In Reply:-We appreciate the interest of the Purdue Frederick Company in our study and thank Dr. Welch for her comments. We agree with Dr. Welch that a great number of variables can produce contamination of PI solution and bottle caps. Indeed, this was the basic premise of our article, [1]which demonstrated that multiple-use povidone-iodine (PI) bottles in normal use can become contaminated with bacteria. Because contamination of PI solution has previously been reported and because PI is a widely used disinfectant for skin preparation before initiation of epidural anesthesia, we undertook our study to assess the frequency with which bacterial contamination occurs, rather than to identify possible sources of contamination.

As noted by Dr. Welch, in our estimate of the prevalence of contamination (40% of bottles), we did not distinguish between microorganisms isolated from the inside of the bottle cap and those isolated from the solution itself because we considered both to be potential sources of patient infection. Unless the cap is completely removed, the PI solution must come into contact with the inside of the cap when the solution is being dispensed. Finding bacteria on the inside surface of the bottle cap is disturbing for two reasons. First, that the presence of bacteria on the inside of the cap offers the potential to introduce organisms into the PI solution, and eventually the patient's skin. Second, one would expect that during previous contact of the cap with the PI solution, these organisms would have been eradicated. Even if results for the four contaminated bottle caps (three contaminated with Staphylococcus epidermidis, one with Stenotrophomonas (xanthomonas/pseudomonas) are considered separately from the four contaminated PI solutions (two contaminated with Staphylococcus haemolyticus, one with Staphylococcus epidermidis, and one with Bacillus) the rate of contamination (10% of bottles in use) is still disturbingly high. Although Dr. Welch suggests that there is no adequate explanation for the presence of Bacillus in the setting we describe, Bacillus species are recognized members of the normal flora and can be found in specimens from many body areas, including the skin and respiratory, gastrointestinal, and genitourinary tracts. A brief observation of a woman in the midst of childbirth clearly illustrates just how easily contamination of the lower back can occur.

Dr. Welch states that our study did not demonstrate that microorganisms on the caps contaminate the PI solution. We remind her that the organisms we describe were isolated from the inside of the caps in question, and not the external surfaces. We contend that the potential for contamination is real. Slime-producing coagulase-negative staphylococci are well suited to growth on a plastic cap surface and could easily contaminate the solution. The potential for this risk was supported by our finding of a case in which Staphylococcus haemolyticus was isolated from the patient's back and the bottle cap.

Although Dr. Welch states that data on file at the Purdue Frederick Company demonstrates that the organisms that we isolated will not survive in PI solution for more than 15-30 s, our findings and those of several other researchers suggest otherwise. Certainly, the manufacturers of Betadine (Purdue Frederick Company) should be aware of the previous reports of contamination of PI solution with Pseudomonas species, [3]including the report of the Morbidity and Mortality Weekly Reports that described contamination of unopened bottles of PI solution that necessitated a voluntary recall. [4]Despite Dr. Welch's assurances that PI solution cannot support bacterial growth, there are numerous reports that trace clinical sequelae to contaminated antiseptic solutions. [5] 

Dr. Welch indicates that there is confusion in our study between microbial contamination and support of growth. We agree that direct inoculation of bacteria into PI solution is a more definitive way to establish whether bacterial multiplication has occurred or whether the organisms are merely tolerant. We are in the process of just such an analysis. However, regardless of whether the bacteria we isolated from PI solution were undergoing active replication, the organisms isolated in our study were clearly viable and were able to multiply once plated. There are no data to indicate that bacterial growth would not occur if a patient's skin or other tissues were similarly inoculated.

Dr. Welch implies that bacterial contamination of the swabs is from environmental sources other than the bottle of PI solution used. She states that “the authors apparently did not test the new sponge sticks used to sample the patients' backs to see whether they may have already been contaminated.” This is not the case. We tested 20 sponges (10 were plated right out of the sterile kit, and 10 were transported to the microbiology laboratory to verify the integrity of our sterile transport system.) All 20 were found to be sterile.

Dr. Welch suggests that our findings are not clinically significant, as demonstrated by the paucity of infectious sequelae after skin disinfection with PI solution. Again, we cannot disagree more strongly. There are a growing number of case reports describing infection after the use of neuraxial analgesia. Optimum skin disinfection is not the only prevention, but it is a key step in decreasing the risk of infection associated with these techniques. Because many patients have epidural catheters that remain in situ for long periods of time, the initial disinfection becomes even more critical.

We agree totally that multiple-use bottles should be handled carefully. However, our results demonstrated that a significant number of multiple-use PI bottles become contaminated in normal use. We do not believe that this experience is limited to our hospital.

Single-use packets of PI solution are very inexpensive and convenient. Our findings suggest they may also be more effective than solution from multiple-use bottles for skin disinfection and eliminate concerns regarding possible contamination of multiple-use containers. We therefore recommend single-use preparations when effective skin disinfection is critical.

David J. Birnbach, M.D.

Deborah J. Stein, M.D.

Odessa Murray, M.T.

Daniel M. Thys, M.D.

Emilia M. Sordillo, M.D., Ph.D.

St. Luke's-Roosevelt Hospital Center; College of Physicians and Surgeons of Columbia University; New York, NY;

(Accepted for publication August 31, 1998.)

Birnbach DJ, Stein DJ, Murray O, Thys DM, Sordillo EM. Povidone iodine and skin disinfection before initiation of epidural anesthesia. Anesthesiology 1998; 88:668-72
Berkelman RL, Lewin S, Allen JR, et al. Pseudobacteremia attributed to contamination of povidone-iodine with Psedomonas cepacia. Ann Intern Med 1981; 95:32-6
Craven DE, Moody B, Connolly MG, et al. Pseudobacteremia caused by povidone-iodine solution contaminated with Pseudomonas cepacia. N Eng J Med 1981; 305:621-3
Contaminated povidone iodine solution-Texas. MMWR Morb Mortal Wkly Rep 1989; 38:133-4
Parrott PL, Terry PM, Whitworth EN, et al. Pseudomonas aeruginosa peritonitis associated with contaminated poloxamer-iodine solution. Lancet 1982; 2:683-5