Lack of casual association between non-alcoholic fatty liver disease and bone mineral density: A mendelian randomization analysis

doi:10.21203/rs.3.rs-2313462/v1

Background

Many observational studies focus on the relationship between Nonalcoholic fatty liver disease (NAFLD) and bone mineral density (BMD). However, the conclusions are controversial and the causal relationship between NAFLD and BMD remains unclear.

Method

A bi-directional two-sample Mendelian randomization (MR) analysis was performed to investigate the potential causal links between NAFLD and BMDs. We applied genetic variants as instrumental variables obtained from the Genetic Factors for Osteoporosis (GEFOS) dataset and several published genome-wide association studies (GWASs). We obtained summary statistics for heel (H) BMD (n = 426,824), femoral neck (FN) BMD (n = 32,735), lumbar spine (LS) BMD (n = 28,498), ultra-distal forearm (UF) BMD (n = 21,907), and total body (TB) BMD (n = 56,284) from some GWAS meta-analyses. Additionally, the NAFLD GWAS included 377,988 individuals of European ancestry which consist of 4,761 NAFLD cases and 373,227 control cases. We used inverse variance weighted (IVW), four supplemental methods, and several sensitivity analyses to estimated and cross-validate the potential causal relationship in the present MR analysis.

Results

The sensitivity analyses do not find any violation of the MR assumptions. We found that NAFLD has no causal association with H-BMD (beta − 0.017; 95%CI -0.0458,0.0117; p = 0.2461), FN-BMD (beta − 0.0166; 95%CI -0.1592,0.1259; p = 0.8191), LS-BMD (beta − 0.021; 95%CI -0.1475,0.1055; p = 0.7446), UF-BMD (beta − 0.0524; 95%CI -0.1726,0.0679; p = 0.3935), TB-BMD (beta − 0.0596, 95%CI -0.1236,0.0044; p = 0.0678). Similarly, reverse MR analysis provided little support for a causal effect of BMDs on NAFLD.

Conclusion

This MR study found no evidence to support a bi-directional causality between NAFLD and BMD.

Bone mineral density

Nonalcoholic fatty liver disease

Mendelian randomization analysis

Bone mineral density (BMD) is one of the major bone biology indicators, and which also has been considered as the main diagnostic criteria of osteoporosis from 1990s¹. Because of a decreased BMD and the consequent osteoporosis, the increase of associated risks in fragility fractures almost affects all skeletons of human body². A previous study demonstrated that the spine fracture and hip fracture have strong relationship with the decreased spine and hip BMD, respectively³. In addition, the population size and the medical costs in fragility fractures of other skeleton sites is non-negligible⁴. Timely preventive measures are essential since low BMD is believed to be a contributing factor in one-third of fall-related deaths worldwide and has nearly doubled over the past 20 years in terms of death and disability-adjusted life years⁵.Therefore, based on the nature of fragility fractures related to decreased BMD, it is necessary to detect BMD in diverse skeletal site for patients with risk factors of lower BMD.

Nonalcoholic fatty liver disease (NAFLD) is the commonest pathologic cause of chronic liver disease in the world. According to the statistics, the incidence of NAFLD is up to ~30% of general adults, and nearly 100% of severe obesity patients⁶. According to clinical research, the relationship between NAFLD and BMD is still inconclusive. A meta-analysis included five cross-sectional studies demonstrated that there are no significant difference in BMD between patients with and without NAFLD⁷. However，a recent study indicated that the NAFLD is negatively linked with BMD in 1980 adults⁸, and a previous study of 6089 participants demonstrated an independently unfavorable relationship between NAFLD and BMD in the general U.S. population⁹. Conversely, several previous studies support the view that NAFLD is related to increased BMD^10,11. Disentangling the associations of NAFLD and BMD is of great public health and clinical importance.

Confined by potential methodological limitations, such as reverse causal link and residual confounding, traditional observational and cross-sectional study designs are unable to conclude causality regarding the role of NAFLD in the development of osteoporosis and the value change of BMD. Another approach is the Mendelian randomization (MR) design, a recent powerful and successful methodology, which uses genetic variants (single nucleotide polymorphisms, SNPs) as instrumental variables (IVs) or proxies for an exposure to infer causality of an exposure-outcome association¹². The MR works by the theory that gametogenesis randomly distributes genetic diversity from parent to offspring, shielding the genotype-phenotype relationship from the bias induced by confounding factors reported in observational studies and reverse causality¹³. Fortunately, numerous genome-wide association studies (GWASs) have been finished since the growing attention given to NAFLD and BMD, and the genetic discoveries associated with both diseases have been made public in recent years, which could provide the foundation for MR analysis.

Therefore, we used the bi-directional two-sample MR approach to analyze the potential causal links between NAFLD. In the MR approach, the IVs of the exposure and the outcome come from two independent GWAS for much greater statistic efficiency. Furthermore, the application of bi-directional MR design made the results of the present study more robust to confounding factors and reverse causation. According to the best of our knowledge, there are no MR studies focus on the casual relationship between NAFLD and BMD so far.

Study design

The valid MR analysis must fulfill the three core assumptions: 1) The IVs of genetic variant are strongly associated with exposure. 2) The selected IVS are not associated with any known confounders. 3) The IVs can only affect the risk of outcome dependently through exposure¹³. This study performed the bi-directional MR analysis in two stages: NAFLD was investigated as exposure while BMD were investigated as outcome in the first stage, whereas the second stage was reversed (Fig. 1).

Data Sources

To investigate a more reliable and comprehensive conclusion of the causal relationship between NAFLD and BMD, we extracted the IVs for NAFLD from a recent GWAS study¹⁴. This was a large-scale meta-analysis conducted using the UK Biobank. This study included 377,988 individuals of European ancestry which consist of 4,761 NAFLD cases and 373,227 control cases. Furthermore, we extracted the IVs from related GWAS summary statistics of ultra-distal forearm BMD (UF-BMD)¹⁵, which could be accessed at https://www.ebi.ac.uk/gwas/downloads/summary-statistics. The sample size of UF-BMD GWAS is 21,907 in European, and the number of SNPs is 11,312,319. The IVs of heel BMD (H-BMD) ¹³, lumbar spine BMD (LS-BMD) and femoral neck BMD (FN-BMD) were extracted from one GWAS summary statistics¹⁶, which could be downloaded from Genetic Factors for Osteoporosis (GEFOS) Consortium (http://www.gefos.org/). The GWAS of total body BMD (TB-BMD) was derived from a larger-scale meta-analysis conducted using the GEFOS database in European¹⁷. The details of these GWASs were shown in Table.1.

Table 1

Detailed information of studies and GWASs used for MR analysis
GWAS	Years	Sample size (case/control)	Number of SNPs	Population	Author	PMID
NAFLD	2022	377,988(4,761/373,227)	9,211,209	European	Cameron J Fairfield	34535985
Heel BMD	2019	426,824	13,705,641	European	Morris JA	30598549
Femoral neck BMD	2015	32,735	10,586,900	European	Hou-Feng Zheng	26367794
Lumbar spine BMD	2015	28,498	10,582,867	European	Hou-Feng Zheng	26367794
Ultra-distal forearm BMD	2020	21,907	11,312,319	European	Surakka I	33097703
Total body BMD	2018	56,284	16,162,733	European	Medina-Gomez C	29304378

BMD: Bone mineral density; NAFLD: Nonalcoholic fatty liver disease; GWAS: Genome-wide association study; SNP: Single nucleotide polymorphism.

Selection And Validation Of Ivs

According to the three assumptions for MR analysis, independent single nucleotide polymorphisms (SNPs) related with exposures (NAFLD or BMDs) to the level of genomic significance (p < 5*10^− 8) were selected as instrumental variables without effects of linkage disequilibrium (LD, clumping r² = 0.001 and kb = 10,000). Furthermore, we deleted the SNPs associated with the confounding factors with outcome by querying from the PhenoScanner database (http://www.phenoscanner.medschl.cam.ac.uk/)¹⁸. We also removed SNPs with palindrome alleles to prevent chain ambiguity issues. The phenotypic variation explained by SNPs was tested as follows: R² = 2* beta² *(1-EAF) * EAF/SD², with EAF = effect allele frequency and beta = the effect of each SNP on the exposures¹⁹. The strength of the association between IVs and the exposure factors was calculated by F statistic (F= (N-k-1)/k * R²/(1-R²), with N = the sample size of GWAS, k = the number of included IVs). Finally, the genetic IVs with F statistics > 10 indicated strong statistic power were included in MR analysis.

Mr Analysis

We performed MR analysis with five methods, and used the inverse variance weighed (IVW) method to assess the bi-directional relationship between NAFLD and five BMD as the primary statistical approach. The IVW method was considered to provide the most reliable causal estimates in the absence of directional pleiotropy²⁰. When the heterogeneity existed we used the random-effects IVW, and there are no heterogeneity we used the fixed-effects model²¹. Furthermore, we had supplemented the results using the other four methods, including the MR-Egger regression, the weighted median, the simple mode, and the weighted mode²². A p value < 0.05 was considered significant.

The sensitivity analysis also performed in this bi-directional MR study. The heterogeneity and horizontal pleiotropy test were used to assess stability of MR analysis results²³. We further performed MR pleiotropy residual sum (Egger-intercept test) and outlier (MR-pleiotropy residual sum and outlier, MR-PRESSO) to evaluate the presence of horizontal pleiotropy which was significant with p value < 0.05. Moreover, Cochran’s Q test was used to calculate the heterogeneity by MR-Egger regression and IVW method, the existence of significant heterogeneity was reflected by p value < 0.05²⁴. In addition, the “leave-one-out” sensitivity analysis was used to remove the potential impact of the SNPs²⁵. All the analyses were two-sided and performed in R software (version 4.0.1) with “TwoSampleMR” package (version 0.5.6).

After verification, the final data of IVs enrolled in our MR analysis and all the F-statistics of the included SNPs were > 10, which indicated that the selected IVs were powerful enough to eliminate potential bias and the datils were shown in supplementary Table 1.

Stage 1 Effect Of Nafld On Bmds

In the first stage, we investigated the casual effect of NAFLD on H-BMD, FN-BMD, LS-BMD, UF-BMD, and TB-BMD in the two-sample MR analysis. The outliers of IVs detected by the MR-PRESSO method were removed, details were shown in supplementary Table 2. Finally, the valid instrumental variables were included in the MR analyses, and the details of SNPs were shown in supplementary Table 3. After the outliers were removed, the IVW results indicated that NAFLD had no causal effect on BMDs [H-BMD related analysis: beta (95% confidence interval, CI)= -0.017 (-0.0458,0.0117), p = 0.2461; FN-BMD related analysis: beta (95% CI)= -0.0166(-0.1592,0.1259), p = 0.8191; LS-BMD related analysis: beta (95% CI)= -0.021 (-0.1475,0.1055), p = 0.7446; UF-BMD related analysis: beta (95% CI)= -0.0524 (-0.1726,0.0679), p = 0.3936; TB-BMD related analysis: beta (95% CI)= -0.0596 (-0.1236,0.0044), p = 0.0678]. Similarly, there are no casual links between NAFLD and BMD in MR analyses with other four methods (Table.2 and Figure.2), and the scatter plots of the MR estimates for the association of NAFLD with H-BMD, FN-BMD, LS-BMD, UF-BMD, and TB-BMD were shown in supplementary Fig. 1. The heterogeneity tests (including MR-Egger and IVW method) showed there was heterogeneity among selected IVs in the analysis of NAFLD and FN-BMD, which was further analyzed by a multiplicative random effects IVW method²⁶. The MR pleiotropy test showed no horizontal pleiotropy (p > 0.05, Table 2). Although the leave-one-out sensitivity test detected two potentially influential IVs (TB-BMD related test: rs182611493 and rs73001065; supplementary Fig. 2), the phenotype of these SNPs does not influence TB-BMD. Hence, we did not exclude the two SNPs from analysis. In total, no MR analyses supported the notion that NAFLD had a significant causal effect on BMD.

Table 2

The effect of NAFLD on BMDs
Exposure	outcome	No. of IVs	Heterogeneity test		Pleiotropy test	MR results
Exposure	outcome	No. of IVs	MR-Egger Cochran’s Q (p)	IVW Cochran’s Q (p)	Egger-intercept (p)	Method	Beta (95% CI)	P
NAFLD	H-BMD	5	0.7964 (0.8503)	7.7217 (0.1023)	-0.0065 (0.0782)	MR Egger	0.0439 (-0.0060,0.0978)	0.1830
						Weighted median	-0.0071 (-0.0348,0.0206)	0.6155
						IVW	-0.0170(-0.0458,0.0117)	0.2461
						Simple mode	-0.0122(-0.0557,0.0313)	0.6121
						Weighted mode	-0.0051(-0.0309,0.0208)	0.7200
NAFLD	FN-BMD	5	10.7866 (0.0129)	11.1289 (0.0252)	-0.0054 (0.7778)	MR Egger	0.0371(-0.3410,0.4153)	0.8596
						Weighted median	-0.0236(-0.1499,0.1027)	0.7143
						IVW	-0.0166(-0.1592,0.1259)	0.8191
						Simple mode	-0.1322(-0.3619,0.0974)	0.3221
						Weighted mode	-0.0475(-0.1889,0.0939)	0.5460
NAFLD	LS-BMD	5	5.8720 (0.1180)	6.6203 (0.1574)	-0.0093 (0.5801)	MR Egger	0.0701(-0.2499,0.3902)	0.6965
						Weighted median	0.0144(-0.1098,0.1387)	0.8198
						IVW	-0.0210(-0.1475,0.1055)	0.7446
						Simple mode	0.0082(-0.1416,0.1581)	0.9193
						Weighted mode	0.0097(-0.1197,0.1390)	0.8906
NAFLD	UF-BMD	6	6.7337 (0.1506)	6.7458 (0.2402)	-0.0014 (0.9366)	MR Egger	-0.0395(-0.3656,0.2866)	0.8240
						Weighted median	-0.0521(-0.1881,0.0838)	0.4522
						IVW	-0.0524(-0.1726,0.0679)	0.3936
						Simple mode	-0.0734(-0.2795,0.1327)	0.5162
						Weighted mode	-0.0548(-0.1946,0.0851)	0.4777
NAFLD	TB-BMD	6	4.6342 (0.3269)	5.6519 (0.3416)	-0.0079 (0.4017)	MR Egger	0.0103(-0.1496,0.1702)	0.9056
						Weighted median	-0.0740(-0.1548,0.0068)	0.0725
						IVW	-0.0596(-0.1236,0.0044)	0.0678
						Simple mode	-0.1028(-0.2192,0.0137)	0.1443
						Weighted mode	-0.0811(-0.1660,0.0038)	0.1199

MR: Mendelian randomization; BMD: Bone mineral density; NAFLD: Nonalcoholic fatty liver disease; FN- femoral neck; LS- lumbar spine; UF- ultra-distal forearm; TB- total body; IVW: inverse variance weighed; SNP: Single nucleotide polymorphism; CI: confidence interval.

Stage 2 Effect Of Bmds On Nafld

In the second stage, the influence of BMDs (including heel, femoral neck, lumbar spine, ultra-distal forearm, and total body) on NAFLD was studied. There are no outliers detected by the MR-PRESSO method (p > 0.05, supplementary Table 3). As shown in Table 3 and Fig. 3, all the five methods in MR analyses suggested that BMD in different skeletal sites had no causal effect on NAFLD in the reverse direction. The scatter plots of the MR analyses for the association of BMDs with NAFLD were shown in supplementary Fig. 3. There was no evidence of significant heterogeneity (p > 0.05) or horizontal pleiotropy (p > 0.05) in the MR analyses of BMDs on NAFLD. Additionally, the leave-one-out sensitivity test detected no potentially influential IVs (supplementary Fig. 4, Fig. 5, and Fig. 6).

Table 3

The effect of BMDs on NAFLD
Exposure	outcome	No. of IVs	Heterogeneity test		Pleiotropy test	MR results
Exposure	outcome	No. of IVs	MR-Egger Cochran’s Q (p)	IVW Cochran’s Q (p)	Egger-intercept (p)	Method	OR (95% CI)	p
H BMD	NAFLD	501	114.2462 (0.9999)	114.3311 (0.9999)	0.0005 (0.771)	MR Egger	1.0094(0.8893,1.1459)	0.8844
						Weighted median	0.9987(0.8800.1.1333)	0.9837
						IVW	1.0251(0.9525,1.1032)	0.5086
						Simple mode	1.0013(0.7826,1.2811)	0.9917
						Weighted mode	0.9807(0.8516,1.1293)	0.7863
FN BMD	NAFLD	20	4.2456 (0.9996)	4.4750 (0.9997)	-0.0129 (0.6378)	MR Egger	1.2299(0.5249,2.8814)	0.6396
						Weighted median	0.9722(0.7805,1.2112)	0.8018
						IVW	1.0027(0.8525,1.1793)	0.9741
						Simple mode	0.9826(0.6948,1.3896)	0.9218
						Weighted mode	0.9686(0.6941,1.3516)	0.8531
LS BMD	NAFLD	21	2.5204 (0.9999)	3.2029 (0.9999)	0.0158 (0.4190)	MR Egger	0.7690(0.4562,1.2962)	0.3365
						Weighted median	0.9469(0.7928,1.1308)	0.5467
						IVW	0.9505(0.8254,1.0945)	0.4805
						Simple mode	0.9084(0.6645,1.2418)	0.5537
						Weighted mode	0.9184(0.6785,1.2432)	0.5878
UF BMD	NAFLD	11	1.0131 (0.9994)	1.3073 (0.9994)	-0.0103 (0.6007)	MR Egger	1.0378(0.7416,1.4524)	0.8335
						Weighted median	0.9548(0.8007,1.1385)	0.6063
						IVW	0.9533(0.8312,1.0934)	0.4943
						Simple mode	0.9691(0.7548,1.2441)	0.8103
						Weighted mode	0.9572(0.7760,1.1808)	0.6918
TB BMD	NAFLD	81	18.8287 (0.9999)	19.2218 (0.9999)	0.0043 (0.5325)	MR Egger	0.9167(0.7175,1.1712)	0.4886
						Weighted median	0.9750(0.8512,1.1168)	0.7149
						IVW	0.9854(0.8966,1.0831)	0.7609
						Simple mode	1.0213(0.7859,1.3271)	0.8752
						Weighted mode	0.9607(0.7772,1.1876)	0.7121

MR: Mendelian randomization; BMD: Bone mineral density; NAFLD: Nonalcoholic fatty liver disease; FN- femoral neck; LS- lumbar spine; UF- ultra-distal forearm; TB- total body; IVW: inverse variance weighed; SNP: Single nucleotide polymorphism; OR: odds ratio; CI: confidence interval.

In this bi-directional two-sample MR study, we used summary statistics from different GWASs to analyze the casual links between NAFLD and BMDs. To the best of our knowledge, this MR analysis is the first one to investigate the casual relationship between NAFLD and BMDs with five MR methods. The present MR study demonstrated that there are no causality exists between NAFLD and BMDs with different skeletal sites (including heel, femoral neck, lumbar spine, ultra-distal forearm and total body BMD).

The casual links of NAFLD and BMDs is controversial to date, which included three viewpoints. 1) NAFLD is related to increased BMD, 2) NAFLD is connected with decreased BMD, 3) there are no relationship between NAFLD and BMD. A prospective study declared that patients with non-alcoholic steatohepatitis had higher lumbar BMD than did healthy controls, and the non-alcoholic steatohepatitis is advanced histological phenotype of NAFLD ¹⁰. Additionally, a cross-sectional study declared that the total body BMD was higher in participants with NAFLD of the population in Caucasian²⁷. However, several studies showed an inverse links between NAFLD and low BMD. A retrospective study demonstrated that the NAFLD was associated with lower BMD in the American adults during the period of 2005-2014²⁸, and a longitudinal cohort study suggested that NAFLD was associated with an increased incidence of low total body BMD²⁹. Moreover, a data were extracted from five cross-sectional studies in meta-analysis showed that no significant difference in BMD between patients with and without NAFLD⁷. Similarly, another meta-analysis revealed that there are no significant differences in BMD at different skeletal sites between individuals with and without NAFLD³⁰. In addition, a recent study also suggested that there are no association between NAFLD and lumbar spine BMD³¹. However, the evidence on available does not support for the inference of a causal relationship between NAFLD and BMD. Our results corroborated the above two meta-analysis and one observational study, and further demonstrated that there are no causal relationships between NAFLD and BMD.

The reason of this result could be associated with the NAFLD patients’ behavior factors, which included physical exercise, smoking, and the intake of fat foods. The NAFLD patients with fat foods intake often led to weight gain. It seems logical to hypothesize that being overweight, which is a typical characteristic of the patients with NAFLD³², may prevent bone loss, presumably through increased mechanical stress by overweight and subsequently accelerated cortical bone formation³⁰. Similarly, a previous study indicated that BMD loss at lumbar spine was decreased in obese and overweight women, and that body mass index had no effect BMD loss at the hip³³. In addition, recent studies focused on the serum levels in patients with NAFLD of several bone turnover biomarkers including osteocalcin³⁴, which is closely related to BMD in general population, although the osteocalcin level does not showed significant correlation with BMD in NAFLD patients³⁰. In total, the mechanical stress and patient’s behavior factors might partly elucidate the reason why there are no causality between NAFLD and BMD, further study need to focus on the changes of BMD in patients with NAFLD.

The present study used a bi-directional two-sample MR design has several significant advantages, such as it is more resistant to confounding factors and robust to reverse causation, and the direction of causal link can be clarified. Additionally, owing to the size of sample in this study is largest to date, which can get greater statistical power. The population of this MR study is European, which could further avoid the heterogeneity in ethnic lines. However, there are some potential limitations that should be considered. First of all, due to all the population of included GWASs are European, application of the conclusion on the other races is un-known in this study. Secondly, the sex ratio, age, and the severity of NAFLD are un-known in the GWAS, and we cannot perform stratified analysis for sex, age, and the severity of NAFLD. Thirdly, the leave-one-out-tests eliminate two SNPs (TB-BMD related test: rs182611493 and rs73001065), but the phenotypes of these SNPs do not influence TB-BMD and we do not exclude the two SNPs from IVs. Fourth, based on the nature of MR analysis which is limited by the GWASs of BMDs and NAFLD, future MR study was needed in a newer and larger GWAS.

In conclusion, the observed association between NAFLD and BMD on different skeletal sites should not be interpreted as causation based on our bi-directional two-sample MR study. The result of our MR study argues for more careful monitoring and evaluation of BMD among patients with NAFLD.

NAFLD: non-alcoholic fatty liver disease

BMD: Bone mineral density

MR: Mendelian randomization

H: Heel

FN: Femoral neck

LS: Lumbar spine

UF: ultra-distal forearm

TB: Total body

SNP: Single nucleotide polymorphism

GWAS: Genome-wide association study

IV: Instrumental instrument

OR: Odds ratio

CI: Confidence interval

IVW: Inverse variance weighted

MR-PRESSO: MR-pleiotropy residual sum and outlier

Availability of data and materials

The summary statistics of GWAS for NAFLD are derived from a GWAS meta-analysis can be accessed at https://www.ebi.ac.uk/gwas/downloads/summary-statistics;

Full GWAS summary statistics for H-BMD, LS-BMD, FN-BMD, UF-BMD, and TB-BMD are publicly available through http://www.gefos.org/.

Acknowledgements

The authors sincerely thank the GEFOS consortium, and Cameron J Fairfield et.al for providing summary statistics.

Funding

This work was supported by the National Natural Science Foundation of China (No.81860396).

Author information

Authors and Affiliations

Xinjiang Medical University, Urumqi, China

Munire·Hudabaierdi & Lei Cui

Department of Joint Surgery, First Affiliated Hospital, Xinjiang Medical University, Urumqi, China

Fei Wang & Abudousaimi·Aimaiti

Department of Orthopaedics, the Affiliated Hospital of Southwest Medical University, Luzhou, Sichuan, China

Yan-shi Liu

Department of Hepatobiliary Surgery, Fengjie County People's Hospital of Chongqing, Chongqing, China

Zong-ding Wang

Contributions

Study conception and design: MH and FW; data analyses: MH, FW, YL, and ZW; draft preparation: MH, FW, YL, and ZW; supervision of the study: LC and AA; All authors reviewed the manuscript.

Corresponding author

Correspondence to Lei Cui and Abudousaimi·Aimaiti.

Ethics declarations

Ethics approval and consent to participate

Not applicable. Ethical approval and informed consent for studies included in the analyses was provided in the original publications.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Compston, J. E., McClung, M. R. & Leslie, W. D. Osteoporosis. The Lancet 393, 364–376 (2019).
Stone, K. L. et al. BMD at Multiple Sites and Risk of Fracture of Multiple Types: Long-Term Results From the Study of Osteoporotic Fractures. J. Bone Miner. Res. 18, 1947–1954 (2003).
Marshall, D. & Wedel, H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. 312, 6 (1996).
Leslie, W. D. et al. Direct costs of fractures in Canada and trends 1996-2006: A population-based cost-of-illness analysis. J. Bone Miner. Res. 26, 2419–2429 (2011).
Sànchez-Riera, L. et al. The global burden attributable to low bone mineral density. Ann. Rheum. Dis. 73, 1635–1645 (2014).
Lonardo, A., Mantovani, A., Targher, G. & Baffy, G. Nonalcoholic Fatty Liver Disease and Chronic Kidney Disease: Epidemiology, Pathogenesis, and Clinical and Research Implications. Int. J. Mol. Sci. 23, 13320 (2022).
Upala, S., Jaruvongvanich, V., Wijarnpreecha, K. & Sanguankeo, A. Nonalcoholic fatty liver disease and osteoporosis: a systematic review and meta-analysis. J. Bone Miner. Metab. 35, 685–693 (2017).
Xie, R. & Liu, M. Relationship Between Non-Alcoholic Fatty Liver Disease and Degree of Hepatic Steatosis and Bone Mineral Density. Front. Endocrinol. 13, 857110 (2022).
Umehara, T. Nonalcoholic fatty liver disease with elevated alanine aminotransferase levels is negatively associated with bone mineral density: Cross-sectional study in U.S. adults. PloS One 13, e0197900 (2018).
Kaya, M. et al. Increased bone mineral density in patients with nonalcoholic steatohepatitis. World J. Hepatol. 5, 627 (2013).
Boonchaya-anant, P., Hardy, E., Borg, B. B. & Burshell, A. L. Bone mineral density in patients with nonalcoholic steatohepatitis among end-stage liver disease patients awaiting liver transplantation. Endocr. Pract. Off. J. Am. Coll. Endocrinol. Am. Assoc. Clin. Endocrinol. 19, 414–419 (2013).
Burgess, S., Butterworth, A. & Thompson, S. G. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet. Epidemiol. 37, 658–665 (2013).
Lawlor, D. A., Harbord, R. M., Sterne, J. A. C., Timpson, N. & Davey Smith, G. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat. Med. 27, 1133–1163 (2008).
Fairfield, C. J. et al. Genome-Wide Association Study of NAFLD Using Electronic Health Records. Hepatol. Commun. 6, 297–308 (2022).
Surakka, I. et al. MEPE loss-of-function variant associates with decreased bone mineral density and increased fracture risk. Nat. Commun. 11, 4093 (2020).
Zheng, H.-F. et al. Whole-genome sequencing identifies EN1 as a determinant of bone density and fracture. Nature 526, 112–117 (2015).
Medina-Gomez, C. et al. Life-Course Genome-wide Association Study Meta-analysis of Total Body BMD and Assessment of Age-Specific Effects. Am. J. Hum. Genet. 102, 88–102 (2018).
Kamat, M. A. et al. PhenoScanner V2: an expanded tool for searching human genotype-phenotype associations. Bioinforma. Oxf. Engl. 35, 4851–4853 (2019).
Shim, H. et al. A multivariate genome-wide association analysis of 10 LDL subfractions, and their response to statin treatment, in 1868 Caucasians. PloS One 10, e0120758 (2015).
Holmes, M. V., Ala-Korpela, M. & Smith, G. D. Mendelian randomization in cardiometabolic disease: challenges in evaluating causality. Nat. Rev. Cardiol. 14, 577–590 (2017).
Hemani, G., Bowden, J. & Davey Smith, G. Evaluating the potential role of pleiotropy in Mendelian randomization studies. Hum. Mol. Genet. 27, R195–R208 (2018).
Burgess, S. & Thompson, S. G. Multivariable Mendelian randomization: the use of pleiotropic genetic variants to estimate causal effects. Am. J. Epidemiol. 181, 251–260 (2015).
Bowden, J. et al. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat. Med. 36, 1783–1802 (2017).
Zhang, S. et al. CLU rs2279590 polymorphism contributes to Alzheimer’s disease susceptibility in Caucasian and Asian populations. J. Neural Transm. Vienna Austria 1996 122, 433–439 (2015).
Burgess, S. & Thompson, S. G. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur. J. Epidemiol. 32, 377–389 (2017).
Bowden, J. et al. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat. Med. 36, 1783–1802 (2017).
Li, B. T. et al. Association Between Liver Fat and Bone Density is Confounded by General and Visceral Adiposity in a Community-Based Cohort. Obes. Silver Spring Md 29, 595–600 (2021).
Zhai, T., Chen, Q., Xu, J., Jia, X. & Xia, P. Prevalence and Trends in Low Bone Density, Osteopenia and Osteoporosis in U.S. Adults With Non-Alcoholic Fatty Liver Disease, 2005-2014. Front. Endocrinol. 12, 825448 (2021).
Shen, Z. et al. Increased risk of low bone mineral density in patients with non-alcoholic fatty liver disease: a cohort study. Eur. J. Endocrinol. 182, 157–164 (2020).
Mantovani, A. et al. Systematic review with meta-analysis: non-alcoholic fatty liver disease is associated with a history of osteoporotic fractures but not with low bone mineral density. Aliment. Pharmacol. Ther. 49, 375–388 (2019).
Du, J., Ma, Y., Lang, H., Huang, C. & Zhang, X. The association of nonalcoholic fatty liver disease with bone mineral density in type 2 diabetes. Eur. J. Med. Res. 27, 143 (2022).
Younossi, Z. et al. Global burden of NAFLD and NASH: trends, predictions, risk factors and prevention. Nat. Rev. Gastroenterol. Hepatol. 15, 11–20 (2018).
Leslie, W. D., Morin, S. N., Majumdar, S. R. & Lix, L. M. Effects of obesity and diabetes on rate of bone density loss. Osteoporos. Int. 29, 61–67 (2018).
Jia, L. & Cheng, M. Correlation analysis between risk factors, BMD and serum osteocalcin, CatheK, PINP, β-crosslaps, TRAP, lipid metabolism and BMI in 128 patients with postmenopausal osteoporotic fractures. Eur. Rev. Med. Pharmacol. Sci. 26, 7955–7959 (2022).

No competing interests reported.

Lack of casual association between non-alcoholic fatty liver disease and bone mineral density: A mendelian randomization analysis

Status:

Version 1

Abstract

Background

Method

Results

Conclusion

Figures

Introduction

Materials And Methods

Study design

Data Sources

Selection And Validation Of Ivs

Mr Analysis

Results

Stage 1 Effect Of Nafld On Bmds

Stage 2 Effect Of Bmds On Nafld

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1