Tải bản đầy đủ - 0 (trang)
4 Conflicting Conclusions from Physical Activity Questionnaire, Diary and Physical Fitness Data

4 Conflicting Conclusions from Physical Activity Questionnaire, Diary and Physical Fitness Data

Tải bản đầy đủ - 0trang

10



Objective Monitoring and the Challenge of Defining Dose/Response. . .



303



10.4.1 Proponents of Vigorous Physical Activity

Morris and his associates [8] had 16,682 male executive-class civil servants recollect their leisure activities on the preceding Friday and Saturday. Observers then

noted periods >5 minutes/day that the subjects had ascribed to active recreation,

“keeping fit,” or “vigorous getting about” (activities demanding near maximal

effort), as well as bouts of heavy work lasting >15 minutes (gardening, building

or moving heavy objects) and stair climbing, all regarded as demanding peak

energy outputs >31 kJ/minute. The first 214 subjects who sustained a clinical

attack of ischaemic heart disease were less likely to have recorded such vigorous

activity than their peers. Vigorous physical activity also protected against fatal

heart attacks, but no protection was seen from moderate activity. Those engaging in

much vigorous exercise seemed to gain more benefit (risk ratio 0.18) than those

doing some vigorous exercise (risk ratio 0.42–0.55), although it could be argued

that participation in vigorous physical activity was serving as a marker of some

other favorable personal characteristic.

A second study of male executive civil servants [9] found 9 % of employees

reporting that they had often participated in vigorous sports, undertaken considerable amounts of cycling, and/or rated their walking pace as >6.4 km/hour over the

preceding 4 weeks. This sub-group of employees experienced less than a half as

many non-fatal and fatal heart attacks as their peers, over a 9-year follow-up.

However, no protection was observed unless the sport or exercise was reported as

vigorous.

Paffenbarger and his associates [10] questioned a large sample of male Harvard

alumni on the number of city blocks walked, the number of stairs climbed, and the

intensity and duration of any sport involvement during a typical week. Based on this

information, they estimated gross weekly leisure energy expenditures. Setting the

risk of cardiovascular disease in the least active group as 1.00, benefit was seen

from walking (<5 vs. >15 km/week, relative risk 0.67), stair-climbing ((<20

vs. >55 flights/week, 0.75), sports (none vs. moderate sports play >4.5 METs,

0.63), and a large total energy expenditure (<2 to 8 vs. 8 to >14 MJ/week, 0.70).

The data were interpreted as showing a need to spend more than 8 MJ/week in order

to enhance cardiovascular prognosis, although risk ratios did not differ greatly

between participants with estimated expenditures of 2–4 MJ/week (0.63) and

12–14 MJ/week (0.68). An 8-year follow-up of the same population examined

health outcomes in relation to changes in physical activity. The baseline risk was

set as that found in subjects who failed to meet the 8 MJ/week standard in either

survey. There was an insignificant increase of cardiovascular risk if physical

activity had diminished over the 8 years, but in those whose physical activity had

increased, the cardiac risk diminished to a similar level to that seen in subjects who

had remained active over the survey (risk ratios for moderate sports, 0.71 vs. 0.69,

and for weekly energy expenditures >8 MJ, 0.74 vs. 0.79).

Thus, the studies of Morris pointed the need for periods of activity with an

intensity >31 kJ/minute, and Paffenbarger’s interpretation of his observations



304



Roy J. Shephard



suggested the need to accumulate a gross energy expenditure >8 MJ/week in order

to enhance cardiovascular prognosis.



10.4.2 Other Questionnaire Studies

Blair and colleagues [7] conducted an exhaustive review of published articles

relating to physical activity and/or fitness and health outcomes (morbidity from

coronary heart disease, stroke, cardiovascular disease and cancer and/or mortality

from cardiovascular disease, cancer, or all causes); 67 papers published to August

2000 were given detailed review. Forty-nine articles based upon self-reports

divided subjects into at least three activity categories, and nine studies categorized

subjects in terms of physical fitness as assessed during a treadmill or cycle

ergometer test.

In terms of the questionnaire data, Blair and his associates [7] concluded that

with a few exceptions that included a study of mortality in very old men [11] and

papers on the risks of breast [12] and testicular [13] cancer, most studies showed an

inverse relationship between health risk and estimates of habitual physical activity.

Nevertheless, because of the differing methods of measuring and categorizing

physical activity and differing health outcomes, it was difficult to combine information from the various studies and specify an overall dose-response relationship.

The apparent shape of the dose/response curve varied widely between studies, and

was particularly inconsistent in women.



10.4.3 Critique of Questionnaire Data

The use of questionnaire data to determine dose/response relationships is limited by

differing methodologies, differing health outcomes, and above all by the limited

accuracy of questionnaire assessments of physical activity.



10.4.3.1



Differing Methodologies



Inter-study differences in the methods of measuring and categorizing levels of

physical activity make it impossible to combine data from the various published

sources in order to gain an overall impression of dose/response relationships. Some

investigators have estimated gross energy expenditures per week, dividing their

sample of subjects into three to six categories of habitual physical activity, albeit

with differing cut-points between categories. Others have used questionnaire

responses to separate subjects into three or four arbitrary physical activity categories, with different authors using differing methods of classification. Some have

considered both leisure and occupational activities, and some investigators have



10



Objective Monitoring and the Challenge of Defining Dose/Response. . .



305



focussed their attention upon the reported frequency rather than the intensity of

physical activity.



10.4.3.2



Differing Health Outcomes



Many observers have focussed upon all-cause and/or cardiovascular mortality, but

others have examined such outcomes as cancer and stroke. It seems unlikely that the

shape of the dose/response curve will be identical for differing health outcomes,

although there is at present little information on this question. For some outcomes

such as rectal cancer, there is little relationship between habitual physical activity

and disease incidence or mortality. In the case of fat accumulation and the metabolic syndrome, one might anticipate a relatively linear dose-response relationship,

with fat loss proportional to any increase in the level of daily energy expenditure. In

the case of anxiety and depression, a very modest increase of physical activity

might be sufficient to cause arousal and thus an enhancement of mood state, but in

order to counter sudden cardiac death it might be necessary to condition the body to

the high levels of energy expenditure sometimes required in an emergency.



10.4.4 Lack of Absolute Accuracy of Questionnaires

The most important factor limiting the interpretation of questionnaire data is

probably the limited absolute accuracy of the estimates of habitual physical activity. Although it is generally accepted that questionnaires can make a useful three or

four level categorization of physical activity patterns, it is also widely acknowledged that the absolute estimates of energy expenditure may have a three-fold error

relative to “gold standard” measurements. Moreover, the magnitude of these gross

errors varies between individuals, and it is likely to differ systematically between

those reporting high and low levels of physical activity. Partly as a consequence of

such errors, and partly as a consequence of the limited number of cardiac incidents

when quite a large sample of subjects are followed for 5–10 years, the confidence

limits to the estimated relative risks associated with various levels of physical

activity are very broad, and this makes it difficult to determine the shape of the

dose/response relationship. This is well illustrated by the data of Haapanen and

colleagues [14, 15], who examined the risks of coronary heart disease over a

10-year follow-up of 1340 men and 1500 women aged 35–63 years (Table 10.2).



10.4.5 Aerobic Fitness Dose/Response Data

Blair et al. [7] identified nine studies that related simple indices of aerobic fitness

(treadmill endurance times or heart rates at specific exercise intensities) to health



306



Roy J. Shephard



Table 10.2 Relative risk of

coronary heart disease over

10-year follow-up of middleaged adults, showing broad

confidence intervals [15]



Activity measure

Relative risk

Men

Total energy expenditure

High

1.00

Moderate

1.33

Low

1.98

Intensity

Vigorous >1/week

1.00

Vigorous <1/week

1.42

Women

Total energy expenditure

High

1.00

Moderate

0.73

Low

1.25

Intensity

Vigorous >1/week

1.00

Vigorous <1/week

1.13



Confidence intervals



0.78–2.27

1.22–3.23



0.92–2.17



0.38–1.39

0.72–2.15



0.62–2.07



outcomes; five of the nine reports were drawn from the Aerobics Center in Dallas,

TX, where he was then working. All of these studies showed cardiovascular risk

and/or all-cause mortality decreasing with increasing aerobic fitness, commonly as

much as three to four fold, [16]. The overall gradient was steeper for fitness than for

questionnaire data, probably reflecting the greater accuracy in classification of the

subjects when using aerobic fitness data.

Most analyses divided subjects into fitness tertiles, quartiles or quintiles. The

trend seemed to demonstrate the largest health benefit on moving from the least fit

to the next fitness quartile, but despite very large sample sizes, the confidence limits

to estimates of risk ratios and thus the shape of the dose/response relationship were

again very broad, as exemplified by the studies of Ekelund et al. on 3106 men [16]

and Blair et al. on 10,224 men and 3120 women [17] (Table 10.3).



10.5



New Dose/Response Information from Objective

Monitoring



Objective monitoring in principle provides a much finer gradation of physical

activity patterns than occupational classifications, physical activity questionnaires

or even simple measures of physical fitness. Does this imply the likelihood of

obtaining further detail on the shape of dose/response relationships between physical activity and health outcomes? Critical issues are sample size, range of physical

activity within the chosen sample, ability to measure the most relevant components

of overall activity, the nature of the outcome measures, and overall confidence

levels.



10



Objective Monitoring and the Challenge of Defining Dose/Response. . .



Table 10.3 Dose/response

relationships between

estimates of aerobic fitness

quartile and cardiovascular

[16] or all-cause mortality

[17]



307



Fitness measure

Risk ratio

Confidence limits

Heart rate in submaximal treadmill test (all men)

Q1 (least fit)

2.21

1.65–3.25

Q2

1.56

0.68–2.44

Q3

1.30

0.49–2.11

Q4 (most fit)

0.26

0.00–0.62

Treadmill endurance time

Men

Q1 (least fit)

3.44

2.05–5.77

Q2

1.37

0.76–2.50

Q3

1.46

0.81–2.63

Q4

1.17

0.63–2.17

Q5 (most fit)

1.00

Women

Q1 (least fit)

4.65

2.22–9.75

Q2

2.42

1.09–5.37

Q3

1.43

0.60–3.44

Q4

0.76

0.27–2.11

Q5 (most fit)

1.00



10.5.1 Sample Size and Range of Physical Activity

Because of the tedious work involved in checking pedometer and accelerometer

records for the presence of various types of artifact, the number of subjects recruited

for objective studies of physical activity and health has to date been relatively

small, particularly when compared with the occupational questionnaire and fitnessbased surveys noted above. Even with a 5–10 year period of observation, the

number of cardiac or all-cause deaths has thus been too few to permit the calculation of accurate statistics, and it has been necessary to consider the physical activity

data in relation to surrogate measures of health such as arterial stiffness or the

ultrasonic index of bones.

A related issue, applicable to all methods of classifying physical activity, is to

ensure that there are an adequate number of active individuals within the sample.

This is becoming progressively more difficult in sedentary urban populations. In a

study of 57 eighty-year-old women, Gerdhem and associates [18] found no significant correlations between accelerometry measurements of physical activity and

balance, muscle strength or bone density. However, there was little chance that a

significant relationship would be seen, since only 8 of the 57 subjects were

engaging in moderate or vigorous physical activity. This negative conclusion was

quickly reversed by recruiting a larger population of 152 men and 206 women aged

50–80 who were followed for 10 years; in the second study, annual bone loss was

0.6 % less in those who were classified as active relative to those who were inactive

[19]. The larger group also showed benefits in terms of balance, although habitual

physical activity still showed no relationship with muscle strength or gait velocity.



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

4 Conflicting Conclusions from Physical Activity Questionnaire, Diary and Physical Fitness Data

Tải bản đầy đủ ngay(0 tr)

×