Article Title: Complement factors and reticular pseudodrusen in intermediate age-related macular degeneration staged by multimodal imaging BMJ Open Ophthalmology Jan 2020, 5 (1) e000361; DOI: 10.1136/bmjophth-2019-000361
E-letter sender: Anne M Lynch, Brandie D Wagner, Alan G Palestine, Jennifer L Patnaik, Ashley A Frazier-Abel, Marc T Mathias, Frank S Siringo, V. Michael Holers, Naresh Mandava
E-Letter Title: Response to letter entitled “The Controversial Role of Inhibitory Complement Factors in Age-Related Macular Degeneration”
February 12, 2020
Dear Editor,
We appreciate the comments from Shiwani et al. regarding our manuscript entitled “Complement factors and reticular pseudodrusen in intermediate age-related macular degeneration staged by multimodal imaging.” As reported in the manuscript, (1) we found significantly elevated levels of C3a, Ba, C5a and C2 in cases with the intermediate phenotype of age-related macular degeneration (AMD) compared with cataract controls without AMD. We also found significantly lower levels of Factor B, Factor H, Factor I, C1q, iC3b/C3b, and C5 in the AMD cases compare...
Article Title: Complement factors and reticular pseudodrusen in intermediate age-related macular degeneration staged by multimodal imaging BMJ Open Ophthalmology Jan 2020, 5 (1) e000361; DOI: 10.1136/bmjophth-2019-000361
E-letter sender: Anne M Lynch, Brandie D Wagner, Alan G Palestine, Jennifer L Patnaik, Ashley A Frazier-Abel, Marc T Mathias, Frank S Siringo, V. Michael Holers, Naresh Mandava
E-Letter Title: Response to letter entitled “The Controversial Role of Inhibitory Complement Factors in Age-Related Macular Degeneration”
February 12, 2020
Dear Editor,
We appreciate the comments from Shiwani et al. regarding our manuscript entitled “Complement factors and reticular pseudodrusen in intermediate age-related macular degeneration staged by multimodal imaging.” As reported in the manuscript, (1) we found significantly elevated levels of C3a, Ba, C5a and C2 in cases with the intermediate phenotype of age-related macular degeneration (AMD) compared with cataract controls without AMD. We also found significantly lower levels of Factor B, Factor H, Factor I, C1q, iC3b/C3b, and C5 in the AMD cases compared with controls.
All of our complement results were corrected for multiple comparisons testing using the false-discovery rate (FDR). The FDR approach to multiple comparisons is a commonly used method and is typically less conservative than a Bonferroni correction. The Bonferroni adjustment controls the familywise type I error rate, whereas the FDR controls the type I error rate only amongst those null hypotheses that were rejected. Both approaches assume complete independence between tests and are therefore more conservative than needed given the tests for complement factors are likely correlated (2). In response to the questions raised by Shiwani et al, we completed the analysis using the Bonferroni statistical correction. Following the application of the Bonferroni statistical correction, we found C2 was no longer significant and Factor B was now of borderline statistical significance (see Table 1).
We agree with Shiwani et al that the low levels of the complement inhibitors need further investigation and as stated in the manuscript, we would hope to explore this finding further when we examine genetic polymorphisms of the complement pathways in this cohort. We also hope that moving forward there will be more investigation of the role of the complement system in patients with the early or intermediate phenotype of AMD so that interventions can be introduced before the patient develops the advanced stages of AMD.
References
1. Lynch AM, Palestine, A.G, Wagner, B.D, Patnaik, J.L, Frazier-Abel, A.A, Mathias, M.T, Siringo, F.S, Holers, V.M, Mandava, N. Complement factors and reticular pseudodrusen in
intermediate age-related macular degeneration staged by multimodal imaging. BMJ Open Ophthalmology Jan 2020, 5 (1) e000361; DOI: 10.1136/bmjophth-2019-000361
2. Westfall, P. and S. Young (1993). Resampling Based Multiple Testing. New York, John Wiley and Sons, Inc.
Table 1. Complement Levels for Cases with Intermediate AMD Versus Controls.
* OR corresponding to change in 10 units of complement factor
†OR = Odds ratio from the multivariable logistic regression with 95% Confidence Intervals (CI) after adjusting for age and family history of AMD.
‡ P-values obtained from multivariate logistic regression including age and family history of AMD and adjusted for multiple comparisons using FDR
§ P-values obtained from multivariate logistic regression including age and family history of AMD and adjusted for multiple comparisons using the Bonferroni statistical correction
We read with interest the role of complement in age-related macular degeneration (AMD) explored by Lynch and colleagues1. The authors have expertly examined the numerous components of the complement system and provided much novel data, particularly in the context of intermediate AMD.
Their results are congruent with that of the literature, especially with regard to lower levels of C3a, C5a and Ba in the setting of AMD2. Furthermore, the results of the FILLY trial have shown positive effects of the inhibition of C3 in limiting geographic atrophy and thus lend an element of support to the conclusions of this study3. We hope that the subsequent phase 3 trials – DERBY and OAKS – will lend further support and allow research such as this to translate to clinical benefit for patients4.
However, we do feel that there may be an element of multiple hypothesis testing evident in this study that increases the risk of generating some erroneous results. We feel a Bonferroni statistical correction would have been an appropriate tool to add further robustness to the positive findings that have been demonstrated. We especially feel that the results relating to the inhibitory factors of the complement system do not tend to correlate with some of the findings in the literature. Although Lynch et al have identified lower levels of Factor B, Factor H and Factor I in the AMD arm of the study, there are numerous reports in the literature reporting the opposite...
We read with interest the role of complement in age-related macular degeneration (AMD) explored by Lynch and colleagues1. The authors have expertly examined the numerous components of the complement system and provided much novel data, particularly in the context of intermediate AMD.
Their results are congruent with that of the literature, especially with regard to lower levels of C3a, C5a and Ba in the setting of AMD2. Furthermore, the results of the FILLY trial have shown positive effects of the inhibition of C3 in limiting geographic atrophy and thus lend an element of support to the conclusions of this study3. We hope that the subsequent phase 3 trials – DERBY and OAKS – will lend further support and allow research such as this to translate to clinical benefit for patients4.
However, we do feel that there may be an element of multiple hypothesis testing evident in this study that increases the risk of generating some erroneous results. We feel a Bonferroni statistical correction would have been an appropriate tool to add further robustness to the positive findings that have been demonstrated. We especially feel that the results relating to the inhibitory factors of the complement system do not tend to correlate with some of the findings in the literature. Although Lynch et al have identified lower levels of Factor B, Factor H and Factor I in the AMD arm of the study, there are numerous reports in the literature reporting the opposite results where levels of these factors are higher in AMD patient populations, albeit not in settings with identical severity of disease2,5,6.Whether the results presented in this study constitute a true finding in the setting of intermediate AMD or whether this represents a statistical type 1 error, it would be difficult to ascertain.
Overall, this study shines more light on this intricate topic with clinically significant results. We hope this will stimulate further research and discussion in the field.
References
1. Lynch AM, Palestine AG, Wagner BD, et al. Complement factors and reticular pseudodrusen in intermediate age-related macular degeneration staged by multimodal imaging. BMJ Open Ophthalmol 2020;5:e000361. doi:10.1136/bmjophth-2019-000361
2. Scholl HPN, Issa PC, Walier M, et al. Systemic Complement Activation in Age-Related Macular Degeneration. Toland AE, ed. PLoS One 2008;3:e2593. doi:10.1371/journal.pone.0002593
3. Apellis Pharmaceuticals. Apellis Pharmaceuticals Announces That APL-2 Met Its Primary Endpoint in a Phase 2 Study in Patients with Geographic Atrophy, an Advanced Form of Age-Related Macular Degeneration Statistically Significant Slowing of Disease Progression Seen at 12 Months. 2017. Accessed January 26, 2020.
4. Apellis Pharmaceuticals. Recruiting patients with geographic atrophy secondary to AMD for phase 3 trials | Apellis APL-2 Opthalmic Trial. . Accessed January 26, 2020.
5. Hecker LA, Edwards AO, Ryu E, et al. Genetic control of the alternative pathway of complement in humans and age-related macular degeneration. Hum Mol Genet 2010;19:209-215. doi:10.1093/hmg/ddp472
6. Silva AS, Teixeira AG, Bavia L, et al. Plasma levels of complement proteins from the alternative pathway in patients with age-related macular degeneration are independent of Complement Factor H Tyr402His polymorphism. Mol Vis 2012;18:2288-2299.
Dear Editor,
We read with interest the role of complement in age-related macular degeneration (AMD) explored by Lynch and colleagues1. The authors have expertly examined the numerous components of the complement system and provided much novel data, particularly in the context of intermediate AMD.
Their results are congruent with that of the literature, especially with regard to lower levels of C3a, C5a and Ba in the setting of AMD2. Furthermore, the results of the FILLY trial have shown positive effects of the inhibition of C3 in limiting geographic atrophy and thus lend an element of support to the conclusions of this study3. We hope that the subsequent phase 3 trials – DERBY and OAKS – will lend further support and allow research such as this to translate to clinical benefit for patients4.
However, we do feel that there may be an element of multiple hypothesis testing evident in this study that increases the risk of generating some erroneous results. We feel a Bonferroni statistical correction would have been an appropriate tool to add further robustness to the positive findings that have been demonstrated. We especially feel that the results relating to the inhibitory factors of the complement system do not tend to correlate with some of the findings in the literature. Although Lynch et al have identified lower levels of Factor B, Factor H and Factor I in the AMD arm of the study, there are numerous reports in the literature reporting the opposite...
Show MorePages