Pseudomonas aeruginosa–Related Postoperative Endophthalmitis Linked to a Contaminated Phacoemulsifier

Cataract extraction is a commonly performed operation in the United States, with 1.3 million procedures performed annually. Postoperative endophthalmitis is a potentially devastating complication, with an incidence of approximately 0.1% to 0.3%.^1,2 We report herein an evaluation of an outbreak of postoperative endophthalmitis.

This outbreak was investigated by the North Carolina Statewide Program for Infection Control and Epidemiology. Patient cultures were performed in a hospital microbiology laboratory. Environmental cultures were processed as follows. Medication vials and water samples were processed by filtration using a disposable filter (MSI Savur Analytical filter; Osmonics, Minnetonka, Minn) placed on either sheep blood or D/E-neutralizing agar. Fluids too viscous for filtration (eg, Keri Lotion) were plated directly onto suitable media. Environmental surfaces (eg, sink drains) were cultured using a sterile swab premoistened with trypticase soy broth and then plated directly onto appropriate agar.

To determine the relatedness of Pseudomonas aeruginosa isolates, we performed pulsed-field gel electrophoresis using standard techniques.

The North Carolina Statewide Program for Infection Control and Epidemiology was notified on October 14, 1999, about a cluster of cases of postoperative endophthalmitis. Preliminary data obtained via telephone revealed that 5 of 7 patients who underwent a cataract extraction with phacoemulsification and implantation of an intraocular lens on October 11, 1999 (outbreak day), had developed postoperative endophthalmitis (Table 1). Patients were seen postoperatively with severe pain in the involved eye, conjunctival hyperemia, and decreased visual acuity.

All infected patients had undergone cataract extraction in the right eye, and all noninfected patients had undergone cataract extraction in the left eye (P = .048, 2-tailed Fisher exact test). All cataract procedures were performed using automated phacoemulsification with a phacomachine (AMO Diplomax; Allergan Inc, Irvine, Calif) that used a peristaltic pump, followed by implantation of a Starr foldable silicone intraocular lens. All patients received 0.3% ofloxacin drops, 3 times per day, postoperatively.

Following notification of this outbreak, we recommended that no additional operations be performed. Multiple environmental cultures were obtained on days 9 and 14 after the outbreak (Table 2). The P aeruginosa isolates from 4 patients and 3 environmental sites were compared using pulsed-field gel electrophoresis (Figure 1). All patient isolates and the strain isolated from the phacoemulsifier's internal tubing were identical. The strains of P aeruginosa isolated from the scrub sink and its drain were identical, but differed by multiple bands from the patient strains.

The fluid pathways of the phacoemulsifier are shown in Figure 2A. The outflow tubing is connected via a T fitting to a pressure transducer and air valve system that is housed within the machine (Figure 2B). Using methylene blue, we investigated conditions that might lead to retrograde flow from the internal tubing to the handpiece. Disrupting the sealed connection of the internal tube from the pressure transducer was effective in producing almost immediate reflux that reached the handpiece within 2 minutes.

Postoperative endophthalmitis is a potentially devastating complication of cataract surgery, with an incidence of 0.08% to 0.30%.^1,2 Multiple P aeruginosa–related outbreaks of postoperative endophthalmitis following cataract extraction are linked to use of contaminated multidose vials or irrigating solutions and to an intrinsically contaminated intraocular lens; recently, 2 European outbreaks were linked to use of a contaminated phacoemulsifier.^3,4

We demonstrated that all infecting strains of P aeruginosa were identical by pulsed-field gel electrophoresis. Patient 1's culture did not yield P aeruginosa, but only the aqueous humor was cultured. In patients with endophthalmitis, cultures of the aqueous humor are less sensitive than cultures of the vitreous. We believe that the additional organisms cultured from patient 3 represented contaminants.

We believe that the phacoemulsifier was the source of this outbreak. All patients who underwent an operation on their right eye developed infection, while all patients who underwent an operation on their left eye were uninfected. The operating room was set up in such a way that the phacoemulsifier was placed on the patient's right side. Thus, when the left eye was operated on, the tubing for the phacoemulsifier was stretched more taut. This may have led to gravity backflow of contaminated fluid during operations on the right eye and might explain the significantly different infection rates between left and right eyes.

Contamination of the internal tubing of 3 different models of phacoemulsifiers that used a peristaltic pump was first demonstrated in 1986.⁵ Under certain vacuum settings, air was introduced into the aspiration catheter from the machine, and during this venting process, previously sequestered fluids in the machine could be regurgitated back into the aspiration catheter, resulting in potential transmission of pathogens. de Kasper and colleagues⁶ investigated the contamination rates of automated evacuation systems equipped with an internal vacuum control manifold compared with a system equipped with a modified external vacuum control manifold that was sterilized between patients; 2+ to 4+ bacterial growth was found in all specimens from the internal vacuum control manifold system. They emphasized that "the observed contamination of intraocular surgery machines is not related to a specific company but must be attributed to the technical constituent of an inbuilt and inaccessible internal vacuum control manifold that exists in a wide variety of types of automated intraocular surgery machines of several producers."⁶(p690)

Using methylene blue, we demonstrated the possibility of retrograde flow when the internal tubing was disconnected from the transducer. Although the machine would not pass normal preoperative checks with this malfunction, it seems plausible that a small leak might have a similar or intermittent result, and the machine could still pass the check yet lead to retrograde flow. This might occur with a leak at one of the fittings, a small defect in the transducer, or an intermittently sticking air valve.

To our knowledge, this represents the third outbreak of postoperative endophthalmitis due to P aeruginosa linked to use of a contaminated phacoemulsifier and the first report from North America. All manufacturer's recommendations regarding machine maintenance were followed. These 3 outbreaks linked to a contaminated phacoemulsifier suggest that this device may need to be redesigned to reduce the likelihood of contamination or that additional disinfection methods may need to be routinely used.

None of the authors has any financial interest in the company that manufactured the phacoemulsifier described in this article.

We thank Marlene Durand, MD, for reviewing the manuscript.

Corresponding author and reprints: David J. Weber, MD, MHA, MPH, University of North Carolina at Chapel Hill, Campus Box 7030, Burnett-Womack, Room 547, Chapel Hill, NC 27599-7030 (e-mail: dweber@unch.unc.edu).