Effects of Body Weight on Bleeding and Ischemic Events in Patients Undergoing Percutaneous Coronary Intervention　― From the CREDO-Kyoto Registry Cohort-2 ―

Abstract

Background: The effect of body weight (BW) on bleeding and ischemic events has not been adequately evaluated in real-world percutaneous coronary intervention (PCI) practice.

Methods and Results: 12,690 consecutive patients undergoing first PCI in the CREDO-Kyoto registry cohort-2 were divided into 3 groups according to tertiles of BW stratified by sex (male; Tertile 1 [<60.0 kg], 2 [60.0–68.0 kg], and 3 [>68.0 kg], and female; Tertile 1 [<47.9 kg], 2 [47.9–55.8 kg], and 3 [>55.8 kg]). Cumulative 5-year incidences of the primary bleeding (GUSTO moderate/severe) and ischemic (myocardial infarction/ischemic stroke) endpoints increased incrementally with decrease in BW in both strata (male Tertiles 1, 2, and 3: 13.7%, 10.3%, and 8.0%, P<0.001, and 13.9%, 11.3%, and 10.2%, P<0.001; female Tertiles 1, 2, and 3: 17.9%, 12.9%, and 10.1%, P<0.001, and 17.9%, 12.9%, and 10.1%, P<0.001). Compared with Tertile 3, the adjusted risks of Tertile 1 for the primary bleeding and ischemic endpoints remained significant in the female stratum (hazard ratio (HR): 1.45, 95% confidence interval (CI): 1.14–1.87, P=0.003, and HR:1.49, 95% CI:1.13–1.95, P=0.004), but not in the male stratum (HR:1.10, 95% CI:0.92–1.32, P=0.31, and HR:1.06, 95% CI:0.90–1.27, P=0.47).

Conclusions: Cumulative incidences of bleeding and ischemic events increased incrementally as BW decreased in both men and women. The adjusted risks of underweight relative to overweight for bleeding and ischemic events were significant only in women.

In the well-known obesity paradox, obese patients have a lower mortality risk in chronic disease conditions, particularly heart failure, even though obesity is associated with a higher cardiovascular risk.^1,2 It has been reported that the obesity paradox is also observed in patients undergoing percutaneous coronary intervention (PCI).^3–5 In those reports, the body mass index (BMI) was used to evaluate the association between obesity and cardiovascular outcomes.^1–6 However, the effect of body weight (BW) on bleeding or ischemic events has not been adequately evaluated in real-world PCI practice and could have important clinical implications in Japan, where underweight is more common for Japanese PCI patients. BW might affect the pharmacokinetics of antithrombotic drugs, and underweight patients might have a higher risk for bleeding events with standard doses of antithrombotic drug aimed at protecting patients from thrombotic events after PCI. Furthermore, BW rather than BMI is more practical and convenient for use in real-world clinical practice.

Thus, we aimed to evaluate the influence of BW on the long-term risks for bleeding and ischemic events in a large Japanese multicenter observational PCI database.

Methods

Study Population

The Coronary Revascularization Demonstrating Outcome study in Kyoto (CREDO-Kyoto) PCI/coronary artery bypass grafting (CABG) registry cohort-2 is a multicenter registry enrolling consecutive patients undergoing first coronary revascularization procedures in 26 Japanese centers between January 2005 and December 2007 (Supplementary Appendix A).⁷ The relevant review board or ethics committee at all participating centers approved the research protocol. Because of retrospective enrollment, written informed consent from patients was waived; however, we excluded patients who refused participation in the study when contacted for follow-up. This strategy is concordant with the guidelines for epidemiological studies issued by the Ministry of Health, Labor and Welfare of Japan. The design and patient enrollment of the CREDO-Kyoto PCI/CABG registry cohort-2 have been described previously.⁷

Among the 15,939 patients enrolled in the registry, 13,058 were enrolled in the PCI arm of the registry, excluding 99 patients who refused study participation and 2,782 patients who underwent CABG. After further excluding 368 patients with missing baseline BW data, the present study population consisted of 12,690 patients undergoing their first coronary revascularization with PCI. The study population was divided into 3 groups according to the tertiles of BW at the index PCI stratified by sex (male stratum: Tertile 1: <60.0 kg [n=3,022], Tertile 2: 60.0–68.0 kg [n=3,077], and Tertile 3: >68.0 kg [n=3,042], female stratum: Tertile 1: <47.9 kg [n=1,184], Tertile 2: 47.9–55.8 kg [n=1,185], and Tertile 3: >55.8 kg [n=1,180]) (Figure 1).

Figure 1.

Study flow chart. CABG, coronary artery bypass grafting; CREDO-Kyoto, Coronary Revascularization Demonstrating Outcome study in Kyoto; PCI, percutaneous coronary intervention.

The recommended antiplatelet regimen included aspirin (≥81 mg daily), which was administered indefinitely, and thienopyridine (200 mg of ticlopidine or 75 mg of clopidogrel daily), which was administered for at least 3 months. The duration of dual antiplatelet therapy (DAPT) was left to the discretion of each attending physician. Persistent discontinuation of DAPT was defined as withdrawal of aspirin or thienopyridine for ≥2 months.

Definitions

Baseline clinical characteristics were defined previously.⁷ High bleeding risk (HBR) was defined according to the ARC-HBR criteria with modification, because some of the ARC-HBR criteria were not exactly captured in this registry.⁸ Details of the modification are described in the Supplementary Methods.

The primary bleeding endpoint was major bleeding defined as a Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded arteries trial (GUSTO) moderate/severe bleeding event.⁹ Bleeding was further classified as intracranial, gastrointestinal or access site bleeding. The primary ischemic endpoint was a composite of myocardial infarction (MI) or ischemic stroke. MI was adjudicated by the definition of the Arterial Revascularization Therapies Study.¹⁰ Stroke during follow-up was defined as ischemic or hemorrhagic stroke requiring hospitalization with symptoms lasting >24 h. The secondary endpoints included MI, ischemic stroke, all stroke, all-cause death, cardiac death, non-cardiac death, and major adverse cardiovascular events (MACE), which indicated a composite of death, MI, or stroke. The details of other endpoints are described in the Supplementary Methods. All definitions of the endpoints were reported previously,⁷ and all clinical events were adjudicated by an independent clinical event committee.

Data Collection and Follow-up

The demographic, angiographic, and procedural data from hospital charts or databases in each participating center according to the prespecified definitions were collected by experienced clinical research coordinators in the study management center (Supplementary Appendix B). We collected the follow-up data on clinical events from the hospital charts in the participating centers, or by contacting the patients or referring physicians.

Statistical Analysis

Categorical variables are expressed as number and percentages and compared by chi-square test. Continuous variables are expressed as mean±standard deviation or median and interquartile range, and were compared by analysis of variance or Kruskal-Wallis test based on their distribution. Cumulative incidence was estimated by the Kaplan-Meier method, and differences were assessed by the log-rank test. Cumulative incidence of the first primary bleeding event was also estimated when stratified by either before or after DAPT discontinuation. We used Cox proportional hazard models to estimate the hazard ratio (HR) and 95% confidence interval (CI) of Tertile 1 and Tertile 2 relative to Tertile 3 for clinical endpoints, adjusting the clinically relevant factors for patient characteristics and medications. We selected 10 risk-adjusting variables for the primary bleeding endpoint, 12 risk-adjusting variables for the ischemic endpoints (the primary ischemic endpoint, MI, ischemic stroke, and stroke), and 19 risk-adjusting variables for all-cause death, cardiac death, non-cardiac death, and MACE, as indicated in Table 1. The continuous variables were dichotomized by clinically meaningful reference values, except for age, which was included as a continuous variable. BW and the risk-adjusting variables were simultaneously included in the Cox proportional hazard model. In the Cox proportional hazard model, we developed dummy code variables for Tertile 1 and Tertile 2 with Tertile 3 as the reference. To distinguish the periprocedural events after PCI from the non-periprocedural events, we also performed a landmark analysis at 30 days for the primary bleeding and ischemic endpoints. As a sensitivity analysis, we constructed Cox proportional hazard models for the primary bleeding and ischemic endpoints with DAPT discontinuation as a time-updated covariate together with the same risk-adjusting variables as in the main analysis. In addition, the relationships between uncategorized BW and the primary endpoints and all-cause death were examined with a restricted cubic spline regression model with 3 knots, adjusted by the multivariable Cox proportional hazard model. BW was displayed between the upper and lower 95% CIs. Furthermore, we performed 2 additional analysis: (1) using BMI, and (2) divided into 5 groups according to the quintiles of BW (male stratum: Quintile 1: <55.8 kg [n=1,817], Quintile 2: 55.8–61.3 kg [n=1,837], Quintile 3: 61.4–66.2 kg [n=1,822], Quintile 4: 66.3–72.9 kg [n=1,805], and Quintile 5: >73.0 kg [n=1,860]; female stratum: Quintile 1: <44.4 kg [n=711], Quintile 2: 44.4–49.4 kg [n=702], Quintile 3: 49.5–53.9 kg [n=697], Quintile 4: 54.0–59.9 kg [n=690], and Quintile 5: >60.0 kg [n=749]), for the primary bleeding and ischemic endpoints. The restricted cubic spline regression model was performed by a physician (Y.Y.) using Stata version 15 (Stata Corp., College Station, TX, USA). Multivariable Cox proportional hazard model with a time-updated covariate was performed by a physician (K.Y.) using the EZR,¹¹ and all other analyses were performed using JMP 14.0 software program (SAS Institute, Inc., Cary, NC, USA). All reported P-values are two-tailed, and P<0.05 indicated statistical significance.

Table 1. Baseline Characteristics of the Subjects According to Sex

Variables	Male stratum (n=9,141)				Female stratum (n=3,549)
	Tertile 1 BW <60.0 kg	Tertile 2 BW 60.0–68.0 kg	Tertile 3 BW >68.0 kg	P value	Tertile 1 BW <47.9 kg	Tertile 2 BW 47.9–55.8 kg	Tertile 3 BW >55.8 kg	P value
Clinical characteristics
Age (years)*,†,‡	71.8±9.2	66.4±9.6	60.7±10.9	<0.001	76.6±9.2	72.3±9.9	68.9±9.2	<0.001
Age ≥75 years	1,296 (42.9%)	640 (20.8%)	323 (10.6%)	<0.001	769 (64.9%)	523 (44.1%)	341 (28.9%)	<0.001
BMI (kg/m2)	20.9±2.2 (2,987)	23.8±1.6 (3,063)	27.0±2.7 (3,031)	<0.001	19.8±2.1 (1,158)	23.1±1.8 (1,167)	27.1±3.0 (1,171)	<0.001
Acute MI†,‡	1,032 (34.1%)	1,138 (37.0%)	1,085 (35.7%)	0.07	456 (38.5%)	417 (35.2%)	356 (30.2%)	<0.001
Hypertension*,‡	2,350 (77.8%)	2,488 (80.9%)	2,574 (84.6%)	<0.001	960 (81.1%)	1,010 (85.2%)	1,053 (89.2%)	<0.001
Diabetes mellitus†,‡	1,032 (34.1%)	1,112 (36.1%)	1,257 (41.3%)	<0.001	367 (31.0%)	443 (37.4%)	571 (48.4%)	<0.001
On insulin therapy	215 (7.1%)	206 (6.7%)	211 (6.9%)	0.81	87 (7.4%)	104 (8.8%)	156 (13.2%)	<0.001
Current smoking†,‡	1,083 (35.8%)	1,184 (38.5%)	1,358 (44.6%)	<0.001	116 (9.8%)	142 (12.0%)	172 (14.6%)	0.002
Heart failure*,†,‡	675 (22.3%)	495 (16.1%)	448 (14.7%)	<0.001	356 (30.1%)	265 (22.4%)	210 (17.8%)	<0.001
Prior MI†,‡	400 (13.2%)	310 (10.1%)	288 (9.5%)	<0.001	128 (10.8%)	102 (8.6%)	114 (9.7%)	0.19
Prior stroke†,‡	416 (13.8%)	307 (10.0%)	241 (7.9%)	<0.001	137 (11.6%)	114 (9.6%)	102 (8.6%)	0.053
Prior hemorrhagic stroke	63 (2.1%)	51 (1.7%)	37 (1.2%)	0.03	20 (1.7%)	20 (1.7%)	13 (1.1%)	0.4
Prior ischemic stroke	366 (12.1%)	268 (8.7%)	213 (7.0%)	<0.001	120 (10.1%)	98 (8.3%)	94 (8.0%)	0.13
PVD†,‡	359 (11.9%)	240 (7.8%)	157 (5.2%)	<0.001	63 (5.3%)	68 (5.7%)	57 (4.8%)	0.62
AF*,†,‡	290 (9.6%)	228 (7.4%)	242 (8.0%)	0.006	119 (10.1%)	106 (8.9%)	87 (7.4%)	0.07
CKD (eGFR <60 mL/min/1.73 m2)	1,220 (40.4%)	1,063 (34.5%)	971 (31.9%)	<0.001	569 (48.1%)	472 (39.8%)	443 (37.5%)	<0.001
Moderate CKD (eGFR 30–59mL/min/1.73 m2)*,†,‡	919 (30.4%)	869 (28.2%)	838 (27.5%)	0.04	418 (35.3%)	381 (32.2%)	359 (30.4%)	0.04
Severe CKD (eGFR <30 mL/min/1.73 m2)*,†,‡	301 (10.0%)	194 (6.3%)	133 (4.4%)	<0.001	151 (12.8%)	91 (7.7%)	84 (7.1%)	<0.001
eGFR <30 mL/min/1.73 m2, not on dialysis	129 (4.3%)	96 (3.1%)	69 (2.3%)	<0.001	83 (7.0%)	58 (4.9%)	61 (5.2%)	0.054
Dialysis	172 (5.7%)	98 (3.2%)	64 (2.1%)	<0.001	68 (5.7%)	33 (2.8%)	23 (1.9%)	<0.001
Mild anemia (Hb 11–12.9 g/dL for men and 11–11.9 g/dL for women)	834 (27.6%)	549 (17.8%)	330 (10.8%)	<0.001	291 (24.6%)	220 (18.6%)	157 (13.3%)	<0.001
Severe anemia (Hb <11 g/dL)*,‡	436 (14.4%)	197 (6.4%)	109 (3.6%)	<0.001	357 (30.2%)	197 (16.6%)	147 (12.5%)	<0.001
Platelets <100 ×109/L*,‡	71 (2.3%)	48 (1.6%)	28 (0.9%)	<0.001	21 (1.8%)	12 (1.0%)	9 (0.8%)	0.06
LDL (mg/dL)	111±33 (2,455)	118±34 (2,506)	121±36 (2,491)	<0.001	119±36 (977)	123±37 (975)	128±37 (991)	<0.001
LDL <70 mg/dL	215 (8.8%)	137 (5.5%)	146 (5.4%)	<0.001	53 (5.4%)	50 (5.1%)	30 (3.0%)	0.02
LVEF	56.7±13.9 (2,522)	58.4±12.6 (2,558)	58.7±12.3 (2,545)	<0.001	58.9±14.5 (974)	60.6±13.1 (986)	61.4±12.1 (981)	<0.001
LVEF ≤40%	337 (13.4%)	238 (9.3%)	226 (8.9%)	<0.001	134 (13.8%)	88 (8.9%)	64 (6.5%)	<0.001
Mitral regurgitation grade 3/4	133 (4.4%)	99 (3.2%)	60 (2.0%)	<0.001	91 (7.7%)	68 (5.7%)	51 (4.3%)	0.002
COPD	104 (3.4%)	85 (2.8%)	101 (3.3%)	0.27	33 (2.8%)	63 (5.3%)	71 (6.0%)	<0.001
Liver cirrhosis*,‡	97 (3.2%)	94 (3.1%)	52 (1.7%)	<0.001	39 (3.3%)	26 (2.2%)	20 (1.7%)	0.03
Malignancy*,‡	390 (12.9%)	260 (8.4%)	200 (6.6%)	<0.001	110 (9.3%)	102 (8.6%)	91 (7.7%)	0.39
ARC-HBR	1,605 (53.1%)	1,110 (36.1%)	790 (26.0%)	<0.001	815 (68.8%)	585 (49.4%)	453 (38.4%)	<0.001
Procedural characteristics
No. of target lesions	1 (1–2)	1 (1–2)	1 (1–2)	0.2	1 (1–2)	1 (1–2)	1 (1–2)	0.42
	1.47±0.76	1.49±0.76	1.45±0.75		1.46±0.77	1.45±0.71	1.44±0.76
Target of proximal LAD	1,707 (56.5%)	1,750 (56.9%)	1,771 (58.2%)	0.36	693 (58.5%)	695 (58.7%)	657 (55.7%)	0.25
Target of unprotected LMCA	127 (4.2%)	113 (3.7%)	84 (2.8%)	0.009	39 (3.3%)	52 (4.4%)	24 (2.0%)	0.005
Target of CTO	351 (11.6%)	361 (11.7%)	441 (14.5%)	0.001	94 (7.9%)	126 (10.6%)	117 (9.9%)	0.07
Total no. of stents	1 (1–2)	1 (1–2)	1 (1–2)	0.29	1 (1–2)	1 (1–2)	1 (1–2)	0.69
	1.89±1.26 (2,836)	1.88±1.24 (2,895)	1.84±1.20 (2,847)		1.84±1.23 (1,084)	1.85±1.22 (1,089)	1.83±1.25 (1,097)
Total stent length (mm)	30 (18–51)	28 (18–51)	28 (18–51)	0.6	28 (18–47)	28 (20–50)	28 (18–46)	0.07
	40.4±30.0 (2,836)	40.2±29.6 (2,895)	39.6±29.1 (2,847)		37.9±27.6 (1,083)	39.5±28.6 (1,089)	38.5±29.5 (1,096)
Minimum stent size	2.90±0.43 (2,836)	2.95±0.46 (2,895)	3.01±0.48 (2,847)	<0.001	2.81±0.39 (1,083)	2.82±0.40 (1,089)	2.86±0.42 (1,096)	0.007
Approach				0.009				0.15
Radial	778 (26.0%)	848 (27.6%)	864 (28.4%)		259 (21.9%)	276 (23.3%)	302 (25.6%)
Femoral	1,862 (61.6%)	1,878 (61.0%)	1,876 (61.7%)		792 (66.9%)	778 (65.7%)	728 (61.7%)
Brachial	356 (11.8%)	337 (11.0%)	286 (9.4%)		127 (10.7%)	120 (10.1%)	141 (12.0%)
Unknown	26 (0.9%)	14 (0.5%)	16 (0.5%)		6 (0.5%)	11 (0.9%)	9 (0.8%)
Medications at hospital discharge
Thienopyridine	2,953 (97.7%)	2,994 (97.3%)	2,971 (97.7%)	0.52	1,133 (95.7%)	1,151 (97.1%)	1,153 (97.7%)	0.02
Ticlopidine	2,670 (88.4%)	2,711 (88.1%)	2,672 (87.8%)	0.83	1,015 (85.7%)	1,017 (85.8%)	1,040 (88.1%)	0.15
Clopidogrel	278 (9.2%)	279 (9.1%)	288 (9.5%)	0.86	117 (9.9%)	131 (11.1%)	107 (9.1%)	0.27
Aspirin	2,979 (98.6%)	3,036 (98.7%)	3,013 (99.0%)	0.21	1,167 (98.6%)	1,161 (98.0%)	1,166 (98.8%)	0.24
Warfarin	282 (9.3%)	244 (7.9%)	265 (8.7%)	0.15	89 (7.5%)	85 (7.2%)	90 (7.6%)	0.91
Statins†,‡	1,225 (41.4%)	1,566 (51.7%)	1,808 (60.1%)	<0.001	524 (46.2%)	654 (57.1%)	793 (67.9%)	<0.001
β-blockers‡	859 (28.4%)	952 (30.9%)	1,057 (34.7%)	<0.001	347 (29.3%)	356 (30.0%)	338 (28.6%)	0.76
ACEI/ARB‡	1,654 (54.7%)	1,812 (58.9%)	1,887 (62.0%)	<0.001	656 (55.4%)	682 (57.6%)	754 (63.9%)	<0.001
Proton pump inhibitors	817 (27.0%)	738 (24.0%)	703 (23.1%)	0.001	358 (30.2%)	345 (29.1%)	301 (25.5%)	0.03

Categorical variables are presented as numbers and percentages. Continuous variables are presented as mean and standard deviation or median and interquartile range depending on their distribution. For the variables with missing values, the numbers of patients evaluated are indicated in parentheses. We modified the ARC-HBR definitions because some of the criteria were not captured exactly in this registry. Details of the modification are described in the Supplementary Methods. *Potential risk-adjusting variables for the multivariate Cox proportional hazard model for the primary bleeding endpoint. ^†Potential risk-adjusting variables for the multivariate Cox proportional hazard model for the ischemic endpoint (the primary ischemic endpoint, MI, ischemic stroke, and stroke). ^‡Potential risk-adjusting variables for the multivariate Cox proportional hazard model for all-cause death, cardiac death, non-cardiac death, and MACE. ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; ARC-HBR, the Academic Research Consortium for High Bleeding Risk; BMI, body mass index; BW, body weight; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CTO, chronic total occlusion; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; LAD, left anterior descending; LDL, low density lipoprotein; LMCA, left main coronary artery; LVEF, left ventricular ejection fraction; MI, myocardial infarction; PVD, peripheral vascular disease.

Results

Baseline Characteristics and Medications

The distributions of BW and BMI for the current study population are shown in Figure 2. At baseline, the mean BW and BMI were 64.7±11.0 kg and 23.9±3.3 kg/m², respectively, in the male stratum, and 52.3±9.7 kg and 23.3±3.8 kg/m², respectively, in the female stratum.

Figure 2.

Distribution of body weight and body mass index. The BMI population consisted of 12,577 patients after excluding 113 patients whose baseline height data were missing. BMI, body mass index; CREDO-Kyoto, Coronary Revascularization Demonstrating Outcome study in Kyoto; PCI, percutaneous coronary intervention.

Underweight patients were older, and more often had comorbidities such as heart failure, chronic kidney disease, anemia, low ejection fraction, mitral regurgitation, and liver cirrhosis in both the male and female strata. Hypertension and diabetes were more common in overweight patients in both strata. Underweight male patients more often had a history of MI and stroke, peripheral vascular disease, atrial fibrillation, thrombocytopenia, and malignancy compared with the female stratum. HBR were more prevalent in underweight patients in both the male and female strata. Regarding medication at hospital discharge, underweight patients were less often treated with statins, β-blockers or angiotensin-converting enzyme inhibitor/angiotensin receptor blocker (Table 1).

Primary Bleeding Endpoint Through 5 Years

Median follow-up duration after the day of the index PCI procedure for the surviving patients was 5.3 (interquartile range 4.6–6.1) years. Complete 1-, 3- and 5-yerar clinical follow-up information was obtained for 98.6%, 96.5%, and 69.0% of patients, respectively.

The cumulative incidence of persistent DAPT discontinuation at 5 years was not significantly different across the tertiles of BW in the male stratum (Tertiles 1, 2, and 3: 71.3%, 69.7%, and 69.6%, log-rank P=0.25), but was significantly lower in overweight patients in the female stratum (Tertiles 1, 2 and 3: 74.7%, 73.4%, and 69.4%, log-rank P=0.004) (Supplementary Figure 1A,B).

The cumulative 5-year incidence of the primary bleeding endpoint increased incrementally with decreasing BW in both the male and female strata (male stratum: Tertiles 1, 2, and 3: 13.7%, 10.3%, and 8.0%, log-rank P<0.001; female stratum: Tertiles 1, 2, and 3: 17.9%, 12.9%, and 10.1%, log-rank P<0.001) (Figures 3A,4A). After adjusting confounders, the risk of Tertile 1 relative to Tertile 3 for the primary bleeding endpoint remained significant in the female stratum (HR 1.45, 95% CI 1.14–1.87, P=0.003), but was no longer significant in the male stratum (HR 1.10, 95% CI 0.92–1.32, P=0.31) (Table 2). In both the male and female strata, the risk of Tertile 2 relative to Tertile 3 was not significant for the primary bleeding endpoint (HR 1.08, 95% CI 0.90–1.29, P=0.42, and HR 1.26, 95% CI 0.98–1.62, P=0.07) (Table 2). In the sensitivity analysis incorporating DAPT discontinuation during follow-up as a time-updated covariate, the result for the primary bleeding endpoint was fully consistent with that in the main analysis in both the male and female strata (Supplementary Table 1). Restricted cubic spline models without categorization of BW demonstrated that the adjusted risk for the primary bleeding endpoint tended to increase linearly with decreasing BW in both the male and female strata.

Figure 3.

Kaplan-Meier curves for the clinical events through 5 years in the male stratum. (A) Primary bleeding endpoint, (B) primary ischemic endpoint, and (C) all-cause death. GUSTO, Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries; MI, myocardial infarction; PCI, percutaneous coronary intervention.

Figure 4.

Kaplan-Meier curves for the clinical events through 5 years in the female stratum. (A) Primary bleeding endpoint, (B) primary ischemic endpoint, and (C) all-cause death. GUSTO, Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries; MI, myocardial infarction; PCI, percutaneous coronary intervention.

Table 2. Clinical Outcomes Through 5 Years

Variables	BW Tertile 1					BW Tertile 2					BW Tertile 3
	No. of patients with event*	HR [95% CI]				No. of patients with event*	HR [95% CI]				No. of patients with event*	HR [95% CI]
		Crude	P value	Adjusted	P value		Crude	P value	Adjusted	P value
Male stratum (n=9,141)	BW <60.0 kg (n=3,022)					BW 60.0–68.0 kg (n=3,077)					BW >68.0 kg (n=3,042)
Primary bleeding endpoint (GUSTO moderate/severe bleeding)	365 (13.7%)	1.75	<0.001	1.10	0.31	295 (10.3%)	1.31	0.002	1.08	0.42	228 (8.0%)	Ref.
		[1.49– 2.07]		[0.92– 1.32]			[1.10– 1.56]		[0.90– 1.29]
Primary ischemic endpoint (MI/ischemic stroke)	372 (13.9%)	1.39	<0.001	1.06	0.47	325 (11.3%)	1.13	0.14	1.01	0.92	290 (11.2%)	Ref.
		[1.19– 1.62]		[0.90– 1.27]			[0.96– 1.32]		[0.86– 1.19]
MI	212 (7.7%)	1.11	0.3	0.96	0.71	204 (7.0%)	1.00	0.96	0.97	0.78	203 (7.1%)	Ref.
		[0.91– 1.34]		[0.77– 1.19]			[0.83– 1.22]		[0.79– 1.19]
Ischemic stroke	173 (6.8%)	2.04	<0.001	1.25	0.11	130 (4.6%)	1.40	0.01	1.10	0.5	94 (3.3%)	Ref.
		[1.59– 2.63]		[0.95– 1.65]			[1.08– 1.83]		[0.84– 1.45]
Stroke	212 (8.4%)	2.00	<0.001	1.27	0.056	176 (6.2%)	1.51	<0.001	1.22	0.11	118 (4.2%)	Ref.
		[1.60– 2.51]		[0.99– 1.63]			[1.20– 1.92]		[0.96– 1.55]
All-cause death	741 (26.0%)	3.35	<0.001	1.59	<0.001	385 (13.3%)	1.58	<0.001	1.11	0.23	246 (8.7%)	Ref.
		[2.91– 3.88]		[1.35– 1.88]			[1.35– 1.86]		[0.94– 1.33]
Cardiac death	302 (11.1%)	2.86	<0.001	1.63	<0.001	146 (5.0%)	1.28	0.049	0.90	0.48	115 (4.0%)	Ref.
		[2.31– 3.55]		[1.25– 2.16]			[1.00– 1.63]		[0.67– 1.21]
Non-cardiac death	439 (16.7%)	3.79	<0.001	1.58	<0.001	239 (8.7%)	1.85	<0.001	1.24	0.054	131 (4.9%)	Ref.
		[3.13– 4.63]		[1.29– 1.96]			[1.50– 2.29]		[0.996– 1.54]
Death/MI/stroke	934 (32.6%)	2.33	<0.001	1.34	<0.001	610 (20.8%)	1.38	<0.001	1.08	0.28	448 (15.7%)	Ref.
		[2.08– 2.60]		[1.17– 1.52]			[1.23– 1.56]		[0.94– 1.23]
Female stratum (n=3,549)	BW <47.9 kg (n=1,184)					BW 47.9–55.8 kg (n=1,185)					BW >55.8 kg (n=1,180)
Primary bleeding endpoint (GUSTO moderate/severe bleeding)	190 (17.9%)	1.88	<0.001	1.45	0.003	144 (12.9%)	1.33	0.02	1.26	0.07	113 (10.1%)	Ref.
		[1.49– 2.38]		[1.14– 1.87]			[1.04– 1.71]		[0.98– 1.62]
Primary ischemic endpoint (MI/ischemic stroke)	157 (15.2%)	1.67	<0.001	1.49	0.004	117 (10.8%)	1.16	0.26	1.13	0.4	105 (9.5%)	Ref.
		[1.31– 2.15]		[1.13– 1.95]			[0.89– 1.51]		[0.86– 1.48]
MI	101 (9.6%)	1.90	<0.001	1.80	<0.001	76 (6.9%)	1.35	0.08	1.34	0.1	58 (5.2%)	Ref.
		[1.38– 2.64]		[1.27– 2.57]			[0.96– 1.91]		[0.94– 1.92]
Ischemic stroke	66 (6.8%)	1.47	0.04	1.16	0.47	46 (4.4%)	0.94	0.77	0.87	0.51	51 (4.7%)	Ref.
		[1.02– 2.12]		[0.78– 1.74]			[0.63– 1.40]		[0.57– 1.31]
Stroke	85 (8.8%)	1.51	0.01	1.26	0.2	61 (5.8%)	1.00	0.99	0.96	0.84	64 (5.8%)	Ref.
		[1.09– 2.09]		[0.89– 1.81]			[0.70– 1.42]		[0.67– 1.39]
All-cause death	316 (28.4%)	3.57	<0.001	2.13	<0.001	178 (15.7%)	1.85	<0.001	1.43	0.009	101 (9.2%)	Ref.
		[2.87– 4.49]		[1.65– 2.78]			[1.45– 2.37]		[1.09– 1.89]
Cardiac death	171 (16.3%)	3.75	<0.001	2.12	<0.001	110 (9.8%)	2.25	<0.001	1.56	0.03	51 (4.7%)	Ref.
		[2.76– 5.18]		[1.45– 3.15]			[1.62– 3.16]		[1.05– 2.35]
Non-cardiac death	145 (14.5%)	3.40	<0.001	2.16	<0.001	68 (6.5%)	1.44	0.049	1.32	0.14	50 (4.7%)	Ref.
		[2.48– 4.73]		[1.53– 3.10]			[1.00– 2.08]		[0.91– 1.93]
Death/MI/stroke	389 (34.8%)	2.62	<0.001	1.73	<0.001	252 (22.1%)	1.57	<0.001	1.29	0.02	170 (15.3%)	Ref.
		[2.19– 3.14]		[1.41– 2.13]			[1.29– 1.91]		[1.05– 1.59]

*Cumulative incidence. Number of patients with events was counted through 5 years. Cumulative incidence was estimated by the Kaplan-Meier method, and censored at 5 years. Hazard ratios with 95% CI are expressed relative to Tertile 3. The covariates for the multivariable analysis are indicated in Table 1. BW, body weight; CI, confidence interval; HR, hazard ratio; GUSTO, Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries; MI, myocardial infarction.

In the 30-day landmark analysis, the cumulative incidence of the primary bleeding endpoint within 30 days and beyond 30 days increased incrementally with decreasing BW in both the male and female strata (Supplementary Table 2). The adjusted risk of Tertile 1 relative to Tertile 3 for the primary bleeding endpoint was significant within 30 days, but not beyond 30 days in the female stratum, but was no longer significant either within or beyond 30 days in the male stratum (Supplementary Table 2).

The cumulative 5-year incidence of the primary bleeding endpoint in underweight patients was high both before and after DAPT discontinuation (Supplementary Figure 2). The data on intracranial bleeding, gastrointestinal bleeding, and access site bleeding are shown in Supplementary Figure 3. The incidence of access site bleeding was significantly higher in underweight patients than in overweight patients in the female stratum. In the analysis using BMI, the results were consistent with those using BW in both strata (Supplementary Table 3). In the analysis divided into 5 groups according to the quintiles of BW, the cumulative 5-year incidence of the primary bleeding endpoint increased incrementally with decreasing BW in both strata (Supplementary Figures 4,5).

Primary Ischemic Endpoint Through 5 Years

The cumulative 5-year incidence of the primary ischemic endpoint increased incrementally with decreasing BW in both the male and female strata (male stratum: Tertiles 1, 2, and 3: 13.9%, 11.3%, and 10.2%, log-rank P<0.001; female stratum: Tertiles 1, 2, and 3: 15.2%, 10.8%, and 9.5%, log-rank P<0.001) (Figures 3B,4B). After adjusting confounders, the risk of Tertile 1 relative to Tertile 3 for the primary ischemic endpoint remained significant in the female stratum (HR 1.49, 95% CI 1.13–1.95, P=0.004), but was no longer significant in the male stratum (HR 1.06, 95% CI 0.90–1.27, P=0.47) (Table 2). In both the male and female strata, the risk of Tertile 2 relative to Tertile 3 was not significant for the primary ischemic endpoint (HR 1.01, 95% CI 0.86–1.19, P=0.92, and HR 1.13, 95% CI 0.86–1.48, P=0.4) (Table 2). In the sensitivity analysis incorporating the status of DAPT discontinuation during follow-up, the result for the primary ischemic endpoint was fully consistent with that in the main analysis in both the male and female strata (Supplementary Table 1). Restricted cubic spline models without categorization of BW demonstrated that the adjusted risk for the primary ischemic endpoint tended to increase linearly with decreasing BW in both the male and female strata (Figure 5).

Figure 5.

Restricted cubic spline regression models showing the effect of body weight (BW) on the primary bleeding endpoint, primary ischemic endpoint, and all-cause death. (A) Male stratum, and (B) female stratum. The reference was the median value of baseline BW (male stratum: 64.0 kg; female stratum: 52.0 kg). The horizontal axis of BW is displayed between the upper and lower 95% confidence intervals. GUSTO, Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries; HR, hazard ratio; MI, myocardial infarction.

In the 30-day landmark analysis, the cumulative incidence of the primary ischemic endpoint within 30 days and beyond 30 days increased incrementally with decreasing BW in both the male and female strata (Supplementary Table 2). The adjusted risk of Tertile 1 relative to Tertile 3 for the primary ischemic endpoint was significant within 30 days, but not beyond 30 days in the female stratum, but was no longer significant either within or beyond 30 days in the male stratum (Supplementary Table 2).

In the analysis using BMI, the results were consistent with those using BW in both strata (Supplementary Table 3). In the analysis divided into 5 groups according to the quintiles of BW, the cumulative 5-year incidence of the primary ischemic endpoint increased incrementally with decreasing BW in both strata (Supplementary Figures 4,5).

Secondary Endpoints Through 5 Years

The cumulative 5-year incidence of all-cause death increased incrementally with decreasing BW in both the male and female strata (male stratum: Tertiles 1, 2, and 3: 26.0%, 13.3%, and 8.7%, log-rank P<0.001; female stratum; Tertiles 1, 2, and 3: 28.4%, 15.7%, and 9.2%, log-rank P<0.001) (Figures 3C,4C). In the male stratum, the adjusted risk of Tertile 1 relative to Tertile 3 for all-cause death remained significant (HR 1.59, 95% CI 1.35–1.88, P<0.001), but the adjusted risk of Tertile 2 relative to Tertile 3 was no longer significant (HR 1.11, 95% CI 0.94–1.33, P=0.23) (Table 2). In the female stratum, the adjusted risks of Tertile 1 and Tertile 2 relative to Tertile 3 for all-cause death remained significant (HR 2.13, 95% CI, 1.65–2.78, P<0.001, and HR 1.43, 95% CI 1.09–1.89, P=0.009, respectively) (Table 2). Restricted cubic spline models without categorization of BW demonstrated incrementally higher hazards in underweight patients for all-cause death in both the male and female strata.

The results for the other secondary endpoints such as MI, ischemic stroke, all stroke, cardiac death, non-cardiac death, and MACE are shown in Table 2.

Discussion

The main findings of this observational study in patients who underwent PCI were as follows. (1) Underweight patients had higher incidences of both bleeding and ischemic events, which increased incrementally with decreasing BW. (2) Being underweight was associated with adjusted risks for bleeding and ischemic events in women, but not in men. (3) Underweight was associated with adjusted mortality risk in both men and women.

Previous studies have shown that underweight patients have an increased event rate for bleeding among those undergoing PCI.^4,6 In line with those reports, underweight patients had a higher incidence of bleeding events in both men and women in the present study. Approximately 50% of the underweight male patients and 70% of the underweight female patients had HBR based on the ARC-HBR criteria. A higher bleeding event rate in underweight patients was likely to be due to higher prevalence of comorbidities related to HBR. It has also been reported that being underweight could be associated with malnutrition, cachexia, chronic inflammation, amyloidosis, frailty, and undiagnosed malignancy, leading to higher risk of major bleeding.^12,13 Ndrepepa et al reported that the risk of bleeding after PCI was significantly higher in women compared with men after matching for age and BMI.¹⁴ In line with that report, the adjusted risk of underweight patients for major bleeding was not significant in men, but remained significant in women in this study. One of the reasons for the excess bleeding risk in underweight women, but not men, might be related to the lower BW cutoff in women. Unmeasured confounders for the increased bleeding risk might be more prevalent in female patients with very low BW. In current practice, the use of antiplatelet therapy after PCI is fundamentally based on the “one-dose-fits-all” principle. However, it has been reported that underweight patients taking antiplatelet drugs have been associated with lower maximal platelet aggregation and higher active metabolite exposure in some pharmacodynamic studies.^6,15,16 Therefore, we might need to consider adjusting the doses of antithrombotic drugs and/or the intensity of the treatment regimen according to BW in order to prevent ischemic events without increasing bleeding. Ohya et al reported that low-dose prasugrel could be a treatment option after PCI in HBR patients such as those with low BW or the elderly.¹⁷ Therefore, dose adjustment for antiplatelet drugs or short duration of DAPT after PCI might be considered in underweight patients, particularly female patients. The results of this study support the Japanese version of HBR including low BW as a major criterion proposed in the Japanese Circulation Society (JCS) guidelines focused update on antithrombotic therapy.¹⁸

BMI has been used to define underweight in previous studies.^1–6 However, the influence of BW on bleeding or ischemic events was consistent with that of BMI in this study. In fact, dose adjustment criteria of antithrombotic drugs such as prasugrel, and direct oral anticoagulants adopt BW, not BMI. BW rather than BMI would be more practical and convenient for use in real-world clinical practice.

It has been reported that underweight patients who undergo PCI also shown have a higher risk of periprocedural bleeding.^6,19 In the current analysis, underweight patients had a higher risk for bleeding event mainly within 30 days, and the incidence of access site bleeding was significantly higher in underweight patients than in overweight patients in the female stratum. McDonagh et al reported that radial-approach PCI was particularly safer than femoral approach PCI in underweight patients,²⁰ so transradial PCI should be considered as an important option to reduce bleeding complications, especially in underweight patients.

In the current analysis, ischemic events occurred incrementally more frequently as BW decreased, although it has been reported that obesity (BMI >30–35 kg/m²) has a higher risk of ischemic events.⁶ This difference in results is likely related to a higher prevalence of advanced age patients and comorbidities such as peripheral vascular disease, atrial fibrillation, chronic kidney disease, and history of MI and stroke in the underweight patients in the present study. In addition, it could be affected by the very low prevalence of obese patients (BMI >30 kg/m²) in the current study population, which differed greatly from previous studies in non-Asian patients.²¹ Further research is needed to elucidate the relationship between BW and cardiovascular outcomes in those populations.

In the current analysis, underweight patients had much higher mortality risk in both men and women, consistent with the previous studies.^3–5 Malnutrition, cachexia, chronic inflammation, and frailty prevalent in underweight patients might be associated with higher risk of death, as well as higher risk of bleeding events. Veronese et al reported that both underweight and obesity were associated with higher mortality risk in patients without a healthy lifestyle such as high levels of physical activity, healthy diet, and never smoking, whereas only obesity was associated with higher mortality risk in patients with a healthy lifestyle.²²

Study Limitations

First, in this observational study, we could not deny residual unmeasured confounders in underweight female patients. Second, 1st-generation drug-eluting stents, and older antiplatelet agents were used in a large proportion of patients, which are no longer used in current PCI practice. None of the patients in the present study used prasugrel or non-vitamin K oral anticoagulant, for which dose adjustment is considered in underweight patients. Third, we only measured BW at the time of the index procedure, and did not consider changes in BW. In addition, we did not have any data on waist circumference or waist-to-hip ratio, which are related to the amount of body fat. Finally, we did not have information on malnutrition, cachexia, gastrointestinal disease, depression, and frailty, which are related to underweight and clinical outcomes.

Conclusions

Cumulative incidences of bleeding and ischemic events increased incrementally as BW decreased in both men and women. The adjusted risks of underweight relative to overweight for bleeding and ischemic events were significant only in women.

Acknowledgments

We appreciate the support and collaboration of the co-investigators participating in the CREDO-Kyoto PCI/CABG Registry Cohort-2. We are indebted to the clinical research coordinators for data collection.

Data Availability

The deidentified participant data will not be shared.

Disclosures

The CREDO-Kyoto PCI/CABG registry cohort-2 was supported by the Pharmaceuticals and Medical Devices Agency in Tokyo, Japan. Dr. Kimura has served on the advisory board of Abbott Vascular and Terumo, and reports research grants from Abbott Vascular. All the other authors report that they have no relationships relevant to the contents of this paper to disclose.

IRB Information

Kyoto University Graduate School and Faculty of Medicine, Ethics Committee (reference no. e42).

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-20-0343

References

• 1.   Di Angelantonio E, Bhupathiraju SN, Wormser D, Gao P, Kaptoge S, de Gonzalez AB, et al. Body-mass index and all-cause mortality: Individual-participant-data meta-analysis of 239 prospective studies in four continents. Lancet 2016; 388: 776–786.
• 2.   Oreopoulos A, Padwal R, Kalantar-Zadeh K, Fonarow GC, Norris CM, McAlister FA. Body mass index and mortality in heart failure: A meta-analysis. Am Heart J 2008; 156: 13–22.
• 3.   Oreopoulos A, Padwal R, Norris CM, Mullen JC, Pretorius V, Kalantar-Zadeh K. Effect of obesity on short- and long-term mortality postcoronary revascularization: A meta-analysis. Obesity 2008; 16: 442–450.
• 4.   Lamelas P, Schwalm JD, Quazi I, Mehta S, Devereaux PJ, Jolly S, et al. Effect of body mass index on clinical events after acute coronary syndromes. Am J Cardiol 2017; 120: 1453–1459.
• 5.   Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, Allison TG, et al. Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: A systematic review of cohort studies. Lancet 2006; 368: 666–678.
• 6.   Rocca B, Fox KAA, Ajjan RA, Andreotti F, Baigent C, Collet JP, et al. Antithrombotic therapy and body mass: An expert position paper of the ESC Working Group on Thrombosis. Eur Heart J 2018; 39: 1672–1691.
• 7.   Kimura T, Morimoto T, Furukawa Y, Nakagawa Y, Kadota K, Iwabuchi M, et al. Long-term safety and efficacy of sirolimus-eluting stents versus bare-metal stents in real world clinical practice in Japan. Cardiovasc Interv Ther 2011; 26: 234–245.
• 8.   Urban P, Mehran R, Colleran R, Angiolillo DJ, Byrne RA, Capodanno D, et al. Defining high bleeding risk in patients undergoing percutaneous coronary intervention: A Consensus Document from the Academic Research Consortium for High Bleeding Risk. Circulation 2019; 140: 240–261.
• 9.   Topol E, Califf R, Vandewerf F, Armstrong PW, Aylward P, Barbash G, et al. An international randomized trial comparing 4 thrombolytic strategies for acute myocardial-infarction. N Engl J Med 1993; 329: 673–682.
• 10.   Serruys PW, Unger F, Sousa JE, Jatene A, Bonnier H, Schonberger J, et al. Comparison of coronary-artery bypass surgery and stenting for the treatment of multivessel disease. N Engl J Med 2001; 344: 1117–1124.
• 11.   Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant 2013; 48: 452–458.
• 12.   Knudtson MD, Klein BEK, Klein R, Shankar A. Associations with weight loss and subsequent mortality risk. Ann Epidemiol 2005; 15: 483–491.
• 13.   Hernandez JL, Riancho JA, Matorras P, Gonzalez-Macias J. Clinical evaluation for cancer in patients with involuntary weight loss without specific symptoms. Am J Med 2003; 114: 631–637.
• 14.   Ndrepepa G, Schulz S, Neumann FJ, Byrne RA, Hoppmann P, Cassese S, et al. Bleeding after percutaneous coronary intervention in women and men matched for age, body mass index, and type of antithrombotic therapy. Am Heart J 2013; 166: 534–540.
• 15.   Erlinge D, Ten Berg J, Foley D, Angiolillo DJ, Wagner H, Brown PB, et al. Reduction in platelet reactivity with prasugrel 5 mg in low-body-weight patients is noninferior to prasugrel 10 mg in higher-body-weight patients. J Am Coll Cardiol 2012; 60: 2032–2040.
• 16.   Riesmeyer JS, Salazar DE, Weerakkody GJ, Ni L, Wrishko RE, Ernest CS, et al. Relationship between exposure to prasugrel active metabolite and clinical outcomes in the TRITON-TIMI 38 substudy. J Clin Pharmacol 2012; 52: 789–797.
• 17.   Ohya M, Shimada T, Osakada K, Kuwayama A, Miura K, Murai R, et al. In-hospital bleeding and utility of a maintenance dose of prasugrel 2.5 mg in high bleeding risk patients with acute coronary syndrome. Circ J 2018; 82: 1874–1883.
• 18.   Nakamura M, Kimura K, Kimura T, Ishihara M, Otsuka F, Kozuma K, et al. JCS 2020 guideline focused update on antithrombotic therapy in patients with coronary artery disease. Circ J 2020; 84: 831–865.
• 19.   Numasawa Y, Kohsaka S, Miyata H, Kawamura A, Noma S, Suzuki M, et al. Impact of body mass index on in-hospital complications in patients undergoing percutaneous coronary intervention in a Japanese real-world multicenter registry. PLoS One 2015; 10: 17.
• 20.   Wang TY, Chen AY, Roe MT, Alexander KP, Newby LK, Smith SC, et al. Comparison of baseline characteristics, treatment patterns, and in-hospital outcomes of Asian versus non-Asian white American’s with non-ST-segment elevation acute coronary syndromes from the CRUSADE quality improvement initiative. Am J Cardiol 2007; 100: 391–396.
• 21.   McDonagh JR, Seth M, LaLonde TA, Khandewal AK, Wohns DH, Dixon SR, et al. Radial PCI and the obesity paradox: Insights from blue cross blue shield of michigan cardiovascular consortium (BMC2). Catheter Cardiovas Interv 2016; 87: 211–219.
• 22.   Veronese N, Li YP, Manson JE, Willett WC, Fontana L, Hu FB. Combined associations of body weight and lifestyle factors with all cause and cause specific mortality in men and women: Prospective cohort study. BMJ 2016; 355: I5855.