disabled by the presence of an HME.77 Others have systems and parallel circle circuits. An opportunity exists
investigated failure of disconnection alarms in the presence for manufacturers to produce rounded ®lter housing
of HMEs. Pryn and Crosse86 investigated the relationship without a gas analyser port to reduce some elements of
between the pressures generated in breathing systems during this hazard.
the inspiratory phase of positive pressure ventilation and
devices attached to the circuit. The increased airways
resistance associated with the con®guration of circuit End-tidal carbon dioxide estimation
attachments, especially HMEs, enabled increases in airway When asked the hypothetical question `Which monitor
pressure between the HME and the ventilator. These would you retain if you were allowed to retain only a single
increases in pressure were detected by the low-pressure monitor?' most anaesthetists choose the carbon dioxide
alarm as normal ventilatory cycles. Nine different heat and monitor. Hardman and co-workers measured the end-tidal
moisture exchanges were compared in the same breathing carbon dioxide values on either side of breathing ®lters.49
system. Those with 15-mm male connectors generate the They found that the PE¢CO2 was signi®cantly lower at the
highest pressures on disconnection (1.1 kPa), suf®cient to machine side of the ®lter in both ventilated and spon-
stimulate the disconnect alarm. They warn that low-pressure taneously breathing groups (P=0.00001). The measurement
alarms should be adjusted to the circuit and ventilator error induced by the inclusion of the breathing system ®lter
con®guration. was signi®cantly greater in spontaneously breathing patients
The problem was reinvestigated several years later by one (P=0.0004). The source of the error is presumably as a result
of these original authors and it was found that the situation of the small pressure gradient across the ®lter membrane.
had deteriorated further.26 The incidence of disconnection Expired gases mix with some fresh gas on the machine side
alarms not being been adjusted appropriately for the of the ®lter, whereas on the patient side, the sample is of
ventilators in use had increased. Bengtsson and Johnson pure expired gases. Although not a major hazard of using a
reported two cases of ventilator disconnection resulting ®lter, this effect may be clinically signi®cant in certain
from defective HME (Table 4).14 Collins and Vaghadia situations such as raised intracranial pressure. During
warn that during spontaneous ventilation, incorporating the carbon dioxide insuf¯ation for laparoscopic surgery, inter-
gas analyser port in the ®lter housing obscures disconnec- pretation of the end-tidal carbon dioxide is further con-
tion from the anaesthetic gases, with a risk of intraoperative founded by intra-peritoneal insuf¯ation of carbon dioxide.
awareness.31 This problem is not averted by changing the site of the gas
Disconnection from the breathing circuit at a site distal to analyser port but by avoiding the use of a ®lter in these
the gas sampling line, renders PE¢CO2 unreliable for detect- circumstances.
ing circuit disconnection. When disconnection occurs Cruz and co-workers expressed disquiet at the practice of
between the ®lter and the circuit during spontaneous estimating arterial carbon dioxide pressure (PaCO2) by
ventilation, the carbon dioxide waveform is still present, capnography during laparoscopic surgery in pregnant
rendering a false sense of security. We might assume that women, as a result of studies on pregnant ewes.32 During
other components of the gas monitor would detect this event carbon dioxide insuf¯ation, there was evidence of both
and warn us that disconnection has occurred, but we maternal and fetal acidosis if PE¢CO2 was used to guide
generally rely on an audible signal. Aside from observing ventilation. Amos and co-workers concluded from the
changes in the bag movement, circuits not monitored by retrospective analysis of the anaesthetic records of seven
inspiratory and expiratory gas volumes will provide no pregnant patients who underwent laparoscopic surgery, that
clues. Monitor manufacturers'1 default settings for inspired caution should be used in interpreting end-tidal carbon
N2O and volatile agents are ±1 or 0. For inspired oxygen, it dioxide when considering non-obstetric laparoscopic sur-
is 18% for most manufacturers bar one, at 30%. These gery in pregnant women.4 Bhavani-Shankar investigated the
default settings mean that no audible alarm will alert the possible consequences of this observation and he and his
anaesthetist to this type of disconnection. Kennedy and colleagues have disputed the signi®cance of these ®nd-
French report just such an incident.61 Disconnection ings.18 Nonetheless, the use of ®lters when arterial carbon
between the circuit and the ®lter during spontaneous dioxide levels are critical should be considered carefully.
ventilation is best monitored electronically by detecting
changes in expired gas volume within the circuit. Expired
volume measurements are only available for circle systems. Conclusions
There is therefore a risk of intraoperative awareness unless Newer ®lters and HMEs are highly effective, and while
the PE¢CO2 records directly from the circuit, rather than for many are more ef®cient than required for protection against
the ®lter or HME housing. Several manufacturers make co- standards for biohazard protection, they are not all equally
axial circle systems that incorporate gas analysis tubing in ef®cient. It would seem that the use of HMEs and ®lters is
their structure, or incorporate a sampling port on the Y not always indicated and they should not be used indis-
connector of the circuit. There is no reason that similar criminately. Including a ®lter or an HME in a circuit has a

260
 Hazards and dangers associated with HME/®lters

potentially signi®cant impact on morbidity and mortality. 2 Adams AP. Symposium on mishap or negligence. Breathing
The ef®cacy of all ®lters is not fully established. Their system disconnections. Br J Anaesthesia 1994; 73: 46±54
3 Allred EN, Bleecker ER, Chaitman BR, et al. Short-term effects of
ability to clearly reduce the incidence of transmission of
carbon monoxide exposure on the exercise performance of
infection in clinical studies may not match up to their subjects with coronary artery disease. N Engl J Med 1989; 321:
performance in the laboratory. Until the ®ltration ef®ciency 1426±32
of all ®lters is published against a known standard, there can 4 Amos JD, Schorr SJ, Norman PF, et al. Laparoscopic surgery
be no conclusion to this problem. Not all the potential during pregnancy: a word of caution. Am J Surg 1996; 171: 435±7
hazards of incorporating ®lters and HMEs in anaesthetic 5 Association of Anaesthesists of Great Britain and Ireland. Blood
circuits have been fully explored. The contribution of an Borne Viruses and Anaesthesia: an update, January 1996
6 Atkinson MC, Girgis Y, Broome IJ. Extent and practicalities of
HME as opposed to a pure ®lter in creating a microenviron-
®lter use in anaesthetic breathing circuits and attitudes towards
ment in circle absorber systems that predisposes to accu- their use: a postal survey of UK hospitals. Anaesthesia 1999; 54:
mulation of carbon monoxide and other toxic metabolites 37±41
has not been fully investigated. 7 Barbara J, Chabane MH, Leynadier F, Girard F. Retention of
HMEs and ®lters are a hazard during gas induction either airborne latex particles by a bacterial and viral ®lter used in
directly through gas diversion or indirectly through absorber anaesthesia apparatus. Anaesthesia 2001; 56: 231±4
desiccation. They render pointless the use of non-applied 8 Barker SJ, Tremper KK. The effect of carbon monoxide
inhalation on pulse oximetry and transcutaneous pO2.
masks for pre-oxygenation of routine cases, except at high
Anesthesiology 1987; 66: 677±9
fresh gas ¯ows. They should not be used in the presence of 9 Barnes S, Normoyle D. Failure of ventilation in an infant due to
active humidi®cation, nebulized drugs, copious secretions, increased resistance of a disposable heat and moisture
or pulmonary oedema. They should be clearly visible and exchanger. Anesth Analg 1996; 83: 193
accessible at all times in order to detect contamination, 10 Barton RM. Detection of expiratory antibacterial ®lter occlusion.
obstruction, or disconnection. There is an argument for not Anesth Analg 1993; 77: 197
incorporating the gas analyser port in the ®lter, but rather 11 Baum J, Sachs G, v.d. Driesch C, Stanke HG. Carbon monoxide
generation in carbon dioxide absorbents. Anesth Analg 1995; 81:
incorporating it into the circuit. Filters constructed with a
144±6
smooth contour in a clear housing are preferable to angular 12 Baxter AD. Low and minimal ¯ow inhalational anaesthesia. Can J
constructions if a disconnection hazard is to be reduced. If Anaesth 1997; 44: 643±53
you cannot see the ®lter, there is a strong argument not to 13 Baxter PJ, Kharasch ED. Rehydration of desiccated Baralyme
use one. When there is a relative or absolute contraindica- prevents carbon monoxide formation from des¯urane in an
tion to ®lters, they should be omitted and the circuit should anesthesia machine. Anesthesiology 1997; 86: 1061±5
be discarded between cases, irrespective of cost. 14 Bengtsson M, Johnson A. Failure of a heat and moisture
exchanger as a cause of disconnection during anaesthesia. Acta
Industrial protective masks and air purifying ®lters are
Anaesthesiol Scand 1989; 33: 522±3
clearly identi®ed by name and colour, speci®c to their 15 Berry A, Nolte F. An alternative strategy for infection control of
purpose. A clearly identi®able distinction between a ®lter anesthesia breathing circuits: a laboratory assessment of the Pall
and an HME or an HMEF with or without bactericidal HME Filter. Anesth Analg 1991; 72: 651±5
properties, would be useful so that their separate and distinct 16 Berry PD, Sessler DI, Larson, MD. Severe carbon monoxide
functions can be used appropriately. Some manufacturers poisoning during des¯urane anesthesia. Anesthesiology 1999; 90:
have moved in this direction. The biological ®ltration 613±6
17 Beydon L, Tong D, Jackson N, Dreyfuss D. Correlation between
capability for particles, bacteria and viruses, in both wet and
simple clinical parameters and the in vitro humidi®cation
dry suspension should be clearly apparent. Dynamic characteristics of ®lter heat and moisture exchangers. Groupe
changes in function in terms of changes in resistance with de Travail sur les Respirateurs. Chest 1997; 112: 739±44
saturation, and duration of effective usage should be 18 Bhavani-Shankar K, Steinbrook RA, Brooks DC, Datta S. Arterial
included in the product insert, as should the humidi®cation to end-tidal carbon dioxide pressure difference during
ef®ciency at different ¯ow rates. For longer procedures, laparoscopic surgery in pregnancy. Anesthesiology 2000; 93:
some mechanism to indicate the degree of membrane 370±3
19 Bisinotto FM, Braz JRC, Martins RH, GregoÂrio EA, Abud TMV.
saturation or transmembrane pressure gradient would be
Tracheobronchial consequences of the use of heat and moisture
useful. The internal dead space and the nature of the HME in exchangers in dogs. Can J Anesth 1999; 46: 897±903
terms of whether it is hydrophobic, hydroscopic, and 20 Bissonnette B, Sessler DI. Passive or active inspired gas
electrostatic or a pleated micropore type would also be humidi®cation increases thermal steady-state temperatures in
helpful, as would the speci®c indications for choosing one anesthetized infants. Anesth Analg 1989; 69: 783±7
device over another. 21 Body SC, Philip JH. Gram-negative rod contamination of an
ohmeda anesthesia machine. Anesthesiology 2000; 92: 911
22 Booker PD. Equipment and monitoring in paediatric anaesthesia.
Br J Anaesth 1999; 83: 78±90
References 23 Brown ES, Bakamjian V, Seniff A. Performance of absorbents:
1 Aarhus D, Soreide E, Holst-Larsen H. Mechanical obstruction in effect of moisture. Anesthesiology 1959; 20: 613±61
the anesthesia delivery-system mimicking severe broncho-spasm. 24 Buckley PM. Increase in resistance of in-line breathing ®lters in
Anaesthesia 1997; 52: 989±97 humidi®ed air. Br J Anaesth 1984; 56: 637±43

261
 Lawes

25 Buckley RG, Aks SE, Eshom JL, Rydman R, Schaider J, Shayne P. 44 Gentz BA, Malan TP Jr. Renal toxicity with sevo¯urane: a storm
The pulse oximetry gap in carbon monoxide intoxication. Ann in a teacup? Drugs 2001; 61: 2155±6
Emerg Med 1994; 24: 252±5 45 Goddard JM, Bennett HR. Filters and Ayre's T-piece. Anaesthesia
26 Campbell RM, Sheikh A, Crosse MM. A study of the incorrect 1996; 51: 605
use of ventilator disconnection alarms. Anaesthesia 1996; 51: 46 Grodin WK, Epstein MAF, Epstein RA. Mechanisms of halothane
369±70 absorption by dry soda lime. Br J Anaesth 1982; 54: 555±9
27 Caplan R, Vistica MF, Posner KL, Cheney FW. Adverse 47 Grodin WK, Epstein MAF, Epstein RA. Soda lime absorption of
anesthetic outcomes arising from gas delivery equipment: a iso¯urane and en¯urane. Anesthesiology 1985; 62: 60±4
closed claims analysis. Anesthesiology 1997; 87: 741±8 48 Han JN, Liu YP, Ma S, et al. Effects of decreasing the frequency of
28 Chant K, Kociuba K, Munro R, et al. Investigation of possible ventilator circuit changes to every 7 days on the rate of
patient-to-patient transmission of hepatitis C in a hospital. NSW ventilator-associated pneumonia in a Beijing hospital. Resp Care
Public Health Bull 1994; 5: 47±51 2001; 46: 891±6
29 Christopher K, Saravolatz LD, Bush TL, Conway WA. The 49 Hardman JG, Curran J, Mahajan RP. End-tidal carbon dioxide
potential role of respiratory therapy equipment in cross- measurement and breathing system ®lters. Anaesthesia 1997; 52:
infection. Am Rev Resp Dis 1983; 128: 271±5 646±8
30 Cohen IL, Weinberg PF, Fein IL, Rowinski GS. Endotracheal tube 50 Heat and moisture exchange ®lter obstruction. Australian
occlusion associated with the use of heat and moisture Therapeutic Device Bulletin, October 1990; 67
exchangers in the intensive care unit. Crit Care Med 1988; 16: 51 Hedley RM, Allt-Graham J. Heat and moisture exchangers and
277±9 breathing systems. Br J Anaesth 1994; 73: 227±36
31 Collins L, Vaghadia H. Patient disconnect. Can J Anaesth 1998; 45: 52 Hedley RM, Allt-Graham J. Comparison of the ®ltration
1135 properties of heat and moisture exchangers. Anaesthesia 1992;
32 Cruz AM, Sutherland LC, Duke T, Townsend HGG, Ferguson 47: 414±20
JG, Crone LAA. Intra-abdominal carbon dioxide insuf¯ation in 53 Heinsen A, Brendtsen F, Fomsgaard A. A phylogenetic analysis
the pregnant ewe. Anesthesiology 1996; 85: 1395±402 elucidating a case of patient-to-patient transmission of hepatitis
33 Cunningham DD, Huang S, Webster J, Mayoral J, Grabenkort C virus during surgery. J Hosp Infect 2000; 46: 309±13
RW. Sevo¯urane degradation to compound A in anaesthesia 54 Henriksson BA, Sundling J, Hellman A. The effect of a heat and
breathing systems. Br J Anaesth 1996; 77: 537±43 moisture exchanger on humidity in a low-¯ow anaesthesia
34 Da Fonseca JM, Wheeler DW, Pook JA. The effect of a heat and system. Anaesthesia 1997; 52: 144±9
moisture exchanger on gas ¯ow in a Mapleson F breathing 55 Hess D. Filters and anesthesia breathing circuits: can we cut
system during inhalational induction. Anaesthesia 2000; 55: 571±3 costs without harm? J Clin Anesth 1999; 11: 531±3
35 Daggan R, Zefeiridis A, Steinberg D, et al. High quality ®ltration 56 Higuchi H, Adachi Y, Arimura S, Kanno M, Satoh T. Compound
allows reuse of anesthesia breathing circuits resulting in cost A. Concentrations during low-¯ow sevo¯urane anesthesia
savings and reduced medical waste. J Clin Anesth 1999; 11: 536±9 correlate directly with the concentration of monovalent bases
36 Davis K, Evans SL, Campbell RS, et al. Prolonged use of heat and in carbon dioxide absorbents. Anesth Analg 2000; 91: 434±9
moisture exchangers does not affect device ef®ciency or 57 Hogarth I. Anaesthetic machine and breathing system
frequency rate of nosocomial pneumonia. Crit Care Med 2000; contamination and the ef®cacy of bacterial/viral ®lters. Anaesth
28: 1412±8 Intens Care 1996; 24: 154±63
37 Dreyfuss D, Djedaini K, Gros I, et al. Mechanical ventilation with 58 Iotti GA, Olivei MC, Palo A, et al. Unfavorable mechanical effects
heated humidi®ers or heat and moisture exchangers: effects on of heat and moisture exchangers in ventilated patients. Intensive
patient colonization and incidence of nosocomial pneumonia. Am Care Med 1997; 23: 399±405
J Respir Crit Care Med 1995; 151: 986±92 59 ISO 9360-1:2000 Anaesthetic and respiratory equipment. Heat
38 du Moulin CG, Saubermann AJ. The anaesthesia machine and and moisture exchangers (HMEs) for humidifying respired gases
circle system are not likely to be sources of bacterial in humans Part 1: HMEs for use with minimum tidal volumes of
contamination. Anesthesiology 1977; 47: 353±358 250 ml
39 Eger EI 2nd, Ionescu P, Laster MJ, Weiskopf RB. Baralyme 60 Janshon GP, Dudziak R. Interactions of dry soda lime with
dehydration increases and soda lime dehydration decreases the en¯urane and sevo¯urane. Clinical report on two unusual
concentration of compound A resulting from sevo¯urane anesthesias. Anaesthesist 1997; 46: 1050±3
degradation in a standard anesthetic circuit. Anesth Analg 1997; 61 Kennedy RR, French RA. A breathing circuit disconnection
85: 892±8 detected by anesthetic agent monitoring. Can J Anaesth 2001; 48:
40 Fang Z, Eger EI 2nd, Laster MJ, Chortkoff BS, Kandel L, Ionescu P. 847±9
Carbon monoxide production from degradation of des¯urane, 62 Kleemann PP. Humidity of anaesthetic gases with respect to low
en¯urane, iso¯urane, halothane, and sevo¯urane by soda lime ¯ow anaesthesia Anaesth Int Care 1994; 22: 396±408
and Baralyme. Anesth Analg 1995; 80: 1187±93 63 Le Bourdelles G, Mier L, Fiquet B, et al. Comparison of the
41 Frink EJ jr, Nogami WM, Morgan SE, Salmon RC. High effects of heat and moisture exchangers and heated humidi®ers
carboxyhemoglobin concentrations occur in swine during on ventilation and gas exchange during weaning trials from
des¯urane anesthesia in the presence of partially dried carbon mechanical ventilation. Chest 1996; 110: 1294±8
dioxide absorbents. Anesthesiology 1997; 87: 308±16 64 Leijten D, Rejger V, Mouton R. Bacterial contamination and the
42 Funk W, Gruber M, Wild K, Hobbhahn J. Dry soda lime effect of ®lters in anaesthetic circuits in a simulated patient
markedly degrades sevo¯urane during simulated inhalation model. J Hosp Infect 1992; 21: 51±60
induction. Br J Anaesth 1999; 82: 193±198 65 Lentz RE. CO poisoning during anaesthesia poses puzzle. J Clin
43 Garibaldi RA, Britt MR, Webster C, Pace NL. Failure of bacterial Monit 1995; 11: 67±71
®lters to reduce the incidence of pneumonia after inhalation 66 Lerman J, Davis P, Welborn L, et al. Induction, recovery, and
anaesthesia. Anesthesiology 1981; 54: 364±8 safety characteristics of sevo¯urane in children undergoing

262
 Hazards and dangers associated with HME/®lters

ambulatory surgery: a comparison with halothane. Anesthesiology 90 Ruzicka JA, Hidalgo JC, Tinker JH, Baker MT. Inhibition of volatile
1996; 84: 1332±40 sevo¯urane degradation product formation in an anesthesia
67 Littlewood K, Durbin CG jr. Evidenced-based airway circuit by a reduction in soda lime temperature. Anesthesiology
management. Resp Care 2001; 46: 1392±405 1994; 81: 238±44
68 Lloyd G, Howells J, Liddle C, Klineberg PL. Barriers to hepatitis 91 Seropian MA, Robins B. Smaller than expected sevo¯urane
C transmission within breathing systems: ef®cacy of a pleated concentrations using the sevotec 5 vaporizer at low ®ll states
hydrophobic ®lter. Anaesth Int Care 1997; 25: 235±8 and high fresh gas ¯ows. Anesth Analg 2000; 91: 834±6
69 Lumley J, Gaya H, Holdcroft A, Darlow HM, Adams DJ. 92 Sigston PE, Jenkins AMC, Jackson EA, Sury MR J, Mackersie AM,
Expiratory bacterial ®lters. Lancet 1976; 2: 22±3 Hatch DJ. Rapid inhalation induction in children: 8% sevo¯urane
70 Manthous CA, Schmidt GA. Resistive pressure of a condenser compared with 5% halothane. Br J Anaesth 1997; 78: 362±5
humidi®er in mechanically ventilated patients. Crit Care Med 93 Smith CE, Otworth JR, Kaluszyk GSW. Bilateral tension
1994; 22: 1792±5 pneumothorax due to a defective anesthesia breathing circuit
71 Marcus R, Wandless J, Thompson J. Filters and Ayre's T-piece. ®lter. J Clin Anesth 1991; 3: 229±34
Paediatr Anaesth 1999; 9: 93 94 Smith SM. Chemical and biological weapons. Implications for
72 McEwan AI, Dowell L, Karis JH. Bilateral tension pneumothorax anaesthesia and intensive care. Br J Anaesth 2002; 89: 306±24
caused by a blocked bacterial ®lter in an anaesthesia breathing 95 Snowdon SL. Hygiene standards for breathing systems? Br J
circuit. Anesth Analg 1993; 76: 440±2 Anaesth 1994; 72: 143
73 McGinley J, Briggs L, Carey, M. Inhalation induction with 96 Soro M, Belda FJ, Paredes I, CorteÂs-Uribe A, Bonome C, Alvarez
sevo¯urane. Can J Anaesth 1997; 44: 1218 F. Effect of canister size and gas ¯ow rate on the exponential
74 McGowan P, Skinner A. Preoxygenation: the importance of a time-course of soda lime dehydration in anaesthesia circuits. Br J
good face mask seal. Br J Anaesth 1995; 75: 777±8 Anaesth 1999; 82: 72±3
75 Mebius C. Heat and moisture exchangers with bacterial ®lters: a 97 Stacey MRW, Asai T, Wilkes A, Hodzovic I. Obstruction of a
laboratory evaluation. Acta Anaesthesiol Scand 1992; 36: 572±6 breathing system ®lter. Can J Anaesth 1996; 43: 1276
76 Middleton VA, VanPoznak A, Artusio JF, et al. Carbon monoxide 98 Stevens J. Breathing system ®lters. Anaesthesia 1999; 54: 90
accumulation in closed circle anesthesia systems. Anesthesiology 99 Subayi L. Chergui K. Beydon L. Filtres echangeurs de chaleur et
1965; 26: 715±9 d'humidite pour le conditionnement des gaz inspires en
77 Milligan KA. Disablement of a ventilator disconnect alarm by a anesthesie-reanimation de l'adulte. Ann Francaises Anesth
heat and moisture exchanger. Anaesthesia 1992; 47: 279 Reanimation 1998; 17: 699±708
78 Moon RE. Cause of CO poisoning, relation to halogenated 100 Thwaites AJ, Smith I, Ip-Yam PC, Board P. Sevo¯urane for
agents still not clear. J Clin Monit 1995; 11: 66±7 dif®cult tracheal intubation. Br J Anaesth 1998; 81: 103±05
79 Myers RA, Britten JS. Are arterial blood gases of value in 101 Tibbles PM, Perrotta PL. Treatment of carbon monoxide
treatment decisions for carbon monoxide poisoning? Crit Care poisoning: a critical review of human outcome studies
Med 1989; 17: 139±42 comparing normobaric oxygen with hyperbaric oxygen. An Int
80 Nardell EA. Dodging droplet nuclei: reducing the probability of Med 1994; 24: 269±76
nosocomial tuberculosis transmission in the AIDS era. Am Rev 102 Vegfors M, Lennmarken C. Carboxyhaemoglobinaemia and pulse
Respir Dis 1990; 142: 501±3 oximetry. Br J Anaesth 1991; 66: 625±6
81 Neft MW. Goodman JR. Hlavnicka JP. Veit BC. To reuse your 103 Vezina DP, Trepanier CA, Lessard MR, Gourdeau M, Tremblay
circuit: the HME debate. AANA J 1999; 67: 433±9 C. Anesthesia breathing circuits protected by the DAR
82 Paris ST, Cafferkey M, Tarling M, et al. Comparison of Barrierbac S breathing ®lter have a low bacterial contamination
sevo¯urane and halothane for outpatient dental anaesthesia in rate. Can J Anaesth 2001; 48: 748±54
children. Br J Anaesth 1997; 79: 280±4 104 Villafane, MC, Cinnella G, Lofaso F, et al. Gradual reduction
83 Pelosi P, Solca M, Ravagnan I, et al. Effects of heat and moisture of endotracheal tube diameter during mechanical ventilation
exchangers on minute ventilation, ventilatory drive, and work of via different humidi®cation devices. Anesthesiology 1996; 85:
breathing during pressure-support ventilation in acute 1341±9
respiratory failure. Crit Care Med 1996; 24: 1184±8 105 Walton JS, Fears R, Burt N, Dorman HB. Intraoperative
84 Prados W. A dangerous defect in a heat and moisture exchanger. breathing circuit obstruction caused by Albuterol nebulization.
Anesthesiology 1989; 71: 804 Anesth Analg 1999; 89: 650±1
85 Prasad KK, Chen L. Complications related to the use of a heat 106 Wilkes AR, Benbough JE, Speight SE, Harmer M. The bacterial
and moisture exchanger. Anesthesiology 1990; 72: 958 and viral ®ltration performance of breathing system ®lters.
86 Pryn SJ, Crosse MM. The role of ventilator and breathing system Anaesthesia 2000; 55: 458±65
accessories. Anaesthesia 1989; 44: 978±81 107 Wilkes AR, Ferguson RA, Mecklenburgh JS. Ability of breathing
87 Rathgeber J, Kietzmann D, Mergeryan H, et al. Prevention of system ®lters to prevent liquid contamination of breathing
patient bacterial contamination of anaesthesia circle systems: a systems. Br J Anaesth 1998; 80: 550
clinical study of the contamination risk and performance of 108 Wilkes AR, Hampson MA, Mecklenburgh JS. A method of testing
different heat and moisture exchangers with electret ®lter the compliance of 15 and 22 mm conical connectors on heat and
(HMEF). Eur J Anaesthesiol 1997; 14: 368±73 moisture exchangers to ISO 5356-1: 1987. J Med Eng Technol
88 Ricard JD, Le Miere E, Markowicz P, et al. Ef®ciency and safety of 1997; 21: 147±50
mechanical ventilation with a heat and moisture exchanger 109 Wilkes AR. Heat and moisture exchangers. Structure and
changed only once a week. Am J Respir Crit Care Med 2000; 161: function. In: Branson RD, Pererson BD, Carson KD, eds.
104±9 Humidi®cation: Current Therapy and Controversy. Respiratory Care
89 Russell WJ, Webb RK, Van der Walt JH, Runciman WB. The Clinics of North America, 1998; 4: 261±79
Australian incident monitoring study. Problems with ventilation: 110 Wilkes AR. Measuring the ®ltration performance of breathing
an analysis of 2000 incident reports. Anaesth Intensive Care 1993; system ®lters using sodium chloride particles. Anaesthesia 2002;
21: 617±20 57: 162±8

263
 Lawes

111 Wilkes AR. Mecklenburgh JS. Predicting the moisture-conserving 116 Williams DJ, Stacey MR. Rapid and complete occlusion of a heat
performance of breathing system ®lters with different and moisture exchange ®lter by pulmonary edema (clinical
anaesthetic breathing systems. Br J Anaesth 1997; 79: 681±2 report). Can J Anaesth 2002; 49: 126±31
112 Wilkes AR. Stevens AJ. Guidance on the use of breathing system 117 Woehlck H, Kharasch ED. Carbon dioxide absorber moisture
®lters. Anaesthesia 2000; 55: 187 content. Anesthesiology 1997; 87: 1590±1
113 Wilkes AR. The ability of breathing system ®lters to prevent 118 Woehlck HJ, Dunning M 3rd, Connolly LA. Reduction in the
liquid contamination of breathing systems: a laboratory study. incidence of carbon monoxide exposures in humans undergoing
Anaesthesia 2002; 57: 33±9 general anesthesia. Anesthesiology 1997; 87: 228±34
114 Wilkinson KA, Cranston A, Hatch DJ, Fletcher ME. Assessment 119 Woehlck HJ, Dunning M 3rd, Nithipatikom K, Kulier AH, Henry
of a hygroscopic heat and moisture exchanger for paediatric use. DW. Mass spectrometry provides warning of carbon monoxide
Anaesthesia 1991; 46: 296±9 exposure via tri¯uoromethane. Anesthesiology 1996; 84:
115 Wille B. Hygiene measures for anaesthesia and ventilator 1489±93
equipment. Krankenhaus-HygieneInfektionsverhutung 1989; 11: 120 Woehlck HJ. Severe intraoperative CO poisoning: should apathy
17±21 prevail? Anesthesiology 1999; 90: 353±4

264