Methods

We reviewed charts of consecutive epiphora patients who underwent external or endoscopic DCR surgery at Severance Hospital, Yonsei University College of Medicine between January 2007 and March 2010. All patients underwent standard irrigation of the lacrimal system and probing to localize the site of lacrimal obstruction or canalicular stenosis. Nasal endoscopy was performed in all patients to detect the presence of nasal septal or middle turbinate hypertrophy. All patients underwent standard external or endoscopic DCR surgery with silicone intubation. Systemic third-generation cephalosporin was orally administered postoperatively for 5 days. Before tube removal, slit-lamp examinations were performed to assess the inflammatory signs and discharge of pus material around the tubes. The tubes were removed from the nasal space with an endoscope after cutting the loop between the upper and lower puncta.

All removed silicone tubes were placed in trypticase soy broth (Oxoid Inc, Ottawa, Ontario, Canada) and transported immediately to a microbiology laboratory for processing. In the laboratory, Gram stain preparations were made for all samples. Samples were then plated onto blood agar (incubated in 5% carbon dioxide at 35 C), MacConkey agar (incubated in oxygen at 35 C), or chocolate agar (incubated in 5% carbon dioxide at 35 C) for aerobic growth and in thioglycollate broth (incubated anaerobically in carbon-nitrogen atmosphere at 35 C) to isolate anaerobes. All plates were examined daily for 4 days. For fungal isolation, samples were plated onto BBL Mycosel Agar (Becton Dickinson, Franklin Lakes, New Jersey, USA) and incubated for 2 weeks. Routine laboratory methodology was used for species identification.

The identified organisms were analyzed and correlated to the duration of the silicone tube placement, the presence of inflammatory signs such as conjunctival injection and pus discharge, endoscopic findings of membranous obstruction of the nasal ostium, preoperative canalicular stenosis and preoperative nasal space narrowing attributable to septal deviation/hypertrophy, or hypertrophy of the middle turbinate. Failure was regarded as recurrence of epiphora, including both anatomical failure of the surgical fistula proven with syringing and endoscopic examination and functional failure showing recurrent epiphora regardless of patent lacrimal passage on syringing.

Data were analyzed using SAS software version 9.1.3 (SAS Institute Inc, Cary, North Carolina, USA). The χ ² test, Fisher’s exact test (The R Foundation for Statistical Computing, Vienna, Austria), and the Mann-Whitney U test were used to test for statistical significance, and P values less than .05 were considered significant.

Results

A total of 39 silicone tubes removed from 33 patients were cultured: 34 (87.2%) external DCR cases and 5 (12.8%) endoscopic DCR cases. Extubation was performed at a median of 5 months (range, 0.840 months). Preoperative canalicular stenosis was present in 11 cases (28.2%), and preoperative nasal endoscopy detected 12 cases (30.7%) of nasal space narrowing attributable to septal deviation/hypertrophy or hypertrophy of the middle turbinate. Eleven cases (28.2%) had histories of endoscopic revision attributable to narrowing of nasal ostium with granulation tissue or membranous obstruction of the ostium. Eight cases (20.5%) exhibited inflammatory signs with pus discharge at the time of extubation. In 3 of them, symptoms disappeared immediately after the removal of silicone tube, and 5 of them were determined as failure at final assessment. Final surgical failure with persistent epiphora at last follow-up was observed in 5 cases (12.8%), 2 of which were anatomical failure and 3 of which were functional. Of the 5 failed cases, 4 cases had preoperative severe canalicular stenosis and required distal canaliculotomy at the time of external DCR with a long postoperative intubation period. Three of them had a very narrow nasal space attributable to septal hypertrophy or deviation.

Culture provided positive results in 37 cases (94.9%). A total of 52 isolates were identified: 73.1% (38/52) gram-positive bacteria, 23.1% (12/52) gram-negative bacteria, and 3.8% (2/52) fungi. In total, 64.1% (25/39) exhibited simple infection with 1 cultured organism and 35.9% (14/39) showed mixed infection with more than 1 cultured organism.

Among the gram-positive organisms, Staphylococcus aureus was isolated most frequently, being isolated in 28 cases. Among those, 22 cases (78.6%) were methicillin-resistant Staphylococcus aureus (MRSA). Diphtheroids were detected in 5 of 39 eyes (12.8%), and coagulase-negative Staphylococcus , including Staphylococcus epidermidis, was isolated from 4 of 39 eyes (10.3%) ( Table ).

Most of the gram-negative organisms were Pseudomonas aeruginosa , which was found in 5 of 39 eyes (12.8%), followed by Stenotrophomonas maltophilia , which was found in 3 of 39 eyes (7.7%). Other gram-negative strains included Klebsiella pneumoniae , Moraxella catarrhalis , and Serratia marcescens .

Fungal organisms were isolated in only 2 of 37 cases, 1 of which was Aspergillus sp and the other Fusarium sp.

The association was observed between pus discharge and each bacterial strain. Discharge was related with pseudomonas infection (χ ² test, P = .049). Particularly, pseudomonas infection was more frequently found in cases of final surgical failure ( P = .010). Other bacterial strains showed no statistically significant relation with discharge (MRSA, P = .709; other strains, P > .05).

The median time of tube placement was significantly longer in patients with postoperative discharge (11 months) than in patients without discharge (4 months) (Mann-Whitney U test, P < .001). The time of tube placement was also significantly longer in cases with Pseudomonas aeruginosa than in those with other bacteria (Mann-Whitney U test, P = .001). In addition, the duration of tube placement was significantly longer in patients who underwent endoscopic revision for narrowing or membranous obstruction of the nasal ostium than in patients without revision (Mann-Whitney U test, P = .001).

All 5 cases with pseudomonal infection had histories of endoscopic revision for membranous obstruction, which were 45.5% of endoscopic revision cases. Three of them were final surgical failure cases. The rate of pseudomonas infection was significantly higher in cases with revision than in those without revision (Fisher exact test, P = .001). Final surgical failure was significantly related with canalicular stenosis (Fisher’s exact test, P = .017), pus discharge at extubation (Fisher’s exact test, P < .001), history of endoscopic revision (Fisher’s exact test, P = .001), and pseudomonal infection (Fisher’s exact test, P = .010).