Assessing The Relationship Between High-sensitivity C-reactive Protein And Kidney Function Employing Mendelian Randomization in The Japanese Community-based J-MICC Study

RESULTS 

Table 1 shows the basic characteristics of CRP and eGFR datasets. The mean ages of participants were not significantly different between the CRP dataset (55.5; standard deviation [SD], 9.6) and the eGFR dataset (55.1; SD, 9.2), and almost half of the subjects were women in both datasets (CRP: 64.7% and eGFR: 53.4%). Median and interquartile range [IQR] of hs-CRP levels and eGFR was 0.04 mg=dL (IQR, 0.02–0.08) and 77.2 mL=min=1.73 m2 (IQR, 68.7–86.7).

CLICK HERE TO GET NATURAL ORGANIC CISTANCHE EXTRACT WITH 25% ECHINACOSIDE AND 9% ACTEOSIDE FOR KIDNEY FUNCTION

Supportive Service Of Wecistanche-The largest cistanche exporter in the China:

Email:wallence.suen@wecistanche.com 

Whatsapp/Tel:+86 15292862950

Shop For More Specifications Details:

https://www.xjcistanche.com/cistanche-shop

Associations between instrumental variables and baseline hs-CRP 

Two SNPs (rs3093077 and rs1205) in IVCRP were significantly associated with ln(hs-CRP), but not for the other two SNPs (rs1130864 and rs1800947) (Table 2). All three SNPs in IVAsian were associated with ln(hs-CRP). Combining these SNPs, both IVCRP and IVAsian had an F-statistic >10 (14.8 and 22.5, respectively), which indicated that two IVs met a criterion for relevance assumption. Four SNPs in IVCRP and three SNPs in IVAsian explained 3.4% and 3.9% of the variation in hs-CRP, respectively. 

Conventional analysis for the association between hs-CRP and eGFR 

Before the MR analysis, we performed the conventional statistical analysis for the cross-sectional association between hs-CRP and eGFR. Among 1,598 participants who were available on both hs-CRP and eGFR, ln(hs-CRP) was strongly associated with ln(eGFR) (β = −0.015; 95% confidence interval [CI], −0.024 to −0.007; P = 3.26 × 10−4 ), after adjustment for sex, age, and study sites. The scatter plot for the association between ln(hs-CRP) and ln(eGFR) is shown in figure 2.

Two-sample 

MR analysis Using the IVCRP, genetically determined hs-CRP was not significantly associated with eGFR in the IVW method (estimate per 1 unit increase in ln(hs-CRP) = 0.000; 95% CI, −0.019 to 0.020; P = 0.97) (Figure 1, black blocks). Consistent with the result in the IVW method, no causal relationship was found in the WM method (estimate per 1 unit increase in ln(hs-CRP) = −0.003; 95% CI, −0.019 to 0.014, P = 0.77) and the MR-Egger method (estimate per 1 unit increase in ln(hs-CRP) = −0.008; 95% CI, −0.058 to 0.042; P = 0.75). The intercept estimated in the MR-Egger method was likely to be zero (estimate, 0.003; 95% CI, −0.011 to 0.016; P = 0.71). The scatter plot using IVCRP is provided as eFigure 3A

Figure 1. The results of a two-sample MR study using two different IVs (IVCRP and IVAsian). Effect sizes on the causal relationship between hs-CRP (X) and eGFR (Y) were estimated in three different MR methods. Estimates on eGFR are shown per 1 unit increase of log(hs-CRP). Red blocks and solid lines indicate the estimates and 95% CI using IVAsian, while black blocks and solid lines indicate the estimates and 95% CI using IVCRP. CI, confidence interval; eGFR, estimated glomerular filtration rate; hs-CRP, high-sensitivity C-reactive protein; IV, instrumental variable; MR, Mendelian randomization.

DISCUSSION

We assessed causality between genetically determined inflammation and kidney function employing MR approaches in a Japanese population. In this study, we used four and three SNPs as different genetic instruments (IVCRP and IVAsian). Neither of the two instrumental variables for hs-CRP was associated with eGFR levels in a two-sample MR analysis. These results suggested no significant causal relationship between hs-CRP and eGFR in this population. 

We found that IVCRP was not significantly associated with eGFR, which indicates no causal relationship between genetically determined inflammation and kidney function. In this study, we used four SNPs (rs3093077, rs1205, rs1130864, and rs1800947) within the CRP gene as instrumental variables. A previous study reported that these four SNPs were selected as a set of tagging SNPs in the CRP gene.17 One of the previous MR studies in Caucasians reported that genetically determined CRP was not significantly associated with creatinine-based eGFR (β = 0.004; 95% CI, −0.01 to 0.02). Interestingly, this previous study used the same SNP set (IVCRP) as in this study, and the effect size of IV on eGFR was similar to that observed in the present study. Therefore, this insignificant association between genetically determined CRP levels and eGFR is likely to be consistent across different ethnic groups.

The SNPs in IVCRP were originally selected among individuals of European descent. Moreover, it is well known that the CRP level in the Asian population was lower than that of Caucasians.31 Therefore, we tried to develop the IV specific for Asian people and selected three CRP-associated SNPs (rs3093068, rs7553007, and rs7310409) found in previous GWAS in Asian populations.18–20 However, no evidence was found that IVAsian was associated with eGFR in the study population.

IV requires the following three key assumptions: 1) relevance assumption (IV is associated with exposure), 2) exclusion restriction assumption (IV affects the outcome only through the exposure), and 3) exchangeability assumption (the effect of outcome is not confounded). Regarding relevance assumption, in this study, we restricted to only three SNPs in the robust selection process, thereby F-statistics of IVAsian were relatively small (F-statistic = 22.5). Although this value barely satisfied the assumption of IV (F-statistic >10), the results were likely to be empirically verifiable for relevance assumption. Considering that we adopted the top significantly associated SNPs as IVs of CRP, which were not associated with renal functions, the contributions of genetically determined CRP on the risk of renal disease may be relatively limited compared to the multiple risk factors of this complex disease.32,33

Another key assumption for MR analysis is the exclusion restriction assumption.34 This methodological review provided multiple scenarios violating this assumption (eg, inadequate phenotype definition and time-varying exposure). For scenarios of inadequate phenotype definition and measurement error, we used hs-CRP levels as an exposure variable. This is a clear definition of exposure and can lead to less measurement error compared with questionnaire-based phenotyping. For a scenario for the presence of LD, we carefully excluded either one of the SNPs in LD, which seems to have adequately addressed the problem. The scenarios of time-varying exposure and reverse causality are closely related and are usually worrisome for retrospective case-control studies, where data on the exposure is collected after diagnosis of the outcomes. As above, we adequately addressed these problems, or our analysis has no substantial possibility of meeting these scenarios. In addition to these scenarios, horizontal pleiotropy can violate the exclusion restriction assumption. IVAsian consisted of SNPs not only in the CRP gene but in HNF1A. Although previous studies suggested the SNP in HNF1A may have other pleiotropies,35–37 sensitivity analyses (MR-Egger and WM methods) in this study indicate that the IVW estimate was not biased by the average horizontal pleiotropic effect (known as directional pleiotropy).

The main strength of this study is that we used the IV specific for Asian populations. Given that the CRP level is different in each population, the construction of IVAsian might provide a meaningful approach to causal inference in the Asian population. The present study also has limitations to be discussed. First, we created IVAsian in this study but investigated the association only in a Japanese population. Therefore, it remains unclear whether this result is consistent across Asian groups. Further studies are needed to examine this association in other Asian populations. Second, the number of SNPs used in IVs was small. Both of the two IVs met the relevance assumption of IV in MR but were relatively weak. Given the relatively low variance of blood CRP levels and the existence of several other determinants of CRP in human environments, such as bacterial infections or other inflammatory diseases not derived from inherited CRP levels, the contribution of genetically determined blood CRP levels may be limited in the development of human renal disease. Potentially, the bigger the number of SNPs in IV, the bigger the explained variance of the exposure. On the contrary, with an increment of SNPs in IV, the pleiotropic effects will also increase. In a recent paper, a researcher suggested that there was no need to exclude SNPs with pleiotropic effects.37,38 In this study, however, we prioritized selecting SNPs within the CRP gene over the explanatory rate (ie, the number of SNPs in IV). In the future study, the selection of IV will be more important. Third, the study sample size was relatively large in Asian populations, but this sample size may lead to limited power for two-sample MR (eMaterials 1). Therefore, the result needs to be validated in a larger dataset. In addition, it is difficult to conclude no causality between hs-CRP and eGFR because the estimated coefficient in the conventional analysis was included in the confidence intervals of the MR methods. In conclusion, the present MR analyses investigated the causal relationship between hs-CRP and kidney function. Our twosample MR analyses with two different IVs did not support a causal effect of hs-CRP on eGFR in this population.

ACKNOWLEDGMENTS 

The authors would like to thank Kyota Ashikawa, Tomomi Aoi, and other members of the Laboratory for Genotyping Development, Center for Genomic Medicine, RIKEN, for help with the genotyping. The authors are grateful to Yoko Mitsuda and Keiko Shibata at the Department of Preventive Medicine, Nagoya University Graduate School of Medicine, for their technical assistance, and also grateful to Dr. Hideo Tanaka at Kishiwada City Public Health Center and Dr. Nobuyuki Hamajima at the Department of Healthcare Administration, Nagoya University Graduate School of Medicine for taking care of the J-MICC Study as former principal investigators.

Committee members of this Consortium (J-MICC Study Group): Kenji Wakai,3 Kenji Takeuchi,3 Asahi Hishida,3 Takashi Tamura,3 Keitaro Matsuo,6,7 Keitaro Tanaka,9 Katsuyuki Miura,10 Yoshikuni Kita,10 Sadao Suzuki,4 Toshiro Takezaki,12 Hiroki Nagase,13 Haruo Mikami,13 Hiroaki Ikezaki,14 Kiyonori Kuriki,15 Ritei Uehara,16 Kokichi Arisawa,17 and Hiroto Narimatsu19 (Affiliations in the author list, except 19Cancer Prevention and Cancer Control Division, Kanagawa Cancer Center, Research Institute, 1-1-2 Nakaonaga, Asahi-ku, Yokohama 241-0815 Japan) Funding: This study was supported by Grants-in-Aid for Scientific Research for Priority Areas of Cancer [grant number: 17015018] and Innovative Areas [grant number: 221S0001] and by JSPS KAKENHI Grants [grant numbers: 16H06277, 15H02524, 19K21461, and 19K10659] from the Japanese Ministry of Education, Culture, Sports, Science, and Technology. This study was supported in part by funding for the BioBank Japan Project from the Japan Agency for Medical Research and Development since April 2015, and the Ministry of Education, Culture, Sports, Science and Technology from April 2003 to March 2015.

Author Contributions (names must be given as initials): The authors' responsibilities were as follows-K.W. supervised this collaborative cohort study; R.F., A.H., T.N., K.M., H. Ito, Y. Nishida, C.S., Yasuyuki Nakamura, T.C., S.S., M.W., R.I., T. Takezaki, H.M., Yohko Nakamura, H. Ikezaki, M.M., K.K., N.K., D.M., K.A., S.K., M.T., T. Tamura, Y.K., T.K., Y.M., M.K., K.T., K.W., and J-MICC Study Group conducted the research in each study site; R.F. organized data from each study; A.H. and M.N. organized data for genetic analysis; R.F. conceptualized and analyzed data; R.F. wrote original draft; A.H., T.N., K.M., T.C., and K.W. reviewed the manuscript critically for important intellectual content; R.F. had primary responsibility for final content; and all authors read and approved the final manuscript. Conflicts of interest: Dr Nakatochi reports grants from Boehringer Ingelheim outside the submitted work.

APPENDIX A. SUPPLEMENTARY DATA 

Supplementary data related to this article can be found at https:== doi.org=10.2188=jea.JE20200540. 

REFERENCES

1. Dehghan A, Kardys I, de Maat MP, et al. Genetic variation, Creactive protein levels, and incidence of diabetes. Diabetes. 2007; 56:872–878.

2. Sesso HD, Buring JE, Rifai N, Blake GJ, Gaziano JM, Ridker PM. C-reactive protein and the risk of developing hypertension. JAMA. 2003;290:2945–2951. 

3. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med. 1997;336:973–979. 

4. Mihai S, Codrici E, Popescu ID, et al. Inflammation-related mechanisms in chronic kidney disease prediction, progression, and outcome. J Immunol Res. 2018;2018:2180373. 

5. Kubo S, Kitamura A, Imano H, et al. Serum albumin and highsensitivity C-reactive protein are independent risk factors of chronic kidney disease in middle-aged Japanese individuals: the circulatory risk in communities study. J Atheroscler Thromb. 2016;23:1089– 1098. 

6. Fox ER, Benjamin EJ, Sarpong DF, et al. The relation of C–reactive protein to chronic kidney disease in African Americans: the Jackson Heart Study. BMC Nephrol. 2010;11:1. 

7. Lee JE, Choi SY, Huh W, Kim YG, Kim DJ, Oh HY. Metabolic syndrome, C-reactive protein, and chronic kidney disease in nondiabetic, nonhypertensive adults. Am J Hypertens. 2007;20:1189– 1194. 

8. Ross R. Atherosclerosis-an inflammatory disease. N Engl J Med. 1999;340:115–126. 

9. Zhang L, Shen ZY, Wang K, et al. C-reactive protein exacerbates epithelial-mesenchymal transition through Wnt=β-catenin and ERK signaling in streptozocin-induced diabetic nephropathy. FASEB J. 2019;33:6551–6563. 

10. Chen N, Wan Z, Han SF, Li BY, Zhang ZL, Qin LQ. Effect of vitamin D supplementation on the level of circulating highsensitivity C-reactive protein: a meta-analysis of randomized controlled trials. Nutrients. 2014;6:2206–2216.