Between September 2007 and February 2008, 150 consecutive patients undergoing shoulder surgery were enrolled in the present prospective randomized study. All procedures were performed at one institution by one of three surgeons (J.L.K., G.W.N., and S.M.G.). One hundred and thirty-seven of the 150 procedures were done entirely arthroscopically. Four of the patients had shoulder arthroplasties, none of which were revision in nature. Institutional review board approval was obtained, and every patient gave informed consent to participate in the study. This study is registered at ClinicalTrials.gov under registry number NCT00829023. Patients were excluded if they had an open wound or a current infection or were chronically immunosuppressed. The study included eighty-four men and sixty-six women, and the patients ranged from seventeen to seventy-nine years of age.
Patients
Overall, the patients were relatively healthy. In the povidone-iodine group, three patients were heavy smokers (>1.5 packs per day), three patients had diabetes mellitus, none had rheumatoid arthritis, and one had a history of alcoholism or hepatitis. In the DuraPrep group, one patient was a heavy smoker, four had diabetes mellitus, none had rheumatoid arthritis, and four had a history of alcoholism or hepatitis. In the ChloraPrep group, none of the patients were heavy smokers, six had diabetes mellitus, two had rheumatoid arthritis, and two had a history of alcoholism or hepatitis.
The patients were instructed to shower the day before surgery. Otherwise, no specific home cleansing protocol was undertaken prior to surgery. Thirty-seven (25%) of 150 patients reported voluntary shaving of the axillary hair prior to enrollment in the study. All of these patients were women; twelve were in the povidone-iodine group, nine were in the ChloraPrep group, and sixteen were in the DuraPrep group.
Preoperative antibiotics were administered to all 150 patients enrolled in the study. Patients undergoing arthroscopic or open soft-tissue procedures received 2 g of cefazolin prior to surgery unless they had an allergy to penicillin, in which case they received 900 mg of clindamycin. Patients who had prosthetic components implanted received the same protocol with the addition of 1 g of vancomycin. Antibacterial-impregnated barriers (Ioban; 3M Healthcare, St. Paul, Minnesota) were used only for patients in whom prosthetic implants were placed. In each of these patients, all cultures were obtained prior to application of the antibacterial-impregnated barrier.
Skin Preparation Methods and Culture Analysis
Each shoulder was prepared with one of three commonly used surgical skin-preparation agents: ChloraPrep (2% chlorhexidine gluconate and 70% isopropyl alcohol; Enturia, El Paso, Texas), DuraPrep (0.7% iodophor and 74% isopropyl alcohol; 3M Healthcare), or povidone-iodine scrub and paint (0.75% iodine scrub and 1.0% iodine paint; Tyco Healthcare Group, Mansfield, Massachusetts). The agent used for each patient was chosen by opening a sealed, randomly assigned envelope that indicated the agent to be used. Each shoulder was prepared according to the manufacturer's instructions by the attending surgeon. Fifty patients were included in each of the three surgical preparation groups.
For the first twenty patients, culture specimens were obtained, both before and after the surgical preparation agent was applied, from three different sites: the anterior aspect of the shoulder (1 cm lateral to the tip of the coracoid process), the posterior aspect of the shoulder (2 cm distal to and 1 cm medial to the posterolateral border of the acromion), and the axilla. The purpose of obtaining cultures before skin preparation in this group was to define the native bacteria present around the shoulder. For the next 130 patients, culture specimens were obtained, after the surgical preparation agent was applied, from two sites: a combined anterior-posterior site and the axilla. These sites were chosen to represent sites of common surgical wounds. All cultures were obtained with dry, sterile, cotton-tipped swabs by the attending surgeon. The culture swabs were sealed, placed in a sterile container, and immediately transported to the microbiology laboratory at our institution for aerobic and anaerobic cultures. Cultures were monitored for seven days since Propionibacterium acnes requires an average of 5.1 days before first growth is detected8.
Statistical Analysis
Sample-size requirements were based on the findings of a prospective study evaluating positive cultures from the foot and ankle following surgical skin preparation9. On the basis of the assumption that a 30% difference in positive culture rates would be clinically significant, the number of patients required to achieve 80% power at alpha = 0.05 was forty-five patients per group. Descriptive statistics were calculated for all variables of interest. Continuous measures were summarized with use of means and standard deviations, whereas categorical measures were summarized with use of counts and percentages. Generalized linear models with pairwise comparisons were run to assess differences in culture rates between sites (anterior-posterior and axillary) and among treatments (ChloraPrep, DuraPrep, and povidone-iodine). Chi-square analyses were used to evaluate other associations such as the relationship between the presence of axillary hair and positive cultures. Significance was defined as p < 0.05.
Source of Funding
External funding of this study was obtained from Enturia, the company that manufactures ChloraPrep. This funding was used exclusively for microbiology expenses and the company was not involved with the organization or analysis of the data.
Prior to surgical skin preparation, bacteria grew on culture from 95% of the anterior shoulder sites, 90% of the posterior shoulder sites, and 70% of the axillary sites. The most common organism isolated overall was coagulase-negative Staphylococcus (forty isolates) followed by Propionibacterium acnes (twenty-eight isolates), Corynebacterium (fourteen isolates), and Micrococcus (seven isolates) (Fig. 1). On the average, 2.1 different organisms grew from each culture site prior to surgical skin preparation.
The overall rate of positive cultures following skin preparation was 31% in the povidone-iodine group, 19% in the DuraPrep group, and 7% in the ChloraPrep group. The positive culture rate in the ChloraPrep group was lower than that in the povidone-iodine group (p < 0.0001) and the DuraPrep group (p = 0.01). The positive culture rate in the DuraPrep group was lower than that in the povidone-iodine group (p = 0.05) (Table I).
The positive culture rate for coagulase-negative Staphylococcus was 19% in the povidone-iodine group, 4% in the DuraPrep group, and 2% in the ChloraPrep group. Both DuraPrep and ChloraPrep were more effective than povidone-iodine (p < 0.001 for both), and there was no difference between DuraPrep and ChloraPrep for this organism (p = 0.41). The positive culture rate for Propionibacterium acnes was 15% in the povidone-iodine group, 12% in the DuraPrep group, and 7% in the ChloraPrep group. There was no significant difference with regard to the positive culture rate for the DuraPrep group compared with the povidone-iodine group (p = 0.53) or the ChloraPrep group (p = 0.23). There was also no difference with regard to the positive culture rate for ChloraPrep compared with povidone-iodine (p = 0.07) for this organism. The positive culture rates for organisms other than coagulase-negative Staphylococcus and Propionibacterium acnes were too small to permit meaningful statistical analysis (Table I).
In the povidone-iodine group, bacteria grew on culture of specimens obtained from 22% of the anterior-posterior sites and 40% of the axillary sites. In the DuraPrep group, bacteria grew on culture of specimens obtained from 26% of the anterior-posterior sites and 12% of the axillary sites. In the ChloraPrep group, bacteria grew on culture of specimens obtained from 10% of the anterior-posterior sites and 4% of the axillary sites. With regard to the anterior-posterior site, the positive culture rate in the ChloraPrep group was lower than that in the DuraPrep group (p = 0.05). Neither ChloraPrep nor DuraPrep differed significantly from povidone-iodine (p = 0.12 and p = 0.64, respectively). With regard to the axillary site, both the ChloraPrep and DuraPrep groups had significantly lower rates of positive cultures than did the povidone-iodine group (p = 0.001 and p = 0.004, respectively). ChloraPrep did not differ significantly from DuraPrep (p = 0.16) (Table II).
Following surgical skin preparation, bacteria were isolated from 19.3% of the anterior-posterior sites compared with 18.6% of the axillary sites. Propionibacterium acnes was isolated from 16.7% of the anterior-posterior sites (twenty-five of 150 isolates) and 6% of the axillary sites (nine of 150 isolates). Coagulase-negative Staphylococcus was isolated from 2% of the anterior-posterior sites (three of 150 isolates) and 15.3% of the axillary sites (twenty-three of 150 isolates). Propionibacterium acnes was cultured at a higher rate at the anterior-posterior site than the axillary site following surgical skin preparation (p = 0.004).
Thirty-seven (25%) of 150 patients reported that they had shaved the axillary hair prior to surgery. Of the patients who shaved the axillary hair, the rate of positive cultures after skin preparation was 27%. Of those who had not shaved the axillary hair, the rate of positive cultures after skin preparation was 35%. No significant difference between the groups was detected with respect to the rates of positive cultures after skin preparation (p = 0.40).
No postoperative infection had developed in any of the patients in this study at a minimum follow-up of ten months after surgery.
Note: The authors thank Alex Kiss, PhD, and acknowledge Eugene Lautenschlager (deceased) for their assistance with the statistical analysis and study design.