Two Physical Activity Measures: Paffenbarger Physical Activity Questionnaire Versus Aerobics Center Longitudinal Study as Predictors of Adult-Onset Type 2 Diabetes in a Follow-Up Study

Abstract

Purpose:

To compare 2 self-report methods of measuring weekly minutes of physical activity based on the Aerobics Center Longitudinal Study (ACLS) questionnaire and question 6 of the Paffenbarger Physical Activity Questionnaire (PPAQ) to determine the better predictor of adult-onset type 2 diabetes mellitus (T2DM).

Design:

An observational, prospective study.

Setting:

Survey data from the Adventist Health Study-2 (AHS-2) collected between 2002 and 2006 (baseline) and the Psychosocial Manifestations of Religion Sub-Study (PsyMRS), an AHS-2 subset collected 1 to 4 years later.

Patients:

Nine thousand eight hundred seventy-three male and female participants aged 23 to 106 years (mean, 63 years). Three hundred eighty participants reported adult-onset T2DM at follow-up.

Measures:

Question 6 from the PPAQ and a question adopted from the ACLS were assessed at baseline. Incident diabetes was defined as participants who reported receiving treatment for adult-onset T2DM in the last 12 months in the PsyMRS and not at baseline.

Analysis:

Multivariate logistic regression analyses controlled for age, gender, ethnicity, education, body mass index (BMI), diet, and sedentary activity. Each exposure variable was compared to nonexercisers.

Results:

The PPAQ (odds ratio [OR]: 0.998; 95% confidence interval [CI]: 0.997-1.000) and the ACLS (OR: 0.999; 95% CI: 0.998-1.001) exhibited similar likelihood of predicting incident adult-onset T2DM in a healthy, mixed-gender population when controlling for several confounders.

Conclusions:

The demonstrative nomenclature of the PPAQ may be more effectual in capturing physically active individuals than the ACLS notwithstanding generalizability and response bias limitations.

Keywords

diabetes mellitus jogging Paffenbarger Physical Activity Questionnaire running vigorous activity walking

Purpose

A dose relation exists between physical activity and chronic disease reduction, leading to the development of a national guideline endorsed by the American College of Sports Medicine and the American Heart Association for healthy adults to accumulate at least 20 minutes of moderate physical activity 5 days a week or 30 minutes of vigorous activity 3 days a week to promote and maintain health.¹ Moderate activity corresponds to a metabolic equivalent task (MET) score of 3 to 6. Vigorous activity is defined as jogging or an activity that causes a rapid increase in breathing and a significant increase in heart rate,¹ requiring an MET score of above 6. The physiological adaptations that occur during and after physical activity may explain some of the benefits.^2,3

During physical activity, insulin-driven glucose transporters mobilize glucose, thus upregulating glycogen synthase required for glycogen synthesis.⁴ An increase in muscle mass results in growth of insulin-sensitive muscle fibers and capillary proliferation essential for upholding insulin sensitivity.⁵ Physiological effects of physical activity may hasten or prevent the onset of adult-onset type 2 diabetes mellitus (T2DM), substantiating the value of measuring the physical activity and T2DM relationship. Epidemiologists who study diabetes risk in large sample populations may prefer the inherent cost-efficiency of previously validated physical activity self-report questionnaires to identify strategies that significantly modify diabetes risk.

The favored forms of physical activity in the US population are walking, running, and jogging.⁶ The association between physical activity and incident diabetes was measured in runners (n = 33 060) and walkers (n = 15 945) over 6.2 years. Activity was measured in MET hours/day (METh/d) and compared with incident diabetes that was identified by self-report and physician diagnosis.⁷ Running decreased diabetes risk by 12.1%. Furthermore, moderate amounts of walking are associated with protection against diabetes compared to minimal amounts. In a meta-analysis, the relative risk of diabetes incidence was 0.70 (0.58-0.84) for regular walking (typically 2.5 h/wk of brisk walking) compared with almost no walking.⁸

The protective effects from physical activity were not significantly different for runners and walkers (12.1% and 12.3% per METh/d).⁸ In a cohort study of 70 102 surveyed healthy women,⁹ 8 common physical activities including walking were assessed using equivalent energy expenditures measured in MET score duration per week. Walking and vigorous physical activity had similar magnitudes in risk reduction.⁹

Self-reported moderate-to-vigorous physical activity may be associated with a lower risk of diabetes.^7,10
–12 The relationship between vigorous activity and diabetes appears to be independent of body mass index (BMI) in postmenopausal women¹³ as well as in normal-weight and overweight men.¹⁴ Risk reduction was greatest in those with hypertension, a positive family history of T2DM, and high BMI categories.¹⁵ Obese men had to increase their frequency of vigorous physical activity at least 5 times a week to reduce diabetes risk.¹⁴ Risk of diabetes was lowered in black women who participated in brisk walking for at least 5 hours per week.¹⁶ Native American, Japanese American, and Caucasian men and women who engaged in strenuous sport activity significantly reduced their diabetes risk.¹⁷ Vigorous level of activity was protective for males across the ethnic groups, and brisk walking was protective for Caucasian men.¹⁷ The protective effect of physical activity was removed for black women who spent 5 or more hours watching TV per week.¹⁶

There may be a threshold above which further walking does not reduce risk. In postmenopausal women, walking 10 to 15 METh/week, or up to 2 to 3 hours of weekly walking, was protective without further reduction in disease risk above these thresholds.¹⁸ On the other hand, analysis of postmenopausal multiethnic cohorts revealed an inverse, dose–response relationship for walking and vigorous physical activity in Caucasian women.¹⁸

The benefits of vigorous activity have been demonstrated in healthy males of normal-weight,¹⁹, overweight, and obese men.¹⁴ In the Physician’s Health Study of nearly 21 000 men,¹⁴ weekly vigorous activity 1 to 3 times a month or at least twice a week reduced diabetes risk up to 43% in normal-weight men. Overweight men benefitted from any form of vigorous activity frequency above weekly, with a linear trend of greater protection and increased weekly minutes of vigorous activity. Obese men were not protected against diabetes until nearly daily bouts of vigorous activity were reached. The incidence of diabetes decreased 41% in high-risk men who increased their activity level from the lowest quartile to the highest (<500 to >2000 kcal/wk).¹⁵

In epidemiological surveys, it is important to identify previously validated self-report questions regarding physical activity that most efficiently predict incident diabetes. As walking, jogging, or running activities are common and relate to moderate and vigorous intensity, questions regarding such activity may be appropriate survey questions. The purpose of the current study is to determine the better physical activity measure for capturing physical activity status in those identified with incident adult-onset T2DM. Physical activity was assessed by 2 previously validated physical activity measures: firstly, a question from the Aerobics Center Longitudinal Study (ACLS)²⁰ that measured walking, jogging, or running activity,^21,22 and secondly, a question assessing the frequency of engagement in vigorous activities such as brisk walking, jogging, and bicycling for periods of time long enough to work up a sweat, increase heart rate, or cause one to become out of breath which was adopted from question 6 of the Paffenbarger Physical Activity Questionnaire (PPAQ).²³

Methods

Design

The current investigation is an observational, prospective follow-up study that accessed archival data from the Adventist Health Study-2 (AHS-2)²⁴ collected between 2002 and 2006 and the Psychosocial Manifestations of Religion Sub-Study (PsyMRS), a substudy of the AHS-2 that is part of the Biopsychosocial Religion and Health Study (BRHS),²⁵ collected 1 to 4 years later. The AHS-2 database included approximately 97 000 Seventh-day Adventist (SDA) church members living in the United States and Canada who were 30 years of age and older and fluent in English, whereas the PsyMRS is a subpopulation of over 11 000 people who also completed the AHS-2 questionnaire at time 1.²⁵

Sample

The majority of SDAs follow a health-conscious lifestyle including abstinence from alcohol, drugs, and tobacco as well as reduction in dietary animal meat and caffeine intake.²⁴ Participants were recruited by targeting 2 major English-speaking congregation groups in SDA churches throughout the United States and Canada. The groups were comprised of 1000 predominantly black congregations, and 3500 other congregations made up of mostly Caucasian members. Females made up 65.1% of the sample.²⁵

The BRHS consisted of 2 cohorts: the PsyMRS (n = 10 988) that measured the psychosocial components of religious involvement as well as cumulative risk exposure (traumatic life events and other sources of stress) and the Biological Manifestations of Religion (Bio-MRS) substudy (n = 508) that collected additional biological and cognitive measures in a clinic visit. The sampling strategy of the original researchers of the AHS-2 database was to create 2 substudy populations from the original AHS-2 database (PsyMRS and Bio-MRS; collectively referred to as the BRHS) to contain ease of data collection and cost of continued surveying over time.²⁵

Each study was accepted by the institutional review board of Loma Linda University. A consent agreement was on the first page of each study questionnaire that was mailed to study participants. The consent stated, “Your completion of this survey and returning it in the enclosed self-addressed postage-paid envelope tells us that you’ve agreed to participate in the study.” Incident adult-onset T2DM cases were determined from self-reported diabetes diagnoses in the PsyMRS 2006 to 2007 questionnaire. The PsyMRS subsample was composed of 62.7% females and 32.8% males. With regard to ethnicity, 59.5% of the population was self-defined as white and 34.2% as black.

Measures

Both physical activity measures (the ACLS and PPAQ) were part of the PsyMRS questionnaire that was given at follow-up to a subpopulation of over 11 000 people who were also queried using the AHS-2 questionnaire at baseline.²⁵ New cases of adult-onset T2DM were defined as occurring among 380 participants that reported affirmatively to 2 concomitant dichotomous questions in the PsyMRS questionnaire (time 2) that queried whether the participant had ever been diagnosed with adult-onset T2DM and whether the participant had received treatment for the condition in the last 12 months. The respondent also had to respond negatively to the same questions in the AHS-2 data set (were not diagnosed and treated for adult-onset T2DM) at time 1. After adjusting for missing data, there were 380 new cases of adult-onset T2DM in the PsyMRS database and 9493 who did not report adult-onset T2DM in AHS-2 or PsyMRS. We adjusted for a number of control variables, including gender, ethnicity, age, BMI, level of education, hours of television watching to represent time spent in sedentary activities, and diet status.

The ACLS variable was generated from 3 questions from the ACLS.²⁰ A question that stated: “Do you walk, run, or jog as part of a physical activity program (include these same activities when they are performed on exercise machines)?” was used to capture those who participated in such activity. The questions “How many of these ‘walk’ or ‘run’ or ‘jog’ workouts do you usually do per week?”, and “What is your average time spent in each ‘walk’ or ‘run’ or ‘jog’ exercise session (excluding rest stages)?” were used to measure frequency and duration, respectively, of those who participated in walking, jogging, and running activities.

The PPAQ measure was constructed from 2 questions adopted from the PPAQ²³ that stated: “How many times a week do you usually engage in regular vigorous activities such as brisk walking, jogging, bicycling, etc.: long enough, or with enough intensity to work up a sweat, get your heart thumping, or get out of breath?” and “On average, how many minutes do you exercise each session?” The original Paffenbarger questions were combined into a single measure of frequency and duration of vigorous physical activity.

The possible responses of physical activity duration were none, 10 minutes or less, 11 to 20 minutes, 21 to 30 minutes, 31 to 40 minutes, 41 to 50 minutes, 51 to 60 minutes, and more than 1 hour. The possible responses for frequency of physical activity were less than once per week, 1 time per week, 2 times per week, 3 times per week, 4 times per week, 5 times per week, and 6 times per week. “Never engage in activities this vigorous” and “none” were exclusive responses for vigorous physical activity frequency and duration, respectively.

Analysis

Descriptive values were reported as mean (standard deviation) and frequency (percentage). Analysis of variance and χ² tests were utilized to test for between-group differences for continuous and categorical data, respectively. The criterion for statistical significance was set at P < .05. Ethnicity was delineated into 2 separate groups: blacks versus whites and other ethnicities versus whites. Diet was categorized into 5 groups: nonvegetarian, vegan, pesco-vegetarian, lacto-vegetarian, and semi-vegetarian. Education level served as a marker of socioeconomic status and was divided into grade school, high school, some college, associate’s degree, bachelor’s degree, master’s degree, and doctoral degree. Hours of watching television represented sedentary behavior and consisted of initially 6 categories: no TV, less than 1 hour per day, 1 hour per day, 2 hours per day, 3 to 4 hours per day, and more than 5 hours per day. The first 2 categories, “no TV” and “less than 1 hour per day,” were collapsed into 1 category, “less than 1 hour per day.” Each category of daily hours of television watching was compared to the “less than 1 hour per day” category.

Each of the physical activity measures, the ACLS measure and the PPAQ measure, were further transformed and recoded into a continuous variable of minutes of physical activity (duration) per week (frequency). The midpoint of each duration category was computed and transformed into a new duration variable. Each number was further multiplied by the midpoint of each frequency value of each physical activity category and coded to create 2 new continuous variables: physical activity minutes per week based on the ACLS questionnaire and physical activity minutes per week based on the PPAQ.

Incident adult-onset T2DM was regressed onto each new continuous measure: physical activity minutes per week based on the ACLS questionnaire or physical activity minutes per week based on the PPAQ questionnaire. To preserve less than a 1% reduction in power, multiple imputation of 40 data sets was executed.²⁶ Logistic regression was performed using data from the original data set, after the 40 imputed data sets were created,²⁶ and pooled data across all 40 imputed data sets. After accounting for missing data, there were 380 participants with new-onset T2DM. Gender, ethnicity, age, BMI, level of education, hours of television watching per day, and diet status were included in the regressions as control variables. The interaction of BMI and diabetes risk was performed. The P value was set at .05. Statistical analyses were completed using SPSS version 20 software.

Results

A total of 9873 participants (33% black and 67% nonblack) were included in the final analysis. The percentage of participants with incident adult-onset T2DM versus those without adult-onset T2DM according to black, white, and other ethnicity is 5.5% versus 94.5%, 3.0% versus 97.0%, and 0.3% versus 96.5%, respectively (Table 1). Adult-onset T2DM was predominant in older, black obese individuals when compared to individuals who did not develop diabetes. A greater proportion of participants with incident adult-onset T2DM followed nonvegetarian diets. Individuals who developed adult-onset T2DM had lower levels of education, were more likely to watch television ≥5 h/d, and spend less time in both measures of physical activity than their counterparts who did not report a diagnosis of diabetes.

Table 1.

Descriptive Characteristics of Study Sample. ^a,b,c,d,e

Variable	Diabetes (n = 380)	No Diabetes (n = 9493)
Age, years^f	65.7 (12.2)	60.6 (13.7)
Gender
Female	265 (4.0%)	6400 (96.0%)
Male	115 (3.6%)	3093 (96.4%)
Ethnicity^f
Black	180 (5.5%)	3077 (94.5%)
White	178 (3.0%)	5811 (96.8%)
Other	22 (3.5%)	605 (96.5%)
Body mass index^f kg/m²	30.6 (7.0)	26.7 (5.7)
Diet
Nonvegetarian^f	226 (5.2%)	4087 (94.8%)
Vegan^f	18 (2.2%)	810 (97.8%)
Pesco-vegetarian^f	27 (2.7%)	988 (97.3%)
Lacto-vegetarian^f	87 (2.9%)	2915 (97.1%)
Semi-vegetarian	13 (2.6%)	487 (97.4%)
Education^f
Grade school	7.3 (4.3%)	162 (95.7%)
High school	81 (5.2%)	1489 (94.8%)
Some college	122 (4.7%)	2466 (95.3%)
Associate degree	45 (3.9%)	1099 (96.1%)
Bachelor degree	56 (2.4%)	2251 (23.7%)
Master’s degree	44 (3.0%)	1415 (97.0%)
Doctoral degree	25 (3.9%)	611 (96.1%)
Hours of TV/day^f
Less than 1	64.1 (2.5%)	2524 (97.5%)
1-2	162.4 (3.5%)	4428 (96.5%)
3-4	117 (5.4%)	2068 (94.6%)
≥5	36.3 (7.7%)	437 (92.4%)
PPAQ^f min/wk	60.5 (80.0)	81.8 (91.0)
ACLS^g min/wk	64.8 (83.9)	75.1(83.8)

Abbreviations: ACLS, Aerobics Center Longitudinal Study; PPAQ, Paffenbarger Physical Activity Questionnaire.

^aAnalysis of variance for mean differences and χ² tests for percentages.

^bDiet: Each contrast is a measurement of each specified diet versus all other diets.

^cHours of TV/day: Each category was compared to less than 1 hour of TV/day category.

^dn = 9873.

^eValues are mean (SD) or frequency (percentages).

^fP < .001 for comparisons between diabetes and no diabetes.

^gP < .05.

Results of the multivariate logistic regression analysis of each physical activity variable are shown in Table 2. Black ethnicity, ethnicity other than black or white, increasing age, BMI, and diet were associated with an increased risk of adult-onset T2DM. Gender and daily hours of sedentary activity were not significant risk factors. The 2 different questionnaire groups were similar and differed by a total of 1 average minute of physical activity per week. The group that answered the ACLS questionnaire reported an average of 70 minutes of physical activity per week at baseline, whereas the group that answered the PPAQ questionnaire reported an average of 71 minutes of physical activity per week at baseline. The PPAQ measure was significantly associated with protection against adult-onset T2DM (Table 2), whereas the ACLS variable was not significantly associated with decreased risk of adult-onset T2DM. The confidence intervals of the physical activity measures overlap and are related to diabetes risk reduction. The ACLS variable was nonsignificant (P = .363), whereas the PPAQ measure was a significant determinant in predicting adult-onset T2DM (P = .008). A 0.2% reduction in the risk of adult-onset T2DM occurs for every weekly self-reported minute using the PPAQ exists and is equivalent to a 6% reduction in the risk of adult-onset T2DM per 30-minute weekly increase in activity when compared to individuals who do not report such activity. The interaction of BMI and each physical activity measure and adult-onset T2DM was nonsignificant (Table 3).

Table 2.

Risk of Incident Adult-Onset Type 2 Diabetes According to Questionnaire. ^a,b,c

	Model 1			Model 2
Variable	OR	95% CI	P	OR	95% CI	P
PPAQ min/wk^d	0.998	0.997-1.000	.008
ACLS min/wk				0.999	0.998-1.001	.363
Female (reference, male)	1.074	0.851-1.356	.548	1.055	0.836-1.332	.652
Ethnicity (reference, white)
Blacks^e	1.972	1.569-2.480	.000	1.955	1.555-2.458	.000
Other^d	1.604	1.011-2.546	.045	1.586	1.000-2.517	.050
Age, years^e	1.040	1.031-1.049	.000	1.041	1.032-1.050	.000
Level of education	0.987	0.935-1.041	.627	0.983	0.932-1.037	.536
BMI, kg/m^2e	1.079	1.063-1.095	.000	1.080	1.064-1.096	.000
Nonvegetarian (reference, vegetarian)	1.497	1.191-1.882	.001	1.512	1.203-1.900	.000
Hours of television/week (reference, <1 hour TV/d)	1.045	0.956-1.143	.334	1.050	0.960-1.149	.282

Abbreviations: ACLS, Aerobics Center Longitudinal Study; BMI, body mass index; CI, confidence interval; OR, odds ratio; PPAQ, Paffenbarger Physical Activity Questionnaire.

^aModel 1: Association of diabetes risk and PPAQ. Gender, ethnicity, age, BMI, level of education, hours of television watching per day, and diet status were included in the regressions as control variables.

^bModel 2: Association of diabetes risk and ACLS. Gender, ethnicity, age, BMI, level of education, hours of television watching per day, and diet status were included in the regressions as control variables.

^cN = 9873.

^dP < .05.

^eP < .001 for comparisons between adult-onset type 2 diabetes and no diabetes (multivariate logistic regression).

Table 3.

Interaction of Physical Activity Measure and BMI.^a

Interaction	Odds Ratio	Significance
PPAQ and BMI	1.00	0.618
ACLS and BMI	1.00	0.194

Abbreviations: ACLS, Aerobics Center Longitudinal Study; BMI, body mass index; PPAQ, Paffenbarger Physical Activity Questionnaire.

^aN = 9873.

Discussion

After analyses of the overall model statistics, our main finding was that the PPAQ and the ACLS had similar likelihood of capturing incident adult-onset T2DM in a healthy population of men and women when controlling for several lifestyle-based factors. The physical activity measures share confidence intervals illuminating the possibility that participants may have answered affirmatively to both physical activity measures; therefore, the sample population mean may be computed from a shared population. Participants may have identified with both physical activity measures because both physical activity measures queried walking and jogging activities. There are subtle differences between the measures. The terminology of the PPAQ specifically queried “brisk” walking versus “walking,” and the ACLS queried “regular” physical activity, whereas the ACLS measure did not specify the regularity of exercise. The PPAQ may therefore capture persons who engage in structured physical exercise as part of their lifestyle, whereas the ACLS may capture those who are physically active in more general activities of daily life.

In the multivariate analysis, we found that black ethnicity, ethnicity other than black or white, age, and BMI were risk factors for new-onset diabetes consistent with previous literature. Parallel with published work, a pronounced racial disparity exists among black individuals.²⁷ In agreement with prior studies, BMI^12,14,28 remained a relevant risk factor for diabetes.

Although individuals with diabetes were more sedentary when compared to those who did not become diabetic (Table 1), sedentary activity did not significantly contribute to an increased risk of adult-onset T2DM (Table 2). In line with the current study, hours of television viewing was not predictive of insulin resistance, a precursor of diabetes, in a population of men and women with a positive family history of diabetes.²⁹ Such data are in contrast to a nearly 3-fold increase in disease risk in a prospective study of healthy women³⁰ who watched more than 40 hours of weekly television.

In a subanalysis, black ethnicity, increased age, BMI, and diet status remained significant predictors of incident diabetes. Gender did not contribute significantly to the risk of diabetes, in contrast to national, age-adjusted data.^31,32 Concurrent with published data, age and diabetes are positively correlated with a linear increase.³¹

In an interventional study, moderate physical activity was measured via self-report (eg, walking, jogging, and running) and predicted reduction of diabetes in participants with impaired fasting glucose and impaired glucose tolerance when combined with dietary modification.^33

–36 Lifestyle interventions of exercise and saturated fat reduction resulted in weight loss and improved insulin sensitivity. Measurement of walking, jogging, and running without dietary restriction did not predict reduction of diabetes in a healthy population.

In a systematic review of cohort and cross-sectional studies of 301 221 participants, it has been found that incident diabetes occurred in 9367 moderate exercisers.⁸ Unlike the current study, Jeon and colleagues quantified physical activity using MET scores and only included studies that controlled for vigorous activity. Significance was maintained for persons who participated regularly in moderate activity (MET = 3.0-6.0) versus sedentary activity and regular walking (≥2.5 h/wk brisk walking) of at least 2 hours a week when compared with almost no walking. Similar associations were found for men and women, in the US and Europe.

Limitations

There are several limitations of the current study. Although the cumulative incidence rate of diabetes (25%) is comparable to a gender-mixed nondiabetic Finish population measured in 1987 and 1992 from a National registry,³⁷ generalizability is limited as the sample belonged to a single church denomination that recommends a healthy lifestyle. Furthermore, geographical differences within the sample were not measured. Although SDAs generally follow a similar health-conscious lifestyle, meaningful differences may exist from region to region that effect diabetic trends.³¹ Such geographical data were not collected. Therefore, a multilevel model could not be created to explore clustering by geographical region. The researchers did not control for effects that may exist between different zip codes within the same state, different states within the United States, and between urban versus rural communities in the United States and Canada. Although the researchers controlled for socioeconomic status, individual response may vary within similar socioeconomic levels depending upon environmental factors that differ from region to region.

The study measured the risk of diabetes in self-reported exercisers and did not include an intervention. Therefore, the diabetes incidence rate may not generalize to intervention studies that have reported nearly half the incidence rate (11%) of new-case diabetes after a lifestyle intervention.³⁴ When compared to a global sample of previously and newly diagnosed cases of diabetes measured in 2008, nearly twice as many men (9.8%) and women (9.2%) were identified as having diabetes³¹ when compared to the current study. Men shared a similar preponderance of diabetes (9.6%) that remained unaltered in the presence of a positive family history, in a cohort drawn from 6 states and 2 US metropolitan areas.¹²

Cardiovascular disease³⁸ and certain metabolic disease processes such as metabolic syndrome,³⁷ endocrine dysfunction,³⁹ and a family history of diabetes⁴⁰ reduce insulin sensitivity and are associated with heightened diabetes risk. The presence of diabetes in family members was not measured at baseline; therefore, the findings of this study may not be generalized to high-risk populations.

The physical activity measures were not quantified into discrete MET hours per day. Construct validity bias is especially relevant because similar nomenclature exists in both questionnaires. Walking and brisk walking were measured in the ACLS and PPAQ, respectively, and jogging activities are included in both questionnaires.

Data were collected from previously validated^21
–23 self-report physical activity questionnaires with inherent self-report bias. Validation of diabetes diagnosis was previously attempted in this population.^21
–23 Questions regarding treatment of diabetes were time-specific, and respondents may not have had accurate recall or misunderstood the question. Cases of diabetes may have occurred that were not diagnosed at the time of the study, limiting detection in participants who did not seek medical care. Without biometric measurements, or conversion of weekly minutes of walking, jogging, and running activity to MET hours per day, physical activity intensity was not quantified, limiting the comparability of our findings to published data that associate physical activity intensity to diabetes risk reduction. Additionally, physical activity was measured at follow-up; therefore, changes in exercise habits and subsequent weight fluctuations in response to diabetes diagnoses may be related and were not analyzed.

However, the sample also exhibited several strengths. The population consisted almost exclusively of nonsmokers. A total of 9751 of 9796 participants reported that they have either never smoked on a regular basis or answered “0 to 1 cigarette in the last year.” Cigarette smoking is a strong confounder of the activity–diabetes link, as smoking is associated with less physical activity and increased incidence of diabetes.^41,42 In addition, the sample followed a wide range of diets, from vegan to omnivorous, and the results were consistent with previous data showing protective effects of vegetarian diets on diabetes.⁴³

Conclusion

The previously validated Paffenbarger index has been a widely accepted self-reported measure of vigorous activity in epidemiological studies. The PPAQ and the ACLS were similarly effective in predicting incident adult-onset T2DM in healthy male and female participants when each activity was compared to no activity and multiple lifestyle-related confounders were controlled.

SO WHAT?

What is already known on this topic?

Because of the dose–response relationship of physical activity and chronic disease, the American College of Sports Medicine and the American Heart Association endorse expenditure at least 20 minutes of moderate physical activity 5 days a week or 30 minutes of vigorous physical activity 3 days weekly to promote and maintain health.¹ Physiological adaptations occur during exercise that modify T2DM risk,² lending credence to self-report measurement of PA activity when determining adultonset T2DM risk in population studies.

What does this article add?

Both physical activity measures were equally likely to capture incident adult-onset T2DM in a population of healthy men and women.

What are the implications for health promotion practice or research?

The PPAQ may be better at capturing vigorous PA because the terminology of the question is more descriptive and visceral lessening ambiguity between the 2 modes of activity, the question defines the physiological effect of vigorous activity, and the term “regular” is used that may have captured those who exercise on a consistent basis.

The PPAQ may be more adept at capturing vigorous activity for several reasons. Firstly, the PPAQ used descriptive terms before each mode of activity (eg, brisk walking versus walk). Furthermore, the PPAQ defined the physiological effect of vigorous activity in distinguishable terms (eg, work up a sweat, get your heart thumping, or get out of breath), whereas the ACLS uses the general term “workout” after each mode of activity (walk, jog, run). The more visceral description used to describe vigorous activity may have lessened ambiguity between the 2 modes of activity. Finally, the use of “regular” in the PPAQ denotes a measure of how consistently vigorous activity was performed. The ACLS does not specify consistency of walking, jogging, or running activity. Participants may have overreported vigorous activity if the activity was performed on a regular basis.

Clinical correlation of question 6 from PPAQ with direct measures of diabetes (eg, glycated hemoglobin levels)⁴⁴ and biometric measures of physical activity intensity will enhance comparison of our data to other published data. A second analysis that measures the significance of the interaction of ethnicity with BMI and physical activity using question 6 of the PPAQ in black individuals may elucidate whether question 6 is more effective in predicting diabetes in black individuals as BMI changes. Such data may lend credence to the utilization of question 6 as a screening tool for diabetes and related complications in epidemiological studies.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Haskell

Lee

Pate

. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation. 2007;116(9):1081–1093.

Mayer-Davis

D’Agostino

Jr Karter

. Intensity and amount of physical activity in relation to insulin sensitivity: the Insulin Resistance Atherosclerosis Study. JAMA. 1998;279(9):669–674.

Rockl

Witczak

Goodyear

. Signaling mechanisms in skeletal muscle: acute responses and chronic adaptations to exercise. IUBMB Life. 2008;60(3):145–153.

Perseghin

Price

Petersen

. Increased glucose transport-phosphorylation and muscle glycogen synthesis after exercise training in insulin-resistant subjects. N Engl J Med. 1996;335(18):1357–1362.

Goodyear

Kahn

. Exercise, glucose transport, and insulin sensitivity. Annu Rev Med. 1998;49:235–261.

Sherwood

Jeffery

. The behavioral determinants of exercise: implications for physical activity interventions. Annu Rev Nutr. 2000;20:21–44.

Williams

Thompson

. Walking versus running for hypertension, cholesterol, and diabetes mellitus risk reduction. Arterioscler Thromb Vasc Biol. 2013;33(5):1085–1091.

Jeon

Lokken

van Dam

. Physical activity of moderate intensity and risk of type 2 diabetes: a systematic review. Diabetes Care. 2007;30(3):744–752.

Sigal

Rich-Edwards

. Walking compared with vigorous physical activity and risk of type 2 diabetes in women: a prospective study. JAMA. 1999;282(15):1433–1439.

10.

Schulze

. Primary prevention of diabetes: what can be done and how much can be prevented? Annu Rev Public Health. 2005;26:445–467.

11.

Gill

Cooper

. Physical activity and prevention of type 2 diabetes mellitus. Sports Med. 2008;38(10):807–824.

12.

Reis

Loria

Sorlie

Park

Hollenbeck

Schatzkin

. Lifestyle factors and risk for new-onset diabetes: a population-based cohort study. Ann Intern Med. 2011;155(5):292–299.

13.

Folsom

Kushi

Hong

. Physical activity and incident diabetes mellitus in postmenopausal women. Am J Public Health. 2000;90(1):134–138.

14.

Siegel

Sesso

Bowman

Lee

Manson

Gaziano

. Physical activity, body mass index, and diabetes risk in men: a prospective study. Am J Med. 2009;122(12):1115–1121.

15.

Helmrich

Ragland

Leung

Paffenbarger

Jr . Physical activity and reduced occurrence of non-insulin-dependent diabetes mellitus. N Engl J Med. 1991;325(3):147–152.

16.

Krishnan

Rosenberg

Palmer

. Physical activity and television watching in relation to risk of type 2 diabetes: the Black Women’s Health Study. Am J Epidemiol. 2009;169(4):428–434.

17.

Steinbrecher

A EE

Grandinetti

Nigg

Kolonel

Maskarinec

. Physical activity and risk of type 2 diabetes among Native Hawaiians, Japanese Americans, and Caucasians: the multiethnic cohort. J Phys Act Health. 2012;9(5):634–641.

18.

Hsia

Allen

Macera

Kronenfeld

. Physical activity and diabetes risk in postmenopausal women. Am J Prev Med. 2005;28(1):19–25.

19.

Knab

Shanely

Corbin

Jin

Sha

Nieman

. A 45-minute vigorous exercise bout increases metabolic rate for 14 hours. Med Sci Sports Exerc. 2011;43(9):1643–1648.

20.

Kohl

Blair

Paffenbarger

Jr Macera

Kronenfeld

. A mail survey of physical activity habits as related to measured physical fitness. Am J Epidemiol. 1988;127(6):1228–1239.

21.

Singh

Tonstad

Abbey

Fraser

. Validity of selected physical activity questions in white Seventh-day Adventists and non-Adventists. Med Sci Sports Exerc. 1996;28(8):1026–1037.

22.

Singh

PN FG

Knutsen

Lindsted

Bennett

. Validity of a physical activity questionnaire among African-American Seventh-day Adventists. Med Sci Sports Exerc. 2000;33(3):468–475.

23.

Paffenbarger

Jr Wing

Hyde

. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol. 1978;108(3):161–175.

24.

Butler

Fraser

Beeson

. Cohort profile: the Adventist Health Study-2 (AHS-2). Int J Epidemiol. 2008;37(2):260–265.

25.

Lee

Morton

Walters

. Cohort profile: the Biopsychosocial Religion and Health Study (BRHS). Int J Epidemiol. 2009;38(6):1470–1478.

26.

Graham

Olchowski

Gilreath

. How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prev Sci. 2007;8(3):206–213.

27.

Chatterjee

Brancati

Shafi

. Non-traditional risk factors are important contributors to the racial disparity in diabetes risk: the atherosclerosis risk in communities study. J Gen Intern Med. 2014;29(2):290–297.

28.

Bombelli

Facchetti

Sega

. Impact of body mass index and waist circumference on the long-term risk of diabetes mellitus, hypertension, and cardiac organ damage. Hypertension. 2011;58(6):1029–1035.

29.

Ekelund

Brage

Griffin

Wareham

; ProActive UK Research Group. Objectively measured moderate and vigorous-intensity physical activity but not sedentary time predicts insulin resistance in high-risk individuals. Diabetes Care. 2009;32(6):1081–1086.

30.

. Sedentary lifestyle and risk of obesity and type 2 diabetes. Lipids. 2003;38(2):103–108.

31.

Danaei

G FM

. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: a systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2.7 million participants. Lancet. 2011; 378(9785):31–40.

32.

Yang

Weng

. Prevalence of diabetes among men and women in China. N Engl J Med. 2010;362(12):1090–1101.

33.

Knowler

Barrett-Connor

Fowler

. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346(6):393–403.

34.

Tuomilehto

Lindström

Eriksson

. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med. 2001;344(18):1343–1350.

35.

Kosaka

Noda

Kuzuya

. Prevention of type 2 diabetes by lifestyle intervention: a Japanese trial in IGT males. Diabetes Res Clin Pract. 2005;67(2):152–162.

36.

Ramachandran

Snehalatha

Mary

. The Indian Diabetes Prevention Programme shows that lifestyle modification and metformin prevent type 2 diabetes in Asian Indian subjects with impaired glucose tolerance. Diabetologia. 2006;49(2):289–297.

37.

Lindstrom

Tuomilehto

. The diabetes risk score: a practical tool to predict type 2 diabetes risk. Diabetes Care. 2003;26(3):725–731.

38.

Faerch

K WDR

Tabák

A G

. Trajectories of cardiometabolic risk factors before diagnosis of three subtypes of type 2 diabetes: a post-hoc analysis of the longitudinal Whitehall II cohort study. Lancet Diabetes Endocrinol. 2013;1(1):43–51.

39.

Liu

Zhang

Tang

Chen

Wang

. Endocrinological analysis of endothelium-dependent vasodilation in middle-aged patients with impaired glucose tolerance during prediabetes mellitus. Exp Ther Med. 2014;7(3):697–702.

40.

InterAct

Scott

Langenberg

. The link between family history and risk of type 2 diabetes is not explained by anthropometric, lifestyle or genetic risk factors: the EPIC-InterAct study. Diabetologia. 2013;56(1):60–69.

41.

Patja

Jousilahti

Valle

Qiao

Tuomilehto

. Effects of smoking, obesity and physical activity on the risk of type 2 diabetes in middle-aged Finnish men and women. J Intern Med. 2005;258(4):356–362.

42.

Will

Galuska

Ford

Mokdad

Calle

. Cigarette smoking and diabetes mellitus. Evidence of a positive association from a large prospective cohort study. Int J Epidemiol. 2001;30(3):540–546.

43.

Tonstad

Stewart

Oda

Batech

Herring

Fraser

. Vegetarian diets and incidence of diabetes in the Adventist Health Study-2. Nutr Metab Cardiovasc Dis. 2013;23(4):292–299.

44.

Bennett

Guo

Dharmage

. HbA(1c) as a screening tool for detection of type 2 diabetes: a systematic review. Diabet Med. 2007;24(4):333–343.