Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 24.
Published in final edited form as: Arterioscler Thromb Vasc Biol. 2024 Dec 5;45(2):323–331. doi: 10.1161/ATVBAHA.124.321988

Human genetic evidence to inform clinical development of interleukin-6 signaling inhibition for abdominal aortic aneurysm

Stephen Burgess 1,2,3, Héléne T Cronjé 1,3, Emil deGoma 4, Yung Chyung 4, Dipender Gill 3,✉
PMCID: PMC7617413  EMSID: EMS202956  PMID: 39633572

Abstract

Background

Abdominal aortic aneurysm (AAA) represents a significant cause of mortality, yet no medical therapies have proven efficacious. The aim of the current study was to leverage human genetic evidence to inform clinical development of interleukin-6 (IL6) signaling inhibition for treatment of AAA.

Methods

Associations of rs2228145, a missense variant in the IL6R gene region, are expressed per additional copy of the C allele, corresponding to the genetically-predicted effect of IL6 signaling inhibition. We consider genetic associations with AAA risk in the AAAgen consortium (39,221 cases, 1,086,107 controls) and UK Biobank (1963 cases, 365,680 controls). To validate against known effects of IL6 signaling inhibition, we present associations with rheumatoid arthritis, polymyalgia rheumatica, and severe COVID-19. To explore mechanism specificity, we present associations with thoracic aortic aneurysm, intracranial aneurysm, and coronary artery disease. We further explored genetic associations in clinically-relevant subgroups of the population.

Results

We observed strong genetic associations with AAA risk in the AAAgen consortium, UK Biobank, and FinnGen: odds ratio (OR) 0.91 (95% confidence interval [CI]: 0.90 to 0.92, p = 4×10-30), OR 0.90 (95% CI: 0.84, 0.96, p=0.001), and OR 0.86 (95% CI: 0.82, 0.91, p = 7×10-9) respectively. The association was similar for fatal AAA, but with greater uncertainty due to the lower number of events. The association with AAA was of greater magnitude than associations with coronary artery disease and even rheumatological disorders for which IL6 inhibitors have been approved. No strong associations were observed with thoracic aortic aneurysm or intracranial aneurysm. Associations attenuated towards the null in populations with concomitant rheumatological or connective tissue disease.

Conclusions

Inhibition of IL6 signaling is a promising strategy for treating AAA, but not other types of aneurysmal disease. These findings serve to help inform clinical development of IL6 signaling inhibition for AAA treatment.

graphic file with name EMS202956-f005.jpg

Open in a new tab

Nonstandard Abbreviations and Acronyms

AAA

abdominal aortic aneurysm

CAD

coronary artery disease

CI

confidence interval

CVD

cardiovascular disease

GWAS

genome-wide association study

IL6

interleukin-6

IL6R

interleukin-6 receptor

OR

odds ratio

PP-H

posterior probability hypothesis

Introduction

Abdominal aortic aneurysm (AAA) is defined as a permanent localized dilatation of the abdominal aorta to more than 3cm and has a prevalence of approximately 5% in individuals aged more than 60 years.1 Mortality from ruptured AAA is estimated to be approximately 90%,2,3 and other than for modifying cardiovascular risk factors, no efficacious pharmacotherapies are available for AAA treatment.4 The mainstay of intervention is therefore surgical repair, despite which global AAA-related mortality remains to be at approximately 150,000 to 200,000 deaths per year.5,6

Drug development can be slow, expensive, and inefficient,7 with the high failure rate largely attributable to insufficient efficacy or unacceptable safety profiles.8 As the majority of drug targets are proteins, which are coded for by genes, human genetic data offer the opportunity to dramatically improve the probability of successful drug development.9 At a population level, naturally occurring genetic variation in the gene coding for a drug target protein can be used to study the effect of pharmacologically perturbing that protein.10 This drug target Mendelian randomization paradigm can be used to inform various aspects of clinical development,11 including efficacy, secondary indications, adverse effects, effect heterogeneity, and biomarkers of target engagement.12 In this way, drug target Mendelian randomization has been used to inform on the effects of various cardiovascular disease (CVD) drug targets, including for lipid-lowering, antihypertensive, anti-diabetic, and anticoagulant agents.11

The drug target Mendelian randomization paradigm has been used extensively to study the effect of interleukin-6 (IL6) signaling inhibition on CVD outcomes.13–15 The genetic evidence supporting its efficacy has contributed to the pursuit of IL6 signaling inhibition as a therapeutic target in CVD, with one phase 2 study (NCT06362759) and five phase 3 clinical studies currently ongoing (NCT05021835, NCT06118281, NCT05636176, NCT05485961, NCT06200207).16 However, while there is also genetic support for the efficacy of IL6 signaling inhibition in AAA,17 this has remained limited in nature with key questions yet unanswered. The aim of this work was therefore to leverage the drug target Mendelian randomization paradigm to inform clinical development of IL6 signaling inhibition for AAA. Specifically, we aimed to affirm the genetic evidence for efficacy of IL6 signaling inhibition in AAA (including testing for a per allele dose-response relationship) and fatal AAA, compare against established positive control outcomes and coronary artery disease (CAD), investigate specificity for different types of aneurysmal disease, and explore heterogeneity of effect across population subgroups (including subgroups defined by sex, smoking status, hypertensive status, and prevalent rheumatological or connective tissue disorders). These findings collectively serve to inform clinical development efforts of IL6 signaling inhibition for the treatment of AAA.

Methods

Overview

As the instrument for IL6 signaling inhibition, we leverage a genetic variant in the IL6R gene region that has previously been shown to mimic pharmacological IL6 signaling inhibition.14 To investigate potential clinical effects of inhibiting IL6 signaling, we present associations of this variant with traits and diseases from a variety of data sources, including publicly-available summarized datasets and individual-level data from UK Biobank. The summarized datasets are larger and have more cases, resulting in more precise estimates. The individual-level data analyses allow the exploration of more specific outcomes, such as fatal AAA, and subgroup analyses in specific strata of the population. An overview of all analyses is provided as Figure 1, and a summary of the outcome datasets used in the analysis is provided as Table 1.

Figure 1. Overview of analyses.

Figure 1

Open in a new tab

The Mendelian randomization framework relies on the following core instrumental variable assumptions: 1) the genetic instrument strongly relates to the exposure; 2) there are no confounding pathways linking variant and outcome, and; 3) the genetic instrument affects the outcome only through the exposure and not through independent pathways. IL6: interleukin-6, IL6R: interleukin-6 receptor.

Table 1. List of outcomes and datasets.

Dataset Includes UK Biobank? Outcomes considered
UK Biobank (✓) Any AAA, fatal AAA, polymyalgia rheumatica
AAAgen ✓ Any AAA
FinnGen × Any AAA, rheumatoid arthritis
Michigan Genomics Initiative × Thoracic aortic aneurysm
Aneurysm consortium ✓ Intracranial aneurysm
Coronary artery disease ✓ Coronary artery disease
Rheumatoid consortium ✓ Rheumatoid arthritis
COVID-19 Host Genome Initiative ✓ Severe COVID-19
Global Biobank Meta-analysis Initiative ✓ Replication of sex-specific associations for any AAA
Open in a new tab

AAA: abdominal aortic aneurysm. ✓ = consortium includes participants from UK Biobank, × = consortium does not include participants from UK Biobank

Genetic instrument

We focused on rs2228145 (previously also called rs8192284), a missense variant located at chr1:154426970 on GRCh37, chr1:154454494 on GRCh38 with minor allele frequency of 42% in the European ancestry UK Biobank analytic sample considered here. It is the lead variant associated with IL6 levels in the IL6R gene region (Figure S1). Associations are given per additional copy of the C allele, which is the minor allele and is associated with lower levels of C-reactive protein (0.093 units lower log-transformed C-reactive protein, p < 10-400).18 Hence estimates correspond to the genetically-predicted effect of increasing IL6 signaling inhibition, and reflect the consequences of lifelong lower inflammation levels, with the relative magnitude of reduction in C-reactive protein in adulthood estimated at 9%. We also considered estimates based on an extended instrument consisting of 7 variants in the IL6R gene region and its neighborhood (±100 kilobase pairs) previously demonstrated to be conditionally associated with C-reactive protein levels (Table S1).14

Summarized datasets

The primary outcome was AAA. Associations were obtained from a meta-analysis of 17 individual genome-wide association studies (GWASs) by the AAAgen Consortium including 39,221 cases and 1,086,107 controls.19 We also validated results in FinnGen, a large genome-wide association study in the Finnish population, including 3,869 cases and 381,977 controls. FinnGen did not contribute data to the AAAgen Consortium. We performed colocalization between IL6 and AAA risk using the coloc method,20 taking genetic associations with IL6 from the SCALLOP consortium (N=14,242)21 and using the default priors from the coloc package.22 Colocalization is a statistical approach to assess whether the genetic predictors of two traits overlap (known as colocalization) or are distinct (known as non-colocalization). The coloc method reports two key outputs: the posterior probability of colocalization (PP-H4) and the posterior probability of non-colocalization (PP-H3). High values (close to 1) of PP-H4 indicate colocalization, which is supportive of a causal relationship; high values of PP-H3 indicate non-colocalization, which opposes a causal relationship; low values of both PP-H3 and PP-H4 indicate lack of strong evidence supporting or opposing a causal relationship.23

As positive controls, we present associations with rheumatoid arthritis (GWAS consortium, 35,871 cases, 240,149 controls;24 FinnGen, 15,223 cases, 138,246 controls),25 polymyalgia rheumatica (UK Biobank, 2,460 cases, 433,511 controls),26 and severe COVID-19 (COVID-19 Host Genome Initiative round 7, 18,152 cases, 1,145,546 controls),27 as IL6 signaling inhibition is known from randomized trials to offer therapeutic benefit in these diseases.28–30

For comparison with other aneurysmal and CVDs, we present associations with thoracic aortic aneurysm (Michigan Genomics Initiative, 1,351 cases, 18,295 controls),31 intracranial aneurysm (GWAS Consortium, 10,754 cases, 306,882 controls)32, and CAD (CARDIoGRAMplusC4D, 210,842 cases, 1,167,328 controls).33

For replication of sex-specific findings, we present associations with AAA risk in men and women separately published by the Global Biobank Meta-analysis Initiative.34 We note that this is not a perfect replication, as the Global Biobank Meta-analysis Initiative dataset includes UK Biobank participants. However, we were unable to find an independent dataset with sex-stratified genetic association estimates. The Global Biobank Meta-analysis Initiative contains 515,358 women (1,289 AAA cases) and 456,574 men (5,589 AAA cases) from eight cohorts: Biobank Japan, BioMe (men only), BioVU, Colorado Center for Personalized Medicine (men only), FinnGen, HUNT, Michigan Genomics Initiative, and UK Biobank.

We note that all summarized datasets include UK Biobank participants, with the exception of FinnGen and the Michigan Genomics Initiative.

Individual-level data in UK Biobank

The UK Biobank cohort comprises around 500,000 participants (94% of self-reported European ancestry) aged 40 to 69 years at baseline.35 They were recruited between 2006-2010 in 22 assessment centers throughout the UK, and followed up until November 2022 or their date of death. We performed detailed quality control procedures on UK Biobank participants and on genetic variants as described previously,36 restricting analyses to unrelated participants (that is, more distant than third degree relatives) of European ancestries.

Abdominal aortic aneurysm in UK Biobank was defined as having an International Statistical Classification of Diseases and Related Health Problems (ICD)-10 code of I71.3 or I71.4 (AAA, either ruptured or without mention of rupture), or the equivalent ICD-9 code (441.3, 441.4) in their hospital episode statistics or death certificate. The secondary outcome was fatal AAA, defined as having one of the relevant ICD codes in their death certificate.

Stratification variables

We estimated associations with AAA in UK Biobank in subgroups of the population, stratifying by five separate variables: sex (men vs women), hypertension (defined as systolic blood pressure > 140 mmHg and diastolic blood pressure > 90 mmHg at baseline, or usage of hypertensive medication at baseline), smoking status (ever regular smoker vs never regular smokers), any rheumatologic disorder (present vs absent), and any connective tissue disorder (present versus absent).

A never regular smoker was defined as answering the touchscreen question “In the past, have you ever smoked tobacco?” with the response “I have never smoked” or “Just tried once or twice” as opposed to “smoked occasionally” or “smoked on most or all days” at baseline, and never contradicting this answer at a future survey.

Any rheumatologic disorder was defined as ankylosing spondylitis, anti-neutrophil cytoplasmic antibody-associated vasculitis, Behçet disease, Cogan’s syndrome, giant cell arthritis, relapsing polychondritis, rheumatoid arthritis, sarcoidosis, systemic lupus erythematosus, or Takayasu arteritis, defined using ICD codes and self-reported information (Table S2).

Any connective tissue disorder was defined as Marfan syndrome, Ehlers-Danlos syndrome, or any non-specific connective tissue disorder, defined using ICD codes and self-reported information (Table S2).

Statistical analyses

All associations in UK Biobank were obtained by logistic regression (for disease outcomes) adjusted for age, sex, and 10 genomic principal components of ancestry using an additive genetic model (i.e. coefficients represent log odds ratios per additional copy of the effect allele). We repeated the primary analysis using a factorial model regarding major homozygotes as the reference group, and providing separate estimates for heterozygotes and minor homozygotes. Analyses for the primary instrument are reported as genetic associations per additional C-reactive protein decreasing allele. Analyses for the extended instrument were performed using exposure data on genetic associations with C-reactive protein,18 and implemented using the inverse-variance weighted method to combine the summarized data accounting for correlation between variants, and a fixed-effect model. The genetic correlation matrix was estimated in the analytic sample of UK Biobank participants. All analyses were performed in R (version 4.3.3). All p-values are two-sided.

Ethical approval and participant consent

Individual-level analyses on data from the UK Biobank were approved by its Research Ethics Committee and Human Tissue Authority research tissue bank under application number 7493. For analysis of prior published genetic associations, summary statistics were gathered from studies that had obtained appropriate independent ethical approval and participant consent for analyses and distribution of summary-level data, as described in the original publications.

Results

Demographic information on UK Biobank analytic sample

A total of 367,643 unrelated UK Biobank participants of European ancestries were included in individual-level data analyses. In total, 1,963 individuals had an AAA diagnosis, and 131 had a fatal AAA diagnosis. A total of 198,838 individuals were women (54.1%), 120,051 individuals had hypertension (32.6%), 169,788 individuals were ever-smokers (46.2%), 11,551 had a rheumatological disorder (3.1%), and 505 had a connective tissue disorder (0.1%). Demographic information on participants is provided in Table 2.

Table 2. Demographic information on UK Biobank participants in the analytic sample divided by stratification variables.

Trait Overall Sex Smoking status Hypertension RDs CTDs
Men Women Ever Never Yes No
Sample size (N) 367,643 168,748 198,838 169,788 197,853 120,051 244,971 11,551 505
Age (years) 57.2 57.4 57.0 58.0 56.5 59.9 55.9 59.6 56.1
Sex (% women) 54.1 0 100 48.6 58.8 44.2 59.1 62.2 68.9
AAA events (N) 1,963 1,669 294 1,622 341 1,233 717 116 10
Fatal AAA (N) 131 117 14 117 14 87 44 8 0
BMI (kg/m2) 27.4 27.8 27.0 27.7 27.1 29.2 26.4 28.3 26.7
LDL-c (mmol/L) 3.57 3.49 3.64 3.54 3.59 3.45 3.62 3.52 3.51
ApoB (g/L) 1.03 1.03 1.04 1.03 1.03 1.02 1.04 1.03 1.02
HbA1c (mmol/mol) 35.9 36.3 35.6 36.4 35.5 37.6 35.1 36.6 35.4
Open in a new tab

AAA: abdominal aortic aneurysm, ApoB: apolipoprotein-B, BMI: body mass index, CTD: connective tissue disorders, HbA1c: glycated hemoglobin, RDs: rheumatological disorders, LDL-c: low-density lipoprotein cholesterol.

Genetic evidence for efficacy of IL6 signaling inhibition in AAA

The genetic association with AAA in the AAAgen consortium was an odds ratio (OR) of 0.91 (95% confidence interval [CI]: 0.90 to 0.92, p = 4×10-30) per additional copy of the C allele of rs2228145 (Figure 2). The genetic association in FinnGen was similar: 0.86 (95% CI: 0.82, 0.91, p=7×10-9). The association in UK Biobank was also similar: 0.90 (95% CI: 0.84, 0.96, p=0.001). The genetic association with fatal AAA risk in UK Biobank was similar, but had wider confidence intervals due to the lower number of events: 0.89 (95% CI: 0.69, 1.14, p=0.34). Associations were approximately additive considering genetic subgroups separately in a factorial model, with an OR of 0.91 (95% CI: 0.83, 1.00, p = 0.065) for heterozygotes and 0.80 (95% CI: 0.70, 0.92, p = 0.001) for minor homozygotes (Figure 3).

Figure 2. Genetic associations with outcomes estimated in summarized and individual-level data.

Figure 2

Open in a new tab

Estimates represent odds ratio per additional copy of the C allele for rs2228145, corresponding to the genetically-predicted effect of increasing interleukin-6 signalling inhibition. Positive control outcomes are outcomes for which interleukin-6 inhibitors have proven efficacious. AAA: abdominal aortic aneurysm; CI: confidence interval; Ncases: number of cases; OR: odds ratio; pval: p-value.

Figure 3. Genetic associations with abdominal aortic aneurysm in UK Biobank from per allele and factorial models.

Figure 3

Open in a new tab

Estimates represent odds ratio of an abdominal aortic aneurysm per additional copy of the C allele for rs2228145 (per allele model), or for heterozygotes (AC genotype) and minor allele homozygotes (CC genotype) compared with major allele homozygotes (AA genotype). CI: confidence interval; N: sample size; OR: odds ratio; pval: p-value.

The estimate using the extended instrument of 7 variants in the IL6R gene region was OR of 0.93 (95% CI: 0.91, 0.94) per 0.1 unit lower log-transformed C-reactive protein (i.e. per 10% lower C-reactive protein). All 7 variants provided supportive evidence of a causal effect with the exception of rs12083537, which was an outlier, and hence may be pleiotropic in terms of its genetic association with the outcome (Figure S2). On the same scale, the estimate using the rs2228145 variant only was 0.90 (95% CI: 0.89, 0.92). Given limited benefit in terms of precision, and the possibilities of including pleiotropic variants and overfitting using the extended instrument,38 we performed all further analyses using the rs2228145 variant only.

There was strong evidence of colocalization for the genetic association with IL6 levels and AAA risk at the IL6R gene locus, supportive of a causal relationship: posterior probability of colocalization (PP-H4) = 0.996.

Genetic evidence for efficacy of IL6 signaling inhibition on positive control outcomes

The genetic association with rheumatoid arthritis was 0.94 (95% CI: 0.92, 0.96, p = 2×10-9) in the GWAS consortium, and 0.95 (95% CI: 0.93, 0.98, p = 0.0009) in FinnGen. The association with polymyalgia rheumatica was 0.93 (95% CI: 0.88, 0.98, p = 0.012). The association with severe COVID-19 was 0.98 (95% CI: 0.95, 1.00, p = 0.023). Associations with these positive controls provide evidence that the genetic associations are a reliable guide for the impact of IL6 signaling inhibition in clinical trials.

Genetic evidence for efficacy of IL6 signaling inhibition on other aneurysmal and cardiovascular outcomes

Associations with other aneurysmal diseases were 1.00 (95% CI: 0.92, 1.09, p = 0.95) for thoracic aortic aneurysm, and 0.99 (95% CI: 0.96, 1.03, p = 0.67) for intracranial aneurysm. It appears that genetic evidence for benefit of IL6 signaling inhibition on aneurysm risk is specific to AAA.

The association with CAD was 0.96 (95% CI: 0.95, 0.97, p = 4×10-18). We note that associations with AAA are at least twice as strong as with CAD.

Stratified analyses in UK Biobank

Stratified analyses were performed in UK Biobank participants. The genetic association with AAA risk was slightly stronger in men: 0.89 (95% CI: 0.83, 0.95, p=0.0008) than in women: 0.97 (95% CI: 0.82, 1.14, p=0.69), although there was no statistical evidence for a difference between estimates (p=0.34). Given the lower number of AAA events in women than men (294 in women, 1669 in men), the null estimate in women may reflect low statistical power to detect an association in women rather than a genuine null finding. Similar findings were obtained regardless of hypertension or smoking status. Among individuals with rheumatological or connective diseases, estimates were attenuated towards the null although the confidence intervals still overlapped (Figure 4).

Figure 4. Stratified genetic association estimates with abdominal aortic aneurysm in UK Biobank.

Figure 4

Open in a new tab

Estimates represent odds ratio of an abdominal aortic aneurysm event per additional copy of the C allele for rs2228145, corresponding to the genetically-predicted effect of greater interleukin-6 signalling inhibition. CI: confidence interval; N: sample size; OR: odds ratio; pval: p-value.

A similar difference in estimates between men and women was observed in the Global Biobank Meta-analysis Initiative. Sex-specific associations were available for rs12133641, a variant physically close to rs2228145, and in perfect linkage disequilibrium with rs2228145. The genetic association in men was OR 0.91 (95% CI: 0.87, 0.95, p = 3×10-6), and in women was OR 0.96 (95% CI: 0.88, 1.04, p=0.32). Again, while the estimate was larger in magnitude in men, there was no convincing evidence for a difference in estimates (p = 0.21). We repeat the earlier caution that this is not an independent replication of this finding, as the Global Biobank Meta-analysis Initiative includes UK Biobank participants.

Discussion

We used a genetic variant mimicking the effects of IL6 signaling inhibition in the Mendelian randomization paradigm and identified evidence supporting protective effects on AAA risk. These findings are consistent with previous work,14,17 but make a number of important advances. Firstly, we demonstrate an additive effect for each additional rs2228145 C allele (mimicking IL6 signaling inhibition) on AAA risk reduction, consistent with IL6 signaling driving AAA pathophysiology. Secondly, we show that the magnitude of the Mendelian randomization estimate is similar for risk of fatal AAA as it is for risk of any AAA, supporting similar protective effects on risk of AAA rupture. Thirdly, we find that the Mendelian randomization association was similar in population subgroups stratified by sex, blood pressure, and smoking status, but was attenuated in individuals with AAA related to rheumatological or connective tissue disease, albeit with wider confidence intervals in the diseased subgroups. This supports that protective effects of IL6 signaling inhibition may be specific to AAA arising secondary to atherosclerotic risk factors rather than due to pre-existing rheumatological disease or connective tissue disease. Fourthly, we show that the genetic evidence of effect was specific to AAA and not other types of aneurysmal disease, such as intracranial aneurysm or thoracic aortic aneurysm. The discrepancy between AAA and thoracic aortic aneurysm may relate to smooth muscle cells in the thoracic aorta originating from the neural crest and the somitic mesoderm, while smooth muscle cells in the abdominal aorta originate from the splanchnic mesoderm.39 This distinction could in turn lead to divergent responses to injury across the two sites, and thus a differing role for IL6 signaling in driving pathology.

Collectively, these findings may be directly employed to inform the clinical development of IL6 signaling inhibition for the treatment of AAA. Previous work has employed Mendelian randomization to study potential biomarkers and adverse effects of inhibiting IL6 signaling.40 Clinical trials investigating IL6 signaling inhibition for the treatment of CVD are already underway16 and the insights generated in our current study may be used to inform similar endeavors for AAA. As Mendelian randomization estimates for IL6 signaling inhibition are greater for AAA than for CAD, the beneficial effect for AAA risk may be greater on a relative scale. Indeed, IL6 signaling may represent a disease mechanism common to both AAA and CAD, in that its relevant effects include endothelial cell activation, lymphocyte proliferation and differentiation, increased coagulation, and activation of the hypothalamic-pituitary-adrenal axis.41 Other than addressing general CVD risk factors such as hypertension, dyslipidemia or diabetes mellitus, there are currently no approved pharmacological therapies for the treatment of AAA. The current standard of care is based on monitoring for expansion, with the option of surgical intervention should certain size thresholds be crossed or in the case of rupture, which itself is associated with an approximately 90% mortality rate.2 Thus, the availability of efficacious pharmacological therapies for AAA treatment would represent a notable advance in patient care.

There is already a plethora of data implicating IL6 signaling in AAA pathophysiology. Inhibition of IL6 signaling has been shown to limit progression of AAA in animal models,42 and is also associated with improved survival.43 In humans, IL6 is abundantly expressed in AAA tissue,44 and may even be a source of systemic IL6.45 Previous Mendelian randomization analyses have supported IL6 signaling in AAA risk,14,43 as well as potential effects of IL6 signaling on reducing progression on AAA, although this latter work was limited by low statistical power. Inflammation is a key driver in AAA occurring outside the background of a rheumatological or connective tissue disease,46 with inflammatory cell infiltrates observed in the aneurysm wall,47 and aneurysm mural thrombus.48 It therefore follows that inhibition of IL6 signaling might reduce AAA risk, progression, and rupture.

This work has a number of strengths. Using the Mendelian randomization paradigm, we were able to efficiently generate causal evidence in humans to inform clinical development efforts supporting IL6 signaling inhibition for the treatment of AAA. Specifically, the insights generated here may be used to prioritize the specific type of aneurysmal disease and the target population. To ensure the robustness of our approach, we validated the method with established positive control outcomes where IL6 signaling inhibition has proven efficacious, including for rheumatoid arthritis, polymyalgia rheumatica, and COVID-19. Statistically, we showed that using a biologically validated missense variant as the instrument in Mendelian randomization produced similar estimates to a polygenic cis-instrument, and further through colocalization we generated support that genetic confounding through a variant in linkage disequilibrium was unlikely to be explaining the observed associations.

There are also limitations. The Mendelian randomization paradigm employed in this work considers the cumulative lifetime effect of genetic variation on risk of clinical outcomes in a select population. Caution should therefore be taken when extrapolating these findings to assume the effect of a clinical intervention having a larger effect at a discrete timepoint in life in an entirely different population. In this regard, these analyses were largely limited to European genetic ancestry populations, although previous work supports that similar associations may hold in other genetic ancestry populations.49 Furthermore, these genetic analyses evaluated the risk of developing AAA but do not directly assess the clinical impact of reduced IL6 signaling after AAA has already manifested, the clinical setting in which AAA therapies would actually be applied. Our analyses assume a linear model, and so provide population-averaged estimates. We are not able to detect departures from linearity or to model the shape of the potential effect of IL6 signaling on AAA risk. We did not have access to measurements of aneurysm size, and so we were not able to estimate the predicted effect of IL6 signaling on AAA progression, limiting somewhat the clinical utility of findings for IL6 treatment. Finally, a fundamental limitation of all Mendelian randomization analyses is that it assumes any genetic associations between the instrument and outcome are only occurring through the exposure and not some pleiotropic pathway, which can never be proven. It therefore remains possible that our findings may be biased by such pleiotropic effects.

In conclusion, this Mendelian randomization study finds human causal evidence to support the clinical development of IL6 signaling inhibition for the treatment of AAA, including the specific disease subtypes and target populations to be prioritized. Five phase 3 clinical trials of IL6 signaling inhibition for CVD are already underway,16 and the weight of the supportive evidence for AAA coupled with the unmet need for efficacious medical therapies signposts this as a promising opportunity for clinical investigation.

Supplementary Material

STROBE guidelines
Supplementary material

Highlights.

  • Consistent with previous work, this study identified genetic evidence to support a protective effect of inteluekin-6 (IL6) signaling inhibition on abdominal aortic aneurysm (AAA).

  • There is evidence of an additive effect for each additional rs2228145 C allele (mimicking IL6 signaling inhibition) on AAA risk reduction, consistent with IL6 signaling driving AAA pathophysiology.

  • The magnitude of the Mendelian randomization estimate is similar for risk of fatal AAA as it is for risk of any AAA, supporting similar protective effects on risk of AAA rupture.

  • The Mendelian randomization association was similar in population subgroups stratified by sex, blood pressure, and smoking status, but was attenuated in individuals with AAA related rheumatological or connective tissue disease, albeit with wider confidence intervals in the diseased subgroups.

  • The genetic evidence of effect was specific to AAA and not other types of aneurysmal disease, such as intracranial aneurysm or thoracic aortic aneurysm.

Acknowledgements

This research has been conducted using the UK Biobank Resource under Application Number 7439. The authors acknowledge participants and investigators of the UK Biobank and the other cohort studies incorporated in our work. We thank the cohorts and consortia that made their summary-level data publicly available. Data sources are cited throughout this manuscript. Downloads were performed through the following repositories: IL6, https://www.ebi.ac.uk/gwas/studies/GCST90012005; C-reactive protein, https://www.ebi.ac.uk/gwas/studies/GCST90029070; AAA, https://csg.sph.umich.edu/willer/public/AAAgen2023/; AAA (sex-stratified), https://www.globalbiobankmeta.org/resources; Rheumatoid arthritis (consortium), https://www.ebi.ac.uk/gwas/studies/GCST90132222; Rheumatoid arthritis and AAA (FinnGen), https://www.finngen.fi/en/access_results; Polymyalgia rheumatica, https://www.ebi.ac.uk/gwas/studies/GCST90129454; Severe COVID-19, https://www.covid19hg.org/results/r7/; Thoracic aortic aneurysm, https://www.ebi.ac.uk/gwas/studies/GCST90027266; Intracranial aneurysm, https://cd.hugeamp.org/downloads.html; Coronary artery disease, https://www.ebi.ac.uk/gwas/studies/GCST90132315.

Sources of Funding

This work was supported by Tourmaline Bio.

Footnotes

Disclosures

Sequoia Genetics is a private company that works with investors, pharma, biotech, and academia by performing research that leverages genetic data to help inform drug discovery and development. SB, HTC and DG are employees of Sequoia Genetics, and were supported by Tourmaline Bio to undertake this work. YC and ED are employees and shareholder of Tourmaline Bio. DG has financial interests in several biotechnology companies.

Data and code availability

Summary statistics used in our analyses can be accessed through the citations provided. Requests to access UK Biobank data can be made by bona fide researchers from any sector. More information can be found at https://www.ukbiobank.ac.uk/enable-your-research. Statistical code for the analyses undertaken in this work are available from the corresponding author upon reasonable request. This study is reported using the Strengthening the Reporting of Observational Studies in Epidemiology-MR guidelines (Checklist).37

References

  • 1.Benson RA, Poole R, Murray S, Moxey P, Loftus IM. Screening results from a large United Kingdom abdominal aortic aneurysm screening center in the context of optimizing United Kingdom National Abdominal Aortic Aneurysm Screening Programme protocols. J Vasc Surg. 2016;63:301–304. doi: 10.1016/j.jvs.2015.08.091. [DOI] [PubMed] [Google Scholar]
  • 2.Kent KC. Clinical practice. Abdominal aortic aneurysms. N Engl J Med. 2014;371:2101–2108. doi: 10.1056/NEJMcp1401430. [DOI] [PubMed] [Google Scholar]
  • 3.Centers for Disease Control and Prevention. Underlying Cause of Death 1999-2020. [Internet] Available from: https://wonder.cdc.gov/wonder/help/ucd.html.
  • 4.Chaikof EL, Dalman RL, Eskandari MK, Jackson BM, Lee WA, Mansour MA, Mastracci TM, Mell M, Murad MH, et al. The Society for Vascular Surgery practice guidelines on the care of patients with an abdominal aortic aneurysm. J Vasc Surg. 2018;67:2–77.:e2. doi: 10.1016/j.jvs.2017.10.044. [DOI] [PubMed] [Google Scholar]
  • 5.GBD 2017 Causes of Death Collaborators. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Lond Engl. 2018;392:1736–1788. doi: 10.1016/S0140-6736(18)32203-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Krafcik BM, Stone DH, Cai M, Jarmel IA, Eid M, Goodney PP, Columbo JA, Mayo Smith MF. Changes in global mortality from aortic aneurysm. J Vasc Surg. 2024;80:81–88.:e1. doi: 10.1016/j.jvs.2024.02.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Wouters OJ, McKee M, Luyten J. Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018. JAMA. 2020;323:844–853. doi: 10.1001/jama.2020.1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Paul SM, Mytelka DS, Dunwiddie CT, Persinger CC, Munos BH, Lindborg SR, Schacht AL. How to improve R&D productivity: the pharmaceutical industry’s grand challenge. Nat Rev Drug Discov. 2010;9:203–214. doi: 10.1038/nrd3078. [DOI] [PubMed] [Google Scholar]
  • 9.Hingorani AD, Kuan V, Finan C, Kruger FA, Gaulton A, Chopade S, Sofat R, MacAllister RJ, Overington JP, et al. Improving the odds of drug development success through human genomics: modelling study. Sci Rep. 2019;9:18911. doi: 10.1038/s41598-019-54849-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Gill D, Georgakis MK, Walker VM, Schmidt AF, Gkatzionis A, Freitag DF, Finan C, Hingorani AD, Howson JMM, et al. Mendelian randomization for studying the effects of perturbing drug targets. Wellcome Open Res. 2021;6:16. doi: 10.12688/wellcomeopenres.16544.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Burgess S, Mason AM, Grant AJ, Slob EAW, Gkatzionis A, Zuber V, Patel A, Tian H, Liu C, et al. Using genetic association data to guide drug discovery and development: Review of methods and applications. Am J Hum Genet. 2023;110:195–214. doi: 10.1016/j.ajhg.2022.12.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Daghlas I, Gill D. Mendelian randomization as a tool to inform drug development using human genetics. Camb Prisms Precis Med. 2023;1:e16. doi: 10.1017/pcm.2023.5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Interleukin-6 Receptor Mendelian Randomisation Analysis (IL6R MR) Consortium. Swerdlow DI, Holmes MV, Kuchenbaecker KB, Engmann JEL, Shah T, Sofat R, Guo Y, Chung C, et al. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis. Lancet Lond Engl. 2012;379:1214–1224. doi: 10.1016/S0140-6736(12)60110-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Georgakis MK, Malik R, Gill D, Franceschini N, Sudlow CLM, Dichgans M, INVENT Consortium,CHARGE Inflammation Working Group Interleukin-6 Signaling Effects on Ischemic Stroke and Other Cardiovascular Outcomes: A Mendelian Randomization Study. Circ Genomic Precis Med. 2020;13:e002872. doi: 10.1161/CIRCGEN.119.002872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Georgakis MK, Malik R, Richardson TG, Howson JMM, Anderson CD, Burgess S, Hovingh GK, Dichgans M, Gill D. Associations of genetically predicted IL-6 signaling with cardiovascular disease risk across population subgroups. BMC Med. 2022;20:245. doi: 10.1186/s12916-022-02446-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ridker PM. From RESCUE to ZEUS: will interleukin-6 inhibition with ziltivekimab prove effective for cardiovascular event reduction? Cardiovasc Res. 2021;117:e138–e140. doi: 10.1093/cvr/cvab231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Harrison SC, Smith AJP, Jones GT, Swerdlow DI, Rampuri R, Bown MJ, Folkersen L, Baas AF, Aneurysm Consortium et al. Interleukin-6 receptor pathways in abdominal aortic aneurysm. Eur Heart J. 2013;34:3707–3716. doi: 10.1093/eurheartj/ehs354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Said S, Pazoki R, Karhunen V, Võsa U, Ligthart S, Bodinier B, Koskeridis F, Welsh P, Alizadeh BZ, et al. Genetic analysis of over half a million people characterises C-reactive protein loci. Nat Commun. 2022;13:2198. doi: 10.1038/s41467-022-29650-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Roychowdhury T, Klarin D, Levin MG, Spin JM, Rhee YH, Deng A, Headley CA, Tsao NL, Gellatly C, et al. Genome-wide association meta-analysis identifies risk loci for abdominal aortic aneurysm and highlights PCSK9 as a therapeutic target. Nat Genet. 2023;55:1831–1842. doi: 10.1038/s41588-023-01510-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Giambartolomei C, Vukcevic D, Schadt EE, Franke L, Hingorani AD, Wallace C, Plagnol V. Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics. PLOS Genet. 2014;10:e1004383. doi: 10.1371/journal.pgen.1004383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Folkersen L, Gustafsson S, Wang Q, Hansen DH, Hedman ÅK, Schork A, Page K, Zhernakova DV, Wu Y, et al. Genomic and drug target evaluation of 90 cardiovascular proteins in 30,931 individuals. Nat Metab. 2020;2:1135–1148. doi: 10.1038/s42255-020-00287-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Wallace C. Eliciting priors and relaxing the single causal variant assumption in colocalisation analyses. PLOS Genet. 2020;16:e1008720. doi: 10.1371/journal.pgen.1008720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zuber V, Grinberg NF, Gill D, Manipur I, Slob EAW, Patel A, Wallace C, Burgess S. Combining evidence from Mendelian randomization and colocalization: Review and comparison of approaches. Am J Hum Genet. 2022;109:767–782. doi: 10.1016/j.ajhg.2022.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ishigaki K, Sakaue S, Terao C, Luo Y, Sonehara K, Yamaguchi K, Amariuta T, Too CL, Laufer VA, et al. Multi-ancestry genome-wide association analyses identify novel genetic mechanisms in rheumatoid arthritis. Nat Genet. 2022;54:1640–1651. doi: 10.1038/s41588-022-01213-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Kurki MI, Karjalainen J, Palta P, Sipilä TP, Kristiansson K, Donner KM, Reeve MP, Laivuori H, Aavikko M, et al. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature. 2023;613:508–518. doi: 10.1038/s41586-022-05473-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zorina-Lichtenwalter K, Bango CI, Van Oudenhove L, Čeko M, Lindquist MA, Grotzinger AD, Keller MC, Friedman NP, Wager TD. Genetic risk shared across 24 chronic pain conditions: identification and characterization with genomic structural equation modeling. Pain. 2023;164:2239–2252. doi: 10.1097/j.pain.0000000000002922. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Niemi MEK, Karjalainen J, Liao RG, Neale BM, Daly M, Ganna A, Pathak GA, Andrews SJ, Kanai M, et al. Mapping the human genetic architecture of COVID-19. Nature. 2021;600:472–477. doi: 10.1038/s41586-021-03767-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bijlsma JWJ, Welsing PMJ, Woodworth TG, Middelink LM, Pethö-Schramm A, Bernasconi C, Borm MEA, Wortel CH, ter Borg EJ, et al. Early rheumatoid arthritis treated with tocilizumab, methotrexate, or their combination (U-Act-Early): a multicentre, randomised, double-blind, double-dummy, strategy trial. The Lancet. 2016;388:343–355. doi: 10.1016/S0140-6736(16)30363-4. [DOI] [PubMed] [Google Scholar]
  • 29.RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet Lond Engl. 2021;397:1637–1645. doi: 10.1016/S0140-6736(21)00676-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Spiera RF, Unizony S, Warrington KJ, Sloane J, Giannelou A, Nivens MC, Akinlade B, Wong W, Bhore R, et al. Sarilumab for Relapse of Polymyalgia Rheumatica during Glucocorticoid Taper. N Engl J Med. 2023;389:1263–1272. doi: 10.1056/NEJMoa2303452. [DOI] [PubMed] [Google Scholar]
  • 31.Roychowdhury T, Lu H, Hornsby WE, Crone B, Wang GT, Guo D-C, Sendamarai AK, Devineni P, Lin M, et al. Regulatory variants in TCF7L2 are associated with thoracic aortic aneurysm. Am J Hum Genet. 2021;108:1578–1589. doi: 10.1016/j.ajhg.2021.06.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Bakker MK, van der Spek RAA, van Rheenen W, Morel S, Bourcier R, Hostettler IC, Alg VS, van Eijk KR, Koido M, et al. Genome-wide association study of intracranial aneurysms identifies 17 risk loci and genetic overlap with clinical risk factors. Nat Genet. 2020;52:1303–1313. doi: 10.1038/s41588-020-00725-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, Weeks EM, Wang M, Hindy G, et al. Discovery and systematic characterization of risk variants and genes for coronary artery disease in over a million participants. Nat Genet. 2022;54:1803–1815. doi: 10.1038/s41588-022-01233-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Zhou W, Kanai M, Wu K-HH, Rasheed H, Tsuo K, Hirbo JB, Wang Y, Bhattacharya A, Zhao H, et al. Global Biobank Meta-analysis Initiative: Powering genetic discovery across human disease. Cell Genomics. 2022;2:100192. doi: 10.1016/j.xgen.2022.100192. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, Downey P, Elliott P, Green J, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12:e1001779. doi: 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Astle WJ, Elding H, Jiang T, Allen D, Ruklisa D, Mann AL, Mead D, Bouman H, Riveros-Mckay F, et al. The Allelic Landscape of Human Blood Cell Trait Variation and Links to Common Complex Disease. Cell. 2016;167:1415–1429.:e19. doi: 10.1016/j.cell.2016.10.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Skrivankova VW, Richmond RC, Woolf BAR, Yarmolinsky J, Davies NM, Swanson SA, VanderWeele TJ, Higgins JPT, Timpson NJ, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA. 2021;326:1614–1621. doi: 10.1001/jama.2021.18236. [DOI] [PubMed] [Google Scholar]
  • 38.Burgess S, Zuber V, Valdes-Marquez E, Sun BB, Hopewell JC. Mendelian randomization with fine-mapped genetic data: Choosing from large numbers of correlated instrumental variables. Genet Epidemiol. 2017;41:714–725. doi: 10.1002/gepi.22077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Zhang L, Issa Bhaloo S, Chen T, Zhou B, Xu Q. Role of Resident Stem Cells in Vessel Formation and Arteriosclerosis. Circ Res. 2018;122:1608–1624. doi: 10.1161/CIRCRESAHA.118.313058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Georgakis MK, Malik R, Li X, Gill D, Levin MG, Vy HMT, Judy R, Ritchie M, Verma SS, et al. Genetically Downregulated Interleukin-6 Signaling Is Associated With a Favorable Cardiometabolic Profile: A Phenome-Wide Association Study. Circulation. 2021;143:1177–1180. doi: 10.1161/CIRCULATIONAHA.120.052604. [DOI] [PubMed] [Google Scholar]
  • 41.Hartman J, Frishman WH. Inflammation and Atherosclerosis: A Review of the Role of Interleukin-6 in the Development of Atherosclerosis and the Potential for Targeted Drug Therapy. Cardiol Rev. 2014;22:147. doi: 10.1097/CRD.0000000000000021. [DOI] [PubMed] [Google Scholar]
  • 42.Nishihara M, Aoki H, Ohno S, Furusho A, Hirakata S, Nishida N, Ito S, Hayashi M, Imaizumi T, et al. The role of IL-6 in pathogenesis of abdominal aortic aneurysm in mice. PLoS ONE. 2017;12:e0185923. doi: 10.1371/journal.pone.0185923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Paige E, Clément M, Lareyre F, Sweeting M, Raffort J, Grenier C, Finigan A, Harrison J, Peters JE, et al. Interleukin-6 Receptor Signaling and Abdominal Aortic Aneurysm Growth Rates. Circ Genomic Precis Med. 2019;12:e002413. doi: 10.1161/CIRCGEN.118.002413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Shteinberg D, Halak M, Shapiro S, Kinarty A, Sobol E, Lahat N, Karmeli R. Abdominal aortic aneurysm and aortic occlusive disease: a comparison of risk factors and inflammatory response. Eur J Vasc Endovasc Surg Off J Eur Soc Vasc Surg. 2000;20:462–465. doi: 10.1053/ejvs.2000.1210. [DOI] [PubMed] [Google Scholar]
  • 45.Dawson J, Cockerill GW, Choke E, Belli A-M, Loftus I, Thompson MM. Aortic aneurysms secrete interleukin-6 into the circulation. J Vasc Surg. 2007;45:350–356. doi: 10.1016/j.jvs.2006.09.049. [DOI] [PubMed] [Google Scholar]
  • 46.Brophy CM, Reilly JM, Smith GJ, Tilson MD. The role of inflammation in nonspecific abdominal aortic aneurysm disease. Ann Vasc Surg. 1991;5:229–233. doi: 10.1007/BF02329378. [DOI] [PubMed] [Google Scholar]
  • 47.Kokje VBC, Gäbel G, Koole D, Northoff BH, Holdt LM, Hamming JF, Lindeman JHN. IL-6: A Janus-like factor in abdominal aortic aneurysm disease. Atherosclerosis. 2016;251:139–146. doi: 10.1016/j.atherosclerosis.2016.06.021. [DOI] [PubMed] [Google Scholar]
  • 48.Fontaine V, Jacob M-P, Houard X, Rossignol P, Plissonnier D, Angles-Cano E, Michel J-B. Involvement of the mural thrombus as a site of protease release and activation in human aortic aneurysms. Am J Pathol. 2002;161:1701–1710. doi: 10.1016/S0002-9440(10)64447-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Zhao SS, Gill D. Genetically Proxied IL-6 Receptor Inhibition and Coronary Artery Disease Risk in a Japanese Population. Clin Ther. 2024:S0149-2918(24)00107–3. doi: 10.1016/j.clinthera.2024.04.015. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

STROBE guidelines
EMS202956-supplement-STROBE_guidelines.docx (36.8KB, docx)
Supplementary material
EMS202956-supplement-Supplementary_material.pdf (250.6KB, pdf)

Data Availability Statement

Summary statistics used in our analyses can be accessed through the citations provided. Requests to access UK Biobank data can be made by bona fide researchers from any sector. More information can be found at https://www.ukbiobank.ac.uk/enable-your-research. Statistical code for the analyses undertaken in this work are available from the corresponding author upon reasonable request. This study is reported using the Strengthening the Reporting of Observational Studies in Epidemiology-MR guidelines (Checklist).37

RESOURCES