Purpose
To evaluate the clinical characteristics, bacterial culture, and visual outcome of patients with acute endophthalmitis after cataract surgery.
Design
Retrospective consecutive interventional case series.
Methods
Clinical notes from patients treated for acute endophthalmitis after cataract surgery in a single center from 1996 to 2006 were reviewed. Patients with less than 1 month of follow-up and missing bacterial cultures were excluded. Vitreous biopsy or primary vitrectomy followed by intravitreal injection of vancomycin and ceftazidime (± prednisolone) was performed. Main outcome measures were bacterial culture and final visual acuity.
Results
Bacterial cultures (total 250 cases) showed bacterial growth in 166 cases (66.4%). From these 166 cultures, 89 (53.6%) revealed gram-positive coagulase-negative, 63 (38.0%) other gram-positive, 10 (6.0%) gram-negative, and 4 (2.4%) polymicrobial cultures. Vitreous biopsy with intravitreal antibiotics injection was performed in 225 (90.0%) of cases. Primary vitrectomy with intravitreal antibiotics was performed in 25 eyes (10.0%). Final visual acuity ≥0.5 was achieved in 129 (51.6%) of all cases, 54 (60.7%) of the 89 gram-positive coagulase-negative cultures, 20 (31.7%) of the 63 other gram-positive cultures, 5 (50.0%) of the 10 gram-negative cultures, and 9 (45.0%) of the 20 Staphylococcus aureus cultures. There was no additional effect for treatment by primary vitrectomy or intravitreal prednisolone.
Conclusions
Treatment outcome after endophthalmitis is highly dependent on the causative organism. Treatment outcomes for gram-negative bacteria and S. aureus may be better than previously reported. Prompt treatment of endophthalmitis remains essential and the role of complete primary vitrectomy remains subject to debate.
Cataract surgery is one of the most commonly performed surgical procedures worldwide. Endophthalmitis as the most devastating complication occurs in 0.04 to 0.2 percent of cases. Factors influencing treatment outcome after postcataract endophthalmitis are visual acuity at presentation, bacterial culture results, and the type of treatment. There is no consensus on whether a primary vitrectomy should be done as initial treatment for endophthalmitis after cataract surgery. The Endophthalmitis Vitrectomy Study (EVS) advocates primary vitrectomy only in patients with visual acuity as bad as light perception at presentation, while Kuhn and associates advocate early vitrectomy in more patients. There are only few data about the outcome of endophthalmitis after cataract surgery in relation to the microbial culture, although the causative organism is considered a major determinant in treatment results.
In this study, we evaluate the causative organisms in relation to clinical characteristics and visual outcome of patients with acute postoperative endophthalmitis after cataract surgery in a tertiary referral center within a period of 11 years.
Methods
Clinical records of all patients treated for acute postoperative endophthalmitis at Radboud University Nijmegen Medical Centre between January 1, 1996, and December 31, 2006, were reviewed. The study included patients with onset of symptoms within 6 weeks after cataract surgery referred from other hospitals and from our own department. Patients with endophthalmitis following secondary lens implantation or any combined surgery for vitreoretinal disease or glaucoma were excluded.
All patients underwent a complete ophthalmologic evaluation at presentation, including initial best-corrected visual acuity (BCVA) by Snellen charts, slit-lamp examination, indirect ophthalmoscopy, and Goldmann applanation tonometry. Additional data obtained from the medical records included gender, age, laterality of the involved eye, time from surgery to endophthalmitis symptoms, treatment method, results of bacterial cultures from vitreous taps, final BCVA, and length of follow-up.
On admission, either a vitreous biopsy or a primary vitrectomy was performed. This was followed by intravitreal injection of antibiotics with or without prednisolone. The exact choice of treatment was at the discretion of the eye surgeon. If patients were sent back to their referring ophthalmologist after initial treatment, outcome data were obtained by contacting the referring ophthalmologist.
Results of bacterial cultures were grouped as “negative” (no growth), “gram-positive coagulase-negative,” “other gram-positive,” “gram-negative,” or “polymicrobial.” The gram-positive coagulase-negative group consisted of coagulase-negative staphylococci only . All other gram-positive bacteria formed the group of “other gram-positive.” “Polymicrobial” was defined as the isolation of more than 1 species of bacteria.
For statistical analysis a multivariate logistic regression was used to investigate the association of visual acuity outcome with the causative bacteria, presenting visual acuity, time from surgery to signs of endophthalmitis, and type of treatment.
Results
During the study interval 286 eyes of 284 patients were treated for acute endophthalmitis after cataract surgery. Excluded were patients with follow-up of less than 1 month (n = 14), patients with presenting visual acuity (VA) of no light perception (n = 3), patients with presenting visual acuity of >0.40 (n = 4), patients who were enrolled in a randomized multicenter trial (n = 2), patients who received only topical treatment (n = 7), patients in whom a bacterial culture was taken but no result was obtained (n = 2), and patients in whom no vitreous biopsy was performed (n = 4; 1 primary vitrectomy without a bacterial culture performed, 1 failed vitreous biopsy, and 2 eyes that were judged to be in too bad shape to perform a vitreous biopsy). Eventually, 250 eyes of 248 patients were included in our present study and data were used for further analysis.
The mean age of the study participants was 74 years (range, 19–97). They were referred by 97 ophthalmic surgeons from 33 hospitals (0.25 cases/surgeon/year). In 136 of the 250 cases (54.4%), signs of endophthalmitis occurred within 3 days after cataract surgery, and 203 of the 250 cases (81.2%) had signs of endophthalmitis within 7 days after cataract surgery (range, 0–31 days; mean, 5 days). Visual acuity at presentation was <0.05 in 228 of the 250 eyes (91.2%), with 94 of the 250 eyes (37.6%) presenting with light perception only (LP). Mean follow-up was 14 months (range, 1 month to 9 years; median, 6 months). More detailed characteristics are shown in Table 1 .
| Characteristics | n (%) |
|---|---|
| Gender | |
| Male | 101 (40.4) |
| Female | 149 (59.6) |
| Time from surgery to signs of endophthalmitis | |
| <3 days | 136 (54.4) |
| 4 to 7 days | 67 (26.8) |
| >7 days | 47 (18.8) |
| Presenting visual acuity | |
| 0.05–0.40 | 22 (8.8) |
| Counting fingers | 20 (8.0) |
| Hand motions | 114 (45.6) |
| Light perception | 94 (37.6) |
Visual Acuity Outcomes
Overall, 129 of the 250 cases (51.6%) had final visual acuities of ≥0.50 at final follow-up; 172 of the 250 cases (68.8%) had final visual acuities of ≥0.20. Forty-three of the 250 cases (17.2%) had final visual acuities of worse than or equal to light perception only, including 10 cases (4.0%) with no light perception. In 20 cases (8.0%) the affected eye had to undergo evisceration.
Bacterial Cultures and Final Visual Acuity
Of the 250 bacterial cultures, 166 (66.4%) showed bacterial growth and 84 (33.6%) were negative for bacterial growth. Of the 166 positive bacterial cultures, 89 (53.6%) were gram-positive coagulase-negative, 63 (38.0%) were other gram-positive, 10 (6.0%) were gram-negative, and 4 (2.4%) were polymicrobial. The results of the bacterial cultures with corresponding final visual acuities are shown in Table 2 .
| Eyes | VA ≤LP | VA HM – <0.50 | VA ≥0.50 | ||||
|---|---|---|---|---|---|---|---|
| Bacterial Cultures/Causative Organisms | n | n | % | n | % | n | % |
| Negative bacterial cultures | 84 | 5 | 6.0 | 30 | 35.7 | 49 | 58.3 |
| Positive bacterial cultures | |||||||
| Gram-positive coagulase-negative | 89 | 4 | 4.5 | 31 | 34.8 | 54 | 60.7 |
| Other gram-positive | 63 | 32 | 50.8 | 11 | 17.5 | 20 | 31.7 |
| Staphylococcus aureus | 20 | 6 | 30.0 | 5 | 25.0 | 9 | 45.0 |
| Streptococcus pneumoniae | 12 | 12 | 100.0 | 0 | 0 | 0 | 0 |
| Viridans group streptococci | 11 | 4 | 36.4 | 3 | 27.3 | 4 | 36.4 |
| β-hemolytic streptococcus | 9 | 8 | 88.9 | 0 | 0 | 1 | 11.1 |
| Group A | 2 | 2 | 100.0 | 0 | 0 | 0 | 0 |
| Group C | 1 | 1 | 100.0 | 0 | 0 | 0 | 0 |
| Group G | 6 | 5 | 83.3 | 0 | 0 | 1 | 16.7 |
| Enterococcus species | 3 | 1 | 33.3 | 0 | 0 | 2 | 66.7 |
| Diphtheroid gram-positive rods | 3 | 1 | 33.3 | 2 | 66.7 | 0 | 0 |
| Abiotrophia species | 2 | 0 | 0 | 1 | 50.0 | 1 | 50.0 |
| Gemella morbillorum | 1 | 0 | 0 | 0 | 0 | 1 | 100.0 |
| Peptostreptococcus species | 1 | 0 | 0 | 0 | 0 | 1 | 100.0 |
| Propionibacterium acnes | 1 | 0 | 0 | 0 | 0 | 1 | 100.0 |
| Gram-negative | 10 | 2 | 20.0 | 3 | 30.0 | 5 | 50.0 |
| Proteus mirabilis | 3 | 1 | 33.3 | 0 | 0 | 2 | 66.7 |
| Haemophilus influenzae | 3 | 0 | 0 | 2 | 66.7 | 1 | 33.3 |
| Pseudomonas aeruginosa | 2 | 1 | 50.0 | 0 | 0 | 1 | 50.0 |
| Achromobacter xylosoxidans | 1 | 0 | 0 | 0 | 0 | 1 | 100.0 |
| Acinetobacter lwoffii | 1 | 0 | 0 | 1 | 100.0 | 0 | 0 |
| Polymicrobial | 4 | 0 | 0 | 3 | 75.0 | 1 | 25.0 |
| Streptococcus milleri and gram negative rod | 1 | 0 | 0 | 0 | 0 | 1 | 100.0 |
| Diphtheroid gram-positive rod and Propionibacterium acnes | 1 | 0 | 0 | 1 | 100.0 | 0 | 0 |
| Viridans group streptococci and P. acnes | 2 | 0 | 0 | 2 | 100.0 | 0 | 0 |
For statistical analysis the polymicrobial group was excluded. Success was defined as a final visual acuity of ≥0.50. The “gram-positive coagulase-negative” group had a significantly greater probability of success than the “other gram-positive” group (odds ratio 2.6; 95% confidence interval [CI] from 1.2 to 5.7, P = .015). The group that was negative for bacterial growth tended to be associated with a greater probability of success than the “other gram-positive” group (odds ratio 2.1; 95% CI from 1.0 to 4.6, P = .055).
Presenting Visual Acuity and Visual Acuity Outcomes
Final VA was ≥0.50 in 18 of the 22 cases (81.8%) with presenting VA ≥0.05 to 0.40, 15 of the 20 cases (75.0%) with presenting VA of counting fingers, 76 of the 114 cases (66.7%) with presenting VA of hand motions, and 20 of the 94 cases (21.3%) with presenting VA of light perception only (LP).
A presenting VA better than LP was significantly associated with a final VA ≥0.50. For presenting VA, in comparison with LP, the odds ratio for having a final VA ≥0.50 was 7.2 for hand movements (95% CI from 3.8 to 13.7), 12.0 for counting fingers (95% CI from 2.8 to 51.3), and 15.6 for VA 0.05–0.40 (95% CI from 3.8 to 63.2).
Time From Surgery to Signs of Endophthalmitis and Visual Acuity Outcomes
The percentage of patients with signs of endophthalmitis within 3 days after cataract surgery was 57.1% for no bacterial growth, 38.2% for gram-positive coagulase-negative, 65.1% for other gram-positive, and 100.0% for gram-negative.
Final VA ≥0.50 for cases with signs of endophthalmitis within 3 days versus cases with signs of endophthalmitis more than 3 days after surgery was 58.3% versus 58.3% for negative bacterial cultures, 61.8% versus 60.0% for gram-positive coagulase-negative cultures, and 14.6% versus 63.6% for other gram-positive cultures.
All cases with a culture of S. pneumoniae (n = 12) or β-hemolytic streptococcus (n = 9) presented within 3 days after surgery, and only 1 out of these 21 cases (4.8%) achieved a final VA ≥0.50. For cases with a bacterial culture of S. aureus , final VA was ≥0.50 in 2 of the 9 cases (22.2%) with signs of endophthalmitis within 3 days after surgery, compared to 7 of the 11 cases (63.6%) with signs of endophthalmitis more than 3 days after surgery.
Using multivariate logistic regression, there was no significant association of time from surgery to signs of endophthalmitis with final VA ≥0.50 ( P = .357).
Treatment Features and Final Visual Acuity
Vitreous biopsy followed by intravitreal administration of antibiotics (tap and inject) was performed in 225 of the 250 cases (90.0%). Twenty of the 225 cases (8.9%) treated by tap and inject had a secondary vitrectomy between 2 days and 7 months after initial treatment. Primary vitrectomy with intravitreal administration of antibiotics was performed in 25 of the 250 cases (10.0%). Two of the 25 cases (8.0%) treated by primary vitrectomy had a second vitrectomy. Fifteen of the 94 cases (16.0%) presenting with light perception only at time of diagnosis were treated with primary vitrectomy and intravitreal antibiotics. Vancomycin (1 mg) and ceftazidime (2.25 mg) were used as intravitreal antibiotics in all but 1 patient. In 1 patient vancomycin and gentamycin were used because of an allergy to ceftazidime. Intravitreal prednisolone (2.5 mg) was added to the antibiotics in 186 of the 225 cases (82.7%) treated with tap and inject and in 7 of the 25 cases (28.0%) treated with primary vitrectomy. After initial treatment all cases were treated topically every hour with gentamycin 2.25% and cefazoline 3.33%. Systemic antibiotics were not administered in any of the patients. Differences in treatment outcomes between primary vitrectomy and tap and inject are summarized in Table 3 .
