Parienti JJ, Thibon P, Heller R, Le Roux Y, von Theobald P, Bensadoun H, Bouvet A, Lemarchand F, Le Coutour X, for Members of the Antisepsie Chirurgicale des Mains Study Group . Hand-Rubbing With an Aqueous Alcoholic Solution vs Traditional Surgical Hand-Scrubbing and 30-Day Surgical Site Infection RatesA Randomized Equivalence Study. JAMA. 2002;288(6):722-727. doi:10.1001/jama.288.6.722
Author Affiliations: Departments of Infectious Diseases and Intensive Care Unit (Dr Parienti), Hygiene and Public Health (Drs Parienti, Le Coutour, and Thibon), Abdominal Surgery (Dr Le Roux), Gynaecology and Obstetrics (Dr von Theobald), and Urology (Dr Bensadoun), Côte de Nacre University Hospital Centre, Caen; Department of Hygiene, General Hospital, Colmar (Dr Heller); Department of Anaesthesiology, François Baclesse Oncology Centre, Caen (Dr Bouvet); and Department of General Surgery, General Hospital, Pont-Audemer (Dr Lemarchand), France.
Context Surgical site infections prolong hospital stays, are among the leading
nosocomial causes of morbidity, and a source of excess medical costs. Clinical
studies comparing the risk of nosocomial infection after different hand antisepsis
protocols are scarce.
Objective To compare the effectiveness of hand-cleansing protocols in preventing
surgical site infections during routine surgical practice.
Design Randomized equivalence trial.
Setting Six surgical services from teaching and nonteaching hospitals in France.
Patients A total of 4387 consecutive patients who underwent clean and clean-contaminated
surgery between January 1, 2000, and May 1, 2001.
Interventions Surgical services used 2 hand-cleansing methods alternately every other
month: a hand-rubbing protocol with 75% aqueous alcoholic solution containing
propanol-1, propanol-2, and mecetronium etilsulfate; and a hand-scrubbing
protocol with antiseptic preparation containing 4% povidone iodine or 4% chlorhexidine
Main Outcome Measures Thirty-day surgical site infection rates were the primary end point;
operating department teams' tolerance of and compliance with hand antisepsis
were secondary end points.
Results The 2 protocols were comparable in regard to surgical site infection
risk factors. Surgical site infection rates were 55 of 2252 (2.44%) in the
hand-rubbing protocol and 53 of 2135 (2.48%) in the hand-scrubbing protocol,
for a difference of 0.04% (95% confidence interval, −0.88% to 0.96%).
Based on subsets of personnel, compliance with the recommended duration of
hand antisepsis was better in the hand-rubbing protocol of the study compared
with the hand-scrubbing protocol (44% vs 28%, respectively; P = .008), as was tolerance, with less skin dryness and less skin irritation
after aqueous solution use.
Conclusions Hand-rubbing with aqueous alcoholic solution, preceded by a 1-minute
nonantiseptic hand wash before each surgeon's first procedure of the day and
before any other procedure if the hands were soiled, was as effective as traditional
hand-scrubbing with antiseptic soap in preventing surgical site infections.
The hand-rubbing protocol was better tolerated by the surgical teams and improved
compliance with hygiene guidelines. Hand-rubbing with liquid aqueous alcoholic
solution can thus be safely used as an alternative to traditional surgical
Surgical site infections (SSI) prolong hospital stays1;
they are also among the leading nosocomial causes of morbidity2
and a source of excess medical costs.3 Sterile
gloves contribute to preventing surgical wound contamination,4
but some are permeable to bacteria and all can be damaged during use. Effective
hand antisepsis thus remains crucial. Guidelines on hand and forearm antisepsis
for operating department staff5 aim at removing
transient microorganisms and reducing the resident flora (by at least 2-log).
In France, 2 protocols are recommended for surgical hand preparation:
5 minutes of hand-scrubbing with antiseptic soap; and an optional 1-minute
hand wash with nonantiseptic soap and tap water, followed by 5 minutes of
hand-rubbing with a liquid aqueous alcoholic solution (AAS) alone. In other
European countries, only 3 minutes of hand-rubbing with AAS are recommended.
Many reports suggest that AAS has better and more sustained antimicrobial
activity than traditional scrubs.6- 11
For example, Hobson et al11 compared hand antimicrobial
count after either 3 minutes of hand-scrubbing with solutions containing 7.5%
povidone iodine and 4% chlorhexidine gluconate or 3 minutes of hand-rubbing
with an AAS. Mean logarithmic reduction of colony forming units (CFU) from
baseline was significantly higher in the AAS protocol when compared with povidone
iodine or chlorhexidine gluconate after 1 minute (2.90 log10 vs
1.20 log10 and 1.68 log10 of CFU reduction from baseline),
3 hours (1.58 log10 vs 0.71 log10 and 1.08 log10 of CFU reduction from baseline) and 6 hours (1.94 log10
vs –0.21 log10 and 0.86 log10 of CFU reduction
from baseline) at day 1 and day 2. However, these investigations involved
healthy volunteers or short study periods, with hand microorganism counts
as the primary end point. Poor compliance and the risk of dermatitis were
not taken into account. Because of the need for a very large population sample,
together with the existence of numerous confounding factors and prohibitive
costs, clinical studies comparing the risk of nosocomial infection after different
hand antisepsis protocols are scarce.5,9
To confirm the validity of current recommendations for hand antisepsis
before surgery, we conducted a randomized equivalence study comparing hand-scrubbing
and hand-rubbing protocols with a multiple service crossover experimental
design. The aim was to demonstrate the equivalence of the 2 protocols in terms
of SSI rates.
Six surgical services in France were invited to participate in this
study and all accepted. They consisted of 3 surgical services of a teaching
hospital (Côte de Nacre University Hospital Centre, Caen) and the surgical
services of 3 nonteaching hospitals (François Baclesse Centre, Caen;
General Hospital, Colmar; General Hospital, Pont-Audemer). A 2-month feasibility
study was first conducted in a single surgical service, during which training
requirements were determined; the surgical team switched readily from traditional
hand-scrubbing to hand-rubbing with AAS. Surgeons from the 6 candidate centers
were then invited to discuss the project. The clinical trial started January
1, 2000, and lasted 16 months (May 1, 2001). The following end points were
used: nosocomial SSI rates (main end point), compliance by the surgical teams,
and tolerance of the 2 hand antisepsis protocols by the surgical teams in
real world conditions. The first protocol to be used in each surgical service
was chosen randomly. Each participating surgical service was assigned a 2-digit
random number by using a random number table. Surgical services corresponding
to the 3 higher numbers were assigned to hand-rubbing with AAS and the remaining
3 services were assigned to traditional hand-scrubbing. The alternative antiseptic
product was systematically removed from the services during each period. At
the end of each month, the antiseptic products were switched in a multiple
service crossover design.
Diagnosis for SSI was standardized in accordance with the Centers for
Disease Control and Prevention (CDC) definitions for nosocomial infection.12 For the purpose of this study, however, SSI surveillance
lasted 30 days regardless of prosthesis implantation. It should be noted that
our hygiene department has been involved in SSI surveillance for many years.13
In-hospital SSIs were prospectively diagnosed by a surgeon, infectious
disease specialist, or hygiene specialist on a standard data-collection form.
Postdischarge surveillance was based on chart review of visits and telephone
contacts with the surgeons. If data at 30 days were unavailable, the patient
was contacted by telephone and asked to answer a brief questionnaire on fever
and other potential symptoms of SSI, together with antibiotic use and visits
to an emergency department or to another physician, who was then contacted
to confirm the SSI.
According to CDC guidelines, all SSIs had to be confirmed by the surgeon
or the physician in charge of the patient. Thus, observers of the clinical
outcome could not be blinded to the hand antisepsis protocol. For possible
postdischarge SSIs reported by the patient only, a data validation for SSIs
according to CDC guidelines was performed by investigators who were blinded
to the protocol used by surgeons prior to surgical procedure.
All consecutive patients treated in the 6 participating surgical services
were screened for SSI. Because bacteria on operating department personnel
hands are more likely to affect the outcome in clean and clean-contaminated
surgical wounds (classified according to Altemeier et al14),
our study group decided to exclude from analysis patients in contaminated
or dirty procedure groups, when designing this study. Patients undergoing
a second operation less than 15 days after the first were also excluded, because
this is an independent risk factor for SSI.
The first 2 months after randomization of the services, investigators
checked if the protocols were correctly performed. Therefore, because introduction
of a new protocol and feedback can influence short-term surgical personnel
behavior, compliance observations for the hand antisepsis protocol started
at month 3 and continued until month 16. Twice as many observations were performed
in the new hand-rubbing protocol compared with the traditional hand-scrubbing
protocol. Compliance observers did not belong to the operating department
team but were usually present in the surgical suite. To avoid a Hawthorne
effect, the surgical teams were not informed of the timing of the evaluations.
Only the first hand antisepsis of the day was observed in the 2 study protocols.
In the 3 participating surgical services of our teaching hospital (Côte
de Nacre University Hospital Centre), the surgical personnel (77 subjects)
were asked to estimate the effect of the 2 protocols on their skin. We used
2 10-cm visual analog scales, at month 0 and after 3 crossovers; 0 cm representing
absence of any tolerance problem and 10 cm representing maximal dryness with
chapped hands and desquamation or maximal irritation with erythema, burning
sensation, and abrasion.
The standard surgical scrubbing antisepsis technique was as defined
in CDC guidelines. In particular, at least 5 minutes of systematic hand-scrubbing
was required with a sterile sponge and brush. Hand-scrubbing with antiseptic
solutions containing 4% povidone iodine (Betadine, Asta Medica, Merignac,
France) or 4% chlorhexidine gluconate (Hibiscrub, AstraZeneca, Rueil-Malmaison,
France) had been used for many years by the surgical teams and was thus chosen
as the control protocol for this study.
Hand-rubbing involved a 75% AAS containing propanol-1, propanol-2, and
mecetronium etilsulfate (Stérillium, Rivadis Laboratories, Thouard,
France). We chose this product because it was the only AAS licensed for surgical
antisepsis in France. Prior to the first procedure of the day, or if the hands
were visibly soiled, the surgical team was instructed to use a nonantiseptic
soap (Savon Codex, Rivadis Laboratories, Thouard, France) for a 1-minute hand
wash, including subungual space cleaning with a brush. The hands and forearms
were then rinsed with nonsterile tap water and wiped carefully with nonsterile
paper. The user was instructed to take enough AAS to fully cover the hands
and forearms (at least 5 mL, which represents at least 4 pump strokes), and
to apply it twice for 2 minutes 30 seconds (for a total of 5 minutes) without
drying. As recommended by the manufacturer, users were also instructed to
rub their hands with AAS for 30 seconds when changing gloves. The hand-rubbing
technique was based on the European Norm 1500 from the Association Française
The study was designed to demonstrate the equivalence of hand-scrubbing
and hand-rubbing in preventing SSI (primary end point). We chose an SSI rate
of 4% in the control group, a maximal difference between the protocols of
2% (considered as clinically relevant), a real difference between the 2 protocols
of 0%, a type I error α risk of .05, and 90% power when calculating
the number of patients required. On this basis, at least 4158 patients had
to be enrolled.15 We performed an as-treated
analysis as well as a conservative intent-to-treat analysis computing missing
value equals SSI in the new protocol and missing value equals no SSI in the
control group. The 2 protocols were considered equivalent if the 95% confidence
interval (CI) of the SSI rate difference, calculated according to the Wallenstein
method,16 was within the limits of –2%
to +2% and contained the bound zero in both analyses, the as-treated as well
as in the conservative intent-to-treat. The level of significance for equivalence
was given by the highest P value related to the lowest χ2 value of the continuity-corrected 1-sided test described by Dunnett
and Gent.17 Compliance and tolerance (secondary
end points) were compared using the Fischer exact test for qualitative data
and Mann-Whitney test or a matched t test for quantitative
data as appropriate (2-sided P values). After checking
for the absence of interaction, observations regarding compliance with and
tolerance of hand hygiene protocols from different surgical services and from
different quarters of the study were merged. P<.05
was considered statistically significant. We used EPI-INFO version 6.04dfr
(EPI-INFO, CDC, Atlanta, Ga) for data collection and EPI-INFO and SAS version
6.12 (SAS Institute Inc, Cary, NC) for data analysis.
During the study period (Figure 1),
4823 consecutive patients underwent surgery. Among these, 385 patients underwent
contaminated or dirty-contaminated surgery, and 51 were lost to follow-up
at 30 days (17 in the hand-rubbing group). The remaining 4387 patients (68.5%
of whom underwent clean surgery) were considered for analysis. The characteristics
of these patients are shown in Table 1,
according to the type of protocol. There were 446 gynecological operations
(33%), 383 cesarean deliveries (28%), and 356 breast operations (26%) (n =
1359). Thyroidectomy was the most frequent otolaryngology procedure (n = 346).
Urological indications (n = 540) comprised 384 (71%) bladder or urethral operations
and 85 (16%) kidney operations. Orthopedic procedures (n = 746) comprised
290 (39%) osteosyntheses and 147 (20%) prosthesis insertions.
The surveillance system identified 99 in-hospital and 9 postdischarge
SSIs. The global SSI rate 30 days after surgery was 2.46% (95% CI, 1.81%-3.11%).
The SSI rates for clean and clean-contaminated surgery were 2.03% (95% CI,
1.33%-2.73%) and 3.40% (95% CI, 2.07%-4.73%), respectively. In the hand-scrubbing
protocol, 21 SSIs were superficial, 19 deep, and 8 organ-space (5 unknown).
In the hand-rubbing protocol, 24 SSIs were superficial, 15 deep, and 7 organ-space
(9 unknown). The distribution of these categories of SSI did not differ between
the 2 protocols.
There were 53 of 2135 SSIs (2.48%) in the hand-scrubbing protocol and
55 of 2252 (2.44%) in the hand-rubbing protocol (Table 2). The difference between the SSI rate with hand-scrubbing
and the SSI rate with hand-rubbing with AAS was 0.04% (as treated 95% CI, −0.88%
to 0.96%). In an intention-to-treat analysis, considering that all the 17
patients lost to follow-up in the hand-rubbing group had an SSI and none of
the 34 patients lost to follow-up in the hand-scrubbing group had an SSI (maximal
bias), the rate difference would have been –0.69% (95% CI, −1.67%
to 0.29%). The equivalence of the 2 protocols in preventing SSI was thus accepted.
During the study period, 278 individual compliance assessments were
made of the operating teams (174 in the hand-rubbing group), corresponding
with 160 surgical procedures (102 in the hand-rubbing group). On average,
the first hand-cleansing protocol of the day (Table 3), excluding the simple nonantiseptic hand wash prior to
hand-rubbing, lasted significantly longer in the hand-rubbing group than in
the hand-scrubbing group (mean [SD], 313  seconds vs 287  seconds; P = .01). Scrub nurses complied better with the recommended
duration of hand antisepsis than did surgeons and assistants (56% vs 33%; P<.001). Compliance with the recommended duration of
hand hygiene was poor in both protocols but was significantly better in the
hand-rubbing group than in the hand-scrubbing group (44% vs 28%, respectively; P = .008).
In the hand-rubbing group, 6 of 102 (6%) of first procedures did not
include a simple nonantiseptic hand wash before hand-rubbing (4 in abdominal
surgery, 2 in gynecology). During the days nonantiseptic hand washing prior
to the first handrub with AAS was not performed by the surgical personnel
in the hand-rubbing protocol, 2 SSIs occurred for a total of 34 procedures
(5.9%) compared with 2 SSIs for 56 procedures (3.6%) when the nonantiseptic
hand washing was performed in the same services. Glove changes occurred during
32 of the 102 observed procedures in the hand-rubbing protocol; the recommended
30-second AAS rub before a glove change was complied with in 16 of these 32
cases, not complied with in 10 cases, and 6 were unknown.
A total of 77 operating department staff members were assessed for skin
tolerance at entry to the study and after the first 3 crossovers. Based on
the visual analog scales scores, skin dryness decreased by 0.9 cm (95% CI,
0.5-1.2) after the hand-rubbing periods and increased by 0.4 cm (95% CI, –0.1
to 1.2) after the hand-scrubbing periods (P = .046).
Similarly, skin irritation decreased by 1.5 cm (95% CI, 1.1-1.9) after the
hand-rubbing periods and increased by 0.4 cm (95% CI, 0.2-0.6) after the hand-scrubbing
periods (P = .03). One scrub nurse reported hand
and eye irritation (swelling) when using AAS for hand-rubbing.
To our knowledge, this study is the first randomized trial to compare
hand-rubbing with alcohol-based solution and traditional hand-scrubbing in
the routine surgical setting, with the 30-day SSI rate as the primary end
point. The hand-rubbing with AAS was equivalent to traditional hand-scrubbing
in preventing SSI after clean and clean-contaminated surgery. In addition,
hand-rubbing with AAS improved the tolerance of and compliance with hand antisepsis
protocols, as evaluated for 4 and 14 months, respectively.
This study involved an unselected population of patients undergoing
routine surgery in teaching and nonteaching hospitals. The baseline SSI rates
for abdominal surgery procedures only, with routine 3-month per year surveillance
methods were 45 of 912 (4.9%) on average during 1997-2000.13
The SSI rates we observed in this study are consistent with those reported
elsewhere.18,19 In a cohort of
59 352 patients, Haley et al18 found an
SSI rate of 2.9% and 3.9% in clean and clean-contaminated class of procedures,
respectively. Olson and Lee19 reported an infection
rate of 1.4% in clean and 2.8% in clean-contaminated procedures in 40 915
patients. However, we observed a very low proportion of postdischarge SSIs
(8.3%) compared with recent studies. The percentages of SSIs occurring after
hospital discharge vary from 13.6% to 84.0% according to studies20- 22;
thus, we cannot exclude that possible SSIs were not validated in the absence
of evidence, although they were real.
Any direct comparison between reported studies would be hazardous, as
they differ in several respects, such as the SSI rate surveillance methods,
the study period, and the characteristics of the study population. In our
study, differences in the characteristics of the patients and surgical personnel
in the 2 protocols were minimized by the randomized service crossover experimental
Infection control epidemiology has clearly demonstrated that bacteria
responsible for SSI can be shed from the surgical team's hands, despite standard
Alcohol-based hand disinfection has previously been shown to reduce nosocomial
infection rates and to improve compliance with hand hygiene rules when implemented
throughout a hospital, particularly at the bedside in medical wards,26 although other studies are less favorable.27,28 An 8-month prospective study conducted
in 3 intensive care units showed that the rate of nosocomial infections was
significantly higher after alcohol-based vs scrub-based hand hygiene, possibly
due to poor compliance with hand-rubbing instructions.28
This underscores the need to evaluate new protocols in the routine context.
As previously reported in intensive care unit patients,28
the rate of SSIs was higher when nonantiseptic hand wash was not performed
prior to use of AAS. Although the number of observations is too low to perform
any test, it underscores the importance of the optional simple nonantiseptic
hand wash in the hand-rubbing protocol. However, we cannot exclude that surgical
personnel did not perform previous nonantiseptic hand wash before hand-rubbing
in the case of an emergency procedure with a higher risk for SSI (confounding
We chose each surgical service for initial randomization of the antisepsis
protocol. This allowed us to evaluate the compliance and tolerance of each
protocol during 1-month periods. In addition, the alternative antiseptic was
removed from the surgical service at each crossover, as contamination between
protocols of a study increases the chances of declaring equivalence in the
As reported in previous studies, AAS was, on average, better tolerated
than traditional hand-scrubbings by the surgery team.9,30,31
Less reported are issues regarding compliance of the operating department
personnel to hand antisepsis protocols. For presurgery hand antisepsis, the
critical end point for compliance is not the occurrence of the protocol but
how well and how long it is performed. We observed significantly better compliance
with the duration of hand hygiene in the AAS hand-rubbing group, and this
effect persisted throughout the 14-month evaluation. One possible explanation
is that the necessary duration of hand-rubbing in the AAS-based protocol depends
on the amount of AAS applied, whereas the hand-scrubbing protocol can be foreshortened
by drying the hands with absorbent material. Contrary to studies performed
in nonsurgical settings,26,30
compliance was also improved by the use of AAS in the surgeon or assistant
subgroup; this subgroup had already been identified32
as complying poorly with hand-hygiene regimens. The fact that the rate of
SSI is equivalent in both groups, despite a better compliance in the hand-rubbing
group is not surprising given the low SSI occurrence rate.
A previous French study33 has compared
the costs of the 2 techniques for hand and forearm antisepsis before scheduled
orthopedic surgery. They reported a relative cost of 203 euros per week when
using povidone iodine or chlorhexidine gluconate and 25 euros per week when
using AAS (1 euro-US $1.03). Nonetheless, AAS may not be accepted or tolerated
by all surgical personnel. Thus, we do not believe that AAS should systematically
replace surgical hand antisepsis with traditional hand-scrubbings for economic
reasons. The choice of the technique for hand and forearm antisepsis before
surgery should remain a matter of personal preference among users. In conclusion,
given its equivalence to standard hand-scrubbing in preventing SSI, we consider
that preoperative hand-rubbing with AAS preceded by a nonantiseptic hand wash
is a safe alternative.