Intravitreal air bubbles following intravitreal injections: a comprehensive analysis

Mark Krauthammer, Etty Trabelsi, Elad Moisseiev

Research output: Contribution to journalArticlepeer-review


Purpose: To evaluate the prevalence of residual air bubbles following intravitreal injections of anti-VEGF agents, and to compare two techniques designed to reduce their occurrence. Methods: Patients who received intravitreal injections reported the presence of air bubbles following the procedure, and were followed to determine when they disappeared. Two techniques used to reduce air bubbles prior to injection were compared—tapping the syringe with the needle up (“upwards” technique) or down (“downwards” technique). Rates of residual air bubbles were compared between techniques, and between different drugs. Results: The study included 344 intravitreal injections, 172 injected with each technique. The overall rate of residual air bubbles was 11.3%, with 94.9% resolution by 48 h. The rate was significantly lower with the “downwards” technique (7.5% vs. 15.1%, p = 0.027). It was also significantly lower with ranibizumab injected using pre-filled syringes than with bevacizumab and aflibercept (0% vs. 12.1% and 14.7%, p < 0.0001). A questionnaire revealed patients reported medium levels of discomfort and a high importance of avoiding air bubbles. Conclusions: Residual air bubbles are a common occurrence, likely to be experienced by most patients undergoing repeated injections. This phenomenon may be significantly reduced by using the described “downwards” technique, or pre-filled syringes. [Figure not available: see fulltext.]

Original languageEnglish
Pages (from-to)3697-3702
Number of pages6
JournalGraefe's Archive for Clinical and Experimental Ophthalmology
Issue number12
StatePublished - Dec 2021


  • Air bubbles
  • Anti-VEGF
  • Intravitreal injection
  • Syringe preparation


Dive into the research topics of 'Intravitreal air bubbles following intravitreal injections: a comprehensive analysis'. Together they form a unique fingerprint.

Cite this