Dr Andrew Kemp PhD (Bio), Dr Vanessa Hodgkinson MBBS, BSc, FRCA, FFPMRCA and Miss A Bugg MSc (Microbiology) explore the comparative effect on skin bacterial counts of three surgical skin preparations.
Published in the Operating Theatre supplement of the CSJ August 2018
Surgical site infections (SSIs) occur in approximately 2% – 5% of patients who undergo clean extra-abdominal surgeries, such as thoracic and orthopaedic surgery, and in up to 20% of patients who undergo intraabdominal surgery interventions.[1, 51]
SSIs can lead to increased morbidity and mortality and are associated with prolonged hospital stay and greater hospital costs.
[1, 39, 46, 51]
The Institute for Healthcare Improvement reports that SSIs in the United States increase the length of hospital stay by an average of 7.5 days, at an estimated cost of $130 million to $845 m per year.2 In 2006, SSIs accounted for 14% of healthcareassociated infections in the United Kingdom, resulting in additional costs of between £814 and £6626, depending on severity.
Microbial contamination of the surgical site is a possible factor in the development of an SSI, prevention techniques aim to minimise the presence and spread of microorganisms.
Prevention strategies include antibiotic prophylaxis, antiseptic prophylaxis, hair removal, perioperative glucose control, and maintenance of normothermia.[4, 51] Topical antiseptics may be applied to the skin preoperatively to reduce SSI risk.
The main types of skin antiseptics currently in use are, iodine or iodophor (such as povidone-iodine [PI]), ethyl alcohol, and chlorhexidine gluconate. CHG and PI can be mixed with either alcohol or water, which may have implications for their effectiveness.
Clinical practice guidelines from the National Institute for Health and Clinical Excellence (NICE) recommend that patients shower or bathe with soap the day before – or the day of – surgery, and that iodophor-impregnated surgical drapes be used when incise drapes are required.
They also recommend preparing the skin at the surgical site with antiseptic immediately before incision, but they do not indicate a preference for CHG or PI. As there is a significant difference in the colour of the liquid skin antiseptic, it is impossible to blind the selection of liquid to the surgical team, ward team and research team.
The ‘observer effect’ or ‘Hawthorn effect’, where the act of measuring alters the outcome, is impossible to measure in studies looking at the link between skin disinfectant solutions and patient outcome.
The significance of skin bacteria around surgical wounds, and its direct/ indirect effect as a cause of SSIs have therefore been well documented and recommendations have been made by agencies, such as the United States Centre for Disease Control (USCDC) and NICE in the UK.[39, 40]
However, in recent reviews of the current available literature, the value of the studies used by these organisations has been brought into question.[7,38,49]
The authors’ conclusions in both reviews were similar: “The methodological quality of the studies was mixed. Evidence was drawn from Randomised Controlled Trials (RCTs) and Non-Randomised Trials (NRTs), although the method of randomisation was generally poorly reported. Efforts were made to blind outcome assessors, but patients and surgeons often were not blinded, compromising internal validity.”
They therefore concluded that: “There is insufficient evidence to support one antiseptic over another.”[7,38]
The problems with RCTs in respect to patient outcomes are obvious. It is impossible to isolate the one variable to be tested amongst the many variables that are known to cause SSIs.
There are several reasons why academically the studies may be relevant, but in practice they can do no more than point the reader in the
right direction for clinical practice.
The most important factors to take into consideration when looking at infection control related outcome studies are the extraneous variables that involve the patient. A simple example of this would be to assess the individual patient’s immune system, as an indicator of their potential to fight off potential infections. If these variables are not isolated, when asked for ‘evidence’ in reference to patient outcome, infection control studies should therefore be considered, at best, weak.
The clinical effectiveness of pre-surgical antiseptic showers have been reported in three Randomised Controlled Trials (RCTs) and four cohort studies.[10,16] The clinical effectiveness of antiseptic preparation versus non medicated soap and alcohol or saline was reported in two RCTs published in 2001 and 2005.[17,18] The effectiveness of one antiseptic preparation versus another for reducing bacterial
colonisation and SSIs has been reported in five RCTs and four non-randomised studies published between 2002 and 2011.[19,27] Skin colonisation levels were reported by seven RCTs [11,13,19,21,28] and one observational study.
Studies considered all bacterial species when determining colonisation. Two RCTs reported that pre surgical showering with PI or CHG resulted in a statistically significant reduction in preoperative skin colonisation. One RCT  showed no statistically significant reduction in preoperative colony counts with CHG showering compared with control or placebo.
In a cohort study, twice-daily, five-day topical 4% CHG scrubbing reduced preoperative perineal colonisation four-fold compared with usual hygiene in patients undergoing artificial urinary sphincter placement. One RCT showed a statistically significant reduction in pre surgical bacterial colonisation in patients prepared with 4% CHG in 70% isopropyl alcohol
compared with patients prepared with PI.
Until recently the effects over extended time periods, of ethyl alcohol on its own on skin bacterial counts have also been overlooked. Several “real time” audits of skin flora now show that alcohol, when used on its own, may actually be detrimental to bacterial levels on the skin. Two RCTs [23,28] and one retrospective study  published between 2002 and 2005 focused on the use of iodophor-impregnated incise drapes to prevent surgical wound infection.
The evidence available, although inconclusive, suggests that preoperative showers with an antiseptic agent are effective at reducing bacterial colonisation of the skin and that could be used to reduce SSIs. CHG was primarily used as the ‘antiseptic’ with varying showering regimens and varying compliance rates in those trials.
As the results do remain inconclusive, it is unusual then for section 8A.1. of the USCDC 2017 guidance on prevention of SSIs to state: “Advise patients to shower or bathe (full body) with soap (antimicrobial or non-antimicrobial) or an antiseptic agent on at least the night before the operative day.” This is given a “category 1B – strong recommendation” based on “accepted practice.”
If there is no appropriate RCT evidence, then the authors must clearly be assuming that the antiseptic shower will reduce skin CFU counts, and therefore less bacteria – including skin commensals – at the time when surgical skin preparation is undertaken, will have a positive effect on potential outcome.
This should not be a surprise when it is known that the skin commensal Staphylococcus Epidermidis (which is the most prolific skin bacterial species), has the capacity to evade the human immune system, as well as being resistant to Oxacillin. The dangers
associated with this bacterial species can therefore not be overestimated in a surgical wound.
As the USCDC takes the view that the lower the skin bacterial count the better potential outcome, a study that looks only at bacterial skin counts has to give a good indication of a serious patient risk factor.
Disinfectant products are often mixed with ethyl alcohol, which makes it difficult to form overall conclusions about another active ingredient during any time periods when it can be assumed the alcohol is still effective.
More large, well-conducted RCTs with consistent protocols are still unlikely to provide the evidence we are searching for in respect to the effectiveness of one antiseptic preparation over another where patient outcome is ultimate measure. This is again due to the many other factors affecting outcome that cannot be controlled when undertaking such a trial.
Antiseptic surgical skin preparation
There can be no doubt that antiseptic surgical skin preparation is important as a preventative step in preventing SSIs. The question that is unlikely to ever be answered is: “How much effect does it really have?”
It was this last question, and its inability to be answered, that caused the authors to undertake this study. As the study published in the New England journal which looked at the efficacy of CHG 2% versus PI (42), showed a significant reduction in SSI in the CHG group, the authors felt that this study would form the basis for our own study.
In light of the recent revelations about the accuracy of some reporting on patient outcomes in studies on
CHG, the authors feel that potential inaccuracies did not affect the outcomes of this research.
According to the New England journal study, CHG 2% had a significantly positive effect on SSIs rates and therefore on patient outcome. The study reported that choice of antimicrobial agent and its concentration in solution are the two main factors affecting the number of bacteria on patient skin, and that this is linked to the patient outcome.[40, 41, 43, 44, 45, 46, 47]
Unfortunately, that study did not compare CHG 2% with CHG 0.5% which is the concentration recommended by the CDC and others.
A factor not discussed in the study is the potential that different areas of the body will have different pre-operative bacterial counts, which may also have been affected by pre-operative washing and/ or the type of surgical skin prep.
When comparing the number of SSIs among patients prepared with DuraPrep to those prepared with PI solution in combination with iodophor-impregnated drapes, one study reported no SSIs in either group.
Similarly, one study found no statistically significant difference in the number of SSIs among patients prepared with a one-step iodophorand- alcohol water-insoluble film with or without iodophor-impregnated drapes.
In contrast, one study found statistically significant lower rates of wound infection with the use of Ioban drapes compared with surgeries without drape use (12.1% without drapes vs 3.1% with drapes).
With so much conflicting evidence from patient outcome studies, for the moment at least we should consider the only practical method of determining the effectiveness of pre surgical preparations, is to measure and compare the skin bacterial counts prior to, during and for many hours after application.[40, 41, 43, 44, 45, 46, 47] We should use patient outcome studies as a further validation of the results – and not the other way around.
The authors chose the commonly used area of the longitudinal mid line incision of the abdomen as the area to sample. This area has also been used for many studies looking at levels of bacterial contamination.[5,6,7,8,9,13,14] In addition, and in light of the recent evidence that shows the adverse effect of alcohol with no other active agent skin flora over time, and as some clinicians still use 70% ethyl alcohol as a surgical skin preparation, this was included in the study.
As the comparisons between CHG 2%, CHG 0.5% and PI have already been studied sufficiently to show that 2% Chlorhexidine is most effective and killing skin surface bacteria, in the short term, it was decided that the most relevant comparison would be to study the effect of CHG 2% versus 70% ethyl alcohol on its own and a newly released 5th generation SiQuat with a low ethyl
The authors considered the most relevant times for sampling of bacteria counts during a surgical procedure to be after five minutes, one hour post, two hours post and four hours post. The reason for these times is that aside from the emergency operating theatre, it is highly unlikely that if surgical skin prep is used correctly, “knife to skin” will take place within five minutes of the skin prep being applied. This means that the most relevant time to test initial effectiveness is as the blade cuts through the skin at around five minutes post application.
As the surgical procedure continues it is important to look at the effect of the operating theatre environment, including ‘fallout’ of bacteria onto the skin around the wound edges. It should be noted that this is of particular importance when using the PI compounds as they are de activated by human plasma proteins. Any haemoserous fluid leaking from the wound will negate any initial antimicrobial effect of the solution. Finally, the effect on skin bacterial counts after dressings have been applied with wound edges not closed may also have an impact on outcome.
It was therefore decided that for reduced numbers of skin bacteria to have an effect on SSI outcome, the chosen times would be the most relevant times to test. It is worthy of note that although two hours post application is not necessarily a critical time in respect to wound infection rates, it was chosen to determine if there could be a predictable curve for any increase in bacterial counts.
As already described, there is no reliable evidence to support pre-test showering with an antiseptic, and as this was not standard procedure in the hospital, the subject was not discussed with patients and they were left to shower/ wash as they would normally do.
A total of 300 samples were taken from areas of abdominal skin 2 cm x 1 cm along the mid line running superior to inferior using a sterile technique. These areas were marked with a surgical skin marker to ensure that further samples were taken from the same areas. A sample from every area was taken prior to any skin prep being used and a surface count recorded using a Bacteria Specific Rapid Metabolic Assay (BSRMA) as this has been proven to be a far more reliable and accurate way of counting CFU’s than either culture or Polymerase Chain Reaction (PCR).[52,53]
All areas were prepared for surgery using one of the chosen skin antimicrobial preparations. Due to colour variations in the liquid preparations it was impossible to blind the study. Post application of the surgical skin preparations, swab counts were again taken from each site after five minutes, one hour, two hours and four hours. The graph below shows the average skin bacterial counts for the three types of surgical skin prep.
All CFU counts were averaged from the group counts. It is noteworthy that in every group – and at every time differential – the results of 98% of the samples were within 5% of the mean average count. CHG 2% is effective until around four hours postoperatively, when CFU counts begin to rise again. After an initial 90% reduction in CFU counts at five minutes, the count continues to fall to a low of 91.5% reduction at two hours, before increasing again to a 63% reduction at four hours.
70% ethyl alcohol liquid showed an initial reduction in CFU counts of 49% at five minutes. However, between five minutes and one hour postapplication the CFU counts increased to a level 18% above the pre-treatment counts. This decreased to 5% above pre-treatment levels at two hours post-application and to 3% above the pre-treatment counts, at four hours.
The 5th generation SiQuat demonstrated the most effective microbial reduction at each of the time periods after application. After an initial 97.5% reduction in CFU counts, it continues to fall to 99.5% reduction at four hours.
Both the 5th generation SiQuat and 2% CHG, showed a continuous and prolonged drop in CFU counts from the pre-application levels to four hours after application.
These results demonstrate that 70% ethyl alcohol on its own has a significantly lower efficacy than either of the other two solutions which both contain alcohol.
For surgical procedures of up to two hours in duration, both 2% CHG and the 5th generation SiQuat continued to show good antimicrobial activity on the skin around the surgical wound. At four hours post application it is clear the 2% CHG is losing some of its antimicrobial activity, it may therefore be considered prudent to reapply this solution at around two hours after the original application.
The surgical skin preparation that is shown to be most effective at every test point from application over the course of four hours is the 5th generation SiQuat. Reapplication would not be necessary with this type of skin prep solution.
Further investigation into the reasons why there should be such a significant increase in skin bacterial counts when alcohol is used, has now taken place, the results of which will be published in due course. It is evident that 70% ethyl alcohol needs to be combined with another antimicrobial agent before being considered for use as a surgical skin preparation.
It is an undisputed fact that bacteria at the skin edges during surgery pose a threat to any surgical wound. It is
therefore common sense that the most effective surgical skin preparations will not only kill bacteria effectively at the start of an operation but during, and for as long as possible afterwards.
These results identify the need for studies that use values as well as outcomes to be taken into consideration when determining the most effective products to use as a surgical skin preparation. A much more difficult question to answer is: “What evidence should clinicians be asking for to base decisions about product efficacy in relation to infection rates post surgery?”
It is undeniable that prevention is always better than cure. In the case of surgical site infections, it is also a fact that, due to the continuing rise in the numbers and type of antibiotic resistant strains of bacteria, attaining a ‘cure’ becomes even more difficult to achieve. It is therefore common sense that any – and all – interventions that can reduce the numbers of bacteria that could potentially cause SSI need to be considered and assessed for efficacy.
All surgical procedures differ in duration and the technique for skin preparation differs from clinician to clinician. Further studies may be indicated to be able to determine the most appropriate surgical skin prep based on individual surgical specialities, especially for the highest risk patients and procedures.
Current standard tests for all disinfectants including skin sanitisers are based only on short term antimicrobial activity and effect. In Europe, these tests include EN1500, EN12791, and in the US the ASTM E2755-15 is used, amongst others. These standard tests are all complete within one to five minutes of microbial exposure to the surface chemical agents. They were developed many years ago to test the efficacy of short term disinfectants and sanitisers as there were no long acting chemicals available at that time.
As it’s clear now that persistent disinfectants have a positive effect on reducing bacterial counts over time, these tests are no longer fit for purpose and new tests methods need to be developed. As the two main bodies referred to for guidance on hand cleaning, is it now time for the USCDC and the WHO – as well as the wider infection prevention community to ask themselves the following questions:
Are the current short term standard tests sufficient to test the true effectiveness of antimicrobial activity on skin, or should there be a longer lasting test standard reflecting the risks over time?
Should the CDC, the regulatory bodies and the numerous organisations whose membership are responsible for surgical outcomes, now include long term disinfection as part of their guidance on prevention of surgical infections?
References for this article are available on request.
Published in the Operating Theatre supplement of the CSJ August 2018