Page 97 - 《运动与健康科学》(英文)2024年第2期

TaggedAPTARAEndPA and mortality in type 2 diabetes 219
evidence has low certainty according to Grading of Recom- Kadoorie Biobank, whereas the doseresponse pattern was
mendations, Assessment, Development and Evaluation initially modest and shallow in the UK Biobank. Conversely,
(GRADE) criteria. 12 A very strong curvilinear doseresponse the doseresponse association with cardiovascular mortality
association was observed in a study of 3000 British adults with was noticeably steeper in British than in Chinese adults. A
type 2 diabetes. Compared with no activity, activity below the maximal achievable all-cause mortality reduction of 23% was
recommended level was associated with 26% lower mortality, observed at approximately 35 MET-h/week in Chinese adults,
and meeting recommendations was associated with 35% lower but no upper level of benefit was observed for other outcomes
mortality. 27 Similarly, in an East Asian cohort, PA below and in either cohort.TaggedAPTARAEnd
at the recommended level was associated with 32% and 37% TaggedAPTARAPSeveral factors could explain these differences. The large
28
lower mortality, respectively. In both studies, a similar curvi- number of participants with undiagnosed type 2 diabetes from
linear doseresponse pattern was observed for cardiovascular China reflects different health-care systems, which could also
mortality. This pattern and effect sizes are consistent with data impact detection and coding of cardiovascular outcomes. Simi-
from the general population. 29,30 In our study, PA below the larly, the difference in the use of cardiovascular prophylactics
recommended level had a much weaker association with is reflected in the CVD incidence rates. The UK Biobank and
mortality outcomes. The all-cause mortality risk reduction was China Kadoorie Biobank cohorts are subject to different selec-
6% in the UK Biobank and 13% in the China Kadoorie tion-mechanisms, with a particularly high risk of healthy-
35
Biobank, but cardiovascular mortality was 6% higher in the volunteer bias in the UK Biobank. In high-income Western
China Kadoorie Biobank. PA below recommendations was countries, BMI is strongly linked to a suite of lifestyle risk
36
associated with 20% lower cardiovascular mortality in the UK factors and with poorer socioeconomic circumstances. These
Biobank, but there was no support of a curvilinear pervasive sources of confounding are expected to amplify
doseresponse association. Compared with our study, the East associations between PA and mortality, and residual
Asian cohort was based on long-term average PA from several confounding may remain despite careful statistical modeling. 16
repeated assessments, demonstrating that sustained engage- In contrast, there was no gradient in BMI across leisure-time
ment in PA is needed to maximize health benefits. The study PA in the China Kadoorie Biobank, implying that BMI and its
based in the UK was much smaller in comparison to ours associated network of socioeconomic factors are not a source
and included a less comprehensive approach to minimize of confounding in this context. The minimal attenuation of
confounding. Collectively, our data corroborate previous effect sizes after adjustment for BMI in the China Kadoorie
evidence that adults with type 2 diabetes could achieve impor- Biobank confirms this notion and suggests the association
tant reductions in total and cardiovascular mortality from between PA and all-cause and cardiovascular mortality is inde-
leisure-time PA but suggest that larger amounts than currently pendent of BMI. Finally, adults of Asian origin may have a
distinct pathophysiology and develop type 2 diabetes at a
recommended are needed to achieve those benefits.TaggedAPTARAEnd
TaggedAPTARAPPA, including exercise, improves conventional cardiovas- lower BMI compared with adults of European origin. 37 Our
cular risk markers, including HbA1c, in adults with type 2 data thus provide much-needed representation of the majority
5
diabetes. Effects are dose-dependent with a mean change in of adults living with type 2 diabetes worldwide.TaggedAPTARAEnd
HbA1c of 0.36% with structured weekly exercise of 150 min
or less and 0.89% with more than 150 min per week. 31
TaggedAPTARAH24.3. Strengths and weaknessesTaggedAPTARAEnd
However, translating these effects into hard end-points, such
as mortality and cardiovascular events, have been far less TaggedAPTARAPWe included large population-based samples, and we there-
convincing. The LOOK AHEAD trial compared a weight-loss fore expect participants to represent the level of variation in
and PA-based lifestyle intervention (aiming for 175 min of social conditions and medical treatment given to the majority
moderate-intensity activities per week) with usual care but found of adults with type 2 diabetes in the UK and China. Our statis-
no reduction in cardiovascular morbidity or mortality. 32 In the tical model was informed by a directed acyclic graph to reduce
Nurses’ Health Study and the Health Professionals Follow-up the impact of confounding, and we carefully removed partici-
Study, PA was not associated with non-fatal stroke, coronary pants with mobility limitations or a history of major conditions
heart disease, or myocardial infarction 33 andmore than4 hof other than type 2 diabetes in addition to left-censoring the first
exercise per week was needed to lower a composite of fatal and 3 years of follow-up to reduce confounding from poor health.
non-fatal CVD. 33,34 These results are consistent with our findings Corroborative analysis in patients with type 2 diabetes and a
and suggest PA does not lower rates of non-fatal CVD in adults history of CVD provides wider generalizability to a common
with type 2 diabetes but may instead prevent cardiovascular high-risk group. We highlight the following limitations: (a)
Despite our best efforts to standardize data analysis between
mortality by increasing survival with CVD.TaggedAPTARAEnd
the 2 studies, the leisure-time PA questionnaires were not iden-
tical. Specifically, only the main type of activity was reported
TaggedAPTARAH24.2. Cross-country comparisonTaggedAPTARAEnd
in the China Kadoorie Biobank, which could underestimate
TaggedAPTARAPWe were unable to determine a consistent minimal effective absolute levels of PA in individuals performing multiple activi-
dose or a maximal achievable risk reduction across both ties and, thereby, influence the doseresponse pattern. Addi-
cohorts and mortality outcomes. Any non-zero level of PA tionally, we included light do-it-yourself activities (assigning
was associated with lower all-cause mortality in the China these as 2.25 METs) from the UK Biobank questionnaire