Page 97 - 南京医科大学自然版

第46卷第8期 陈芳怡，王 威，胡馨茹，等. 心理韧性与中国老年人全因死亡率关联强度的异质性：一项前瞻性队列
2026年8月 研究［J］. 南京医科大学学报（自然科学版），2026，46（8）：1197-1210 ·1203 ·

（续表1）
Psychological resilience Standardized
Variable All（n=7 481） P
Low（n=3 214） High（n=4267） difference
BMI（kg/m ，x ± s） 21.05 ± 4.370 20.64 ± 4.530 21.37 ± 4.220 <0.001 0.167
2
Psychological resilience（x ± s） 13.99 ± 3.020 11.17 ± 1.820 16.11 ± 1.740 <0.001 2.777
Activity score［M（P25，P75）］ 0 010.0（5.0，15.0） 0 008.0（4.0，12.0） 0 012.0（8.0，16.0） <0.001 0.564
DD score［M（P25，P75）］ 0 004.0（3.0，6.0） 0 004.0（2.0，5.0） 0 005.0（3.0，6.0） <0.001 0.454
BADL score（x ± s） 17.23 ± 1.980 16.85 ± 2.450 17.52 ± 1.470 <0.001 0.332
Follow⁃up［d，M（P25，P75）］ 1 408.0（858.0，2 482.0） 1 142.0（696.5，2 236.0） 1 672.0（945.0，2 514.0） <0.001 0.237
*Category 1：staff of farming/forestry/animal husbandry/side ⁃ line production/fishery；Category 2：workers/teachers/doctors/military/government
workers/businessmen；Category 3：other. BADL score：base activity of daily living score；BMI：body mass index；DD score：diversity of dietary score；
MCC：multiple chronic conditions.

ience group（all P < 0.05，Figure 4）. 2.50
（95% CI）
2.3 STEPP analysis based on obesity indices 2.25
2.00
ABSI，BRI，WHtR，and BMI were used as covari⁃ 1.75
HR of mortality 0.75
ates to split subpopulations（Table 2）. For ABSI，the 1.50
1.25
HR for high versus low psychological resilience 1.00
showed a clear downward trend as the median ABSI in⁃ 0.50
2
creased，with statistically significant heterogeneity 0.25 χ =238.7 P < 0.001
0.00
among subpopulations（P < 0.001）. From subpopula⁃
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
tions 5 to 8，the HRs were the lowest and their changes Psychological resilience
图 2 基于限制性立方样条分析心理韧性与全因死亡风险
with increasing median ABSI were minimal（Figure
之间的非线性关系
5A）. For BRI，the HR for high versus low psychologi⁃
Figure 2 Non ⁃ linear association between psychological
cal resilience generally showed a stable⁃decreasing⁃in⁃
resilience and all⁃cause mortality risk based on
creasing trend with increasing median BRI，and the
restricted cubic spline analysis
heterogeneity among subpopulations was statistically
significant（P < 0.001）. The risk of all⁃cause mortality generally consistent with that of BRI. The lowest HR
was lowest in subpopulation 5（HR=0.87，95% CI：0.78- was also observed in subpopulation 5 at a median
0.97，Figure 5B）. For WHtR，the overall HR trend was WHtR of 0.53，representing the population that benefit⁃

A Psychological resilience High Low B
60.00 P < 0.001 1.00
P=0.032 P=0.038
（%） 50.00 of （95%CI） 0.90
Cumulative survival rate 30.00 Hazard ratio all⁃cause mortality 0.80
40.00

20.00
0.70
10.00
0.00 0.60
Crude PSM IPW cHR aHR HR_PSM HR_IPW
A：Comparison of all⁃cause mortality rates between high and low resilience groups in crude，PSM⁃adjusted，and IPW⁃adjusted samples. B：Hazard
ratios for the association between psychological resilience and all⁃cause mortality based on univariable，multivariable⁃adjusted，PSM⁃adjusted，and IPW⁃
adjusted Cox models.
图3 比较全因死亡率及心理韧性与全因死亡率之间的关联
Figure 3 Comparison of all⁃cause mortality rate and association between psychological resilience and all⁃cause mortality