Mechanistic interaction concerns how exposures affect the outcome. When investigating mechanisms, synergism is the most mentioned type in the fields of genetic study and pharmacology. Synergism is defined under the framework of sufficient component cause model, which is difficult to be quantified directly. Sufficient cause interaction (SCI) is the only alternative metric to imply the existence of synergism. VanderWeele and Robins provided empirical tests for SCIs. However, this test only assesses the lower bound of SCIs rather than estimate SCIs directly due to the lack of the degree of freedom, which causes low power. To address this issue, in this study, we propose a novel method to estimate the probability of individual with SCI by introducing a new factor named quasi-instrumental variable, which is necessary for the background condition of SCI. We also develop a corresponding hypothesis test and show that it is more powerful than the existing empirical test. We demonstrate this method by applying it to estimate the synergistic effects between intestinal bacteria on the formation of Parkinson's disease.
VanderWeeleT. Explanation in causal inference: methods for mediation and interaction. New York, NY, USA: Oxford University Press, 2015.
2.
RothmanKJ. Causes. Am J Epidemiol1976; 104: 587–592.
3.
RothmanKJGreenlandSLashTL. Modern epidemiology, Vol. 3. Wolters Kluwer Health/Lippincott Williams & Wilkins Philadelphia, 2008.
4.
HausenH. Human genital cancer: synergism between two virus infections or synergism between a virus infection and initiating events?Lancet1982; 320: 1370–1372.
5.
LeiterEHReifsnyderPCFlurkeyK, et al.NIDDM Genes in mice: deleterious synergism by both parental genomes contributes to diabetogenic thresholds. Diabetes1998; 47: 1287–1295.
6.
TallaridaRJ. Drug synergism: its detection and applications. J Pharmacol Exp Ther2001; 298: 865–872.
7.
MazumderACerellaCDiederichM. Natural scaffolds in anticancer therapy and precision medicine. Biotechnol Adv2018; 36: 1563–1585.
8.
Graham-GuryshEGMurthyABMooreKM, et al.Synergistic drug combinations for a precision medicine approach to interstitial glioblastoma therapy. J Controlled Release2020; 323: 282–292.
9.
SuzukiEYamamotoETsudaT. On the link between sufficient-cause model and potential-outcome model. Epidemiology2011; 22: 131–132.
10.
VanderWeeleTJRobinsJM. Empirical and counterfactual conditions for sufficient cause interactions. Biometrika2008; 95: 49–61.
11.
SjölanderALeeWKällbergH, et al.Bounds on sufficient-cause interaction. Eur J Epidemiol2014; 29: 813–820.
12.
LeeW-CWuY-C. Disease attribution to multiple exposures using aggregate data. J Epidemiol2023; 33: 405–409.
13.
SuzukiEYamamotoE. Marginal sufficient component cause model: an emerging causal model with merits?Epidemiology2021; 32: 838–845.
14.
AngristJDImbensGWRubinDB. Identification of causal effects using instrumental variables. J Am Stat Assoc1996; 91: 444–455.
15.
SengLLLiuCTWangJ, et al.Instrumental variable model average with applications in Mendelian randomization. Stat Med2023; 42: 3547–3567.
16.
An-ShunTaiYen-TsungHuangWen-ChungLeeSheng-HsuanLin, et al.Conceptualizing agonistic interaction in a marginal sufficient component cause model: an alternative interpretation for subadditive interaction. J Chinese Stat Assoc2020; 58: 168–198.
17.
HuangYTTaiASChouMY, et al.Six-way decomposition of causal effects: unifying mediation and mechanistic interaction. Stat Med2020; 39: 4051–4068.
18.
LinJHLinKILeeWC, et al.Stochastic approach for mechanistic interaction under longitudinal studies with noninformative right censoring. Stat Med2020; 39: 114–128.
19.
LinS. MSR78 Population attributable fraction based on sufficient causal framework for mediation settings-evaluation of resource distribution on prevention and treatment for stroke. Value Health2022; 25: S365.
20.
LinSHHuangYTYangHI. On identification of agonistic interaction: hepatitis B and C interaction on hepatocellular carcinoma. Stat Med2019; 38: 2467–2476.
21.
MiettinenOS. Causal and preventive interdependence: elementary principles. Scand J Work Environ Health1982; 8: 159–168.
22.
VanderWeeleTJRobinsJM. The identification of synergism in the sufficient-component-cause framework. Epidemiology2007; 18: 329–339.
23.
VanderWeeleTJRobinsJM. Directed acyclic graphs, sufficient causes, and the properties of conditioning on a common effect. Am J Epidemiol2007; 166: 1096–1104.
24.
VanderWeeleTJ. Sufficient cause interactions and statistical interactions. Epidemiology2009; 20: 6–13.
25.
KjRGreenlandSLashT. Modern epidemiology. Bostonrroronlo: Lillie, Brown and Company, 1986.
26.
VanderWeeleTJRichardsonTS. General theory for interactions in sufficient cause models with dichotomous exposures. Ann Stat2012; 40: 2128.
27.
GreenlandS. An introduction to instrumental variables for epidemiologists. Int J Epidemiol2000; 29: 722–729.
28.
RassenJASchneeweissSGlynnRJ, et al.Instrumental variable analysis for estimation of treatment effects with dichotomous outcomes. Am J Epidemiol2008; 169: 273–284.
29.
ChenYBriesacherBA. Use of instrumental variable in prescription drug research with observational data: a systematic review. J Clin Epidemiol2011; 64: 687–700.
30.
HausmanJANeweyWKWoutersenT, et al.Instrumental variable estimation with heteroskedasticity and many instruments. Quant Econom2012; 3: 211–255.
31.
MullahyJ. Instrumental-variable estimation of count data models: applications to models of cigarette smoking behavior. Rev Econ Stat1997; 79: 586–593.
32.
SchultzTPTanselA. Wage and labor supply effects of illness in Cote d'Ivoire and Ghana: instrumental variable estimates for days disabled. J Dev Econ1997; 53: 251–286.
33.
PearlJ. Causal diagrams for empirical research. Biometrika1995; 82: 669–688.
34.
Barrera-GómezJAgierLPortengenL, et al.A systematic comparison of statistical methods to detect interactions in exposome-health associations. Environ Health2017; 16: 74.
35.
LampaELindLLindPM, et al.The identification of complex interactions in epidemiology and toxicology: a simulation study of boosted regression trees. Environ Health2014; 13: 57.
36.
BartelsLM. Instrumental and” quasi-instrumental” variables. Am J Pol Sci1991; 35: 777–800.
37.
TaiA-SHuangY-TYangH-I, et al.G-computation to causal mediation analysis with sequential multiple mediators—investigating the vulnerable time window of HBV activity for the mechanism of HCV induced hepatocellular carcinoma. Front Public Health2022; 9: 757942.
38.
Hill-BurnsEMDebeliusJWMortonJT, et al.Parkinson's disease and Parkinson's disease medications have distinct signatures of the gut microbiome. Mov Disord2017; 32: 739–749.
39.
OseiRYangCCuiL, et al.Isolation, identification, and pathogenicity of Lelliottia amnigena causing soft rot of potato tuber in China. Microb Pathog2022; 164: 105441.
40.
LiKZhangLXueJ, et al.Dietary inulin alleviates diverse stages of type 2 diabetes mellitus via anti-inflammation and modulating gut microbiota in db/db mice. Food Funct2019; 10: 1915–1927.
41.
ZhangYChenLHuM, et al.Dietary type 2 resistant starch improves systemic inflammation and intestinal permeability by modulating microbiota and metabolites in aged mice on high-fat diet. Aging (Albany NY)2020; 12: 9173–9187.
42.
MörklSLacknerSMeinitzerA, et al.Gut microbiota, dietary intakes and intestinal permeability reflected by serum zonulin in women. Eur J Nutr2018; 57: 2985–2997.
43.
VanderweeleTJ. Sufficient cause interactions for categorical and ordinal exposures with three levels. Biometrika2010; 97: 647–659.
44.
TianHBurgessS. Estimation of time-varying causal effects with multivariable Mendelian randomization: some cautionary notes. Int J Epidemiol2023; 52: 846–857.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
