Characteristics of the study patients
Table 1 shows the clinical characteristics of the 360 patients included in the study. The mean age of the patients was 62 ± 13 years. 88.9% of the patients were male. The mean GNRI was 97.9 ± 8.3. The medical history of the patients showed that 34.7% of them were diagnosed with diabetes mellitus and 67.5% with hypertension.
Table 1
Clinical characteristics(n = 360)
|
Mean ± SD
|
|
or median [interquartile range]
|
Age, years
|
62 ± 13
|
Sex, male (%)
|
320 (88.9)
|
BMI, kg/m2
|
25.1 ± 4.2
|
Hb, g/dL
|
13.6 ± 2.1
|
Alb, g/dL
|
3.8 ± 0.5
|
Cr, mg/dL
|
1.4 ± 2.3
|
eGFR, mL/min/1.73m2
|
70.0 ± 27.2
|
UA, mg/dL
|
5.9 ± 1.6
|
Glucose, mg/dL
|
168.8 ± 77.2
|
HbA1c (%)
|
6.3 ± 1.2
|
TG, mg/dL
|
109.6 ± 107.9
|
HDL-C, mg/dL
|
49.2 ± 13.3
|
LDL-C, mg/dL
|
121.5 ± 37.3
|
BNP, pg/mL
|
46.1 [14.4–142.1]
|
GNRI
|
97.9 ± 8.3
|
Disease
|
Number (%)
|
Old myocardial infarction (%)
|
31 (8.6)
|
Coronary spastic angina (%)
|
14 (3.9)
|
Diabetes mellitus (%)
|
125 (34.7)
|
Hypertension (%)
|
243 (67.5)
|
Dyslipidemia (%)
|
261 (72.5)
|
Hemodialysis (%)
|
18 (5.0)
|
BMI: body mass index, Hb: hemoglobin, Alb: albumin, Cr: creatinine, eGFR: estimated glomerular filtration rate, UA: uric acid, HbA1c: hemoglobinA1c, TG: triglyceride, HDL-C: high-density lipoprotein cholesterol, LDL-C: low-density lipoprotein cholesterol, BNP: B-type natriuretic peptide, GNRI: Geriatric Nutrition Risk Index. |
Cumulative Incidence Of All-cause Death And MACE
During the observation periods, 23 deaths, 3 non-fatal myocardial infarctions, 11 hospitalizations for heart failure, and 2 strokes occurred. The cumulative incidence of all-cause death was 7.6% (Fig. 1A). The cumulative incidence of MACE was 11.6% (Fig. 1B).
Results Of Time-dependent ROC Analysis
The time-dependent ROC curve analysis for all-cause death from 3 to 48 months are shown in Fig. 2. The optimal cut-off value of GNRI in all-cause death was 82.7 in the short-term of 3 months, which was the lowest value among all the observation periods. Thereafter, the optimal cut-off value of GNRI remained constant at 89.3 from 6 to 36 months and slightly increased to 90.3 at 48 months.
The time-dependent ROC curve for MACE from 3 to 48 months are shown in Fig. 3. The optimal cut-off value of GNRI was 83.0 in the short-term of 3 months, which was the lowest value among all the observation periods. The optimal cut-off value of GNRI was almost constant at 95 after 6 months. The optimal cut-off values of GNRI for two endpoints were relatively low in the short-term but high in the long-term. The optimal cut-off values of GNRI for all-cause death tended to be lower than those for MACE during the observation periods.
Changes over time (months) in AUC by time-dependent ROC analysis is shown in Fig. 4. The AUC for all-cause death by time-dependent ROC analysis were 0.834, 0.838, 0.861, 0.876, 0.852, and 0.854 at 3, 6, 12, 24, 36, and 48 months, respectively. The AUC for MACE by time-dependent ROC analysis were 0.841, 0.853, 0.863, 0.844, 0.816, and 0.821 at 3, 6, 12, 24, 36, and 48 months, respectively. The AUC by time-dependent ROC analysis for both the end points remained relatively high during the observation periods.
Clinical Outcomes With The Optimal Cut-off Values Of GNRI
The incidence of all-cause death between the two groups, each divided by the optimal cut-off value of GNRI, at 3, 6, 12, 24, 36, and 48 months after hospitalization for AMI is shown by Kaplan–Meier curves (Fig. 5). The incidence of all-cause death in the group below the optimal cut-off value of GNRI was significantly higher than in the group above the optimal cut-off value of GNRI during the observation periods (P < 0.001).
The incidence of MACE between the two groups divided by the optimal cut-off value of GNRI was also shown by Kaplan–Meier curves (Fig. 6). The incidence of MACE in the group below the optimal cut-off value of GNRI was significantly higher than in the group above the optimal cut-off value of GNRI during the observation periods (P < 0.001).