Journal of the American Heart Association ORIGINAL RESEARCH Sex Differences in Modifiable Risk Factors and Severity of Coronary Artery Disease Olivia Manfrini , MD; Jinsung Yoon , PhD; Mihaela van der Schaar , PhD; Sasko Kedev , MD, PhD; Marija Vavlukis , MD, PhD; Goran Stankovic , MD, PhD; Marialuisa Scarpone, MD; Davor Miličić, MD, PhD; Zorana Vasiljevic, MD, PhD; Lina Badimon , PhD; Edina Cenko , MD, PhD; Raffaele Bugiardini , MD BACKGROUND: It is still unknown whether traditional risk factors may have a sex-specific impact on coronary artery disease (CAD) burden. METHODS AND RESULTS: We identified 14 793 patients who underwent coronary angiography for acute coronary syndromes in the ISACS-TC (International Survey of Acute Coronary Syndromes in Transitional Countries; Clini​calTr​ials.gov, NCT01218776) registry from 2010 to 2019. The main outcome measure was the association between traditional risk factors and severity of CAD and its relationship with 30-day mortality. Relative risk (RR) ratios and 95% CIs were calculated from the ratio of the abso- lute risks of women versus men using inverse probability of weighting. Estimates were compared by test of interaction on the log scale. Severity of CAD was categorized as obstructive (≥50% stenosis) versus nonobstructive CAD. The RR ratio for ob- structive CAD in women versus men among people without diabetes mellitus was 0.49 (95% CI, 0.41–0.60) and among those with diabetes mellitus was 0.89 (95% CI, 0.62–1.29), with an interaction by diabetes mellitus status of P =0.002. Exposure to smoking shifted the RR ratios from 0.50 (95% CI, 0.41–0.61) in nonsmokers to 0.75 (95% CI, 0.54–1.03) in current smokers, with an interaction by smoking status of P=0.018. There were no significant sex-related interactions with hypercholesterolemia and hypertension. Women with obstructive CAD had higher 30-day mortality rates than men (RR, 1.75; 95% CI, 1.48–2.07). No sex differences in mortality were observed in patients with nonobstructive CAD. Downloaded from http://ahajournals.org by on October 30, 2023 CONCLUSIONS: Obstructive CAD in women signifies a higher risk for mortality compared with men. Current smoking and dia- betes mellitus disproportionally increase the risk of obstructive CAD in women. Achieving the goal of improving cardiovascular health in women still requires intensive efforts toward further implementation of lifestyle and treatment interventions. REGISTRATION: URL: https://www.clini​caltr​ials.gov; Unique identifier: NCT01218776. Key Words: conventional risk factors ■ diabetes mellitus ■ obstructive coronary artery disease ■ sex differences ■ smoking A lthough the relationship between traditional car- These observations attest that there is a substan- diovascular risk factors and clinical event rates tial void in current understanding of the pathogene- is well established in both women and men,1 it sis of coronary heart disease (CHD) in women. This remains unclear whether the presence of risk factors perceived void has led to considerable research on correlate with the extent of atherosclerosis and mortal- nontraditional risk factors as a cause of ischemia in ity, especially in women. Accordingly, acute coronary women.5 Yet, data to support these new pathogenetic syndromes (ACS) may be caused by ruptures of small hypotheses are scarce,6 and some epidemiologic insignificant rather than severely narrowed plaques,2 studies have suggested that traditional risk factors may and women present more often with nonobstructive confer a greater proportional excess cardiovascular coronary lesions than men at cardiac catheterization.3,4 risk to women than to men.7–9 Correspondence to: Raffaele Bugiardini, MD, Department of Experimental, Diagnostic and Specialty Medicine, University of Bologna, Policlinico Sant’Orsola Malpighi, Padiglione 11, Via Giuseppe Massarenti 9, 40138 Bologna, Italy. E-mail: raffaele.bugiardini@unibo.it Supplementary Materials for this article are available at https://www.ahajo​urnals.org/doi/suppl/​10.1161/JAHA.120.017235 For Sources of Funding and Disclosures, see page 10. © 2020 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. JAHA is available at: www.ahajournals.org/journal/jaha J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172351 Manfrini et al Sex differences in CAD US Surgeon General report of 2006 suggests that CLINICAL PERSPECTIVE the simultaneous presence of smoking with another major risk factor is estimated to quadruple the risk. What Is New? The presence of 2 other risk factors with smoking re- • Little is currently known about the sex-specific sults in ≈8 times the risk of individuals with no risk associations between traditional risk factors factors.10 Given this, different distributions of risk fac- and the degree of coronary artery disease tors in different populations may account for the sex (CAD) with related outcomes. differences in outcome observed in some but not in • We approached this issue by reviewing the other studies. Most of the prior studies on this issue presence of traditional risk factors in 14 793 pa- are meta-analyses.7–9 Lack of individual participant tients who were referred to coronary angiogra- data in such investigations has precluded the under- phy for an acute coronary syndrome. Severity taking of more in-depth analyses. of CAD was categorized as obstructive (≥50% stenosis) versus nonobstructive CAD. We sought to investigate at an individual level whether • Women with obstructive CAD have a roughly there are sex differences in the 4 modifiable traditional 75% greater excess risk of 30-day mortality cardiovascular risk factors and how these differences compared with men; cigarette smoking and may impact the severity of coronary artery disease diabetes mellitus disproportionally increase the (CAD) and outcomes in ACS. According to European risk of obstructive CAD in women. Society of Cardiology (ESC) guidelines,11 the severity of atherosclerosis was dichotomized into obstructive ver- What Are the Clinical Implications? sus nonobstructive CAD. To obtain comparable results • Our findings support development of prevention between women and men, we used a nonparametric strategies in women at a greater level to those balancing strategy by weighting to adjust for differences that exist in men. between nonobstructive and obstructive CAD and • Intense efforts to reduce tobacco use and in- among traditional risk factors. This approach may offer creased screening for prediabetes mellitus combined with more stringent follow-up of insights on the associations between typical risk factors women with a history of gestational diabetes and anatomical CAD burden, which may help to imple- mellitus have great potential to decrease the ment sex-tailored preventive strategies. sex lag in cardiovascular disease mortality in women compared with men. Downloaded from http://ahajournals.org by on October 30, 2023 METHODS Setting and Design The authors declare that all supporting data are avail- Nonstandard Abbreviations and Acronyms able within the article and its supplemental material. Study participants were recruited from 22 tertiary ESC European Society of Cardiology healthcare services of the ISACS-TC (International ISACS-TC International Survey of Acute Survey of Acute Coronary Syndromes in Transitional Coronary Syndromes in Transitional Countries; Clini​calTr​ials.gov, NCT01218776) registry Countries providing advanced medical investigation and treat- KNN k-nearest neighbors ment including percutaneous coronary intervention MONICA Monitoring of Trends and (PCI) and/or cardiac surgery.12,13 The data coordinat- Determinants in Cardiovascular ing center has been established at the University of Disease Bologna. The local research ethics committee from NHIS National Health Interview Survey each hospital approved the study. Because patient RoPR Registry of Patient Registries information was collected anonymously, institutional RR relative risk review boards waived the need for individual informed consent. All data were transferred to the Department of Electrical and Computer Engineering, University of A source of uncertainty merits attention. California, Los Angeles, CA, where final statistical anal- Epidemiologic studies often present the relationship yses were performed. between risk factors and incident CHD controlling for only age and sex, but not for concomitant risk factors. Results may be inconsistent because of the Patient Population inclusion of a larger number of people with multiple The initial study population consisted of White pa- traditional risk factors for whom different risks exist tients who underwent coronary angiography for ACS as compared with people with a single risk factor. The from January 1, 2010, to January 15, 2019, which J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172352 Manfrini et al Sex differences in CAD is consistent with estimates of patients enrolled per from the National Health Interview Survey (NHIS).17 year and per center of similar registries according Individuals who smoked in their lifetime but who were to policies and procedures described by the RoPR not active smokers at the time of the index event (Registry of Patient Registries).14 All hospitals used were classified as former smokers regardless of time the same protocol. There were no differences in since they quit. The remaining individuals were clas- treatment strategy based on whether the patients sified as never-smokers. The sex-specific risks of had diabetes mellitus. The designated physician obstructive CAD were estimated for current smok- collected the registry data at the time of clinical as- ers compared with never-/former smokers, hereafter sessment. All patients presented with chest pain or classified as nonsmokers. The other traditional risk equivalent symptoms, such as dyspnea and fatigue. factors were assessed by designation of medical Appropriateness of inclusion was adjudicated by a history before admission in the database. Definition, cardiology specialist, considering clinical history, therefore, refers to patients with diagnosis of hyper- physical examination findings, ECG, and cardiac cholesterolemia, hypertension, or diabetes mellitus biomarkers. The use of medications given at hospi- by a general practitioner. tal admission and before the index event was noted. We defined prior users of aspirin, statins, clopidogrel, Statistical Analysis angiotensin-converting enzyme inhibitors and/or an- Baseline demographics, clinical characteristics, and giotensin II receptor blockers, and β-blockers as acute (within 24 hours) in-hospital medications and those patients who had taken these medications on clinical outcomes were compared by sex and CAD sta- a regular basis at least for 2 weeks before the onset tus: obstructive versus nonobstructive CAD. Baseline of the qualifying event. All vessels >1.5 mm in diam- characteristics were reported as percentages for cat- eter were graded for stenosis severity.15 Obstructive egorical variables and means with SDs for continuous CAD was defined according to 2013 ESC guidelines variables. Comparisons between groups were made ei- on the management of stable CAD as at least 1 main ther by Pearson chi-square test for baseline categorical branch of the epicardial coronary artery with a ≥50% variables or 2-sample t test for continuous variables. A stenosis.11 Because the definition of obstructive CAD 2-sided P value <0.05 was considered statistically sig- varies between different guidelines or studies and nificant. We used inverse probability of weighting models traditional understanding of obstructive CAD was to assess the relative risk (RR) ratios with their 95% CIs 70%,16 we repeated the analyses on outcomes shift- Downloaded from http://ahajournals.org by on October 30, 2023 for the outcomes of interest (Data S1).18 We had com- ing the cutoff for obstructive CAD at ≥70% steno- plete data on sex, age, CAD status, and 30-day mortal- sis. Patients presenting with coronary artery bypass ity. Some patients had missing data on other variables. grafting (n=261) were excluded. We also excluded We used k-nearest neighbors (KNN) algorithms as im- patients with nonobstructive CAD but prior PCI putation method to treat missing data. The value of "k" in (n=57) as these patients were previously categorized the KNN imputation was=10 (Data S1).19 Estimates were as having obstructive CAD. The final analysis sample compared by test of interaction on the log scale (Data consisted of 14 793 patients with ACS (Figure S1). S1).20 We modeled the RR (women versus men) ratio of obstructive CAD for each risk factor. Findings were Outcome Measures adjusted for demographics, cardiovascular risk factors, There were 2 outcome measures. First, we measured and history of ischemic heart disease or cardiovascular the obstructive CAD rates to investigate the relationship disorders (Table S1). For these analyses, we divided the between severity of disease and traditional risk factors; risk factors into dichotomous variables and grouped the second, we measured the 30-day mortality rates to patients into those with and those without the risk factor evaluate the relationship between severity of CAD (ob- under consideration. We tested each of the risk factors structive versus nonobstructive) and outcomes. As it is in a specific model, excluding the tested risk factor and difficult to discuss 30-day mortality following primary PCI estimated the RR ratios with and without the risk factor in patients with obstructive disease and ST-segment–el- under consideration. We calculated the 30-day mortality evation myocardial infarction (STEMI) with no reference rates by adding variables to the prior model, specifically to how revascularization rate differences come into play, ST-segment shifts in anterior leads at ECG, systolic blood subsidiary analyses were performed by noting the pro- pressure at baseline, heart rate at baseline, serum creati- portion of patients undergoing primary PCI. nine at baseline (mg/dL), and Killip class ≥2. Assessment of Traditional Risk Factors We defined current smokers as individuals who RESULTS smoked cigarettes, cigars, and cigarillos at the time Figure S1 describes patient flow through the ISACS-TC of the index event according to recommendations study. Overall, 2.1% of the eligible patients with ACS J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172353 Manfrini et al Sex differences in CAD were excluded as they presented with prior coronary risk factors were present, 7.7% of women and 9.7% revascularization by coronary artery bypass grafting or of men had obstructive CAD; when 1 of these factors with nonobstructive lesions in patients with prior PCI. was present, 27.9% of women and 29.7% of men had Of the 14 793 enrolled patients, 96.2% had obstructive obstructive CAD; when 1 or 2 of these factors were CAD. present, 65.1% of women and 64.8% of men had ob- structive CAD; and when 1 to 3 of these factors were Patient Baseline Characteristics present, 87.8% of women and 86.2% of men had ob- Baseline clinical characteristics of the study population structive CAD. When adding a fourth risk factor, the in- are displayed in Table S1. Women, whether they had crease was small (92.2% in women and 90.2% in men). obstructive or nonobstructive CAD, were older and had more traditional risk factors, except cigarette smoking, Risk Factors and Obstructive CAD by Sex compared with their male counterparts. At hospital pres- The women-to-men RR ratios for obstructive CAD entation, women more frequently developed heart fail- among risk factors are shown in Figure 2. The RR ratio ure (Killip class ≥2) than men and had lower initial levels among patients without diabetes mellitus was 0.49 of serum creatinine. Women received on average fewer (95% CI, 0.41–0.60) and those with diabetes mellitus revascularization procedures and fewer antiplatelet and was 0.89 (95% CI, 0.62–1.29) (Table 1), with evidence anticoagulant agents compared with men (Table S2). of an interaction by diabetes mellitus status of P=0.002 Sex disparities in treatment were also seen when re- (Table S5). Exposure to smoking (Table 2) shifted the stricting the analysis to patients with STEMI. Women RR ratios from 0.50 (95% CI, 0.41–0.61) in nonsmokers continued to be less likely than men to undergo rep- to 0.75 (95% CI, 0.54–1.03) in current smokers, with an erfusion therapies with fibrinolysis (4.9% versus 6.8%, interaction by smoking status of P=0.018 (Table S5). P=0.0002) However, when women were represented The RR ratios for the absence or presence of hyper- in the catheterization laboratory, they had the same cholesterolemia were 0.56 (95% CI, 0.45–0.70) and interventional therapy as the men (97% versus 97.3%, 0.55 (95% CI, 0.42–0.72), respectively (Table 3). The respectively; P=0.508) (Table S3). Of note, women re- RR ratios for the absence or presence of hypertension ceived more evidence-based therapies before admis- were 0.50 (95% CI, 0.35–0.73) and 0.56 (95% CI, 0.47– sion for ACS, namely aspirin, clopidogrel, β-blockers, 0.68), respectively (Table 4). There were no significant angiotensin-converting enzyme inhibitors and/or angio- sex-related interactions for hypercholesterolemia and Downloaded from http://ahajournals.org by on October 30, 2023 tensin II receptor blockers, and statins (Table S4). hypertension (Table S5). Prevalence of Traditional Risk Factors in Severity of CAD and Outcomes ACS After clinical baseline characteristics were well Overall, among patients with ACS, the prevalence of matched between women and men using inverse those patients with at least 1 of the 4 conventional risk probability of weighting, female sex was associated factors was 92.2% in women with obstructive CAD and with a higher risk of STEMI in patients presenting with 89.9% in those with nonobstructive CAD (Figure 1). obstructive CAD (RR ratio, 1.12; 95% CI, 1.03–1.21). Similar rates were seen in men. When none of these No sex difference in STEMI rates were observed in Figure 1. Cumulative prevalence of traditional risk factors in acute coronary syndromes sorted by coronary artery disease (CAD) status and sex. J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172354 Manfrini et al Sex differences in CAD Figure 2. Female sex and obstructive coronary artery disease (CAD) sorted by the presence or absence of traditional risk factors. Women-to-men relative risk (RR) ratios expressed on a logarithmic scale. Variables used for adjustment are reported in Tables 1 through 4. patients with nonobstructive CAD (RR ratio, 0.92; 95% ratio, 1.75; 95% CI, 1.48–2.07). No sex difference in CI, 0.60–1.43) (Table 5). However, the RRs from the 2 mortality was observed with patients with nonobstruc- subgroups did not significantly differ from each other tive CAD (1.5% versus 1.9%, respectively) (RR ratio, (interaction test, P=0.1913) (Table S6). Among patients 0.79; 95% CI, 0.31–1.74). The interaction test between with obstructive CAD, women had higher 30-day mor- the outcomes of obstructive versus nonobstructive Downloaded from http://ahajournals.org by on October 30, 2023 tality than men (5.8% versus 3.4%, respectively) (RR CAD was highly significant (P=0.038) (Table S7). To Table 1. History of Diabetes Mellitus: Incidence of Obstructive CAD Sorted by Sex (Women Versus Men) Using Inverse Probability of Weighting Analysis Diabetes Mellitus No Diabetes Mellitus Women Men Women Men Characteristics (n=1293) (n=2270) P Value (n=3054) (n=8176) P Value Age, mean±SD, y 64.1±10.6 64.5±10.4 0.3282 60.3±11.9 60.4±11.8 0.4990 Cardiovascular risk factors, % Hypertension 83.5 83.3 0.8775 65.3 65.8 0.6196 Hypercholesterolemia 51.6 51.4 0.9086 41.7 42.4 0.5040 Current smokers 33.7 33.8 0.9516 46.3 46.5 0.8500 Former smokers 10.4 9.7 0.5026 7.7 7.3 0.4712 Clinical history of ischemic heart disease, % Previous angina pectoris 20.3 19.6 0.6145 14.3 14.6 0.6882 Previous myocardial 17.1 16.9 0.8785 12.3 12.1 0.7730 infarction Previous heart failure 5.5 5.4 0.8994 3.5 3.3 0.6004 Clinical history of cardiovascular disease, % Peripheral artery disease 2.8 2.8 1.0000 1.3 1.4 0.6876 Previous stroke 3.8 3.9 0.8822 2.5 2.4 0.7576 Outcomes, % Obstructive CAD, % 96.3 96.7 0.5434 93.9 96.9 <0.0001 RR ratio (95% CI) 0.89 (0.62–1.29) 0.5435 0.49 (0.41–0.60) <0.0001 CAD indicates coronary artery disease; and RR, relative risk. J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172355 Manfrini et al Sex differences in CAD Table 2. Smoking Status: Incidence of Obstructive CAD Sorted by Sex (Women Versus Men) Using Inverse Probability of Weighting Analysis Current Smokers Nonsmokers Women Men Women Men Characteristics (n=1394) (n=5026) P Value (n=2953) (n=5420) P Value Age, mean±SD, y 56.7±10.1 56.8±10.2 0.8585 64.8±11.7 65.0±11.3 0.5535 Cardiovascular risk factors, % Diabetes mellitus 18.7 18.7 1.0000 28.1 28.2 0.9226 Hypertension 64.0 64.5 0.7302 73.9 74.3 0.6896 Hypercholesterolemia 48.5 49.2 0.6437 40.3 41.0 0.5334 Clinical history of ischemic heart disease, % Previous angina pectoris 12.9 13.3 0.6967 17.4 17.6 0.8182 Previous myocardial infarction 11.8 11.5 0.7567 14.5 14.6 0.9014 Previous heart failure 3.1 2.8 0.5530 4.6 4.6 1.0000 Clinical history of cardiovascular disease, % Peripheral artery disease 1.6 1.6 1.0000 1.8 1.8 1.0000 Previous stroke 1.8 1.8 1.0000 3.4 3.4 1.0000 Outcomes Obstructive CAD, % 96.3 97.2 0.0778 93.2 96.5 <0.0001 RR ratio (95% CI) 0.75 (0.54–1.03) 0.0788 0.50 (0.41–0.61) <0.0001 CAD indicates coronary artery disease; and RR, relative risk. clarify the previously identified excess mortality risk Confirmatory Analysis on 30-Day among women with obstructive CAD, we examined Mortality mortality rates in patients who underwent primary PCI. Female sex was still associated with higher risk of 30- Results of 30-day mortality were virtually the same in the day mortality compared with men (RR ratio, 1.84; 95% analysis that restricted the cohort of obstructive CAD to Downloaded from http://ahajournals.org by on October 30, 2023 CI, 1.52–2.23) (Table S8). patients with ≥70% stenosis (Tables S9 and S10). Table 3. History of Hypercholesterolemia: Incidence of Obstructive CAD Sorted by Sex (Women Versus Men) Using Inverse Probability of Weighting Analysis Hypercholesterolemia No Hypercholesterolemia Women Men Women Men Characteristics (n=2025) (n=4584) P Value (n=2322) (n=5862) P Value Age, mean±SD, y 61.1±11.2 61.2±11.3 0.6409 61.3±12.1 61.6±11.8 0.3712 Cardiovascular risk factors, % Diabetes mellitus 27.6 27.7 0.9332 20.6 21.1 0.6162 Hypertension 79.2 79.6 0.7106 61.9 62.3 0.7366 Current smokers 47.9 48.0 0.9402 39.8 39.8 1.0000 Former smokers 10.3 9.7 0.4511 7.1 6.4 0.2502 Clinical history of ischemic heart disease, % Previous angina pectoris 21.8 21.7 0.9276 11.0 11.1 0.8967 Previous myocardial 16.0 16.2 0.8385 11.0 10.9 0.8961 infarction Previous heart failure 5.0 5.0 1.0000 3.1 2.9 0.6281 Clinical history of cardiovascular disease, % Peripheral artery disease 2.2 2.3 0.8018 1.2 1.3 0.7144 Previous stroke 2.7 2.8 0.8204 2.8 2.7 0.8014 Outcomes Obstructive CAD, % 95.3 97.3 <0.0001 93.9 96.5 <0.0001 RR ratio (95% CI) 0.55 (0.42–0.72) <0.0001 0.56 (0.45–0.70) <0.0001 CAD indicates coronary artery disease; and RR, relative risk. J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172356 Manfrini et al Sex differences in CAD Table 4. History of Hypertension: Incidence of Obstructive CAD Sorted by Sex (Women Versus Men) Using Inverse Probability of Weighting Analysis Hypertension No Hypertension Women Men Women Men Characteristics (n=3415) (n=6953) P Value (n=932) (n=3493) P Value Age, mean±SD, y 62.9±11.1 63.2±11.1 0.4263 57.2±11.9 57.3±11.6 0.7736 Cardiovascular risk factors, % Diabetes mellitus 28.5 28.6 0.9156 13.1 13.4 0.8107 Hypercholesterolemia 50.3 50.7 0.7018 30.1 30.4 0.8595 Current smokers 39.7 40.0 0.7694 51.5 51.5 1.0000 Former smokers 9.8 9.1 0.2497 5.2 5.0 0.8034 Clinical history of ischemic heart disease, % Previous angina pectoris 18.6 18.6 1.0000 9.7 9.5 0.8535 Previous myocardial 14.7 14.6 0.8923 9.9 10.1 0.8569 infarction Previous heart failure 4.6 4.5 0.8177 2.6 2.2 0.4696 Clinical history of cardiovascular disease, % Peripheral artery disease 2.0 2.1 0.7363 0.8 0.9 0.7767 Previous stroke 3.5 3.4 0.7924 1.2 1.3 0.8079 Outcomes Obstructive CAD, % 94.0 96.5 <0.0001 95.4 97.6 0.0003 RR ratio (95% CI) 0.56 (0.47–0.68) <0.0001 0.50 (0.35–0.73) 0.0004 CAD indicates coronary artery disease; and RR, relative risk. women even in the absence of obstructive CAD. There DISCUSSION is, therefore, a substantial void in current understanding Downloaded from http://ahajournals.org by on October 30, 2023 This study explored the relationships between risk fac- as to whether there are sex differences in the 4 tradi- tors, sex, and obstructive CAD status on 30-day mortal- tional cardiovascular risk factors and how these differ- ity after an ACS. Our results demonstrate that the excess ences may impact the severity of CAD and its relation risk of death in women compared with men is limited to with outcomes. We approached this issue by review- patients with obstructive CAD. Obstructive CAD is, there- ing the presence of traditional risk factors in 14 793 pa- fore, the most life-threatening event in women, and as tients who were referred to coronary angiography for so, warrants intensified efforts to prevent its occurrence. an ACS. Our data indicate that conventional risk factors Our results also shed light on the relationship between are present at a much higher prevalence than previously traditional risk factors and obstructive CAD in women. thought,25 with only 8% to 10% of patients lacking any of Compared with men, obstructive CAD risk in women is the conventional risk factors for the disease. This overall increased to a greater extent by current smoking and pattern was largely independent of sex and severity of diabetes mellitus. These data raise potential challenges, CAD. Therefore, in contrast to prior suggestions,26 we which warrant further considerations. found that only a small minority of patients with nonob- structive CAD lacks conventional risk factors. Traditional Risk Factors in ACS Cigarette smoking, hypertension, diabetes mellitus, and hypercholesteremia are factors of recognized im- Sex Differences in Severity of CAD and portance in the development of CHD in the general Mortality From ACS population. Prior epidemiological work focusing on the The current study challenges the common belief relationship between risk factors and CHD has required that women with chest pain, whether it be associ- cardiac clinical events as outcomes. However, the term ated with ACS or otherwise, are more likely to have CHD holds multiple mechanisms that may contribute to nonobstructive CAD at angiography. The prevalence ischemic events. These mechanisms do not necessarily varies depending on clinical setting and risk pro- need the presence of obstructive CAD. Abnormal coro- file of the population investigated. Nonobstructive nary reactivity,21 microvascular dysfunction,22,23 and CAD was a relatively uncommon clinical entity in plaque erosion with distal microembolization23,24 are our cohort of patients with ACS. Approximately 5% potential factors able to trigger myocardial ischemia in of women and men undergoing angiography had J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172357 Manfrini et al Sex differences in CAD Table 5. Outcomes Sorted by Sex (Women Versus Men) and CAD Status in Patients With ACS at Index Event Using Inverse Probability of Weighting Analysis Obstructive CAD Nonobstructive CAD (Stenosis ≥50%) (Stenosis <50%) Women Men Women Men Characteristics (n=4119) (n=10 119) P Value (n=228) (n=327) P Value Age, mean±SD, y 61.4±11.9 61.4±11.5 0.8232 60.9±11.7 60.8±12.2 0.9484 Cardiovascular risk factors, % Diabetes mellitus 24.4 24.1 0.7045 20.2 22.1 0.5914 Hypertension 69.7 69.6 0.9063 78.6 77.5 0.7589 Hypercholesterolemia 44.5 44.7 0.8277 39.3 38.4 0.8308 Current smokers 43.3 43.9 0.5129 35.1 34.3 0.8458 Former smokers 7.4 7.8 0.4157 9.9 10.8 0.7337 Clinical history of ischemic heart disease, % Previous angina pectoris 15.3 15.6 0.6541 17.8 18.3 0.8806 Previous myocardial infarction 13.0 13.3 0.6320 12.4 12.1 0.9156 Previous heart failure 3.6 3.8 0.5680 3.8 4.4 0.7279 Clinical history of cardiovascular disorders, % Peripheral artery disease 1.6 1.8 0.4067 1.2 1.3 0.9180 Previous stroke 2.7 2.7 1.0000 3.4 2.9 0.7385 Clinical presentation at admission ST-segment deviation in anterior leads (at ECG), % 20.6 21.0 0.5945 7.2 7.0 0.9280 Systolic BP at baseline, mean±SD, mm Hg 139.6±28.0 139.6±26.6 0.9856 143.2±25.5 143.7±26.1 0.8476 Heart rate at baseline, mean±SD, beats per min 80.1±17.9 80.2±17.9 0.7351 79.1±18.3 79.0±20.7 0.9650 Serum creatinine at baseline, mean±SD, mg/dL 0.99±0.50 1.05±0.60 0.0001 0.99±0.30 1.03±0.50 0.1501 Killip class ≥2, % 16.5 16.5 1.0000 12.6 14.9 0.4425 Downloaded from http://ahajournals.org by on October 30, 2023 Outcomes 30-d Mortality, % 5.8 3.4 <0.0001 1.5 1.9 0.7236 RR ratio (95% CI) 1.75 (1.48–2.07) <0.0001 0.79 (0.31–1.74) 0.7237 STEMI, % 70.7 68.4 0.0064 17.8 18.9 0.7243 RR ratio (95% CI) 1.12 (1.03–1.21) 0.0064 0.92 (0.60–1.43) 0.7238 ACS indicates acute coronary syndrome; BP, blood pressure; CAD, coronary artery disease; RR, relative risk; and STEMI, ST-segment–elevation myocardial infarction. nonobstructive disease. This finding is concord- worth considering. Sex is an important confounder ant with prior work exploring obstructive CAD sta- for cardiovascular disease. Each of the traditional risk tus in myocardial infarction. 27,28 Our results also factors increases the rates of cardiovascular mortality confirm that the excess risk of short-term mortal- and may represent residual confounding. In addition, ity after ACS in women is restricted to those with associations may increase confounding. Smokers obstructive CAD, 29 and demonstrate that women have more adverse cardiovascular risk factors, with obstructive CAD present more often with such as dyslipidemia and hypertension, than never- STEMI compared with their male counterparts. smokers.30 Therefore, nonsmokers may have more Taken together, these findings reinforce the view protection against development of significant CAD that women with obstructive CAD are a vulnerable compared with smokers, independently of smoking group and the growing demand for development of status. This reasoning applies equally well to all risk sex-specific prevention strategies. factors.31 To try to circumvent this issue we matched patients sorted by sex and each individual risk fac- tor using inverse probability of weighting. The weights Sex-Specific Weights of Risk Factors in created a population where the weighted risk factors CAD and control groups were representative of the patient It is difficult to establish the precise sex-specific im- characteristics in the overall population of women and pact of each of the 4 major risk factors on develop- men.32 Balanced covariates, including age, could not ment of significant CAD. Potential confounding is be confounders anymore, a property that would be J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172358 Manfrini et al Sex differences in CAD expected under randomization. On the other hand, presence of other major risk factors for CHD.7 sex and risk factors cannot be randomized. Several potential interacting factors may contribute to the acceleration of CHD risk in women with diabe- Current Smoking and Obstructive CAD in tes mellitus. Diabetes mellitus is more likely to be as- sociated with elevations in both systolic and diastolic Women blood pressure in women than in men.36 Women with Although cigarette smoking is harmful for any sex, diabetes mellitus are more likely to be of low socio- there are some discrepancies between studies in economic status and as so are quite often cigarette demonstrating a different effect of smoking as a risk smokers.37 We tried to circumvent such issues by factor for CHD in women. Some studies have sug- matching patients with inverse probability of weight- gested that smoking has a much larger relative detri- ing. In our study, concurrent traditional risk factors mental impact on CHD in women.9 Other studies have were all well balanced among women and men with shown that smoking has a similar effect on increasing diabetes mellitus. Diabetes mellitus on top of the the risk of CHD in both men and women.8 Conflicting other risk factors equalized the risk for obstructive results between studies may be related to many factors CAD by sex at any age. The higher RRof mortality including definition of smokers, age with consequent after ACS conferred by obstructive CAD in women prevalence of oral contraceptive use, and synergistic compared with men may explain why women with action of smoking with other conventional risk factors. diabetes mellitus have a 2-fold increased risk of myo- In the current study, we followed the NHIS17 definition cardial infarction and death compared with their male of smoker and used inverse probability of weighting counterparts.38,39 Screening for prediabetes mellitus to mitigate much of the differences in age and other combined with more stringent follow-up of women concomitant risk factors. Much more importantly, we with a history of gestational diabetes mellitus has the investigated whether current smoking has a hazard- potential to dramatically reduce the burden of CAD ous effect on the RR ratio of women versus men for and sex differences in outcomes. the association with significant CAD, which is one of the factors contributing to the pathophysiology of ACS. Compared with nonsmokers, women who were Sex Differences in Hypercholesterolemia current smokers had a much greater risk of obstruc- and Hypertension tive CAD with statistical evidence of interaction. This Hypercholesterolemia and hypertension are both Downloaded from http://ahajournals.org by on October 30, 2023 clearly indicates that the harm of smoking differs by well-documented primary risk factors. Both fac- sex. Our study, therefore, adds to the understanding tors increased the risk of obstructive CAD without of the relationship between smoking and CHD events any critical matter to separate risk from women by suggesting an important mechanistic basis: its as- and men. This finding supports the results from the sociation with severe atherosclerotic plaques in the MONICA (Monitoring of Trends and Determinants in coronary arteries. Excess risk of obstructive CAD Cardiovascular Disease) study reporting that the in- in female compared with male smokers might have crease in CHD events with increasing total cholesterol some potential explanations. Women might extract a holds over the entire range of patient characteris- greater quantity of toxic agents from the same num- tics.40 Although we found no difference in the risk for ber of cigarettes than men.33 Plasma levels of estro- obstructive CAD by sex, it should be reminded that gen are lower in smoking than in nonsmoking women, there is a remarkably higher prevalence of hyperten- which may lead to accelerated progression of CAD.34 sion in women. Thus, women would benefit more However, in light of the available evidence, no definite from strategies that prevent and treat hypertension answer can be given. Unfortunately, there is an alarm- at the population not at the individual level. ing trend toward increased smoking in women and, therefore, better methods leading to prevention and Limitations cessation of smoking are needed. Our study has several potential limitations. First, an observational study is potentially open to confound- ing. We minimized this factor by using a study design Diabetes Mellitus and Obstructive CAD in based on matching by inverse probability of weight- Women ing to balance covariate distribution among sexes and There is strong evidence from many studies that risk factors.18 On the other hand, randomized con- women with diabetes mellitus face an increased car- trolled trials are not a viable option as it is unethical to diovascular risk relative to men.35 Still, it is unclear randomly assign women and men to be smokers and whether these observed sex differences in CHD nonsmokers. Second, patients who have had coro- risk are real or attributable to differences between nary angiography do not necessarily represent the men and women with respect to the concomitant general ACS population since those who died before J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.0172359 Manfrini et al Sex differences in CAD hospital admission are missing. Yet, mortality rates CONCLUSIONS among women and men with ACS are similar to those The current study found greater 30-day mortality re- reported in recent studies, which supports the external lated to obstructive CAD in women compared with validity of the study.29 Third, some of the risk factors men. Cigarette smoking and diabetes mellitus dis- were ascertained by the general practitioner, which proportionally increase the risk of obstructive CAD in might have led to error in some individuals. Although women, and as so they are key factors in explaining we acknowledge some potential misclassifications, sex differences in outcomes from ACS. Intense efforts it is unlikely that these misclassifications differentially to reduce tobacco use and increase screening for pre- affect women over men and, thus, are unlikely to diabetes mellitus have potential to decrease the sex modify the sex differences that we found. Fourth, our lag in cardiovascular disease mortality in women com- analysis did not account for potential differences in pared with men. antidiabetic medication use or medication adherence before index event in women versus men. However, examination of a recent study showing that men have a higher risk of metformin- and sulfonylureas-associ- ARTICLE INFORMATION Received April 24, 2020; accepted August 10, 2020. ated myocardial infarction than women suggests that our model-based estimates are not influenced by an- Affiliations tidiabetic drug interactions and are reasonably valid.41 From the Department of Experimental, Diagnostic and Specialty Medicine, Fifth, angiographic evaluations were performed at the University of Bologna, Bologna, Italy (O.M., M.S., E.C., R.B.); Department of Electrical and Computer Engineering, University of California, Los local level and, hence, the reliability of the observa- Angeles, CA (J.Y.); Cambridge Centre for Artificial Intelligence in Medicine, tions, especially as it relates to minimal CAD (stenosis Department of Applied Mathematics and Theoretical Physics and <50%) may be difficult to assess. However, such indi- Department of Population Health, University of Cambridge, Cambridge, United Kingdom (M.v.d.S.); University Clinic of Cardiology, Medical Faculty, vidual characterization of CAD reflects real-world CAD University "Ss. Cyril and Methodius", Skopje, Macedonia (S.K., M.V.); Clinic categorization. Further, we repeated the analyses on of Cardiology, University Clinical Centre of Serbia, Belgrade, Serbia (G.S.); 30-day mortality shifting the cutoff for nonobstructive Medical Faculty, University of Belgrade, Serbia (G.S., Z.V.); Department for Cardiovascular Diseases, University Hospital Center Zagreb, University CAD at 70% diameter stenosis. Estimates were similar of Zagreb, Croatia (D.M.) and Cardiovascular Research Program ICCC, to those seen using a 50% cutoff. As so, misclassifi- IR-IIBSant Pau, Hospital de la Santa Creu i Sant Pau, CiberCV-Institute cation of the severity of CAD in the categories of ob- Carlos III, Barcelona, Spain (L.B.). structive and nonobstructive seems unlikely. Defining Downloaded from http://ahajournals.org by on October 30, 2023 Sources of Funding a cutoff value of 50% or 70% stenosis for obstructive None. CAD has shed its critical importance in decision-mak- Disclosures ing and may be lumped into a common basket labeled None. “intermediate” stenosis, now undistinguished and of uncertain importance. Finally, residual confounding Supplementary Materials Data S1 from concomitance of nontraditional risk factors such Tables S1–S10 as stress, body mass index, family history of CHD, Figure S1 and adherence to healthy lifestyle behaviors cannot be excluded. REFERENCES Strengths 1. Mensah GA, Brown DW, Croft JB, Greenlund KJ. Major coronary Although sometimes used interchangeably, CAD and risk factors and death from coronary heart disease: baseline and follow-up mortality data from the Second National Health CHD are not the same condition. CHD may be actually and Nutrition Examination Survey (NHANES II). Am J Prev Med. but not necessarily a result of CAD. Little is currently 2005;29:68–74. known about the sex-specific associations between 2. Ambrose JA, Winters SL, Arora RR, Haft JI, Goldstein J, Rentrop KP, Gorlin R, Fuster V. Coronary angiographic morphology in myocardial traditional risk factors and the degree of CAD. Our infarction: a link between the pathogenesis of unstable angina and myo- angiographic study on CAD represents an advance cardial infarction. J Am Coll Cardiol. 1985;6:1233–1238. in furthering our understanding of the mechanisms of 3. Bugiardini R, Manfrini O, De Ferrari GM. Unanswered questions for management of acute coronary syndrome: risk stratification of patients vulnerability to traditional risk factors in women and with minimal disease or normal findings on coronary angiography. Arch highlights the ongoing need to accurately account for Intern Med. 2006;166:1391–1395. biologic factors specific to women. 4. Gulati M. Improving the cardiovascular health of women in the nation. Circulation. 2017;135:495–498. Our statistical approach improves on existing stud- 5. Helfand M, Buckley DI, Freeman M, Fu R, Rogers K, Fleming C, ies. Logistic regression ignores the interaction among Humphrey LL. Emerging risk factors for coronary heart disease: a sum- risk factors. Inverse probability of weighting exhibits mary of systematic reviews conducted for the U.S. Preventive Services Task Force. Ann Intern Med. 2009;151:496–507. balance on the covariates and weights the risk of each 6. Lin JS, Evans CV, Johnson E, Redmond N, Coppola EL, Smith N. type of traditional risk factors equally. Nontraditional risk factors in cardiovascular disease risk assessment: J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.01723510 Manfrini et al Sex differences in CAD updated evidence report and systematic review for the US Preventive blood flow and delay to hospital presentation. J Am Heart Assoc. Services Task Force. JAMA. 2018;320:281–297. 2019;8:e011190. DOI: 10.1161/JAHA.118.011190. 7. Kanaya AM, Grady D, Barrett-Connor E. Explaining the sex difference in 24. Reynolds HR, Srichai MB, Iqbal SN, Slater JN, Mancini GB, Feit F, Pena- coronary heart disease mortality among patients with type 2 diabetes Sing I, Axel L, Attubato MJ, Yatskar L, et al. Mechanisms of myocardial mellitus: a meta-analysis. Arch Intern Med. 2002;162:1737–1745. infarction in women without angiographically obstructive coronary ar- 8. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen tery disease. Circulation. 2011;124:1414–1425. M, Budaj A, Pais P, Varigos J, et al. Effect of potentially modifiable 25. Magnus P, Beaglehole R. The real contribution of the major risk factors risk factors associated with myocardial infarction in 52 countries (the to the coronary epidemics: time to end the "only-50%" myth. Arch Intern INTERHEART study): case-control study. Lancet. 2004;364:937–952. Med. 2001;161:2657–2660. 9. Huxley RR, Woodward M. Cigarette smoking as a risk factor for cor- 26. Sara JD, Widmer RJ, Matsuzawa Y, Lennon RJ, Lerman LO, Lerman onary heart disease in women compared with men: a systematic A. Prevalence of coronary microvascular dysfunction among patients review and meta-analysis of prospective cohort studies. Lancet. with chest pain and nonobstructive coronary artery disease. JACC 2011;378:1297–1305. Cardiovasc Interv. 2015;8:1445–1453. 10. Office on Smoking and Health (US). The Health Consequences of 27. Ramanath VS, Armstrong DF, Grzybowski M, Rahnama-Mohagdam S, Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon Tamhane UU, Gordon K, Froehlich JB, Eagle KA, Jackson EA. Receipt General. Atlanta, GA: Centers for Disease Control and Prevention (US); of cardiac medications upon discharge among men and women with 2006. Available at: https://www.ncbi.nlm.nih.gov/books/​ NBK44​324/. acute coronary syndrome and nonobstructive coronary artery disease. Accessed January 4, 2020. Clin Cardiol. 2010;33:36–41. 11. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj 28. Shaw LJ, Shaw RE, Merz CN, Brindis RG, Klein LW, Nallamothu B, A, Bugiardini R, Crea F, Cuisset T, Di Mario C, et al. 2013 ESC guidelines Douglas PS, Krone RJ, McKay CR, Block PC, et al. Impact of eth- on the management of stable coronary artery disease: the Task Force nicity and gender differences on angiographic coronary artery dis- on the management of stable coronary artery disease of the European ease prevalence and in-hospital mortality in the American College Society of Cardiology. Eur Heart J. 2013;34:2949–3003. of Cardiology-National Cardiovascular Data Registry. Circulation. 12. Cenko E, Yoon J, Kedev S, Stankovic G, Vasiljevic Z, Krljanac G, Kalpak 2008;117:1787–1801. O, Ricci B, Milicic D, Manfrini O, et al. Sex differences in outcomes after 29. Smilowitz NR, Mahajan AM, Roe MT, Hellkamp AS, Chiswell K, Gulati STEMI: effect modification by treatment strategy and age. JAMA Intern M, Reynolds HR. Mortality of myocardial infarction by sex, age, and Med. 2018;178:632–639. obstructive coronary artery disease status in the ACTION Registry- 13. Cenko E, van der Schaar M, Yoon J, Manfrini O, Vasiljevic Z, Vavlukis M, GWTG (Acute Coronary Treatment and Intervention Outcomes Kedev S, Miličić D, Badimon L, Bugiardini R. Sex-related differences in Network Registry-Get With the Guidelines). Circ Cardiovasc Qual heart failure after ST-segment elevation myocardial infarction. J Am Coll Outcomes. 2017;10:e003443. https://doi.org/10.1161/CIRCO​ UTCOM​ Cardiol. 2019;74:2379–2389. ES.116.003443. 14. Gliklich RE, Leavy MB, Levy D, Karl J, Campion DM, Taylor T.Registry 30. Alkerwi A, Baydarlioglu B, Sauvageot N, Stranges S, Lemmens P, of Patient Registries (RoPR) Policies and Procedures. Effective Health Shivappa N, Hébert JR. Smoking status is inversely associated with Care Program Research Report No. 41. (Prepared by Outcome DEcIDE overall diet quality: findings from the ORISCAV-LUX study. Clin Nutr. Center under Contract No. HHSA 290-2005-0035-1.) AHRQ Publication 2017;36:1275–1282. No. 12-EHC066-EF. Rockville, MD: Agency for Healthcare Research 31. Sowers JR, Epstein M, Frohlich ED. Diabetes, hypertension, and cardio- and Quality; 2012. Available at: https://effec​ tiveh​ ealth​ care.ahrq.gov/ vascular disease: an update. Hypertension. 2001;37:1053–1059. sites/​defau​lt/files/​pdf/regis​try-of-patie​nt-regis​tries_resea​rch-2012-2_1. 32. Kurth T, Walker AM, Glynn RJ, Chan KA, Gaziano JM, Berger K, Robins Downloaded from http://ahajournals.org by on October 30, 2023 pdf. Accessed January 11, 2020. JM. Results of multivariable logistic regression, propensity matching, 15. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS, Budoff propensity adjustment, and propensity-based weighting under condi- MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD, et al. Coronary Artery tions of nonuniform effect. Am J Epidemiol. 2006;163:262–270. Disease—Reporting and Data System (CAD-RADS): an expert consen- 33. Woodward M, Tunstall-Pedoe H, Smith WC, Tavendale R. Smoking sus document of SCCT, ACR and NASCI: endorsed by the ACC. JACC characteristics and inhalation biochemistry in the Scottish population. J Cardiovasc Imaging. 2016;9:1099–1113. Clin Epidemiol. 1991;44:1405–1410. 16. Levine GN, Bates ER, Blankenship JC, Bailey SR, Bittl JA, Cercek B, 34. Geisler J, Omsjø IH, Helle SI, Ekse D, Silsand T, Lønning PE. Plasma Chambers CE, Ellis SG, Guyton RA, Hollenberg SM, et al. 2011 ACCF/ oestrogen fractions in postmenopausal women receiving hormone re- AHA/SCAI guideline for percutaneous coronary intervention: executive placement therapy: influence of route of administration and cigarette summary: a report of the American College of Cardiology Foundation/ smoking. J Endocrinol. 1999;162:265–270. American Heart Association Task Force on Practice Guidelines and the 35. Huxley R, Barzi F, Woodward M. Excess risk of fatal coronary heart Society for Cardiovascular Angiography and Interventions. Circulation. disease associated with diabetes in men and women: meta-analysis of 2011;124:2574–2609. 37 prospective cohort studies. BMJ. 2006;332:73–78. 17. National Health Interview Survey-Adult Tobacco Use Information 2017. 36. Haffner SM, Valdez R, Morales PA, Mitchell BD, Hazuda HP, Stern MP. Available at: https://www.cdc.gov/nchs/nhis/tobac​co/tobac​co_gloss​ Greater effect of glycemia on incidence of hypertension in women than ary.htm. Accessed January 4, 2020. in men. Diabetes Care. 1992;15:1277–1284. 18. Austin PC, Stuart EA. Moving towards best practice when using inverse 37. Beckles GLA, Thompson-Reid PE; Centers for Disease Control and probability of treatment weighting (IPTW) using the propensity score to Prevention (CDC). Socioeconomic status of women with diabetes— estimate causal treatment effects in observational studies. Stat Med. United States, 2000. MMWR Morb Mortal Wkly Rep. 2002;51:147–148, 2015;34:3661–3679. 159. 19. Beretta L, Santaniello A. Nearest neighbor imputation algorithms: a crit- 38. Huxley RR, Peters SA, Mishra GD, Woodward M. Risk of all-cause mor- ical evaluation. BMC Med Inform Decis Mak. 2016;16(suppl 3):74. tality and vascular events in women versus men with type 1 diabetes: 20. Altman DG, Bland JM. Interaction revisited: the difference between two a systematic review and meta-analysis. Lancet Diabetes Endocrinol. estimates. BMJ. 2003;326:219. 2015;3:198–206. 21. Hammadah M, Kim JH, Al Mheid I, Samman Tahhan A, Wilmot K, 39. Morrish NJ, Wang SL, Stevens LK, Fuller JH, Keen H. Mortality and Ramadan R, Alkhoder A, Khayata M, Mekonnen G, Levantsevych O, causes of death in the WHO multinational study of vascular disease in et al. Coronary and peripheral vasomotor responses to mental stress. J diabetes. Diabetologia. 2001;44:S14–S21. Am Heart Assoc. 2018;7:e008532. DOI: 10.1161/JAHA.118.008532. 40. Jousilahti P, Vartiainen E, Tuomilehto J, Puska P. Sex, age, cardiovas- 22. Wong TY, Klein R, Sharrett AR, Duncan BB, Couper DJ, Tielsch JM, cular risk factors, and coronary heart disease: a prospective follow-up Klein BE, Hubbard LD. Retinal arteriolar narrowing and risk of coro- study of 14 786 middle-aged men and women in Finland. Circulation. nary heart disease in men and women. The Atherosclerosis Risk in 1999;99:1165–1172. Communities Study. JAMA. 2002;287:1153–1159. 41. Wang SH, Chen WJ, Hsu LY, Chien KL, Wu CS. Use of spontaneous re- 23. Cenko E, van der Schaar M, Yoon J, Kedev S, Valvukis M, Vasiljevic Z, porting systems to detect host-medication interactions: sex differences Ašanin M, Miličić D, Manfrini O, Badimon L, et al. Sex-specific treatment in oral anti-diabetic drug-associated myocardial infarction. J Am Heart effects after primary percutaneous intervention: a study on coronary Assoc. 2018;7:e008959. DOI: 10.1161/JAHA.118.008959. J Am Heart Assoc. 2020;9:e017235. DOI: 10.1161/JAHA.120.01723511 SUPPLEMENTAL MATERIAL Downloaded from http://ahajournals.org by on October 30, 2023 Data S1. SUPPLEMENTAL METHODS Inverse probability of weighting We used inverse probability of weighting to balance the distribution of covariates between two patient groups. If e denotes the estimated propensity score (i.e. e=\hat{P}(Z=1 | x), where the patient x is included in patient group 1; then, 1-e = \hat{P}(Z=0 | x)), then the original sample is weighted by the following weights: Z/e+(1−Z)/ 1−e where Z represents the patient group. For instance, women (Z=1) are assigned a weight equal to the reciprocal of the propensity score (1/e), while men (Z=0) are assigned a weight equal to the reciprocal of one minus the propensity score (1/1-e). The weighting procedure for each sample balances the covariate distributions between two patient groups.18 Downloaded from http://ahajournals.org by on October 30, 2023 Nearest neighbor imputation algorithms Nearest neighbor (NN) imputation algorithms are efficient methods to fill in missing data where each missing value on some records is replaced by a value obtained from related cases in the whole set of records. Thus, imputation for clinical features was conducted using the average of measured values from k records (kNN).19 NN algorithms are similarity-based methods that rely on distance metrics and results may change in relation to the similarity measure used to evaluate the distance between recipients and donors. In our work, we used the following norm as metric to evaluate distance: (∑ni=1|xi−yi|p)1/p Before imputation of the recipient Xi, the full set with no missing data C(X) was filtered to select a subset of features relevant to the missing variable to be imputed (Xi_miss). To this end, C(X) was considered as a dataset in the context of a regression problem, where the variable with the missing data (Xmiss) was set as the class variable and the other q variables (X1, X2, …, Xq) as predictors. We also applied the RReliefF algorithm. The set was, therefore, filtered to select a subset Cs(X) ⊂ C(X) where (X1, X2, …, Xs) ⊂ (X1, X2, …, Xq) and s < q. In the present context, we set the number of neighbors for RReliefF equal to 10 and set s as 10 %, 20 % or 30 % of q. As C(X) is invariant to Xi, the filtering step was performed only once before the NN imputation step that, on the contrary was performed separately for each Xi. More specifically, to impute the missing value in i-th column, we find k-nearest neighbor columns from i-th column (in terms of Euclidean distance) and replace the missing value with weighted mean of the k-nearest neighbor columns. Weights are inversely proportional to the Euclidean distance from i-th column. Interaction test The comparison of two estimated quantities, each with its standard error, is a general method that can be applied widely.20 These measures were always analyzed on the log scale because the Downloaded from http://ahajournals.org by on October 30, 2023 distributions of the log ratios tend to be those closer to normal than of the ratios themselves. If the estimates are E1 and E2 with standard errors SE(E1) and SE(E2), then the difference d=E1 - E2 has standard error SE(d)=Ö[SE(E1)2 + SE(E2)2] i.e., the square root of the sum of the squares of the separate standard errors. The ratio z=d/SE(d) gives a test of the null hypothesis that in the population the difference d is zero, by comparing the value of z to the standard normal distribution. The 95% confidence interval for the difference is d-1.96SE(d) to d+1.96SE(d). SUPPLEMENTAL RESULTS Interaction tests In our study, the estimated women-to-men RR ratio for obstructive CAD among nondiabetics was 0.43 (95%CI 0.36– 0.51) and diabetics was 0.89 (0.43–1.83), but are the relative risks from the subgroups significantly different from each other? We show how to answer this question by using the interaction test based on the summary data quoted. (Table S4). We obtained the logs of the odds ratios (relative risks) and their confidence intervals (rows 2 and 4). As 95% confidence intervals were obtained as 1.96 standard errors either side of the estimate, the SE of each log relative risk was obtained by dividing the width of its confidence interval by 2×1.96 (row 6). The estimated difference in log relative risks was d=E1- E2= 0.5696 (row 7) and its standard error 0.1958 (row 8). From these two values, we tested the interaction and estimated the ratio of the relative risks (with confidence interval). The test of interaction was the ratio of d to its standard error: z= 2.9091, which gives p value=0.0018 when we referred it to a table of the normal distribution (row 10). The estimated interaction effect was exp =1.7676 (row 11). The confidence interval for this effect was 1.2042 to 2.5945 on the log scale (row 9). Transforming back to the relative risk scale, we got 1.2042 to 2.5945 (row 12). There was thus good evidence to support different outcome effects of diabetes on obstructive CAD between sexes. A similar approach was used for comparing any other sex difference. (Tables S5, S6, and S9). Downloaded from http://ahajournals.org by on October 30, 2023 Table S1. Baseline characteristics of the overall population sorted by sex and CAD status in patients with acute coronary syndrome at index event. Overall Obstructive CAD Nonobstructive CAD (stenosis ≥50%) (stenosis <50%) Women Men Women Men Women Men Characteristics p value p value p value (n=4347) (n=10446) (n=4119) (n=10119) (n=228) (n=327) Age, mean ± SD, y 65.2 ± 11.2 59.9 ± 11.4 <0.0001 65.4 ± 11.2 59.9 ± 11.4 <0.0001 62.5 ± 11.5 59.8 ± 12.3 0.0077 Cardiovascular risk 4020 (92.5) 9543 (91.4) 0.0208 3814 (92.6) 9245 (91.4) 0.0127 206 (90.4) 298 (91.1) 0.7563 factors (overall), n (%) Diabetes, n (%) 1293 (29.7) 2270 (21.7) <0.0001 1247 (30.3) 2196 (21.7) <0.0001 46 (20.2) 74 (22.6) 0.4872 Hypertension, n (%) 3415 (78.6) 6953 (66.6) <0.0001 3228 (78.4) 6710 (66.3) <0.0001 187 (82.0) 243 (74.3) 0.0288 Hypercholesterolemia, n 2025 (46.6) 4584 (43.9) 0.0027 1929 (46.8) 4463 (44.1) 0.0031 96 (42.1) 121 (37.0) 0.2283 (%) Downloaded from http://ahajournals.org by on October 30, 2023 Current smokers, n (%) 1394 (32.1) 5026 (48.1) <0.0001 1344 (32.6) 4889 (48.3) <0.0001 50 (21.9) 137 (41.9) <0.0001 Former smokers, n (%) 176 (4.0) 983 (9.4) <0.0001 162 (3.9) 937 (9.3) <0.0001 14 (6.1) 46 (14.1) 0.0016 Clinical history of 1255 (28.9) 2819 (27.0) 0.0205 1176 (28.6) 2729 (27.0) 0.0569 79 (34.6) 90 (27.5) 0.0763 ischemic heart disease (overall), n (%) Previous angina pectoris, 757 (17.4) 1583 (15.2) 0.0008 705 (17.1) 1531 (15.1) 0.0038 52 (22.8) 52 (15.9) 0.0456 n (%) Previous MI, n (%) 534 (12.3) 1432 (13.7) 0.0178 504 (12.2) 1398 (13.8) 0.0103 30 (13.2) 34 (10.4) 0.3263 Previous heart failure, n 184 (4.2) 384 (3.7) 0.1185 174 (4.2) 368 (3.6) 0.1070 10 (4.4) 16 (4.9) 0.7795 (%) Clinical history of 201 (4.6) 432 (4.1) 0.1909 194 (4.7) 417 (4.1) 0.1259 7 (3.1) 15 (4.6) 0.3521 cardiovascular disorders (overall), n (%) PAD, n (%) 62 (1.4) 195 (1.9) 0.0486 61 (1.5) 189 (1.9) 0.0946 1 (0.4) 6 (1.8) 0.1063 Previous stroke, n (%) 141 (3.2) 260 (2.5) 0.0146 135 (3.3) 251 (2.5) 0.0121 6 (2.6) 9 (2.8) 0.9311 Clinical presentation at admission STEMI, n (%) 2871 (66.0) 7094 (67.9) 0.0284 2833 (68.8) 7027 (69.4) 0.4369 38 (16.7) 67 (20.5) 0.2521 ST-segment shifts in 816 (18.8) 2212 (21.2) 0.0008 800 (19.4) 2189 (21.6) 0.0283 16 (7.0) 23 (7.0) 0.9942 anterior leads (at ECG), n (%) Downloaded from http://ahajournals.org by on October 30, 2023 Systolic BP at baseline, 140.4 ± 127.7 139.5 ± 26.7 0.0699 140.1 ± 27.8 139.4 ± 26.7 0.1619 145.8±25.4 143 ± 25.9 0.2047 mean ± SD, mmHg Heart rate at baseline, 80.2 ± 18.2 80.2 ± 18.0 0.8447 80.3 ± 18.2 80.2 ± 17.9 0.6824 78.7 ± 17.5 79.8 ± 21.8 0.5134 mean ± SD, bets/min Serum creatinine at 1.0 ± 0.5 1.1 ± 0.7 <0.0001 1.0 ± 0.5 1.1 ± 0.7 <0.0001 0.9 ± 0.3 1.1 ± 0.7 0.0009 baseline, mean ± SD, mg/dl Killip Class ≥2), n (%) 855 (19.7) 1602 (15.3) <0.0001 827 (20.1) 1547 (15.3) <0.0001 28 (12.3) 55 (16.8) 0.1317 BP indicates blood pressure; CAD, coronary artery disease; ECG, electrocardiogram; MI, myocardial infarction, PAD, peripheral artery disease, STEMI= ST- segment elevation myocardial infarction. Table S2. Use of medications and PCI within 24 hours from hospitalization sorted by sex (women versus men) and CAD status in the overall population of patients with acute coronary syndromes. Obstructive CAD Nonobstructive CAD All Patients (stenosis ≥50%) (stenosis <50%) Characteristics Women Men Women Men Women Men p value p value p value (n=4347) (n 10446) (n=4119) (n=10119) (n =228) (n =327) Aspirin, n (%) 4298 (98.9) 10352 (99.1) 0.2189 4071 (98.8) 10028(99.1) 0.1654 227 (99.6) 324 (99.1) 0.4857 Clopidogrel, n (%) 3908 (89.9) 9291 (88.9) 0.0819 3703 (89.9) 9000 (88.9) 0.0889 205 (89.9) 291 (89.0) 0.7278 Unfractionated heparin, n (%) 2411 (55.5) 6073 (58.1) 0.0028 2309 (56.1) 5905 (58.4) 0.0121 102 (44.7) 168 (51.4) 0.1239 LMWH, n (%) 2091 (48.1) 4769 (45.7) 0.0066 1960 (47.6) 4595 (45.0) 0.0184 131 (57.5) 174 (53.2) 0.3229 Downloaded from http://ahajournals.org by on October 30, 2023 Heparins (overall), n (%) 3671 (84.4) 9021 (86.4) 0.0030 3484 (84.6) 8735 (86.3) 0.0083 187 (82.0) 286 (87.5) 0.0837 GP IIb/IIIa inhibitor, n (%) 515 (11.8) 1328 (12.7) 0.1414 511 (12.4) 1326 (13.1) 0.2552 4 (1.8) 2 (0.6) 0.2408 Beta-blockers 3336 (76.7) 8065 (77.2) 0.5421 3132 (76.0) 7773 (76.8) 0.3225 204 (89.5) 292 (89.3) 0.9469 ARBs/ACE-inhibitors, n (%) 3425 (78.8) 8139 (77.9) 0.2378 3235 (78.5) 7873 (77.8) 0.3349 190 (83.3) 266 (81.3) 0.5450 PCI, n (%) 3880 (89.3) 9626 (92.2) <0.0001 3880 (94.2) 9626 (95.1) 0.0278 0 (0.0%) 0 (0.0%) - ACE indicates angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; CAD, coronary artery disease; GP, glycoprotein; LMWH, low molecular weight heparin; PCI, percutaneous coronary intervention. Table S3. Use of medications and reperfusion therapies within 24 hours from hospitalization sorted by sex (women versus men) and CAD status in patients with STEMI. All Patients Obstructive CAD Nonobstructive CAD (stenosis ≥ 50%) (stenosis <50%) Women Men Women Men Women Men Characteristics p value p value p value (n=2871) (n=7094) (n=2833) (n=7027) (n=38) (n=67) Aspirin, n (%) 2843 (99.0) 7045 (99.3) 0.1717 2805 (99.0) 6978 (99.3) 0.1673 38 (100.0) 67 (100.0) 1.0000 Clopidogrel, n (%) 2541 (88.5) 6228 (87.7) 0.3158 2508 (88.5) 6170 (87.8) 0.3113 33 (86.8) 58 (86.6) 0.9686 Unfractionated heparin, n (%) 1604 (55.9) 4110 (57.9) 0.0595 1593 (56.2) 4079 (58.0) 0.0993 11 (28.9) 31 (46.3) 0.0765 LMWH, n (%) 1314 (45.8) 3201 (45.1) 0.5581 1290 (45.5) 3168 (45.1) 0.6837 24 (63.2) 33 (49.3) 0.1699 Heparins (overall), n (%) 2424(84.4) 6175 (87.0) 0.0008 2394 (84.5) 6119 (87.1) 0.0011 30 (78.9) 56 (83.6) 0.5693 Downloaded from http://ahajournals.org by on October 30, 2023 Beta-blockers, n (%) 2137 (74.4) 5422 (76.4) 0.0371 2104 (74.3) 5366 (76.4) 0.0300 33 (86.8) 56 (83.6) 0.6514 ARBs/ACE-inhibitors, n (%) 2203 (76.7) 5503 (77.6) 0.3673 2173 (76.7) 5449 (77.5) 0.3699 30 (78.9) 54 (80.6) 0.8427 Reperfusion therapies Fibrinolysis, n (%) 140 (4.9) 479 (6.8) 0.0001 140 (4.9) 479 (6.8) 0.0002 0 (0.0) 0 (0.0) - PCI, n (%) 2749 (95.8) 6836 (96.4) 0.1613 2749 (97.0) 6836 (97.3) 0.5081 0 (0.0) 0 (0.0) - ACE indicates angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; CAD, coronary artery disease; GP, glycoprotein; LMWH, low molecular weight heparin; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction Table S4. Therapy within 15 days before index event. All Patients Obstructive CAD Nonobstructive CAD Women Men Women Men Women Men Characteristics p value p value p value (n=4347) (n=10446) (n=4119) (n=10119) (n=228) (n=327) Aspirin, n (%) 1291 (29.7) 2613 (25.0) <0.0001 1212 (29.4) 2531 (25.0) <0.0001 79 (34.6) 82 (25.1) 0.0162 Clopidogrel, n (%) 462 (10.6) 928 (8.9) 0.0014 426 (10.3) 896 (8.9) 0.0071 36 (15.8) 32 (9.8) 0.0409 ACE-inhibitors /ARBs, n (%) 2222 (51.1) 3904 (37.4) <0.0001 2100 (51.0) 3766 (37.2) <0.0001 122 (53.5) 138 (42.2) 0.0087 Beta-blockers, n (%) 1657 (38.1) 2844 (27.2) <0.0001 1553 (37.7) 2721 (26.9) <0.0001 104 (45.6) 123 (37.6) 0.0609 Downloaded from http://ahajournals.org by on October 30, 2023 Statins, n (%) 1002 (23.1) 2034 (19.5) <0.0001 949 (23.0) 1976 (19.5) <0.0001 53 (23.2) 58 (17.7) 0.1175 ACE indicates angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; CAD, coronary artery disease Table S5. Interaction test calculations for comparing two estimated risk ratios (relative risks of women versus men) by inverse probability of weighting: diabetes, current smoking, hypercholesterolemia, hypertension for obstructive CAD. Group 1 Group 2 Group 1 Group 2 [Diabetes] [No diabetes] [Current smokers] [Non-smokers] (n = 3563) (n = 11230) (n=6420) (n=8373) 1 RR ratio 0.89 0.49 0.75 0.50 2 log RR ratio -0.1165 -0.7133 -0.2877 -0.6931 3 95% CI for 0.62 – 1.29 0.41 – 0.60 0.54 – 1.03 0.41 – 0.61 RR ratio 4 95% CI for -0.4780 – 0.2546 -0.8916 – -0.5108 -0.6162 –0.0296 -0.8916 – -0.4943 log RR ratio 5 Width of CI 0.7326 0.3808 0.6458 0.3973 6 SE (=width / 0.1869 0.0971 0.1647 0.1014 (2*1.96)) Difference between log relative risk ratios Downloaded from http://ahajournals.org by on October 30, 2023 7 d (=𝑬𝟏 − 𝑬𝟐 ) 0.5968 0.4054 8 SE (d) 0.2106 0.1934 9 CI (d) 0.1840 – 1.0096 0.0263 – 0.7845 10 Test of 2.8338 (p-value: 0.0023) 2.0962 (p-value: 0.0180) Interaction Ratio of relative risk ratios 11 RRR ratio 1.8163 1.4999 (=exp(d)) 12 CI (RRR ratio) 1.2020 – 2.7445 1.0266 – 2.1913 Group 1 Group 2 Group 1 Group 2 [Hypercholesterolemia] [No [Hypertension] [No hypertension] (n=6609) hypercholesterolemia] (n=10368) (n=4425) (n=8184) 1 RR ratio 0.55 0.56 0.56 0.50 2 log RR ratio -0.5978 -0.5798 -0.5798 -0.6931 3 95% CI for 0.42 – 0.72 0.45 – 0.70 0.47 – 0.68 0.35 – 0.73 RR ratio 4 95% CI for -0.8675 – -0.3285 -0.7985 – -0.3567 -0.7550 – -0.3857 -1.0498 – -0.3147 log RR ratio 5 Width of CI 0.5390 0.4418 0.3693 0.7351 6 SE (=width / 0.1375 0.1127 0.0942 0.1875 (2*1.96) Difference between log relative risk ratios 7 d (=𝑬𝟏 − 𝑬𝟐 ) -0.0180 0.1133 8 SE (d) 0.1778 0.2098 9 CI (d) -0.3665 – 0.3305 -0.2979 – 0.5245 Downloaded from http://ahajournals.org by on October 30, 2023 10 Test of -0.1012 (p-value: 0.4597) 0.5400 (p-values: 0.2946) Interaction Ratio of relative risk ratios 11 RRR ratio 0.9822 1.1200 (=exp(d)) 12 CI (RRR ratio) 0.6932 – 1.3917 0.7424 – 1.6896 Table S6. Interaction test: calculations for comparing two estimated RR ratios (women versus men) by inverse probability of weighting: STEMI in obstructive versus nonobstructive CAD in patients with acute coronary syndrome at index event. Group 1 Group 2 [Obstructive CAD] [Nonobstructive CAD] (n =14238) (n= 555) 1 RR ratio 1.12 0.92 2 log RR ratio 0.1133 -0.0834 3 95% CI for RR ratio 1.03 – 1.21 0.60 – 1.43 4 95% CI for log RR ratio 0.0296 – 0.1906 -0.5108 – 0.3577 5 Width of CI 0.1611 0.8685 6 SE (=width / (2*1.96) ) 0.0411 0.2216 Difference between log relative risk ratios Downloaded from http://ahajournals.org by on October 30, 2023 7 d (=𝑬𝟏 − 𝑬𝟐 ) 0.1967 8 SE (d) 0.2253 9 CI (d) -0.2449 – 0.6384 10 Test of Interaction 08730 (p-value: 0.1913) Ratio of relative risk ratios 11 RRR ratio( =exp(d) ) 1.2174 12 CI (RRR ratio) 0.7827 – 1.8934 Table S7. Interaction test: calculations for comparing two estimated RR ratios (women versus men) by inverse probability of weighting: 30-day mortality in obstructive versus nonobstructive CAD in patients with acute coronary syndrome at index event. Group 1 Group 2 [Obstructive CAD] [Nonobstructive CAD] (n =14238) (n= 555) 1 RR ratio 1.75 0.79 2 log RR ratio 0.5596 -0.2357 3 95% CI for RR ratio 1.48 – 2.07 0.31 – 1.74 4 95% CI for log RR ratio 0.3920 – 0.7275 -1.1712 – 0.5539 5 Width of CI 0.3355 1.7251 6 SE (=width / (2*1.96) ) 0.0856 0.4401 Difference between log relative risk ratios Downloaded from http://ahajournals.org by on October 30, 2023 7 d (=𝑬𝟏 − 𝑬𝟐 ) 0.7953 8 SE (d) 0.4483 9 CI (d) -0.0834 – 1.6740 10 Test of Interaction 1.7740 (p-value: 0.0380) Ratio of relative risk ratios 11 RRR ratio( =exp(d) ) 2.2151 12 CI (RRR ratio) 0.9200 – 5.3335 Table S8. Inverse probability of weighting: outcomes sorted by sex (women versus men) in patients with obstructive CAD who underwent primary PCI. Primary PCI Women Men Characteristics p value (n=2641) (n=6547) Cardiovascular risk factors Diabetes, % 22.5 22.1 0.6765 Hypertension, % 65.8 66.2 0.7140 Hypercholesterolemia, % 43.1 43.7 0.5996 Current smokers, % 46.6 47.2 0.6021 Former smokers, % 6.7 7.1 0.4957 Clinical history of ischemic heart disease Previous angina pectoris, % 10.8 11.1 0.6780 Previous myocardial infarction, % 10.2 10.2 1.0000 Previous heart failure, % 2.6 2.6 1.0000 Clinical history of cardiovascular disorders Peripheral artery disease, % 1.7 1.7 1.0000 Previous stroke, % 2.8 2.7 0.7894 Downloaded from http://ahajournals.org by on October 30, 2023 Clinical presentation at admission ST-segment shifts in anterior leads (at ECG), % 29.1 29.6 0.6342 Systolic BP at baseline, mean ± SD, mmHg 137.5 ± 28.2 137.5 ± 27.1 0.9307 Heart rate at baseline, mean ± SD, beats/min 80.0 ± 17.7 80.3 ± 17.9 0.6048 Serum creatinine at baseline, mean ± SD, mg/dl 0.98 ± 0.50 1.04 ± 0.60 0.0001 Killip Class≥2, % 17.0 17.1 0.9082 Outcomes 30-day mortality, % 7.1 4.0 <0.0001 Relative Risk Ratio (95% CI) 1.84 (1.52 – 2.23) <0.0001 BP indicates blood pressure; CAD, coronary artery disease; PCI, percutaneous coronary intervention. Table S9. Inverse probability of weighting: outcomes sorted by sex (women versus men) and CAD status in patients with acute coronary syndrome at index event. Analysis restricted the cohort of obstructive CAD patients having 70% or greater stenosis Obstructive CAD Nonobstructive CAD (stenosis ≥70%) (stenosis <70%) Women Men Women Men Characteristics p value p value (n=4037) (n=10043) (n=310) (n=403) Age, mean ± SD, y 61.4 ± 11.9 61.4 ± 11.5 0.8643 60.9 ± 11.8 60.8 ± 12.3 0.8409 Cardiovascular risk factors Diabetes, % 24.4 24.1 0.7070 20.3 21.7 0.6503 Hypertension, % 69.7 69.6 0.9071 78.9 76.8 0.5048 Hypercholesterolemia, % 44.4 44.6 0.8291 43.1 42.0 0.7687 Current smokers, % 43.4 44.0 0.5165 35.3 35.0 0.9338 Downloaded from http://ahajournals.org by on October 30, 2023 Former smokers, % 7.3 7.8 0.3120 10.0 10.3 0.8956 Clinical history of ischemic heart disease Previous angina pectoris, % 15.2 15.6 0.5535 17.6 17.7 0.9723 Previous myocardial infarction, % 13.0 13.4 0.5274 11.6 11.4 0.9339 Previous heart failure, % 3.6 3.8 0.5707 4.5 4.6 0.9496 Clinical history of cardiovascular disease Peripheral artery disease, % 1.7 1.8 0.6821 0.7 1.2 0.4964 Previous stroke, % 2.8 2.8 1.0000 3.4 2.5 0.4778 Clinical presentation at hospital admission ST-segment shifts in anterior leads (at ECG), % 20.7 21.0 0.6922 9.7 9.7 1.0000 Systolic BP at baseline, mean ± SD, mm Hg 139.7±28.0 139.6 ± 26.6 0.8675 142.0±25.5 142.1±26.4 0.9488 Heart rate at baseline, mean ± SD, beats/min 80.0 ± 17.8 80.2 ± 17.9 0.6810 80.1 ± 18.3 79.6 ± 20.9 0.9488 Serum creatinine at baseline, mean ± SD, mg/dl 0.99 ± 0.5 1.06 ± 0.6 <0.0001 0.99 ± 0.4 1.01 ± 0.5 0.4338 Killip Class ≥2, % 16.4 16.4 0.7726 13.4 14.4 0.7029 Outcomes 30-day mortality, % 5.9 3.4 <0.0001 1.1 1.9 0.3846 Relative Risk Ratio (95% CI) 1.75 (1.48 – 2.08) <0.0001 0.56 (0.15 – 2.08) 0.3903 BP indicates blood pressure; CAD, coronary artery disease. Obstructive CAD was defined as a 70% or more narrowing of the luminal diameter. Downloaded from http://ahajournals.org by on October 30, 2023 Table S10. Interaction test: calculations for comparing two estimated RR ratios (women versus men) by inverse probability of weighting: 30-day mortality in obstructive (stenosis ≥70%) versus nonobstructive CAD in patients with acute coronary syndrome at index event. Group 1 Group 2 [Obstructive CAD] [Nonobstructive CAD] (n=14080) (N=713) 1 RR ratio 1.75 0.56 2 log RR ratio 0.5596 -0.5798 3 95% CI for RR ratio 1.48 – 2.08 0.15 – 2.08 4 95% CI for log RR ratio 0.3920 – 0.7324 -1.8971 – 0.7324 5 Width of CI 0.3404 2.6295 6 SE (=width / (2*1.96) ) 0.0868 0.6708 Downloaded from http://ahajournals.org by on October 30, 2023 Difference between log relative risk ratios 7 d (=𝑬𝟏 − 𝑬𝟐 ) 1.1394 8 SE (d) 0.6764 9 CI (d) -0.1863 – 2.4651 10 Test of Interaction 1.6845 (p-value: 0.0460) Ratio of relative risk ratios 11 RRR ratio ( =exp(d) ) 3.1249 12 CI (RRR ratio) 0.8300 – 11.7647 Figure S1. Study Flow Chart. CABG indicates coronary artery bypass graft; CAD, coronary artery disease; NSTE-ACS, non-ST Downloaded from http://ahajournals.org by on October 30, 2023 elevation acute coronary syndromes; PCI, percutaneous coronary intervention; STEMI, ST- elevation myocardial infarction.