Endophthalmitis Rates Higher With Alcohol-Based Chlorhexidine Than Povidone-Iodine

Endophthalmitis is more likely to occur with use of alcohol-based chlorhexidine (CH) compared with povidone-iodine (PI) as an antiseptic in intravitreal injections (IVI), according to a study published in Ophthalmology Retina. This study is the first to calculate endophthalmitis rates following the use of alcohol-based CH for antisepsis for IVI.

The investigators retrospectively reviewed endophthalmitis rates from patient records at a Canadian group retina practice. Patients (n=14,395) had either received IVI (n=170,952) between May 2019 and October 2022 with 10% PI (n=139,817) or .05% alcohol-based CH (n=31,135) as antiseptic, applied to the ocular surface at the injection site for 30 seconds before injection. Patients were older in the CH group compared with the PI group (mean age, 77.9 years vs 75 years).

Prior to administering CH, doctors had determined that those patients could not tolerate povidone-iodine. The medications administered through IVI included anti-vascular endothelial growth factor (VEGF), steroid, methotrexate, and foscarnet. Patients typically had age-related macular degeneration, diabetic retinopathy, or retinal vein occlusions.

Endophthalmitis occurred in 49 cases (29 PI 20 CH). Four in 5 of these cases occurred following aflibercept, while the rest were after bevacizumab (3.48 odds 95% confidence interval (CI) 2.09-7.24). Patients were 3.1 times more likely to develop endophthalmitis after IVI if they had CH antisepsis instead of PI antisepsis. (95% CI 1.9-5.2). 

Culture positivity presented in just under half of cases. The culture positive group had more loss of ETDRS letters compared with the culture negative group. There was more loss of letters in the oral flora group compared with the culture positive group without oral flora. Staphylococcus epidermidis was the most common cause of endophthalmitis.

“Clinically, injecting physicians should utilize PI for antisepsis and only consider changing to alcohol-based CH if intolerance/’allergy’ develops,” the study authors suggest. “Patients with PI intolerance should be informed of the increased risk of alcohol-based CH antisepsis and be given the option of either continuing with PI antisepsis or switching to alcohol-based CH antisepsis. Another consideration would be switching from 10% PI to the more commonly used 5% PI formulation which has been shown to have a similar rate of post IVI endophthalmitis.”

Limitations of the study include its low number of patients who received CH and higher mean number of injections in the CH group compared with the PI group.