To evaluate acute-onset postoperative endophthalmitis occurring at an academic medical center and to compare rates over the last 25 years at a single institution.
Retrospective, consecutive case series.
Medical records were reviewed for all patients diagnosed with acute-onset postoperative nosocomial endophthalmitis from 2002 through 2009 associated with surgery at Bascom Palmer Eye Institute.
The 8-year frequency of acute-onset postoperative endophthalmitis was 0.025% (14 of 56 672 intraocular surgeries). The rate was 0.028% (8/28 568) for cataract surgery and 0.011% (2/18 492) for pars plana vitrectomy (PPV). Both PPV endophthalmitis cases followed 20-gauge surgery and no cases followed small-gauge, transconjunctival PPV (n = 2262). Three cases occurred following penetrating keratoplasty (3/2788, 0.108%). The most common bacterial isolate was Staphylococcus (n = 7, 50%). Initial treatment involved ocular paracentesis (n = 8, 57%) or vitrectomy (n = 5, 36%), in combination with injection of intraocular antibiotics (n = 14, 100%). Vancomycin and ceftazidime were used in 13 eyes (93%) and intraocular steroids were given initially to 9 eyes (64%). Final visual acuity was ≥20/200 in 9 eyes (64%) and 2 eyes (14%) were no light perception. At this institution since 1984, there has been a statistically significant trend for a decreasing rate of acute-onset postoperative endophthalmitis (1984–1994: 0.09%; 1995–2001: 0.05%; 2002–2009: 0.025%; P < .001).
At a university teaching hospital involving resident, fellow, and faculty surgeons, the frequency of acute-onset postoperative nosocomial endophthalmitis is low, has not increased in the era of sutureless clear corneal cataract surgery, and has steadily decreased when compared to prior time periods from the same institution.
Endophthalmitis is a serious, sight-threatening condition that is classified into 2 broad categories: endogenous or exogenous. Exogenous cases include those occurring postsurgically, those related to antecedent trauma, and those attributable to extension of an extraocular infection into the eye. The reported rate of acute-onset postoperative endophthalmitis varies by surgical procedure; because of a variety of factors including improved surgical technique, routine povidone-iodine use, and appropriate use of improved antibiotics, the rate seems to have declined dramatically over the past century.
More recently, with evolving trends in cataract and vitrectomy surgery toward smaller incisions and sutureless techniques, concern has arisen from some studies about an increased risk of postoperative endophthalmitis.
The purpose of the current study was to determine the frequency of acute-onset postoperative endophthalmitis over the most recent 8-year period, 2002 through 2009, and to compare these data with previous data from the last 25 years from the same institution.
Patients and Methods
The study design was a retrospective, consecutive case series. The medical and microbiological records of all patients who underwent surgery at Bascom Palmer Eye Institute and were diagnosed within 6 weeks of their surgery with postoperative endophthalmitis between January 1, 2002 and December 31, 2009 were reviewed.
The Quality Assurance Committee of our hospital requires that all cases of postoperative endophthalmitis be reported for infection control. Using the constraints of postoperative endophthalmitis occurring within 6 weeks of the surgical procedure, it is unlikely that this study missed patients who developed endophthalmitis and moved or sought care by another institution. However, if such an instance did occur, the rates reported herein would be underreported.
As part of operating room protocol in effect since 2000 at Bascom Palmer Eye Institute, 5% povidone-iodine solution was used to prepare the lids, lashes, and conjunctiva before all surgical procedures. In addition, 10% povidone-iodine solution was used to prepare the skin of the lids and face in the periorbital area before all surgical procedures. Some of the cataract cases in this report have been reported previously. Comparison of the current data with endophthalmitis rates since 1984 was performed using data from 2 previously published manuscripts from the same institution.
Data collected included patient age, gender, date and type of initial surgical intervention, other patient medical conditions, date of endophthalmitis diagnosis, associated exam findings and patient symptoms, date and details of initial and subsequent treatment strategies, microbiological culture results, and follow-up visions and dates. Microbiological analysis was performed as previously reported. Culture and organism identification techniques did not change during the study period.
All statistical analysis was performed using SPSS 17.0 for Windows (SPSS Inc, Chicago, Illinois, USA). Analysis of variance and Student t test were used for visual acuity comparisons among organisms and corticosteroid usage respectively. The Cochran-Armitage trend test was used to analyze frequency trends.
A total of 56 672 intraocular surgeries were performed at Bascom Palmer Eye Institute over the 8-year study period through 2009. During this time, acute-onset postoperative nosocomial endophthalmitis was diagnosed in 14 eyes of 14 different patients ( Table 1 ), yielding a frequency of 0.025%. The patients’ mean age was 60.9 years (median 62, range 19–84). There were 7 male patients (50%) and 11 right eyes (79%). Six patients (43%) identified themselves as white, 7 (50%) as Hispanic, and 1 (7%) as black. Diabetes mellitus affected 4 patients (29%) (Patients 1, 3, 6, and 11); 1 patient (7%) (Patient 2) was using methotrexate and etanercept for management of rheumatoid arthritis, and 1 patient (7%) (Patient 14) was on hemodialysis because of amyloidosis.
|Patient (Eye, Age a )||Year||Surgery||Intraoperative Complications||Days to Diagnosis||Aqueous Culture||Vitreous Culture||Organism Cultured|
|1 (OD, 68)||2003||Phaco/PCIOL||Vitreous loss, RLF||22||NP||+||Staph epi|
|2 (OD, 51)||2003||Phaco/PCIOL||None||30||NP||+||Staph epi|
|3 (OD, 52)||2003||Phaco/PCIOL||Iris prolapse||4||−||+||Staph aureus|
|4 (OD, 82)||2004||Phaco/PCIOL||None||2||NP||+||Strep pneumo|
|5 (OD, 84)||2005||Phaco/PCIOL||None||15||NP||+||Staph epi|
|6 (OD, 69)||2006||Phaco/PCIOL||None||8||NP||+||Staph epi|
|7 (OD, 71)||2006||20-g PPV/SB/EL/SO||None||6||NP||−||None|
|8 (OS, 49)||2007||20-g PPV/PPL/MP/SB/SO||None||6||−||NP||None|
|9 (OS, 19)||2007||Phaco/PCIOL||None||1||NP||−||None|
|10 (OS, 58)||2008||PK & sutured iridoplasty||None||23||+||NP||Strep mitis|
|11 (OD, 61)||2008||PK & ECCE/ACIOL||Vitreous loss||5||NP||+||Staph epi|
|12 (OD, 63)||2009||Removal of PCIOL and insertion of ACIOL||None||25||NP||NP||NP|
|13 (OD, 70)||2009||PK||None||2||+||+||Strep agal|
|14 (OD, 55)||2009||Phaco/PCIOL||None||10||NP||+||Staph epi|
Cataract surgery accounted for the majority of cases of endophthalmitis (8 cases out of 28 568 cataract surgeries, 0.028%), of which 2 had documented intraoperative complications: 1 posterior capsular rent with vitreous loss requiring an anterior vitrectomy and another case with significant intraoperative iris prolapse. The 2 cases of endophthalmitis associated with pars plana vitrectomy (PPV) occurred following 20-gauge vitrectomies in 2006 and 2007 (2/18 492, 0.011%). No cases of endophthalmitis occurred following the 2262 transconjunctival vitrectomies performed (23- or 25-gauge) during the study time period. Three cases of endophthalmitis occurred following penetrating keratoplasty (3/2788, 0.108%), 1 with concurrent cataract surgery complicated by vitreous loss requiring an anterior vitrectomy and 1 with concurrent iridoplasty. Two of the 3 cases of endophthalmitis following penetrating keratoplasty were associated with contaminated donor corneal tissue. In both cases, the same bacterial isolate was recovered from the donor corneal rim and the ocular paracentesis performed when the patients presented with endophthalmitis. One case of endophthalmitis occurred following secondary intraocular lens (IOL) implantation (1/1783, 0.056%); an anterior chamber IOL was inserted with removal of a dislocated posterior chamber IOL assisted by 23-gauge vitrectomy in a previously vitrectomized eye (Patient 12). No cases occurred following glaucoma surgery (n = 5041).
In the perioperative period of the surgery that eventually resulted in endophthalmitis, treating physicians used antibiotics and corticosteroids as they believed appropriate ( Table 2 ). Preoperatively, 4 patients used topical antibiotic eye drops, including a fluoroquinolone (n = 3) or polytrim (n = 1). Intraoperatively, all patients were given antibiotics by various routes: 11 were given topical drops including a fluoroquinolone (n = 8), polytrim (n = 1), gentamicin (n = 1), tobramycin (n = 1), or neomycin and polymyxin B (n = 1); 7 were given subconjunctival injection of cefazolin (n = 3) or gentamicin (n = 4); 1 was given gentamicin (8 μg/mL) in the irrigation fluid during cataract surgery. Additionally, 12 patients were given intraoperative steroids including topical prednisolone acetate (n = 8), topical dexamethasone (n = 2), subconjunctival dexamethasone (n = 6), subconjunctival triamcinolone (n = 1), peribulbar triamcinolone (n = 1), or intravitreal triamcinolone (n = 1). Postoperatively, all patients were given topical antibiotic eye drops, including a fluoroquinolone (n = 10), polytrim (n = 3), or vancomycin and gentamicin (n = 1); additionally, 10 patients were given prednisolone acetate eye drops and 2 patients were given a topical nonsteroidal anti-inflammatory eye drop.
|1||FQ||Subconj: C, D||FQ|
|3||FQ||Top: FQ, PF||FQ|
|4||None||Top: Poly, PF||Poly, PF|
|5||None||Top: FQ, PF||FQ, PF|
|6||None||Top: G, PF. Subconj: G, D. IV: T||V, G|
|7||None||Top: FQ, PF. Subconj: G, D||FQ, PF|
|8||None||Top: FQ, PF. Subconj: G, D||FQ, PF|
|9||FQ||G in irrigation||FQ, PF|
|10||None||Top: FQ, TO, D. Subconj: C, D||FQ, PF, NSAID|
|11||Poly||Top: N, P, D. Peribulbar: T||Poly, PF|
|12||None||Top: FQ, PF. Subconj: C, D||FQ, PF|
|13||None||Subconj: G, T||Poly, PF|
|14||None||Top: FQ, PF||FQ, PF, NSAID|
Clinical Presentation, Management, Microbiology, and Outcomes
At initial evaluation, 11 of the 14 patients (79%) noted increased redness, 12 (86%) noted decreased vision, and 10 (71%) complained of significant pain. Clinical examination revealed a hypopyon in 9 of the 14 eyes (64%), and 4 eyes (29%) had fibrin in the anterior chamber without a hypopyon. Presenting vision was 20/200 or better in 2 patients (14%), counting fingers to hand motions in 8 patients (57%), and light perception in 4 patients (29%) ( Table 3 ). The mean time from surgery to diagnosis with endophthalmitis was 11.4 days (SD 9.8 days, median 7 days, range 1-30 days), with 7 patients (50%) presenting within 1 week of surgery.
|Patient||Visual Acuity at Diagnosis||Management (Intravitreal Medications)||Visual Acuity at Follow-up (Months)|
|1||HM||PPV/PPL (V, C, D)||20/30 (12)|
|2||CF||PPV (V, C, D)||20/20 (20)|
|3||CF||T/I (V, C, D)||20/20 (26)|
|4||LP||PPV (V, C, D), repeat PPV (V, C, D) 13 d later||NLP|
|5||CF||T/I (V, C, D)||20/20 (36)|
|6||CF||T/I (V, C, D)||20/30 (6)|
|7||HM||T/I (V, C, D)||HM (12)|
|8||LP||T/I (V, C)||NLP|
|9||20/100||T/I (V, C)||20/25 (12)|
|10||HM||T/I (V, C)||20/200 (6)|
|11||LP||T/I (V, C), repeat I (V, D) 4 d later||20/200 (3)|
|12||HM||I (V, C)||HM (12)|
|13||LP||PPV/I (V, D), then T/I (V, D) 5 d later||LP (1)|
|14||20/200||PPV/I (V, C, D)||20/25 (1)|
As initial treatment, 8 patients (57%) underwent ocular paracentesis and intraocular injection of antibiotics, 5 patients (36%) underwent vitrectomy with intraocular injection of antibiotics, and 1 (7%) underwent intravitreal injection of antibiotics without ocular paracentesis because of hypotony ( Table 3 ). Intravitreal vancomycin and ceftazidime were given to 13 eyes (93%). Intraocular steroid (dexamethasone) was injected into 9 eyes (64%) as part of initial treatment at the discretion of the treating physician. A second intervention was performed in 3 patients: 1 underwent a second PPV (Patient 4) and 2 underwent a second intravitreal injection of vancomycin and dexamethasone (Patients 11 and 13).
Staphylococcus was isolated from 7 eyes (50%), including 6 cases (43%) of Staphylococcus epidermidis and 1 case (7%) of Staphylococcus aureus . Intraocular cultures were negative in 3 cases (21%). Other organisms isolated were Streptococcus pneumoniae (n = 1), Streptococcus mitis (n = 1), and Streptococcus agalactiae ( Table 1 ).
Patients were followed for a mean of 12 months (range 1–36 months; Table 3 ). Final vision ≥20/200 was achieved in 9 of the 14 eyes (64%), and 7 (50%) achieved ≥20/40. Two eyes (14%) were ultimately hand motions, 1 (7%) was light perception, and 2 (14%) were no light perception.
Comparison With Prior Endophthalmitis Data from the Same Institution
Endophthalmitis rates from 1984 to 1994 and from 1995 to 2001 are compared with the current data (2002–2009) in Table 4 . There is a statistically significant trend for a decreasing rate of endophthalmitis following all types of intraocular surgery, from 0.09% (1984–1994) to 0.05% (1995–2001) to 0.025% (2002–2009) ( P < .001). The rate of endophthalmitis following cataract surgery follows a similarly decreasing trend, from 0.08% (1984–1994) to 0.04% (1995–2001) to 0.028% (2002–2009) ( P = .002). No surgical category demonstrates an increasing trend over these time periods.
|Frequency of Postoperative Endophthalmitis (%) a|
|Surgical Procedure||1984–1994 b||1995–2001 b||2002–2009||P for Trend c|
|Cataract||34/41 654 (0.08)||8/21 972 (0.04)||8/28 568 (0.03)||.002|
|Pars plana vitrectomy||3/6557 (0.05)||2/7429 (0.03)||2/18 492 (0.01)||.09|
|Glaucoma||4/3233 (0.12)||4/1970 (0.20)||0/5041 (0)||.03|
|Penetrating keratoplasty||5/2805 (0.18)||2/2362 (0.08)||3/2788 (0.11)||.46|
|Secondary intraocular lens||5/1367 (0.37)||1/485 (0.21)||1/1783 (0.06)||.05|
|Total||54/58 123 (0.09)||17/35 916 (0.05)||14/56 672 (0.03)||<.001|