Page 93 - 《运动与健康科学》(英文)2024年第2期

TaggedAPTARAEndPA and mortality in type 2 diabetes 215
CVD, cancer, or other major comorbidities at study baseline recommendations were 0.80 (95%CI: 0.581.11) and 0.69
were included. During a median follow-up of 12.4 and (95%CI: 0.480.97) in the UK Biobank and 1.06 (95%CI:
9.7 years after baseline, 1571 and 2351 deaths (392 and 1060 0.761.47) and 0.79 (95%CI: 0.690.92) in the China
deaths from CVD) and 2345 and 4458 major adverse cardio- Kadoorie Biobank. PA at recommended levels was not associ-
vascular events were included from the UK Biobank and ated with cardiovascular mortality in the China Kadoorie
China Kadoorie Biobank, respectively. The age of participants Biobank. There was no association between leisure-time PA
was similar, at 59.5 § 7.2 years and 57.8 § 9.7 years (mean § below or at recommended levels and risk of major adverse
SD), but the distributions of sex, BMI, leisure-time PA, cardiovascular events in either cohort (Table 2). Doing more
diabetes duration, use of preventive medications, and the than 35 MET-h/week yielded HRs from 0.87 (95%CI:
proportion of undiagnosed diabetes were different. More than 0.770.99) to 0.82 (95%CI: 0.710.96) in the UK Biobank.TaggedAPTARAEnd
half of Chinese participants did no PA in their leisure-time
while this was the case for only 9% of British participants. The
TaggedAPTARAH23.3. Stratified and sensitivity analysesTaggedAPTARAEnd
mean differences in BMI between participants with no PA and
2
those exceeding recommendations was 3.4 kg/m in the UK TaggedAPTARAPAdditional adjustment for pharmacological treatment
2
Biobank but only 0.2 kg/m in the China Kadoorie Biobank. slightly attenuated associations in the UK Biobank (Model 4).
More physically active participants had completed more Stratified associations with all-cause mortality are shown in
formal schooling in both cohorts. Descriptive characteristics Fig. 2. Age modified the association in the China Kadoorie
across categories of MET-h/week are presented in Table 1 and Biobank, with an HR for PA exceeding recommendations of
in greater detail in Supplementary Tables 4 and 5 of Supple- 0.95 (95%CI: 0.781.14) among adults <60 years old and
mentary File 1. Distributions of leisure-time PA among cases 0.72 (95%CI: 0.640.81) among those 60 years old (p for
and in all participants are shown in Supplementary Fig. 2 of interaction < 0.001). There was no evidence of effect modifi-
cation by pre-existence of cardiovascular morbidity (effect
Supplementary File 1.TaggedAPTARAEnd
modification p values 0.97 and 0.79).TaggedAPTARAEnd
TaggedAPTARAPThe pattern of results did not change with exclusion of ever
TaggedAPTARAH23.1. Leisure-time PA and all-cause mortalityTaggedAPTARAEnd
smokers and exclusion of individuals with low certainty of
TaggedAPTARAPHigher levels of leisure-time PA were associated with lower type 2 diabetes (Supplementary Table 7 of Supplementary File 1).
all-cause mortality in both cohorts (Table 2). Statistical adjust- There was no association between leisure-time PA and
ment for BMI attenuated associations in the UK Biobank but mortality in 5857 British adults who performed at least a
did not impact effect sizes in the China Kadoorie Biobank. In single 24-h dietary recall. Descriptive characteristics of adults
the UK Biobank, the slope of the doseresponse association with active transportation or occupations are shown in Supple-
was shallow below 15 MET-h/week, accelerated thereafter, mentary Tables 8 and 9 of Supplementary File 1; their associa-
and reached statistical significance at 35 MET-h/week, which tions with outcomes are presented in Supplementary Tables 10
is equivalent to 90 min of walking or 38 min of strenuous
and 11 of Supplementary File 1.TaggedAPTARAEnd
sports per day (Fig. 1, p for non-linearity = 0.51). No upper
level of additional risk reduction was observed within the
TaggedAPTARAH14. DiscussionTaggedAPTARAEnd
exposure distribution. In categorical analyses, PA below
recommendations, compared with no activity, was associated TaggedAPTARAPThe main finding was that PA below and at contemporary
with a slight and uncertain reduction in all-cause mortality, recommendations was associated with lower all-cause and
with an HR of 0.94 (95%CI: 0.791.12). The doseresponse cardiovascular mortality in British and Chinese adults with
association in China Kadoorie Biobank suggested a curvilinear type 2 diabetes, but these reductions were uncertain and incon-
relationship (p = 0.03), with lower mortality for any non-zero sistent across cohorts. There was no association between PA
level of leisure-time PA and no additional risk reduction above and risk of major adverse cardiovascular events in the China
35 MET-h/week. The HR for PA below recommendations was Kadoorie Biobank, and activity equivalent to 90 min of
0.87 (95%CI: 0.681.10). The absolute differences in 10-year walking or 38 min of strenuous sports per day, which far
cumulative mortality, compared with no activity, were 0.2%, exceeds the WHO recommended level, was needed to lower
0.4%, and 0.6% in the UK Biobank and 1.4%, 1.3%, the risk in the UK Biobank.TaggedAPTARAEnd
and 1.6% in the China Kadoorie Biobank for PA below, at,
and exceeding recommendations (Supplementary Table 6 of TaggedAPTARAH24.1. Comparison with other studiesTaggedAPTARAEnd
Supplementary File 1).TaggedAPTARAEnd
TaggedAPTARAPPrevious meta-analyses of the doseresponse association
between PA and mortality in adults with type 2 diabetes have
TaggedAPTARAH23.2. Leisure-time PA and cardiovascular mortality and major
been inconclusive. One suggested a weak, linear
adverse cardiovascular eventsTaggedAPTARAEnd 12
doseresponse association while the other provided some
TaggedAPTARAPContinuous doseresponse curves were supportive of a support for a curvilinear pattern, with a steeper gradient at low
linear relationship with cardiovascular mortality in both activity and diminishing returns at higher levels of activity. 14
cohorts (p 0.67) with a steeper slope in the UK Biobank. None of these meta-analyses provided estimates directly appli-
There was no upper threshold of additional risk reduction in cable to the WHO’s quantitative recommendations of
5
either cohort. The HRs for PA below and above 150300 min of moderate-to-vigorous PA/week, and the