Effect of topical antibiotic therapy on recovery after tonsillectomy in adults

Abstract

OBJECTIVE

Systemic antibiotics given during the first week after tonsillectomy appear to be effective in reducing postoperative morbidity. We assessed the effectiveness of perioperative topical antibiotic rinses in reducing posttonsillectomy morbidity.

METHODS

A randomized, double-blinded, placebo-controlled pilot study of 36 patients undergoing tonsillectomy was used to evaluate the effects of a standard 7-day systemic regimen of perioperative intravenous ampicillin/oral amoxicillin and 2 single-day topical antibiotic regimens: (1) clindamycin (Cleocin) and (2) amoxicillin/clavulanate (Augmentin) and ticarcillin/clavulanate (Timentin).

RESULTS

Mean aerobic and anaerobic oral bacterial counts were decreased in both topical treatment groups compared with the placebo group on the first postoperative day, achieving statistical significance with Augmentin/Timentin (aerobic and anaerobic bacterial counts) and Cleocin (aerobic counts). Significantly less postoperative pain and mouth odor were reported for both Cleocin (P = 0.014 and P = 0.005, respectively) and Augmentin/Timentin (P = 0.026 and P = 0.05, respectively) topical treatment groups when compared with the placebo group.

CONCLUSIONS

Preliminary results indicate a reduction in oral bacterial counts and postoperative morbidity in adult patients receiving topical antibiotics compared with patients receiving placebo; further investigation is warranted.

Tonsillectomy continues to be a commonly performed operation in the pediatric and young adult population. However, the procedure is often associated with significant postoperative morbidity including odynophagia, dysphagia, otalgia, fever, mouth odor, weight loss, and reduced oral intake. These problems may prolong the recovery period and occasionally require hospital admission for intravenous hydration and pain control.

Colonization of the open tonsillar fossa surgical wound with oral bacterial flora may cause significant local inflammatory reaction, exacerbating posttonsillectomy pain. 1 Therefore many otolaryngologists in the past prescribed postoperative oral antibiotics to alleviate posttonsillectomy morbidity, although their use was not specifically recommended in many of the major American otolaryngology textbooks. 2 , 3 In more recent years the efficacy of systemic antibiotics in alleviating certain aspects of posttonsillectomy morbidity has been demonstrated in 2 randomized, placebo-controlled clinical trials in pediatric and adult patients. 1 , 4

Topical administration of antibiotics has been effectively used for many years in colorectal surgery, with proven reduction in postoperative infections. 5 , 6 Grandis et al 7 subsequently showed the potential efficacy of topical antibiotic prophylaxis in preventing wound infections after contaminated head and neck surgery. Clindamycin, given as an oral rinse in this patient population, resulted in a striking reduction in overall oropharyngeal bacterial counts for at least 8 hours after the last dose and appeared to be more effective than parenteral clindamycin at reducing oropharyngeal bacterial counts. None of the 10 patients enrolled in this pilot study had a wound infection. A second study 8 compared perioperative topical clindamycin (for 1 or 5 days), perioperative topical Augmentin/Timentin (amoxicillin plus clavulanate/ticarcillin clavulanate for 1 day) and perioperative parenteral clindamycin (for 1 day) as antibiotic prophylaxis for contaminated head and neck surgery. All 3 topical regimens were well tolerated, safe, and effective in preventing wound infections. Again, the topical antibiotic regimens were more effective at reducing the number of bacteria in the neck viscera at the time of surgery than a commonly prescribed parenteral regimen of clindamycin. Among the topical regimens, Timentin/Augmentin appeared to have a “superior bacteriologic effect” at reducing oropharyngeal bacterial counts for up to 7 days after surgery.

Table 1.

Treatment groups

Group	Antibiotic	Preoperative rinse∗	Intraoperative rinse∗	Postoperative rinse∗	Oral capsules †
1	None	Placebo elixir	Saline	Placebo elixir	Placebo
2	Ampicillin/Amoxicillin ‡	Placebo elixir	Saline	Placebo elixir	Amoxicillin (250 mg)
3	Clindamycin	Clindamycin (1.5 g)	Clindamycin (900 mg)	Clindamycin (1.5 g)	Placebo
4	Augmentin/Timentin	Augmentin (3.75 g)	Timentin (3.1 g)	Augmentin (3.75 g)	Placebo

∗Total volume of 60 mL was “swished” in the oral cavity for 60 seconds and then expectorated or aspirated.

^†Taken 3 times daily for 7 days after surgery.

^‡Ampicillin (1 g) administered intravenously during surgery.

Table 2.

Patient characteristics by treatment group

Group	Treatment	Mean age (y) ± SD	Sex (M/F)
1	Placebo	25.7 ± 7.7	4/6
2	Ampicillin/Amoxicillin	22.1 ± 3.8	4/4
3	Clindamycin	30.6 ± 10.3	3/4
4	Augmentin/Timentin	25.2 ± 7.6	6/5
	TOTAL	25.7 ± 7.8	17/19

Integrating these data from systemic and topical antibiotic studies, we hypothesized that perioperative topical oral antibiotic therapy would alleviate the postoperative recovery period by reducing oral bacterial counts and, hence, the inflammation related to bacterial colonization of the surgical site. To test this hypothesis, we conducted the following placebo-controlled, randomized, double-blind pilot study to obtain preliminary data on the effectiveness of 2 topical antibiotic regimens at reducing posttonsillectomy morbidity.

METHODS

General Study Design

Patients at least 18 years of age with no significant medical conditions (eg, diabetes, chronic lung disease, bleeding disorders) who were scheduled to undergo elective tonsillectomy for chronic or recurrent tonsillitis, history of peritonsillar abscess, or tonsillithiasis were eligible for enrollment into the study. The following exclusion criteria were applied: use of antibiotics within 7 days of surgery, allergy to any of the study medications, and any medical conditions requiring perioperative antimicrobial therapy (eg, mitral valve prolapse). After informed consent was obtained, patients were randomly assigned to 1 of the 4 groups (Table 1) with a computer-generated randomization chart.

After providing oral bacterial culture specimens (see Oral Bacterial Counts) in the preoperative holding area, patients were instructed to “swish and spit” for 1 minute with the assigned mouthwash (placebo or antibiotic). Intraoperatively, after removal of the tonsils, the oral cavity was again irrigated for 1 minute with either placebo (saline solution) or the designated antibiotic solution. A final oral rinse was performed 8 hours after surgery with the assigned mouthwash in a similar fashion as above. Patients were discharged the following morning with capsules (amoxicillin or placebo) to be taken 3 times daily for 1 week and a symptom log to be filled out twice daily (see Postoperative Morbidity Assessment). This study was reviewed and approved by the Walter Reed Army Medical Center Institutional Review Board and Human Use Committee. Informed consent was obtained from all subjects.

Oral Bacterial Counts

Aerobic and anaerobic oral cavity bacterial cultures were obtained with a standardized tongue swabbing technique at 3 times during the study: immediately before surgery, the morning after surgery, and the seventh day after surgery. All cultures were taken in the morning before toothbrushing. Specimens were then plated for quantification of total aerobic and anaerobic oral bacterial counts. Each specimen was diluted in standard brain/heart infusion broth (supplemented with reducing agents for anaerobic cultures) to a concentration of 10⁻⁴. A 0.1-mL aliquot of the dilution was plated onto blood agar (aerobic culture) or Brucella agar (anaerobic culture) to give a final dilution of 10⁻⁵. After incubation, total aerobic and anaerobic colonies were determined for each specimen.

Postoperative Morbidity Assessment

Subjects were instructed to score the following parameters twice daily (morning and evening) during the first 7 postoperative days with a 100-mm visual analog scale (0 mm = none, 100 mm = most severe or greatest): pharyngeal pain, otalgia, mouth odor, ability to tolerate a soft diet, and activity level. The following symptoms were also monitored with a 4-point Likert scale (none, mild, moderate, or severe): bleeding, nausea, vomiting, diarrhea, and skin rash. Patients were asked to record their temperatures twice daily.

Statistical Analysis

Preliminary data analyses were performed with SPSS 8.0 for Windows (SPSS Inc, Chicago, IL). A nonparametric approach was used to analyze the bacterial count data. Kruskal-Wallis 1-way analysis of variance (ANOVA) was used to compare the difference among the groups for the baseline bacterial counts and for the percentage of change at postoperative days 1 and 7. To account for the skewness and unequal variances among the groups, data for the morbidity parameters (eg, pharyngeal pain, mouth odor) were transformed with an arc sine transformation. BMDP5V (BMDP, Berkeley, CA) was used to perform repeated-measures ANOVA (4 treatment groups by 7 study times for all the above morbidity parameters). Post hoc multiple comparisons were conducted only when the overall F test from the ANOVA was significant The Cochran Q test was used to analyze patient ability to tolerate a soft diet (yes/no) over time. The average of the twice daily data was used in the statistical analysis. All tests were 2 tailed with a type I error set at α equal to 0.05 for this pilot study.

Fig. 1

Mean total aerobic (A) and anaerobic (B) bacterial counts (in colony-forming units × 10³) for each treatment group on day 0 (preoperative) and postoperative days 1 and 7.

RESULTS

A total of 36 of the 51 patients who enrolled in the study completed all scheduled visits and returned symptom logs. A skin rash developed in 1 patient randomly assigned to group 2, requiring discontinuation from the study on the second postoperative day. Five patients failed to return their symptom logs, and 9 patients were lost to follow-up (an inherent problem with our highly mobile military population). These nonevaluable patients were generally evenly distributed across the 4 treatment groups. There were no significant differences between the treatment groups with respect to age or sex (Table 2).

Oral Bacterial Counts

Quantitative bacterial counts were available for 33 of the 36 patients who completed the study. The mean aerobic and anaerobic oral bacterial counts for each of the treatment groups on days 0 (preoperative), 1, and 7 are shown in Fig 1. Despite a standardized technique of specimen collection, high interpatient variability in baseline oral bacterial counts was found among patients. Because of the small number of patients per group, this variability in baseline mean oral bacterial counts was also noted between treatment groups. However, as shown in Fig 1A, the Wilcoxon signed-rank test indicated that there was a significant overall decrease in aerobic bacterial counts (P = 0.007) on the first postoperative day compared with baseline counts when all treatment groups were considered. Mean aerobic bacterial counts were less in all 4 treatment groups on day 1 than at baseline (Day 0) counts. The counts were significantly decreased in group 3 (P = 0.039) and group 4 (P = 0.04). Likewise, when all treatment groups were considered, a highly significant reduction in oral anaerobic bacterial counts (Fig 1B) compared with baseline bacterial counts was noted on day 1 (P = 0.0059). A striking effect was seen in group 4 (topical Augmentin/Timentin), which revealed significantly greater reduction in anaerobic counts than that seen in group 1 (P = 0.0034), group 2 (P = 0.016), and group 3 (P = 0.027).

Fig. 2

Mean pharyngeal pain scores (with SE bars) for each treatment group during the first 5 postoperative days. Scores were derived from a visual analog scale and represent the average of morning and evening ratings across all subjects in each group (0 = no pain, 100 = severe pain).

Fig. 3

Mean mouth odor scores (with SE bars) for each treatment group during the first 5 postoperative days. Scores were derived from visual analog scales and represent the average of morning and evening ratings across subjects in each group (0 = no odor, 100 = severe mouth odor).

With respect to day 7 aerobic and anaerobic quantitative cultures, no significant differences were seen among groups or between day 7 to day 0 counts.

Postoperative Morbidity

Although patients were asked to record symptom scores twice daily until their day 7 visit, many patients did not complete diaries for the day preceding this visit or were seen 1 day earlier than the day 7 visit because of scheduling difficulties. However, the statistical model used by the BMDP5V method allows for unbalanced data resulting from missed observations. Figures 2 and 3 present complete data from all patients during the first 5 days of the study period, when patients would be expected to have the most symptoms. 4

Pharyngeal pain. Analysis of pain scores (Fig 2) during the 5-day period revealed significantly less pharyngeal discomfort in both the topical clindamycin and topical Augmentin/Timentin groups than in placebo-treated controls (P = 0.014 and P = 0.026, respectively) or in the active control regimen of ampicillin/amoxicillin (P = 0.024 and P = 0.045, respectively). Unlike a previously published placebo-controlled trial, 1 our study found no significant difference between the active control group (group 2) and placebo-treated patients (group 1) with respect to pharyngeal pain.

Mouth odor. Oral bacterial colonization of the open tonsillar fossa frequently leads to an offensive mouth odor during the recovery period. Patients receiving topical clindamycin (group 3) had significantly less mouth odor during the study period than either placebo-treated (P = 0.005) or ampicillin/amoxicillin-treated (P = 0.021) patients (Fig 3). Similarly, patients receiving the topical Augmentin/Timentin regimen demonstrated significantly less mouth odor than placebo-treated (P = 0.05) but not ampicillin/amoxicillin-treated patients (P = 0.16). Unlike previous studies, 1 , 4 ours detected no significant difference (P = 0.72) between placebo and a regimen of postoperative oral antibiotics (group 2).

Otalgia. Referred otalgia is a common symptom after tonsillectomy. Although mean scores for otalgia were less in all antibiotic treatment groups than in the placebo group, no statistically significant differences were found between groups during the 5-day study period (P = 0.33).

Fever. Postoperative fevers were uncommon in patients during the postoperative period. Only 7 patients in groups 2, 3, and 4 had temperature recordings of 100°F or more during the study period. No significant differences among the 4 treatment groups were noted.

Activity level and diet. No significant differences were noted among treatment groups with respect to activity level (P = 0.63). Patients’ abilities to tolerate a soft diet were significantly increased for all groups (P = 0.04) during the course of the study period; however, such increases were not significantly different among the groups.

Complications. A skin rash developed in 1 patient in the ampicillin/amoxicillin-treated group the day after discharge from the hospital, and that patient was discontinued from the study. No other significant adverse events, including postoperative hemorrhage or infections, were noted in any of the treatment groups.

DISCUSSION

Telian et al 1 were the first investigators to assess the effects of systemic antibiotic therapy on the posttonsillectomy recovery period in a prospective, randomized, well-controlled clinical trial. Using a regimen of perioperative intravenous ampicillin followed by a 7-day course of oral amoxicillin in pediatric patients undergoing tonsillectomy, they observed significantly fewer problems with fever, mouth odor, oral intake, and decreased activity level than did a placebo control group. A second prospective, randomized, placebo-controlled trial conducted by Grandis et al 4 in adult patients confirmed the beneficial effect of systemic antibiotic therapy during the recovery period. Patients receiving antibiotics had fewer days with mouth odor and had a faster return to routine activities and a regular diet. However, no significant improvement in postoperative pain or fever was observed in antibiotic-treated patients in this study.

Systemic antibiotics may reduce postoperative morbidity, as shown in these trials, by reducing oral bacterial colonization of the open tonsillar fossa. 1 , 4 Reduced bacterial counts in the open surgical wound might reasonably be expected to decrease local inflammation, promote healing, and hasten recovery. The safety and efficacy of topical antibiotics in reducing oral bacterial counts and wound infections in contaminated head and neck surgery have been previously demonstrated. 7 – 9 Hence, we hypothesized that direct application of high concentrations of topical antibiotics to the oral cavity would be effective in decreasing oral bacterial counts and, consequently, in decreasing morbidity related to bacterial colonization of the denuded tonsillar fossa. However, the best choice of topical antibiotic for a clinical trial to test this hypothesis was unclear. Although two studies by Grandis et al. 7 , 8 found topical clindamycin to be safe and effective in reducing oral bacterial counts and wound infections, a combination regimen of Augmentin/Timentin appeared to be more effective in maintaining suppression of bacterial counts 7 days after contaminated head and neck surgery. We therefore designed a small prospective, randomized, placebo-controlled pilot study to compare the effects of these 2 topical regimens on total oral aerobic and anaerobic bacterial counts and on selected indicators of postoperative morbidity.

Analysis of patient symptom scores recorded during the first 5 postoperative days revealed significantly lower pharyngeal pain and mouth odor scores in topical antibiotic groups than in patients treated with either placebo or a standard regimen of perioperative ampicillin followed by a 7-day course of oral amoxicillin. However, no significant differences in the treatment groups were noted with respect to otalgia, the incidence of postoperative fever, or the time to resumption of normal diet and activity level.

Interestingly, we failed to detect any statistically significant improvement in morbidity parameters for the “current practice” active control (group 2) compared with the placebo group. Several possible factors could explain this unexpected finding. First, amoxicillin in this adult study was administered as a 250-mg capsule 3 times a day, whereas pediatric patients in the study by Telian et al, 1 received 3 times daily doses of 250 mg in patients weighing 20 kg or more and 125 mg for children weighing less than 20 kg (likely in a pediatric suspension dosage form). This represents a higher dose in the pediatric patients, which may also, in fact, have had direct topical antibiotic effects (as a liquid suspension) on the healing tonsillar fossa. Other possible factors may include altered amoxicillin susceptibility of oral bacteria during the decade between the 2 studies and differences in the instruments used to measure morbidity between studies. Finally, the small number of patients per group in our study severely limits the ability to detect differences between the groups.

Despite the limitations of this small pilot study, we believe our findings warrant further formal clinical trials to definitively assess the usefulness of topical antibiotics in patients undergoing tonsillectomy. Potential advantages of topical over systemic antibiotic therapy may include better patient compliance and less chance of systemic, drug-related adverse events.

We thank Ms Regina Van Brakle and Ms Marie Pe in the Clinical Mcrobiology Laboratory, Walter Reed Army Medical Center, for their expertise in evaluating oral bacterial culture specimens.

References

1 Telian

Handler

Fleisher

. The effect of antibiotic therapy on recovery after tonsillectomy in children. Arch Otolaryngol Head Neck Surg 1986;112:610–5.

Joseph

DJ.

Tonsillectomy and adenoidectomy. In: English

, editor. Otolaryngology. New York: Harper & Row; 1984. p. 24–5.

Kornblut

AD.

Tonsillectomy and adenoidectomy. In: Paparella

Shumrick

, editors. Otolaryngology. 2nd ed. Philadelphia: WB Saunders; 1980. p. 2296–9.

4 Grandis

Johnson

Vickers

. The efficacy of peri-operative antibiotic therapy on recovery following tonsillectomy in adults: randomized double-blind placebo-controlled trial. Otolaryngol Head Neck Surg 1992;106:137–42.

Edmondson

Rissing

JP.

Prophylactic antibiotics in colon surgery. Arch Surg 1983;118:496–501.

6 Nichols

Condon

Gorbach

. Efficacy of perioperative preparation of the bowel. Ann Surg 1973;176:227–32.

7 Grandis

Vickers

Rihs

. The efficacy of topical antibiotic prophylaxis for contaminated head and neck surgery. Laryngoscope 1994;104:719–24.

8 Grandis

Vickers

Rihs

. Efficacy of topical amoxicillin plus clavulanate/ticarcillin plus clavulanate and clindamycin in contaminated head and neck surgery: effect of antibiotic spectra and duration of therapy. J Infect Dis 1994;170:729–32.

9 Elledge

Whidden

Fraker

. The effects of topical oral clindamycin rinses on the bacterial content of saliva in healthy human subjects. Otolaryngol Head Neck Surg 1991;105:836–9.