Three-Dose vs Extended-Course Clindamycin Prophylaxis for Free-Flap Reconstruction of the Head and Neck

Background  Twenty-four hours of perioperative antibiotics provides effective prophylaxis for most head and neck cancer resections. Many reconstructive surgeons have been hesitant to apply this standard to free-flap reconstruction of the head and neck. This prospective clinical trial compared short-course and long-course clindamycin prophylaxis for wound infection in patients with head and neck cancer undergoing free-flap reconstruction.

Methods  Seventy-four patients were randomized to receive short-course (3 doses) or long-course (15 doses) clindamycin perioperatively. Wound infections, fistulas, and other postoperative complications were documented by faculty surgeons who were blinded as to treatment group.

Results  The differences in wound infections and other complications were statistically insignificant. No other independent predictors of wound complications emerged in this series of patients.

Conclusions  Short-course clindamycin is as effective as long-course clindamycin in preventing wound infections after free-flap surgery for head and neck ablative defects.

ABLATIVE SURGERY for upper aerodigestive tract malignancy is designated "clean-contaminated." The need for perioperative antibiotic prophylaxis in this setting has been well established.^1,2 Without perioperative antibiotics, wound infection rates as high as 80% are reported.¹ For typical ablative procedures, the optimal duration of antibiotic prophylaxis has also been defined. Twenty-four hours of perioperative prophylaxis has proved comparable to 5 to 7 days of prophylaxis in preventing wound infections.^3,4

When tumor removal produces a defect requiring flap reconstruction, wound infection rates may be higher. Infection rates rise from approximately 10% for defects closed primarily to more than 20% for defects requiring flap reconstruction.^5-8 Wound infections pose a serious threat to free-flap reconstructions, sharply increasing the risk of vascular compromise and flap loss.^9,10 Free-flap loss can be devastating, and many reconstructive surgeons have been hesitant to adopt short-course (24-hour) antibiotic prophylaxis for this patient group. They argue that most of the studies evaluating prophylaxis for head and neck surgery have not been conducted in patients undergoing reconstruction with free flaps.

This study was designed to clearly address the duration of antibiotic prophylaxis in patients with head and neck cancer requiring free-flap reconstruction. Patients undergoing free-flap reconstruction of cancer defects were randomized to receive either short-course (3 doses) or long-course (15 doses) perioperative clindamycin.

This study was designed as a randomized, prospective, single-blind clinical trial. Patients included in the study were scheduled to undergo surgical ablation of head and neck malignancies with immediate free-flap reconstruction. Patients undergoing secondary reconstruction and those whose tumor did not involve the mucous membranes of the upper aerodigestive tract were excluded. The Institutional Review Board for Use of Human Subjects at the University of Alabama at Birmingham approved the trial, and informed consent was obtained from all participants. No subject experienced an adverse event as a direct result of study participation.

For the purpose of study design, we hypothesized that wound infection rates would be significantly lower in the long-course (15-dose) subgroup. Sample size was estimated at 110 patients on the basis of a projected 15% difference in infection rates between the 2 groups and statistical power of 80%. Study patients were enrolled between January 1, 1998, and April 30, 2001, and were randomized to 2 treatment groups as follows: group 1, short course (clindamycin, 900 mg intravenously every 8 hours for 3 doses); and group 2, long course (clindamycin, 900 mg intravenously every 8 hours for 15 doses).

Antibiotic therapy was initiated immediately preoperatively. Each wound was evaluated daily for 7 days (or until discharge) by a faculty head and neck surgeon who was blinded as to treatment group. Both the head and neck wound and the donor site were scored as follows: wound color: 1, normal; 2, pink; or 3, red or swollen; and drainage (in drains or at wound edges): 1, none; 2, serous; or 3, purulent.

A wound was considered infected if the color became red or the wound was swollen. A pink wound that developed purulent drainage was also considered infected. Other outcomes recorded included fistula formation, flap necrosis, vascular compromise of the flap, and development of distant, unrelated infections. Any complication requiring reoperation was recorded.

Descriptive statistics including mean, median, SD, and proportion were calculated to summarize patient characteristics. Infection rates were summarized for each group and compared with the χ² test or Fisher exact test for categorical variables and a 2-sample t test for continuous variables. All data were analyzed with SAS software (SAS Institute Inc, Cary, NC). An α level of .05 was deemed statistically significant for all tests.

A total of 74 patients were enrolled in the study. Accrual was stopped when annual review of the data disclosed that differences between the study groups were much lower than those projected in initial sample size determinations. On the basis of observed differences, we believed that additional patient accrual would be meaningless. At that time, wound infections had been detected in 4 patients in each treatment group. The wound infection rates were 4 (11%) of 35 and 4 (10%) of 39 in the short-course and long-course groups, respectively (P = .99). With a difference of 1% between the 2 treatment groups, the recalculated estimate of the sample size to demonstrate a 2-sided significance level of .005 with 80% power would require more than 10 000 subjects in each treatment arm.

Treatment group characteristics are detailed in Table 1. Thirty-five patients (47%) were assigned to group 1 (short-course clindamycin) and 39 (53%) to group 2 (long-course clindamycin). There were 28 women (38%) and 46 men (62%), with ages ranging from 21 to 88 years (mean, 62 years). Of the 74 patients, 67 (91%) had stage III or IV disease. The 4 flap types used in these patients were radial forearm, jejunum, rectus, and fibula, by descending order of frequency. Forty-six patients (62%) had no recorded complications. Complications of some type developed in 28 patients (38%). Wound infections were documented in 8 patients (11%) and pharyngocutaneous fistula in 6 patients (8%). Distant infections (including urinary tract infection and enterocolitis) developed in 8 patients (11%).

Table 1 also shows a univariate analysis of coexistent variables. T stage, tumor site, flap type, previous radiotherapy, and patient demographics were analyzed. None of these variables predicted complications in this group of patients (P>.05). In a review of our previous free-flap data (W.R.C., J.R.F., T.H., and G.E.P., unpublished data, 1998), both flap type and tumor site significantly predicted wound complications. In the current study, however, no variable was a significant predictor of wound complications, including antibiotic treatment group.

The most meaningful data from this study are presented in Table 2. The similarity in outcomes between the short-course and long-course groups is obvious. The only statistically significant finding was related to distant infections and not surgical site complications. Women were significantly more likely than men to develop distant other infections (primarily urinary tract infections) (25% vs 2%; P = .004). Postoperative fistula was seen more commonly in men than women, but this association fell short of statistical significance (P = .08).

Microvascular free-flap reconstruction of head and neck surgical defects has become commonplace. In our institution, more than half of the patients undergoing ablative surgery for head and neck squamous cell carcinoma now have reconstruction with free flaps. All who care for these patients have seen the consequences of wound infections causing flap loss. A little redness appears, the fluid in the drains becomes cloudy, and the Doppler signal or the color of the flap changes. Surgical salvage may be successful. If not, the patient has a large pharyngeal defect bathing an exposed carotid artery in a wound that may be previously irradiated and is now newly infected. It is no wonder that surgeons are passionate about their preferred perioperative antibiotic regimen in this setting.

Previous prospective studies evaluating antibiotic prophylaxis in head and neck surgery have shown the following: prophylactic antibiotics are necessary for head and neck defects^1,2; continuing antibiotics beyond 24 to 48 hours is unnecessary^3,4; and parenteral and/or topical antibiotics are effective.¹¹

Some aspects of antibiotic prophylaxis remain controversial. Clinical trials evaluating the need for coverage for gram-negative organisms have yielded conflicting results.^5,12 In the current study, we chose clindamycin alone without gram-negative coverage. This has been our standard for perioperative prophylaxis.

Also controversial is the time course for perioperative antibiotics for patients with free-flap reconstruction. Should antibiotics be given for only 24 hours, or is a longer course warranted in this subgroup? Not all surgeons believe that patients treated with free flaps have been adequately studied in the previous antibiotic trials. Many routinely continue perioperative antibiotics until drains are removed or until discharge from the hospital. This controversy was unresolved among surgeons in our own institution and prompted the current study.

The study design was a prospective, randomized comparison of 2 groups. Group 1 received clindamycin, 900 mg intravenously every 8 hours for 3 doses (24 hours). Group 2 received clindamycin, 900 mg intravenously every 8 hours for 5 days or until discharge. The wounds were graded daily by a faculty surgeon blinded to group assignment. Wound infections occurred in 11% of patients in group 1 and 10% of patients in group 2 (P = .99). Univariate analysis of other variables, including patient demographics, tumor site, stage, previous radiotherapy, and flap type, showed no significant predictors of wound infection or other surgical wound complications within this patient group. Although the study was initially targeted for accrual of 110 patients, accrual was stopped at 74 patients. Interim analysis demonstrated infection rates that were very similar between the 2 groups, making the original sample size determinations irrelevant. We believe that the outcome of this study justifies prophylactic use of clindamycin for 24 hours perioperatively to prevent wound infections for patients undergoing free-flap reconstruction.

Corresponding author and reprints: William R. Carroll, MD, Division of Otolaryngology–Head and Neck Surgery, 1501 Fifth Ave S, Birmingham, AL 35233 (e-mail: William.Carroll@ccc.uab.edu).

Accepted for publication January 24, 2003.

This study was presented at the annual meeting of the American Head and Neck Society, Boca Raton, Fla, May 13, 2002.