CASE REPORT

A 68-year-old Filipino woman with well controlled adult onset diabetes mellitus underwent uncomplicated extracapsular cataract extraction by phacoemulsification followed by placement of a posterior chamber intraocular lens. The skin and eye were prepared with povidone-iodine preoperatively. The patient failed to return for follow up for 3 days, at which time she presented with hand movement vision, marked corneal oedema, and a 4+ anterior chamber cell and flare with hypopyon (Hogan, Kimura, and Thygeson grading⁷). The intraocular pressure was normal. The fundus was not visible but ultrasound failed to demonstrate vitritis. The diagnosis of early postoperative endophthalmitis was made, and anterior chamber and vitreous aspirates were taken for Gram stain and culture. The patient was then treated with: (1) intravitreal vancomycin (1 mg/0.1 ml), ceftazidime (2.25 mg/0.1 ml), and dexamethasone (0.4 mg/0.1 ml); (2) subconjunctival vancomycin (25 mg/0.5 ml) and ceftazidime (100 mg/0.5 ml); (3) intravenous vancomycin (500 mg, every 6 hours) and ceftazidime (1000 mg, every 6 hours); and (4) topical hyoscine (scopolamine) (0.25%, one drop, four times a day) and prednisolone acetate (1%, one drop, every hour). Initial Gram stains were negative. On the sixth day after cataract surgery the streak culture of the anterior chamber aspirate grew a few M morganii colonies sensitive to ceftazidime, ciprofloxacin, and tobramycin (automated identification and sensitivity testing by Microscan, Dade International Inc, West Sacramento, CA, USA). The intravenous vancomycin was discontinued and the patient was placed on fortified, topical tobramycin (14 mg/ml, one drop, every 2 hours). The patient was discharged on the eighth day after cataract surgery on topical, fortified tobramycin (14 mg/ml, one drop, every 2 hours), prednisolone acetate (1%, one drop, every 2 hours), and hyoscine (0.25%, one drop, four times per day), and a 7 day course of oral ciprofloxacin (250 mg, four times a day). Examination just before discharge revealed a vision of brisk hand movement, moderate corneal oedema, and 2+ cell and flare in the anterior chamber with no residual hypopyon. The fundus remained poorly visible.

Over the 2 months following surgery, the patient’s vision gradually improved to 20/80, and the anterior chamber reaction decreased to less than 1+ cell and flare. Antibiotics were discontinued and topical prednisolone acetate 1% was tapered to one drop, four times a day. As the view of the fundus improved, scattered intraretinal haemorrhages were visualised, and thought to represent the well recognised phenomenon of accelerated background diabetic retinopathy following cataract surgery.8 9

Two weeks after examination, when she was found to have a visual acuity of 20/80 and 10 weeks after cataract surgery, the patient presented with counting fingers vision and a 4+ anterior chamber reaction with a small hypopyon, having 2 days earlier abruptly discontinued her topical prednisolone acetate drops. The posterior pole was again poorly visualised. Both rebound uveitis and recurrent endophthalmitis were considered as possibilities. However, given the need for prompt and aggressive treatment of endophthalmitis, the decision was made to perform a pars plane vitrectomy for Gram stain, culture, and intravitreal delivery of vancomycin and ceftazidime, antibiotics chosen for broad spectrum coverage in the event of infection with a second, perhaps initially unidentified, organism. In addition, subconjunctival and intravenous vancomycin and ceftazidime, as well as topical prednisolone acetate, were also given, all at the doses used on initial presentation more than 10 weeks earlier. No capsular abscess or plaque was seen. All Gram stains and cultures were negative. Two weeks after the vitrectomy the patient developed a non-clearing vitreous haemorrhage and was noted to have rubeosis iridis sparing the angle. Intraocular pressure remained normal. A second pars plana vitrectomy was performed revealing intraretinal haemorrhages in all four quadrants consistent with a central retinal vein occlusion. Endolaser panretinal photocoagulation was given intraoperatively. Following surgery and for the ensuing 7 months, the patient’s vision remained poor at counting fingers, despite complete regression of rubeosis and clear media.

COMMENT

Morganella morganii is a rare but usually devastating cause of postoperative endophthalmitis.2-6 In those few reported cases the visual outcome has been poor, the one exception being the patient reported by Smolin,1 who, like our patient, received early and aggressive intervention. However, unlike Smolin’s patient, our patient developed an acute increase in anterior chamber inflammation after abruptly discontinuing her topical prednisolone acetate. While there are many possible causes of delayed, postoperative intraocular inflammation,10 the most likely causes in our patient were recurrent endophthalmitis, either due toM morganii or perhaps a second, initially unidentified, organism, and rebound uveitis in response to the patient abruptly discontinuing topical prednisolone acetate. Given the possibility of endophthalmitis and the importance of prompt treatment in this setting, we chose early pars plana vitrectomy for Gram stain, culture, and to deliver intravitreal antibiotics. Since the presentation and treatment of our patient, the results of the Endophthalmitis Vitrectomy Study (EVS) Group, a randomised trial of immediate pars plana vitrectomy and systemic antibiotics for the treatment of postoperative bacterial endophthalmitis,11 have suggested that systemic antibiotics as used in our patient may not be necessary, and that immediate vitrectomy appears to add no benefit unless the presenting visual acuity is down to light perception. However, the patients reported in this study all had acute, postoperative endophthalmitis, and so these results may not be applicable to the type of delayed, postoperative endophthalmitis considered in our patient.

Most organisms implicated in postoperative endophthalmitis exist as normal ocular flora. In contrast, M morganii is an uncommon isolate from normal ocular flora. Okumoto and associates¹² isolated Proteus species, including M morganii and M mirabilis, from 26 of 1000 (2.6%) normal preoperative human eyes, of which five (0.5%) were M morganii. In fact, M morganii is most typically enteric and, as such, is often associated with urinary tract infections. However, our patient denied urinary or constitutional symptoms suggesting either a urinary tract infection or bacteraemia.

Our patient ultimately developed a vitreous haemorrhage and was noted to have rubeosis 10 weeks after her initial presentation with endophthalmitis. Scattered intraretinal haemorrhages were seen after cataract surgery, and were attributed to accelerated, background diabetic retinopathy. In retrospect, however, these haemorrhages would seem most consistent with a mild central retinal vein occlusion, possibly at the time of the initial endophthalmitis, leading 10 weeks later to neovascularisation of the retina and iris.

In summary, M morganii is an uncommon cause of postoperative endophthalmitis. Prompt recognition and treatment of postoperative endophthalmitis, whether due to M morganiior other more common organisms, can lead to good visual recovery, as observed initially in our patient.