See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/234130618

2013011709373986

Data · January 2013

CITATIONS READS

0 954

4 authors:

Chris Del Mar Paul Glasziou

Bond University Bond University
610 PUBLICATIONS   17,027 CITATIONS    846 PUBLICATIONS   80,501 CITATIONS   

SEE PROFILE SEE PROFILE

Anneliese Spinks Sharon Sanders

The Commonwealth Scientific and Industrial Research Organisation Bond University
105 PUBLICATIONS   4,126 CITATIONS    62 PUBLICATIONS   2,204 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

MEDICAL EDUCATION PROJECT View project

Cochrane reviews View project

All content following this page was uploaded by Chris Del Mar on 02 June 2014.

The user has requested enhancement of the downloaded file.

 EBM IN ACTION

Is isopropyl alcohol swabbing before injection really necessary?

~ Clinical Question
A patient ha~ilng her skin swabbed before an intramuscu-
lar injection asked her general practitioner "What is the
value of cleaning the skin like that?". This prompted the GP
to wonder whether sterilising the skin for one to three sec-
onds with 70% isopropyl alcqhoi conferred any benefit. He
wanted to know if it actually prevented sepsis at the site of
injection.
~3 Search Question
The revised question was: "Compared with no cleansing, does
swabbing the skin with 70% isopropyl alcohol before puncture
with a sterile needle reduce the risk of infection at the injection
site?", q13~e ideal study design to answer this question would
be a randomised controlled trial comparing complication
rates at injection sites swabbed, ornot swabbed, with iso-
propyl alcohol.
Search
We searched two online databases, Gochrane Library and
SUMSearch (<http://sumsearch.uthscsa.edu/search
form4.htm>), using the search terms "injection", "sterile
swab", "isopropyl alcohol", "sepsis", "skin piercing", and
"sldn puncture".
’~,,~ Summary of Findings
Three controlled trials, two of which were randomised,
examined the benefit of isopropyl swabbing before injec-
tion in preventing subsequent infection.
In a randomised single-blind controlled trial, patients
ha~img venesection either had their skin prepared with
isopropyl alcohol (n=93) or had no skin preparation
(n=101).1 There were no statistically significant dif-
ferences in complications at the venepuncture site
between the two groups.
The second randomised controlled trial evaluated the
effects of skin disinfection with isopropyl alcohol versus
no disinfection in patients having intravenous cannu-
ladon. No significant differences were found between
the two groups in intraluminal contamination rates,
colony counts or organisms isolated)
Five seconds of isopropyl alcohol swabbing was com-
pared with no skin cleansing in a cross-over trial of 13
patients with diabetes.3 (Patients with diabetes are suit-
able for this type of study because they have many
injections and may be more prone to infection.) With
cleansing, bacterial counts, estimated by culture, were
reduced by 82%-91%. However, over a period of three
to five months, skin preparation before insulin injection
was omitted every second week and no signs of local or
systemic infection were observed.
~ Outcome
We reported to the GP that the studies showed that omit-
ling skin preparation with 70% isopropyl alcohol did not
increase incidence of infection at the injection site. He has
stopped swabbing patients with isopropyl alcohol before
injection.
Christopher B Del Mar
Professor
Paul P Glasziou
Professor
Anneliese B Spinks
Research Officer
Sharon L Sanders
Research Officer
Centre for General Practice, Medical School
University of Queensland, Herston, QLD
cdelmar@cg p,uq.edu.au
References
1. Sutton CD. White SA, Edwards R, Lewis MH. A prospective controlled trial of
the efficacy of isopropyl alcohol wipes before venesection in surgical patients.
Ann R Coil Surg Engl 1999; 81:183-186.
2. Grabe N, Jakobson CJ, Damm MD, Skin disinfection before intravenous
cannulation. Intraluminal contamination after disinfection with 70% isopropyl
alcohol. Acta Anaesthesiol Scand 1985; 29: 764-766.
3. Koivisto VA, Felig P. Is skin preparation necessary before insulin injection?
Lancet1978; 1: 1072-1075. 13

DISABILITY EVALUATION
l~j~l~]~l 1 s notIt’s
always easy
not always easy to distin "~o
to distinguish ~~
the two y where
-- especially impairment
where impairment in in
oneone
~rofession may be a disability in another. This resource was designed to make the decision
~~llm~gned to make the de(ision
easier
]~’~flon, the compaNon guide toby
theproviding specific evaluation guidelines. Disability Evaluation, the companion guide to the
critical d ffeteoces between
Guides to the Evaluation of Permanent Impairment, helps you clarify the critical differences between
~~~r~ty by show~ degree to which
impairment and disability by showing
imp~rments affectyou how to determine the degree to which impairments affect
ot specific io s.
work performance or specific jobs.
For further information, contact AMPCo: Ph 02 9954 8666 ¯ Fax 02 9956 7644 * Emall sates@ampco.com.au
¯ Plus $7.65
m’°°
$7,65 postage
postage and
andhandling withinAustralia
handlingwithin Australia(price
(price includes
includesGST).
GST).AMA
AMAmembers
membersreceive 10%
receive a 10 % discount
discount

306 MJA Vo1174 19 March 2001

[] Although AFP is the "official journal of 1, Askew DA, GIasziou PP, Del Mar CB. Research output of Aus- venepuncture. This would indicate that the
tralian general practice: a comparison with medicine, infection rate is about 0.001%~ although it
[its] respective clinical college", it is not surgery and public health, MedJAust2001; 175: 77-80, CI
funded by that college but is entirely is obvious there will be some under-report-
dependent on commercial funding for its ing. If the skin were not cleaned and the
infection rate doubled to 0.002%, the
publication. Although editorial processes are Is isopropyl alcohol swabbing studies quoted by Del Mar et al would not
scrupulously independent of advertising, it before injection really
have had enough statistical power to detect
does mean that, by commercial necessity, necessary?
this difference.
AFP is a "reader’s" journal rather than an
In Australia there would be several mil-
"author’s" journal. Given that AFP goes to Ray Burn
lion skin punctures each year. If my estimate
every GP in the country, whether a college General Practitioner, Old Linton Medical Practice, of the incidence of infection and my guess
member or not, its appeal has to be very 153 Comur Street, Yass, NSW 2582 that it might double without swabbing are
rburn @interact.net.au
broad and driven by its readers’ needs. As anywhere near correct, then there would be
original research consistently polls very .To THE EDITOR: The article by Del Mar a considerable increase in morbidity (and
poorly in terms of readership interest, and colleagues gives a straight answer to a perhaps even mortality) if skin preparation
AFP’s continual dedication to offering straight question: swabbing with isopropyl was abandoned.
opportunities for publication of relevant alcohol gives no better result than no swab- I agree that there is no evidence to sup-
original research attests to the strong belief bing at all.1 !ks a result, the doctor asking port skin preparation, but equally there is no
of the journal’s editorial staff in the impor- the question stopped swabbing patients adequate evidence to support its with-
tance of primary care research for the ben- when giving iniections. One wonders about drawal. Perhaps Del Mar and colleagues
efit of our discipline. this, especially as in the same issue of the could tell me iust how big a trial would be
[] During the p,egiod of Askew and col- Journal there is a topically perceptive article~ required to really answer the question
leagues’ study, the original research section on evidence-based medicine by Celermaier. raised. Should the general practitioner who
of AFP had an acceptance rate of about The maior question on swabbing then made the enquiry be unfortunate enough to
41%. For a research journal this is actually b~comes: is there an infection problem at find himself in court defending his inaction
quite high. Although the point is made that injection sites and, if so (as isopropyl aico- in not preparing the skin, I believe he would
only 2.2 original research articles were pub- he1 doesn’t help), are there agents that will not have many expert witnesses queuing up
lished per issue, this was not always easy to reduce the extent of the problem? to support him.
achieve, as AFP is issued monthly and only .Doctors who routinely give iniections
1. Del Mar ca, Glasziou PP, Spinks AB, Sanders ~L, Is iso-
51 research articles were actually submitted such as flu shots note that reasonably vig- propyl alcohol swabbing before injection really necessary?
MedJAust2001; 174: 306. CI
during 1999. From AFP’s perspective, this orous swabbing produces a dirty swab -- so
something is there. Perhaps further evidence
raised the possibility that either the research
was not being done or was not being sub- is needed.
James E Elder,* John B Carlint
mitted for consideration for publication. 1. Del Mar CB, Glasziou PP, Spinks A8, Sanders SL, Is ise-
propyl swabbing before injection really necessary? Med J *Director, Department of Ophthalmology; t Directob
~ The fact that 65% of the original Aust2001; 174: 306. Clinical Epidemio!ogy and Biostatistics Unit, Royal
research articles published had a GP as the 2. Celermajer DS. Evidence-based medicine: hew good is the Children’s Hospital, Parkville, VIC 3052
evidence? Med J Aust 2001 ; 174: 293-295. [3 jbcadin @ unimelb.edu.au
primary author is further testament to
AFP’s commitment to encourage GPs to TO THE EDITOR: Calermajer’s article is a
both write up their work and publish it. In common-sense "reality check" that cautions
the light of this, another interesting study Gordon E Rich about the potential for the methods of evi-
would be to look at the number of articles Clinical Microbiologist, Western Diagnostic Pathology, dence-based medicine to provide reliable
"l~bou~: general practice compared with the 74 McCoy Street, Myaree, WA 6154 guidance on clinical decisions.~ However,
number written by a GP as th~ primary TO THE EDITOR: AS a microbiologist, I was the article by Del Mar et al in the same issue
author. Indeed, Askew’s article itself would amazed at the conclusion reached in the does not exhibit similar caution.2 Although
be grouped in the first category rather than article by Del Mar et al relating to skin we do not disagree with the conclusion of
the second. preparation before injections.1 The basis for Del Mar et ai (ie, that swabbing with iso-
While publication in peer-reviewed jour- skin cleaning with alcohol is common propyl alcohol before iniection is probably
nals helps disseminate knowledge, and sense: the alcohol removes dirt a~d many of unnecessary), we are not convinced that it
thus provides the opportunity for it to have the skin bacteria, even if insufficient time is follows from the evidence cited.
some impact on healthcare practices, I am allowed for any bactericidal effect to pro- The specific question posed by Del Mar
not sure that there is a direct correlation duce a sterile field. et al was whether isopropyl alcohol swab-
between publication and research activity. Any general hospital will see a small bing "reduces the risk of infection at the
However, Askew’s study may encourage number of patients each year with gluteai iniection site". Their literature search
future work to address ways of overcoming abscesses occurring after intramuscular yielded three controlled trials (two ran-
the obstacles that GPs face when seeking injections. I am not aware of any published demised and one crossover) related to this
answers to the relevant questions in their studies assessing the infection rate after question. None of these studies included
clinical lives. It may also encourage them to injections, but in a busy pathology practice more than 200 patients and the incidence of
take the next step of sharing those findings I see two or three patients each year who infections was very low, providing no evi-
with the wider general practice community. develop locaiised thrombophlebitis after dence that swabbing reduced infection

MJA Vo1175 17 September 2001 341

 Like everything else in medicine, we have
risk. But do these studies provide evidence Christopher B Del Mar,* Paul P Glasziou,t
Anneliese B Spinks,* Sharon L Sanders* to do the best we can with what we have,
that in the long run there is no difference in
* Professor and Director, t Associate Professor, with evidence as well as suboptimal
infection risk with or without swabbing? ~:Research Officers, Centre for General Practice, resources, drugs, interventions, hospital
One of the studies cited by Del Mar et aP Medical School, University of Queensland, Herston, beds, numbers of doctors and nurses and
QLD 4006
refers to an earlier article by Darm4 that sug- c.de~rnar @ cgp,uq.edu,au everything else. We make no claims that
gests a complication rate for venepuncture "EBM is unassailable fact". The evidence is
without skin cleaning/swabbing of less than IN REPLY: One of the most exciting merely the best that is available. The pity is
one in 5000. It would be impossible to outcomes of practising evidence-based that such rich sources of research are so
detect differences in incidence around such medicine (EBM) is when something rarely tapped to inform clinical practice.
a low base rate without very large sample counterintuitive emerges from looldng at
sizes. Assuming a true infection rate as high the best available research. Our finding dis- 1, Celernajer DS, Evidence*based medicine: how good is the
evidence? Meal J Aust 2001 ; 174: 293-295.
as 1% (whether swabbed or not), a standard turbs Burn, especially when he reads an 2. Del Ma~ CB, Glasziou PP, Spinks AS, Sanders SL Is iso-
calculation shows that for a study to have article critical of EBM~ and looks at a dis- propyl alcohol swabbing before injection really necessa~’?
Med J Aust 2001 ; 174: 306,
80% power of reiecfing a difference as large coloured alcohol swab with which he has 3. Oann TO. Routine skin preparation before injec~on: an
as a doubling of the rate of infection (ie, to iust scrubbed the skin before an injection. unnecessary procedure. Lancet !969; 2: 96-98,
2%) in the non-swabbed population (at the EBM changes the emphasis away from 4. Reining DR, Jacober S J. Vandenbutg MA, etal, ~Eqe safe~y
of injecting insulin through clothing. Diabetes Care 1997; 20:
95% confidence level) would require about pathophysiological mechanisms and towards 244-247, 13
900 subiects per experimental arm. the empirical. Ideally these agree. Cleaning
The best evidence against the need to the skin sounds as if it should be safer than
swab may be in Dann’s estimate of the rate not doing so. However, swabbing the skin Complementary and alternative
of complications (assuming it is based on a conferred no advantage when tested in three
trials. 2 Why would there be no detectable medicine commonly used by
~
valid study), not on the controlled trials,
difference between swabbing and non- cancer patients
which were unable to answer the real
question. If the unswabbed risk of infection swabbing? Firstly, using alcohol wipes on David Chapman-Smith
the skin for the usual five second8 is
were of the order of one in 5000, it is unlikely to kill bacteria -- at least two min- Editor, The Chiropractic Report, 3080 Yonge Street,
unlikely that any further reduction due to Suite 5065, Toronto, Ontario M4N 3N1, Canada
tcr @ chiropracticreport.com
swabbing would be worthwhile. Another utes would be required (as recommended in
preparing the skin before intrathecal or
concern with the evidence of Del Mar et al intra-articular injections).3 Secondiy, asTO THE EDITOR: The primer on comple-
is that one of the trials reported no infec- Eider, Carlin and Rich indicate, the three mentary and alternative medicine (CAM)
tions and focused entirely on bacterial con- trials were not large enough to rule out rare for cancer patientsI has misinformation on
tamination,~ .which has a tenuous harms from not using alcohol swabs (but chiropractic from an author, Professor
connection with infection risk. this is true of using them, too). Ernst, with estabfished bias. His similar
Our observations on the quality of the Burn, Elder and Carlin and Rich might commentary on chiropractic in the British
evidence used by Del Mar et al highlight the take heart from Dann’s experience: in the MedicalffournaF was dismissed by medical
importance of using appropriate criteria for course of giving about 5000 injections in his leaders in the field of back pain as an
assessing the value of a study before using it general practice, for which he cleaned the embarrassing example of "interprofessional
to support an assertion of evidence-based skin only rarely (to remove obvious dirt), confrontation" under the "guise of scientific
medicine. Published evidence in the form of not a single infection resulted.3 In another objectivity"3 and "an unduly cynical pic-
randomised controlled trials may be inade- study (a crossover trial of 42 people with ture" based on self-ituerest.~
quate to answer some questions, in which diabetes), insulin injections through cloth- Ernst’s article was meant to be about
, case clinical decisions must be based on ing were compared with insulin injections treatments for cancer, but his discussion of
other sources of evidence, including best after alcohol swabbing: after 6890 injections chiropractic hardly mentions the subject. He
scientific judgement. This example high- of each type, there were no cases of ery- commences with an archaic definition of
lights a common problem -- bridging the thema, induration, or abscesses.4 It is also of chiropractic which erroneously implies that
gap between "absence of evidence" and note that dentists routinely give injections chiropractors clalm to cure most diseases,
"evidence of absence" of an effect. It can be without preparing the gain, despite the including cancer. They do not. A contem-
difficult ~to prove a negative. mouth being full of bacteria. What further porary definition of chiropractic by the
evidence would we require to make us World Federation of Chiropractic, of which
1 Cefermaier OS Evidence based medicine: how good is the change our mind? the Chiropractors’ Association of Australia
evidence? MedJAust2001; 174: 293-295.
Celemajer’s criticisms of EBM centre on is a member, is "the diagnosis, treatment
2, Del Mar C8, Glasziou PP, Spinks A8, Sanders SL. Is iso-
propyl alcohol swabbing before injection real~y necessary? the facts that the best available evidence and prevention of mechanical disorders of
Med J Aust 2001 ; 174: 306. might be suboptimal or available for patient the musculoskeletal system, and the effects
3 Sutton CO, White SA, Edwards R, Lewis MH. A prospective
groups that are not identical to the patient of these disorders on the function of the
controlled trial of the efficacy of isopropy! alcohol wipes "5
before venesection [n surgical patients. Ann R Cell Surg Engl in a particular case or involve an area of nervous system and general heahh.
1999; 81: 183-186. medicine less well researched than others.~ Surveys consistently report that most chi-
4 Dann TC Routine skin preparafion before injection: an ropractic patients present with muscu-
unnecessap] procedure. Laecet 1969; 2: 96-98. These make the application of EBM more
5 Grabe N, Jakobson C J, D~rnrn MD. Skin disinfection before tricky, and in our EBM workshops we !oskeletal pain, and when chiropractors trea~
inL~avenou~ cannulation. Intralurninal conLaminaticn aSer dis- explicitly teach approaches to deal with cancer patients this is principally palliative
infection ~)ith 70% isopropyl alcohol Acta Anaesthesiol care for concomitant musculoskeletal pain,
Scand 1985; 29: 764-766. CI these challenges.

MJA Vo1175 17 September 2001

342

View publication stats