In 2010, there was a very interesting study published in the New England Journal Of Medicine comparing Chlorhexidine-Alcohol with Povidone-Iodine for surgical site infections. The background to this study is that the standard skin preparation for surgery in most of North America has been Povidone-Iodine for many years. This is the brown antiseptic solution which leaves a yellow hue on the skin after surgery. A few years back, another solution, Chlorhexidine-Alcohol, began receiving increased attention because there appeared to be fewer infections when it was used as the skin prep for placing central lines.
Since Chlorhexidine appeared to decrease line infections for central lines, it was surmised that perhaps it would lead to an overall decrease in surgical site infections in other surgical procedures, such as orthopaedic, breast, hernia and other surgeries. There was one small obstacle, however. The incidence of wound infections after most clean surgery (eg, breast, hernia, hip, knee, shoulder etc) is quite low, in the neighbourhood of 1-2%. Because of this, it would take an enormous number of patients in a trial to have the power to demonstrate a difference in infection rates. The authors, therefore, decided to choose surgical procedures with a higher incidence of wound infection, commonly called clean contaminated surgery, where the infection rate is upwards of 10%. See the table below. The numbers needed to show a difference would be substantially less. Clean contaminated surgery includes any surgery that involves opening of the bowel or transgression of non sterile body organ or cavity, such as in gynecology, thoracic surgery and urology. But the classic example of clean-contaminated surgery is a bowel resection, either small or large bowel.
The results are quite interesting. And are presented in the tables below. You can see here that there was a statistically significant decrease in surgical site infections with Chlorhexidine 9.5 % vs 16.1%. While the absolute numbers are quite small for deep incisional infections, you can see here that there are also significant findings. When you look specifically at abdominal/intestinal surgery, you can see the numbers show similar findings. For colorectal, biliary, small intestinal and gastroesophageal surgery the risk of surgical site infection is 15.1% vs 22% , 4.6 vs 9.3, 9.8 vs 29.4, 11.5 vs 20.7 respectively all in favour of chlorhexidine as a skin prep. The Authors conclude that “Preoperative cleansing of the patient’s skin with chlorhexidine-alcoholo is superior to cleansing with povidone-iodine for preventing surgical-site infection after clean contaminated surgery.”
The authors, explained their finding based on a number of factors related to Chlorhexidine. More rapid onset of action, Persistent activity of chlorhexidine despite exposure to bodily fluids, and Residual effect on the skin. It seemed to make sense.
It seems to make sense on face value. That is… um… until you think about it a little more deeply. You will recall that the authors chose clean contaminated surgical procedures(i.e., bowel surgery etc) as opposed to clean surgery because the infection rate with clean surgery is so low. The source of infection in clean surgery is the skin, whereas the source of most infections in clean contaminated surgery is the unsterile organ (eg, colon or small bowel). Hence the 2% vs 10% difference in surgical site infections. Now, if you think about it, you might wonder why the skin prep should make a difference in clean contaminated surgery. The higher infection rate in clean contaminated surgery is not because of skin flora, but rather because of intestinal contamination. And it makes no sense for the skin prep to decrease the incidence of surgical site infections in clean contaminated surgery so dramatically. It’s like painting your car cool colours and noting an increase in the horsepower. It just doesn’t make sense.
I’ve discussed this study with numerous colleagues, almost all of whom concede that there is something a little peculiar about these results. But they attribute these irregularities to minor flaws in the design of the study and normal variations. The authors too acknowledge that something is a bit peculiar as well when the write in the final paragraph of the paper that the “overall superior protection afforded by chlorhexidine-alcohol was attributed primarily to a reduction in the rates of superficial and deep incisional infections that were caused mostly by gram-positive skin flora”. But this just doesn’t explain why a skin prep with Chlorhexidine should have decreased the wound infection rate so dramatically for clean- contaminated surgery.
Now, this is not to imply any sinister motives, but it is interesting to note that Chlorhexidine and Povidone Iodine are both made by Cardinal Health and that 8 of the 12 authors have financial ties to Cardinal health. And despite it appearing in such a reputable journal, it is also a bit odd that it is not clear who actually funded this study.
As I said, most people assume the somewhat peculiar outcome is related to imperfect application of the methodology of the Randomized controlled trial. But I would ask you to entertain the notion that perhaps the irregularities in this study are not just a result of imperfect execution of the methodology, or just the vicissitudes of clinical research,… I would ask you to consider whether its possible that there is something inherently problematic about the methodology of an RCT itself which leads to this peculiar outcome?
OK. lets look at another example dealing with a device called the Fistula plug next