ness measurements across three spectral domain optical coherence tomography systems” by Branchini et al. 1 In this investigation, we compared choroidal thickness as mea- sured using 3 different spectral domain optical coherence tomography systems: Cirrus high-definition optical coher- ence tomography (HD-OCT) (Carl Zeiss Meditec Inc, Dub- lin, CA), Heidelberg Spectralis (Heidelberg Engineering, Heidelberg, Germany), and Optovue RTVue (Optovue Inc., Fremont, CA). We agree that intraclass correlation coefficient (ICC) is another measure of similarity. From our data, the ICCs are: Cirrus vs. Spectralis 0.98, Cirrus vs. RTVue 0.97, Spectralis vs. RTVue 0.97, all systems 0.98, P0.0001. These high levels of correlation can be interpreted in a similar manner as Pearson correlations in combination with the data from Table 1 which presents the average choroidal thickness by imaging system and location. Bland and Altman describe in their 1999 article, “Mea- suring agreement in method comparison studies” that if data when plotted on a raw scale shows increased variation with larger values, it is appropriate to use percent (%) difference scale for ease of interpretation. 2 Our data demonstrate this increased spread at higher values of choroidal thickness, which prompted us to use % difference scale in our Bland- Altman plot. With regard to the use of the word “Reproducibility” in our title, previous investigators have used similar titles with varying statistical methods. 3–5 As such, there is no standard definition in the literature for what reproducibility means in terms of statistical methods. We have made our statistical methods clear and have interpreted our results accordingly. JAY S. DUKER, MD LAUREN BRANCHINI, BA CAIO REGATIERI, MD, PHD Boston, Massachusetts References 1. Branchini L, Regatieri CV, Flores-Moreno I, et al. Reproduc- ibility of choroidal thickness measurements across three spec- tral domain optical coherence tomography systems. Ophthal- mology 2012;119:119 –23. 2. Bland JM, Altman DG. Measuring agreement in method com- parison studies. Stat Methods Med Res 1999;8:135– 60. 3. Krzystolik MG, Strauber SF, Aiello LP, et al. Reproducibility of macular thickness and volume using Zeiss optical coher- ence tomography in patients with diabetic macular edema. Ophthalmology 2007;114:1520 –5. 4. Miranda D, Smith SD, Krueger RR. Comparison of flap thick- ness reproducibility using microkeratomes with a second mo- tor for advancement. Ophthalmology 2003;110:1931– 4. 5. Yoneda T, Sumi T, Takahashi A, et al. Automated hyperemia analysis software: reliability and reproducibility in healthy subjects. Jpn J Ophthalmol 2012;56:1–7. Macular Degeneration Genetics Dear Editor: Sawitzke et al 1 reported a new method for assessing copy number of CFHR3 and CFHR1, a common deletion of which we were the first to identify, characterize, and report in a 2006 Nature Genetics paper that demonstrated that the deletion was associated with protection against age-related macular degeneration in heterozygosity and homozygosity. 2 Subsequently, several independent studies have replicated the association. The method presented by Sawitzke et al, 1 using propri- etary copy-number polymorphism probes, produced a 94% genotyping rate, 4% inconsistency with their previous method, and 1% internal inconsistency. The authors re- ported that the deletion of CFHR3 and CFHR1 was some- times found on the same haplotypes as the risk alleles of rs1061170 (Y402H) and rs1410996 (an intronic single nu- cleotide polymorphism [SNP]), which has not been previ- ously reported. As well as the 84.7 kb deletion of CFHR3 and CFHR1, 2 we and other investigators have identified a less common copy number polymorphism of 121.9 kb length, resulting in deletion of CFHR1 and CFHR4. CNP148 as shown in Figure 1 of Sawitzke et al 1 does not accurately reflect CFHR1-CFHR4, as it fails to show deletion of CFHR1. The correct positions of the 2 structural polymorphisms and the CFH-related genes are shown in Figure 1 (available online at http://aaojournal.org). Sawitzke et al used 4 probes to ascertain copy number of the polymorphic region: Probe 1, hs04211013cn, in CFHR3; Probe 2, hs04197581cn, in CFHR1; Probe 3, hs03356469cn, between CFHR1 and CFHR4; Probe 4, hs03345427cn, also between CFHR1 and CFHR4 (Figure 1, available online at http://aaojournal.org). 1 For parsimony, they chose to use only Probe 2 (in CFHR1) to determine copy number of CFHR3- CFHR1, but unfortunately, this probe is also located within the second structural polymorphism, CFHR1-CFHR4, and can- not discriminate between deletion of either CFHR3-CFHR1 or CFHR1-CFHR4. We believe that this mistake underlies the authors’ claim that they found CFHR3-CFHR1 on the same haplotype as the risk alleles of rs1061170 and rs1410996, which has not been observed in other cohorts. The method proposed by Sawitzke et al may be useful if the full data from all 4 Taqman probes was analyzed with reference to the known structural variations in the area. If the authors are interested only in determining carriage of CFHR3-CFHR1, Probe 1 is likely to prove to be a much more useful marker than Probe 2, as it is outside the second region of copy number polymorphism. The authors also made some contradictory statements that deserve clarification. They wrote that their previous method for assessing copy number of CFHR3-CFHR1 al- lowed them “to determine 0 or 2 copy status,” 1 but as the authors later make clear, that method, 3 like that of Spencer et al 4 did not allow determination of “2 copy status,” 1 but only a “not homozygous for the deletion” 4 status. These studies therefore failed to assess the true risk effect associ- ated with CFHR3-CFHR1. 3,4 The authors also wrote that CFHR1 and CFHR3 proteins “are involved in the terminal end of the complement pathway, inhibiting the cascade by interacting with C5. . .yet they do not have complement inhibiting properties.” 1 In addition to competition for CFH binding sites, CFHR1 and CFHR3 have regulatory func- tional activity in the complement system that is independent of interaction with CFH. 5 Letters to the Editor 1287