Br. J. Anaesth.

(1987), 59, 648-650

POSSIBLE ROLE OF VACUUM SYSTEMS AND

COMPRESSED AIR GENERATORS IN CROSS-INFECTION IN
THE ICU
A Radioactive Tracer Study

P. BJERRING AND B. 0BERG

Downloaded from http://bja.oxfordjournals.org/ at Cornell University Library on July 25, 2015

Compressed air for medical use has been shown
recently to be a potential source of infection in the SUMMARY
intensive care unit (ICU) (Bjerring and 0berg, Cross-contamination between a hospital's
1986). A mixture of pathogens, and normal skin vacuum and compressed air systems was demon-
bacteria, including Pseudomonas aeruginosa. Bacil- strated using xenon-133 as a tracer. A xenon-133
lus species, Corynebacterium species, Micrococcusbolus was introduced to the vacuum system in
species and Staphy loco ecus epidermidis were traced the intensive care unit. Calculations based on the
to oil-lubricated air compressors. Surprisingly, no amount of radioactive tracer recovered from the
limitation on the microbial content of the air compressed air outlet at the same location as that
supplied directly to seriously ill patients receiving at which the tracer was introduced indicated that
ventilatory assistance with air-oxygen mixtures 17 % of the tracer had entered the compressed air
can be found in the Pharmacopoeia (Pharma- system. The contamination was caused because
copoea Nordica, 1963), or in the recommendations the vacuum and compressed air systems were
of the British or International Standards Or- located in the same machine room. This could
ganisations. conceivably provide a route for respiratory tract
In many hospitals air compressors are situated contamination in patients receiving ventilatory
very close to vacuum compressors in a poorly assistance with air-oxygen mixtures.
ventilated machine room. As a result the possi-
bility that bacteria may be transferred from the
outlet of the vacuum system to the intake of the in the ICU, at a rate of 42 litre min"1 for 5 min.
compressor used to generate air for medical use The vacuum storage tank contained 1500 litre and
should be considered. To verify this possible route the pipelines approximately 200 litre of air. From
of cross-infection we conducted a study, using the vacuum system xenon-133 was exhausted into
xenon-133 as a tracer, in a hospital in which the the compressor room (50 m3). Room air was drawn
outlets of the vacuum generators were located
1.5 m from the intakes of the air compressors
(fig. 1). Vacuum generator Air compressor

MATERIALS AND METHODS

Vacuum
Air containing 19.3 mCi (713 MBq) of xenon-133 generator outlet Q
was introduced to the hospital's vacuum system, Air compressor intake \
which is used for pharyngeal and tracheal suction

PETER BJERRING, M.D., Department of Dermatology, Marselis-

borg Hospital, DK-8000 Aarhus, Denmark. BJAHNE 0BERG,
Vacuum storage tank Air storage tank
M.D., Department of Anaesthesia, Aarhus Kommunehospital,
DK-8000 Aarhus, Denmark. Accepted for Publication: FIG. 1. The compressor room. Distance from vacuum
August 1, 1986. generator outlet to air compressor intake: 1.5 m.
 VACUUM AND COMPRESSED AIR SYSTEMS: CROSS-CONTAMINATION
1
into the compressors at a rate of 500 litre min"
and stored in a tank (1000 litre) at a pressure of
6.1-8.2 atm. From this tank air was supplied to
the peripheral air outlets throughout the hospital.
For measuring purposes air was drawn at a rate
of 10 litre min"1 from the peripheral outlet at the
same location as the suction intake at which the
xenon-133 was introduced. Radioactivity was
measured continuously by a Xenon monitor 301
(Alnor-MABO, Sweden) calibrated to monitor
xenon-133 in the range 0.1-5000 MBq m"3.
649
exhaust gases from the vacuum system with the
intake for the compressed air generator. The
mixing occurred because the vacuum pump and
air compressor were housed in the same small
room, with only 1.5 m between the outlet of the
vacuum system and the compressed air intake.
The design of the compressor room is not in
conflict with Danish standards, and we have
encountered the same arrangement in other
hospitals. The reason for this is partly to facilitate
servicing and maintenance of the compressors and
partly to counteract the problem of noise in the
hospital environment.

Downloaded from http://bja.oxfordjournals.org/ at Cornell University Library on July 25, 2015

RESULTS This study has demonstrated that it is possible
for a gas to be aspirated (from the bedside) into a
Xenon-133 emerged from the air outlet in the central suction unit and be returned to the same
ICU approximately 16 min after the tracer bolus bedside by way of an air compressor.-A radioactive
had been introduced to the vacuum system. The isotope was used as the tracer, because bacteria
levels of radioactivity are shown in figure 2. After introduced to the suction system might eventually
an initial sharp increase in activity a steady state have reached the lungs of any patients receiving
was attained, allowing calculation of the degree of artificial ventilation. Even though xenon-133 gas
cross-contamination between the vacuum and is more transferrable than a bacteria-containing
compressed air systems. Seventeen per cent of the aerosol, the present study has indicated that
xenon-133 administered was estimated to have compressed air for medical use could be con-
entered the compressed air system. taminated easily by bacteria from pharyngeal and
tracheal suction procedures, and from the opera-
tive drainage of infection. In a previous study we
DISCUSSION were able to demonstrate the presence of patho-
This study has demonstrated the consequences of gens in compressed air used for lung ventilation
an engineering design, which allows the mixing of (Bjerring and 0berg, 1986). In the present study
only a small bolus of xenon-133 was injected
to the vacuum system; in everyday use the system
is contaminated continuously by microorganisms
150
from suction procedures. Over the years the inside
of the pipelines and vacuum storage tanks may
become lined with a film of microorganism-
E containing material (0berg and Bjerring, 1986).
s We believe that the practice of locating vacuum
100- and compressed air generators in the same
machine room may involve a risk of cross-
o infection—particularly in the ICU. Infection has
a been quoted as the most important single factor
m
m determining outcome in the intensive care patient
50-
c
o
(Andersen, 1984) and the airway as the most
c
o frequent source of septicaemia (Dominguez de
X Villota et al., 1983). Whereas a general medical or
surgical patient has a 6% risk of becoming
0-J infected during his stay in hospital, the risk is
18 % in the patient in the ICU (Donowitz, Wenzel
0 15 20 30 and Hoyt, 1982).
Time after bolus injection (min) The hospital in which we performed this
FIG. 2. Activity of xenon-133 detected at the air outlet in the investigation is a centre for the treatment of
ICU. patients with epidemic diseases like meningitis,
650
severe respiratory tract infections, and the ac-
quired immune deficiency syndrome. Thus the
substantial (17%) cross-contamination from the
vacuum system to the intakes of the air generators
could prove particularly dangerous.
Bacteria present in medical air may also reach
pneumatic surgical instruments, which are
powered with 200-300 litre of compressed air per
minute. During their use at least some of this air
is released into the operative field (0berg and
Bjerring, 1985). It is generally accepted that
airborne bacteria are responsible for a large
number of postoperative infections (Hambreus
and Laurell, 1980).
Good engineering practice dictates that the
vacuum pump should exhaust outside the com-
pressor room, well away from the intake to the air
compressor. The fitting of bacterial niters to the
air intakes in the compressor room and to
mechanical ventilators is recommended to counter-
act unnecessary bacterial contamination of the
compressed air being supplied to the seriously ill
patient.
BRITISH JOURNAL OF ANAESTHESIA
ACKNOWLEDGEMENT
The authors wish to thank Dr Med. H. Hvid Hansen, Head of
the Department of Nuclear Medicine, Aarhus University
Hospital, for support during the study.
REFERENCES
Andersen, R. (1984). Infections as a problem in the intensive
care unit. Scand. J. Gastroenterol. (Suppl.), 19, 83.
Bjerring, P., and 0berg, B. (1986). Bacterial contamination of
compressed air for medical use. Anaesthesia, 41, 148.
Dominguez de Villota, E., Algora, A., Rubio, J. J., Roig, M.,
Mosquera, J. M., Galdos, P., and Diez-Balda, V. (1983).
Septicaemia in a medical intensive care unit. Clinical,
biochemical and microbiological data of 109 cases. Intern.
Downloaded from http://bja.oxfordjournals.org/ at Cornell University Library on July 25, 2015
Care Med., 9, 109.
Donowitz, L. G., Wcnzel, R. P., and Hoyt, J. W. (1982). High
risk of hospital-acquired infection in the ICU patient. Crit.
Care Med., 10, 355.
Hambreus, A., and Laurell, G. (1980). Protection of the patient
in the operating suite. J. Hosp. Infect., 1, 15.
Pharmacopoea Nordica (1963). Addendum. Aer Medicinalis.
Oberg, B., and Bjerring, P. (1985). Pneumatic surgical
instruments and postoperative infection. Lancet, 2, 1436.
(1986). Comparison of microbiological contents of
compressed air in two Danish hospitals. Effect of oil and
water reduction in air-generating units. Acta Anaesthcsiol.
Scand., 30, 305.