Editor, W e thank Ziemssen et al. (2007) for their interest and valuable comments on our recent article (Kernt et al. 2007). We agree completely that when using intravitreal anti-vascular endothelial growth factor drugs (VEGF) – especially ‘off-label’– one needs to be aware of possible systemic side-effects. In late 2005, when we started collecting data for our study, few reports on treatment by intravitreal bevacizumab existed and no evidence on safety or possible side-effects with regard to intraocular (IOP) or blood pressure (BP) was available. In order to improve patient safety in this situation, our study was designed to collect relevant safety data in a consecutive small patient group. We apologize for the limited design in terms of patient numbers but do believe, because of the completed follow-up, that our data offer a valuable contribution to the safety of intravitreal bevacizumab. We wish to emphasize our complete agreement on the need for sufficient BP management. This is especially true for patients such as those in our reported patient group who suffer from neovascular age-related macular degeneration (AMD) or from diabetic retinopathy (DR), both usually with significant comorbidity. Systemic hypertension is without doubt one of the major causes for morbidity and mortality in the Western world (Ramsay et al. 1999). Therefore, in our inand out-patients we emphasize tight BP control. Being an ophthalmological referral centre this means frequent referrals to internal medicine or the general practitioner treating the patient. The data shown in our study as ‘baseline’ prove that in many patients BP control remains unsatisfactory, although most patients were on at least one form of BP medication. No changes in BP medication occurred during follow-up, data that were not presented because of space limitations and the focus on IOP in an ophthalmological publication. A question was also raised about whether intraindividual BP variability and diurnal oscillations were considered. (The same is also true for IOP.) Baseline measurements of BP and IOP were always taken at hospitalization (approximately 10:00 hr); the intravitreal bevacizumab injection was performed between 14:00 hr and 15:00 hr. Therefore, a similar diurnal cycle-phase can be assumed for all patients. The BP data were indeed distributed normally; intraindividual information was used by anova analysis. Posthoc paired testing was corrected for multiple testing. In the first 6 hr after injection, BP measurements were meant to accompany IOP measurements and to accustom patients with the procedure rather than to detect any change. All patients were hospitalized for at least 1 day at the time of the study, making any return visits unnecessary for the first 24 hr. From our point of view, the out-patient followup data at 2, 6 and 12 weeks are more relevant because a significant increase in BP after systemic bevacizumab treatment is known to occur after several weeks (Michels et al. 2005; Hurwitz & Saini 2006). We agree with Ziemssen et al. (2007) that the ‘baseline’ measurement may be influenced by stress and several other factors such as diurnal changes. However, the first BP measurement was taken several hours before the surgical treatment. In most single BP measurements – including those taken at home – the physician or hospital has to cope with a similar bias. Moreover, the variance analysis can detect any changes over time and is not dependent on only one baseline. On multiple posthoc paired testing, any changes occurring after several weeks would be clinically relevant, while the observed early decreases in BP may represent some of the effects already discussed. In summary, patient safety is a major concern, especially when the treatment is ‘off-label’ and must include all possible side-effects. Therefore, we thank Ziemssen et al. (2007) for their comments regarding BP and the opportunity to further illustrate some previously unaddressed points. We believe that the limited number of patients included is the main limitation of our study and therefore relatively rare effects cannot be assessed. Nevertheless, we believe that our data show validly that a significant increase in IOP or BP does not occur with intraocular bevacizumab. Still further investigations with larger patient numbers are needed to investigate any relatively rare complications and to assess patient safety completely.