Statistics from Altmetric.com
We read the article by Jonas et al on intravitreal triamcinolone injections for exudative age related macular degeneration with interest.1 The paper stated that visual acuity increased significantly (p<0.001) from 0.16 (SD 0.11) to a mean maximum of 0.23 (0.17). The authors therefore picked the best from one of up to 10 postoperative visual acuity measurements and compared it with a single preoperative visual acuity measurement. This is misleading the reader regarding the true effectiveness of the treatment.
The Macular Photocoagulation Study Group found that the differences in between two repeated tests were one line or more in 13% of cases and the differences were greatest in patients with visual acuity of 20/100 or worse.2 By taking up to 10 postoperative measurements, Jonas et al greatly increased the chances of a positive result. The difference between mean pre-injection 0.16 (20/125 or 6/36) and best mean postoperative 0.23 (20/87 or 6/26) was less than one line on the Snellen chart.
Their table 1 gave the mean visual acuity pre-injection and at various time intervals post-injection. At 1 and 2 months, the p values were 0.04. It was unclear whether the p values were one or two tailed but both were described as not significant (NS) in table 1. Multiple significance testing at each of a number of time points is generally not recommended—if it is done, some kind of adjustment to the p values is needed.3,4 Looking at the results presented in table 1, readers might conclude that triamcinolone had a transient and doubtful beneficial effect on the visual acuity.
The authors go on to further analyse the results into improvements of three and six or more lines. The vision was tested on a Snellen chart which has irregular steps. Three or six lines do not therefore represent a constant change in visual angle (as in a logMAR chart) and therefore the analysis was confusing.
Variations in intraocular pressure of 5 or 6 mm Hg occur diurnally in normal individuals as well as glaucomatous patients.5–7 While there is little doubt that triamcinolone may affect the intraocular pressure, the comparison of the baseline with the highest value (p<0.001) was misleading as was the comparison of the highest value with that at 7 months (p<0.001). Of more interest might be the number of patients who had very high levels (the range extended to 64 mm Hg) and whether these very high intraocular pressures responded to treatment.
The authors’ experience in using triamcinolone is well recognised. We congratulate them on an otherwise excellent piece of work.
We thank Wong and colleagues very much for their interest in our study.1 We agree with them that comparing a single preoperative measurement with the best out of a series of postoperative measurements gives a tendency towards a falsely high increase in visual acuity after the triamcinolone acetonide injection. That there was an increase in visual acuity after the injection in some patients, however, may have been demonstrated in table 1 giving the visual acuity measurements before and after the application of triamcinolone acetonide. At 1 and 2 months, the difference from the preoperative values was significant with a p value of 0.04. Unfortunately, the values are described as non-significant in what is a typographical error in the manuscript. We regret this error. The authors agree with Wong and colleagues, that the effect of triamcinolone acetonide is temporary, and that repeated injections may be necessary. In some patients, repeated intravitreal injections of triamcinolone acetonide were associated with repeated re-increase in visual acuity.2
We also agree with Wong and coworkers that comparing the preoperative intraocular pressure measurement with the highest value during the follow up again falsely increased the number of ocular hypertensive subjects after the injection. In this matter, however, it was not in favour of the treatment since it artificially increased the number of patients with unwanted side effects. Concerning the number of patients with very high intraocular pressures after the injection, a preliminary analysis shows that about 1% of the eyes injected will need filtering surgery (own data).
The authors would like to thank Wong and colleagues for constructively commenting on the study and pointing out, as did the authors, that the intravitreal injection of triamcinolone acetonide as treatment for exudative macular degeneration is still in clinical evaluation and that its therapeutic effect has not been proved so far. This may be even more important as a recent randomised study using 4 mg triamcinolone acetonide did not demonstrate an effect on the risk of loss of visual acuity during the first year of the study in eyes with classic choroidal neovascularisation.3 A significant anti-angiogenic effect, however, was found 3 months after treatment. In view of the anti-angiogenic, anti-inflammatory, and anti-oedematous effects of intravitreal triamcinolone, further randomised trials using different doses of triamcinolone acetonide in eyes with different types of subfoveal neovascularisation may be warranted.