Intended for healthcare professionals

CCBYNC Open access

Rapid response to:


Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults

BMJ 2019; 367 doi: (Published 08 October 2019) Cite this as: BMJ 2019;367:l5367

Rapid Response:

Re: Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults

We write on behalf of our large collaborative international and multi-disciplinary authorship team. In response to Professor Lonardo’s comments we respectfully refer to the statements made in our paper[1], and in our response to Professor Targher. We were and remain open minded to new findings which can change or refine currently-held perceptions in order to improve practice as new evidence emerges. We started this work from an agnostic position given the mixed antecedent literature base.

In our paper, we acknowledge and refer to the widely-cited series that report cardiovascular disease (CVD) is the most common cause of death. We cite the cohort studies which vary in their adjustment for risk factors that have suggested a causal link with surrogate markers and the meta-analyses of these. However, other studies have come to different conclusions as to the magnitude and independence of this association [2]. A major barrier has been the reliance on biopsy for accurate diagnosis and staging of disease, which, of course, is unacceptable as population screening and therefore by definition, the studies have historically included highly selected populations. As use of non-invasive modalities in research settings evolved, so greater numbers of patients have been studied over time. In 4119 patients enrolled in the Multi-Ethnic Study of Atherosclerosis with CT-proven fatty liver and no baseline CVD followed up for median 7.6 years[3], the hazard ratio for incident non-fatal CVD events was 1.74 (95% CI 1.25 – 2.41) after adjustment for age, sex, ethnicity and study site reducing to 1.43 (95% CI 1.00 – 2.03; p=0.05) following further adjustment for diabetes, hypertension, body mass index, lipids and smoking; adjustments that we have included in our study. In 4234 patients enrolled in the same study followed for median 12.2 years, CT-proven hepatic fat was not associated with hard atherosclerotic CVD events, including fatal and non-fatal MI and stroke (standardised hazard ratio 0.96, 95% CI 0.86–1.08; p=0.52) [4]. Notwithstanding differences in study design, case definition and outcome measure and the fact that CT is not used to diagnose fatty liver in routine clinical care, there is dissonance in the literature. Therefore, while no one doubts the CVD risk in multi-morbid NAFLD patients with multiple elements of metabolic syndrome, the question as to whether there is an independent impact of NAFLD per se on CVD outcomes is valid and warrants further examination. Our paper informs this debate but it is not, nor do we claim it to be, the definitive study.

Our paper is the biggest window into the real-life working health care records of people with a recorded diagnosis of NAFLD/NASH. If there were increased cardiovascular risk attributable to NAFLD/NASH alone, then the patients whose management would change would be the patients included in our study because these are the individuals in whom NAFLD/NASH has been diagnosed. We completely accept (and discuss in the paper) that patients with NAFLD/NASH may be in the matched group but the matching of 100:1 dilutes this effect somewhat. If our data result in better coding and risk-stratification of patients with NAFLD, then we would most certainly support this. Methodologically, we accept some shortcomings of real-world data. Nevertheless, and for the reasons stated in the paper, these data show that the risk in patients diagnosed with NAFLD is accounted for by other established risk factors. This is analogous to the pattern seen in patients with pre-diabetes. Our recommendations are entirely in keeping with the EASL guidelines as we fully support the assessment of CVD risk in patients with NAFLD. We were unable to ascribe a risk-multiplier for a coded diagnosis of NAFLD nor are we aware of any validated adjusted scores that do this. We welcome new, well-conducted studies to answer this question definitively but for now, we believe our findings are closest to the day to day clinical practice undertaken by many doctors across Europe. Therefore, we robustly defend our publication, the rationale for the study, the data we present and our conservative interpretation of the implications of our study for clinical practice. The decision to publish in the BMJ is an editorial one and we would not presume to comment on this.

William Alazawi & Naveed Sattar on behalf of all authors

1. Alexander M, Loomis AK, van der Lei J, et al. Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults. BMJ 2019;367:l5367. doi: 10.1136/bmj.l5367 [published Online First: 2019/10/09]
2. Lazo M, Hernaez R, Bonekamp S, et al. Non-alcoholic fatty liver disease and mortality among US adults: prospective cohort study. BMJ 2011;343:d6891. doi: 10.1136/bmj.d6891
3. Zeb I, Li D, Budoff MJ, et al. Nonalcoholic Fatty Liver Disease and Incident Cardiac Events: The Multi-Ethnic Study of Atherosclerosis. J Am Coll Cardiol 2016;67(16):1965-6. doi: 10.1016/j.jacc.2016.01.070 [published Online First: 2016/04/23]
4. Shah RV, Anderson A, Ding J, et al. Pericardial, But Not Hepatic, Fat by CT Is Associated With CV Outcomes and Structure: The Multi-Ethnic Study of Atherosclerosis. JACC Cardiovasc Imaging 2017;10(9):1016-27. doi: 10.1016/j.jcmg.2016.10.024 [published Online First: 2017/03/24]

Competing interests: We are lead senior authors of the article being commented upon

20 October 2019
William Alazawi
Reader in Hepatology
Naveed Sattar and on behalf of co-authors
Barts Liver Centre, Queen Mary, University of London