Preventive Medicine 61 (2014) 14–19

Contents lists available at ScienceDirect

Preventive Medicine

j ourna l homepage: www.e lsev ie r .com/ locate /ypmed

Household physical activity and mortality in older adults: A nationalcohort study in Spain

David Martínez-Gómez ⁎, Pilar Guallar-Castillón, Luz M. León-Muñoz, Fernando Rodríguez-ArtalejoDepartment of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid, Madrid, SpainIdiPAZ–CIBER of Epidemiology and Public Health (CIBERESP), Madrid, Spain

⁎ Corresponding author at: Departamento de MediciFacultad de Medicina, Universidad Autónoma de Madri28029 Madrid, Spain. Fax: +34 91 497 5353.

E-mail address: d.martinez@uam.es (D. Martínez-Góm

0091-7435/$ – see front matter © 2014 Elsevier Inc. All rihttp://dx.doi.org/10.1016/j.ypmed.2014.01.006

a b s t r a c t

a r t i c l e i n f o

Available online 15 January 2014

Keywords:Physical activityMortalitySurvivalSedentary timeHousework

Objective. To examine the association between household physical activity (HPA) and all-cause mortality in acohort of older adults from Spain, and the role of sedentary time on this association.

Method. Prospective cohort study of 2874 individuals aged≥62 years. In 2003, the time spent in HPA and thetime spent seatedwere self-reported. The association of HPAwith all-causemortality through 2011was assessedwith Cox regression.

Results. During the follow-up, 970 participants died. In men, HPA was inversely associated with the risk ofdeath only among thosewith longer sitting time (≥8 h/d): compared to thosewhodid not do HPA, themortality

hazard ratio (HR)was 0.80 (95% confidence interval (CI): 0.60–1.08) and 0.43 (95% CI: 0.27–0.69) for those whospent N0 to 2 h/d and N2 h/d in HPA, respectively (P for trend b 0.001). Inwomen, sitting time did notmodify thestudy association. Thus, compared to women who spent b2 h/d in HPA, the HR for mortality was 0.72 (95% CI:0.56–0.93) and 0.52 (95% CI: 0.39–0.70) for those who spent N2 to 4 h/d, and N4 h/d in HPA, respectively(P for trend b 0.001).

Conclusion. In women, HPA is associated with reduced mortality regardless of sitting time. HPA may alsocontribute to longer survival among men with longer sitting time.

© 2014 Elsevier Inc. All rights reserved.

Introduction

According to the current physical activity (PA) recommendations,older adults should do moderate-intensity PA for at least 150 minthroughout the week or do at least 75 min of vigorous-intensity PAthroughout the week (USDHHS, 2008b; World Health Organization,2010). However, only a small percentage of older adults meet thisrecommendation because most of their PA is performed at light-intensity (Evenson et al., 2012; Troiano et al., 2008). In particular,among older women, household activities (e.g., cleaning, cooking,vacuuming, ironing, watering plants, and grandchild care) might be animportant component of their PA (Chen et al., 2012; Ottenbacheret al., 2012).

Information on the effect of household PA (HPA) on mortality inolder adults is scarce. Only two studies, conducted in Mexican-American (Ottenbacher et al., 2012) and Taiwanese (Chen et al., 2012)individuals, have assessed this association; in both of themHPAwas as-sociatedwith lowermortality, regardless of other types of PA. Yet, thesestudies did not examine whether this association was independent ofsitting time, which is inversely correlated with the time spent at light-

na Preventiva y Salud Pública,d, C/Arzobispo Morcillo, 2, CP.

ez).

ghts reserved.

intensity PA (Owen et al., 2010). This is important because older adultsare the population subgroup with the highest level of sedentariness(Matthews et al., 2008), and longer sitting time is associatedwith highermortality in older adults (Koster et al., 2012).

It is therefore unclear whether HPA plays an independent role inreducing mortality in the elderly, or whether it is only a reflection ofthe associated lower sedentariness. In this study, we aimed (i) toexamine the association between HPA and all-cause mortality in acohort of older adults from Spain and (ii) to assess the role of sedentarytime on this association. We hypothesized that HPA is prospectivelyassociated with reduced mortality in older adults, particularly in thosewho are very sedentary.

Materials and methods

Study design and population

The study design and procedures have been reported elsewhere (Gutiérrez-Fisac et al., 2004; López-García et al., 2008). Briefly, in 2001 we obtained base-line information from a cohort of 4008 persons (2269 women) representativeof the non-institutionalized population aged ≥60 years in Spain. Data werecollected by home-based personal interview and physical examination,performed by trained and certified personnel.

In 2003, an attempt was made to contact the subjects again, and this wassuccessful in 3235 of them (1824 women) (Balboa-Castillo et al., 2011b;López-García et al., 2008). Information was obtained by telephone interview

15D. Martínez-Gómez et al. / Preventive Medicine 61 (2014) 14–19

and the individuals contacted did not differ significantly from those lost tofollow-up in socio-demographic and lifestyle characteristics. Since the mainexposure variable (i.e., HPA) was included only in the 2003 interview, onlyparticipants contacted in 2003 were selected for the current analysis.

The study was approved by the Clinical Research Ethics Committee of the“La Paz” University Hospital in Madrid, Spain.

Main exposure variable

Information on the time spent in household activities was obtainedwith thefollowing question: “How much time do you spend doing household activitieson a typical weekday?” The interviewer always gave several examples ofcommon household tasks (e.g., cleaning, cooking, and washing dishes). Thesame question was asked with reference to a weekend day. The time spent inHPA was calculated as follows: [(household task time on a weekday × 5)+ (household task time on a weekend day × 2)] / 7.

Outcome variable

The outcome was all-cause mortality from 2003 to the end of follow-up onDecember 31, 2011. The number and dates of deaths were obtained by acomputerized search of the National Death Index, which contains informationon the vital status of all residents in Spain.

Potential confounders

In 2003, data were collected on variables related to both HPA andmortality.Specifically, data were obtained on age, sex, the highest educational levelattained (no formal education, primary, and secondary or higher),marital status(single, married, divorced, widowed), and if they lived alone (yes/no).

In addition, body weight in 2003 was self-reported (López-García et al.,2008), whereas height in 2003 was assumed to be the same as in 2001, whenit was measured by a portable wall-mounted stadiometer (KaWe, Asperg,Germany) (Gutiérrez-Fisac et al., 2004). Body mass index (BMI) was calculatedas weight in kg divided by height in m squared, and obesity was defined as BMI≥ 30 kg/m2.

Agility limitation was ascertained with the question: “Do you experienceany difficulty in bending or kneeling?” (Balboa-Castillo et al., 2011a). Mobilitylimitation was defined as an affirmative response to any of the followingquestions: (i) “Do you experience any difficulty in picking up or carrying ashopping bag?”; (ii) “Do you experience any difficulty in climbing one flight ofstairs?”; and (iii) “Doyouexperience any difficulty inwalking several city blocks(a few hundred meters)?” (Balboa-Castillo et al., 2011a).

Study participants also reported tobacco and alcohol consumption (Guallar-Castillón et al., 2001). Leisure-time PAwas evaluatedwith the validated Spanishversion of the PA questionnaire used in the Nurses' Health Study and the HealthProfessionals' Follow-up Study (Martinez-Gonzalez et al., 2005). This question-naire rates participation and time devoted to each activity perweek in 16 differ-ent activities: walking, dancing, stationary bicycling, bicycling outdoors,competitive running, jogging, gardening, skiing, climbing, football, going tothe gym, judo, swimming, tennis, sailing, and other team sports. The annualduration of each activity (N6 months per year, 3–6 months per year, and b

3 months per year) was also ascertained. The total time spent in leisure-timePA was calculated as the sum of times for all activities weighted by the annualperiod of participation in each activity.

Sedentary behavior was estimated by the total time spent seated, based onthe following question referred to leisure time: “About how much time do youspend sitting down on weekdays?” (Balboa-Castillo et al., 2011b). The samequestion was asked with reference to a weekend day. Thus, the number ofhours per day spent seatedwas calculated as follows: [(sitting hours on aweek-day × 5) + (sitting hours on a weekend day × 2)] / 7.

Finally, the following diseases diagnosed by a physician and reported by thestudy participant were recorded: coronary heart disease, stroke, diabetesmellitus, hip fracture and cancer at any site.

Statistical analysis

Of the participants followed until 2003, the following persons wereexcluded from the analysis: 245 who died between 2001 and 2003, 44 wholacked data onHPA, and 72without data on potential confounders. Accordingly,we conducted the analyses with 2874 individuals (1246 men and 1628women).

Firstly, we checked that there was no collinearity between activity variables(i.e., HPA, leisure-time PA, and sitting time). On average, women spent 2.1 h/dmore in household tasks than men. Thus, categories of HPA were definedseparately for men (0 h/d,N0 to 2 h/d, and N2 h/d) and women (≤2 h/d,N2 to4 h/d, and N4 h/d). The association between sex-specific categories of HPAand mortality was summarized with hazard ratios (HRs) and their 95%confidence interval (CI) obtained from Cox regression. Test for linear trendswas also performed by modeling time spent in HPA (h/d) as a continuousvariable. Follow-up duration in number of days was used as the time scale,which started at the date of interview in 2003 and continued until date ofdeath or December 31, 2011.

Four Cox models, with progressive adjustment for potential confounders,were fitted. The first model adjusted for age, sex, and educational level; thesecond model further adjusted for marital status, living alone status, BMI,activity limitations (mobility and agility) and comorbidities (coronary heartdisease, stroke, diabetes mellitus, hip fracture, and cancer); the third modeladditionally adjusted for alcohol consumption, smoking status and leisure-time PA; and the final model additionally adjusted for sitting time.

We also conducted several sensitivity analyses for assessing the robustnessof the main results. We performed stratified analyses by age and education.Similar stratifications were performed by health status (obesity, activity limita-tions and comorbidities) and several lifestyles (smoking, alcohol intake, andleisure-time PA). We tested for interactions of HPA with these covariates byincluding appropriate interaction terms in the Cox models and using likelihoodratio tests.

To examine the role of sitting time on the association between HPA andmortality, we performed Cox regression adjusted for potential confoundersacross groups of HPA and sitting time. We used the cut-off point of 8 h/d, inaccordance with previous research (van der Ploeg et al., 2012), to classifyolder adults with low (b8 h/d) or high (≥8 h/d) levels of sitting time. We alsotestedwhether the study association variedwith sitting timebyusing likelihoodratio tests which compared models with interaction terms (products of HPA bysitting time categories) and models without those terms.

We assessed the assumption of proportionality of hazards both graphicallyand by testing the significance of the interaction between the main exposurevariable and years of follow-up, and we found no evidence of departure fromsuch assumption (P for interaction N 0.2). Analyses were performed withSTATA® version 11.1. All tests were 2-sided and statistical significance was setat P b 0.05 for main effects and P b 0.1 for interactions.

Results

Table 1 shows the baseline characteristics of the study participantsaccording to HPA. During an average follow-up of 7.3 years and20,980 person-years of observation, 970 deaths occurred. Table 2shows the association between HPA andmortality. In analyses adjustedfor socio-demographic variables, obesity, activity limitations and co-morbidities, those men who did household activities for N0 to 2 h/andfor N2 h/d had, respectively, a mortality HR = 0.67 (0.52–0.86) and0.63 (0.45–0.87) compared with men who did no household activities.This association decreased in magnitude and was no longer significantafter further adjustment for lifestyle variables and sitting time. Inwomen, there was a clear inverse dose–response relationship betweenthe time spent in household activities and mortality independently ofall potential confounders. Compared to women who spent less than2 h/d in household tasks, the fully-adjusted mortality HR was 0.72(95% CI: 0.56–0.93) and 0.52 (95% CI: 0.39–0.70) for those who spentN2 to 4 h/d and N4 h/d in HPA (P for trend b 0.001).

To assess the robustness of results, we examined the study associa-tion within groups defined by socio-demographic variables, BMI, activ-ity limitations and comorbidities (Table 3). We found an inverse dose–response relation between HPA and mortality in most groups amongwomen (P for trend b 0.05), but not inmen (all P for trend N 0.05). Sim-ilar sex-specific results (data not shown)were obtained whenwe strat-ified the analyses by smoking (smoking, non-smoking), and alcoholintake (drinker, non-drinker). Likewise, when stratifying by leisure-time PA, HPA was only significantly associated (P for trend b 0.001)with mortality in women with low levels of this PA (bsex-specificmedian in the cohort), although we also found a trend for significance

Table 1Baseline characteristics according to time spent in household physical activity among older adults, Spain 2003.

Time spent in household physical activity

Men Women

0 h/d N0 to 2 h/d N2 h/d ≤2 h/d N2 to 4 h/d N4 h/d

N 550 471 225 713 522 392Age (years) 74.3 ± 7.8 71.5 ± 6.7 71.5 ± 5.9⁎ 77.2 ± 8.3 72.6 ± 6.9 70.0 ± 6.2⁎

Educational level, %No education/primary/secondary or + 47.5/34.6/18.0 40.3/39.6/20.0 44.9/36.1/19.0 59.0/33.4/7.6 55.6/36.1/8.3 53.3/35.7/11.0

Marital status, %Single/married/divorced/widowed 4.2/82.1/0.2/13.5 2.9/85.0/1.1/11.0 4.7/73.6/1.7/20.0 5.9/31.2/1.3/61.6 5.7/47.5/1.7/45.1 4.9/60.2/1.6/33.3

Living alone, % 17.2 39.7 33.7⁎ 26.5 25.8 14.9⁎

Body mass index (kg/m2) 27.8 ± 3.4 27.9 ± 3.8 27.7 ± 3.7 29.0 ± 5.5 28.7 ± 4.6 29.3 ± 4.5Agility limitation, % 57.7 50.0 48.5⁎ 78.4 66.0 69.7⁎

Mobility limitation, % 47.7 31.4 28.4⁎ 73.8 55.3 57.1⁎

Coronary heart disease, % 9.1 3.5 2.0⁎ 6.3 2.1 1.8⁎

Stroke, % 7.2 5.6 2.7⁎ 8.7 2.3 3.1⁎

Diabetes mellitus, % 24.2 20.4 16.7⁎ 22.4 17.6 22.5Hip fracture, % 3.5 1.4 2.8 5.9 3.2 1.0⁎

Cancer, % 1.9 2.1 3.4 3.1 1.7 1.2Smoking, %Never/former/currently 20.9/60.1/19.0 22.2/51.2/26.6 27.6/54.7/17.7 95.2/3.2/1.6 94.6/3.5/1.9 92.8/5.6/1.6

Alcohol drinking, %Never/former/currently 25.6/16.0/58.4 14.3/8.2/77.5 19.8/15.0/65.2 70.1/6.7/23.2 63.4/3.9/32.7 62.7/3.4/33.9

Physically active in leisure-time (min/d) 56.9 ± 51.8 75.5 ± 49.3 74.2 ± 50.5⁎ 31.9 ± 37.7 54.8 ± 49.1 55.6 ± 47.1⁎

Sitting time (h/d) 5.5 ± 3.2 4.3 ± 2.1 4.0 ± 1.8⁎ 6.0 ± 3.4 4.5 ± 2.1 4.1 ± 1.9⁎

For continuous variables, values are mean ± standard deviation.⁎ P b 0.05 across groups within the same gender.

16 D. Martínez-Gómez et al. / Preventive Medicine 61 (2014) 14–19

(P for trend= 0.079) among womenwith high leisure-time PA (≥sex-specific median in the cohort). In fact, the P for the interaction betweenthese variables and HPA was N0.1 in all cases, so that no evidence wasfound that the stratification variables modified the study association.

Fig. 1 shows the association between HPA and mortality by sittingtime categories (≥8 h/d and b8 h/d). In men, HPA was associatedwith lower mortality in those with longer sitting time, but not inthose with less time spent seated (P for interaction = 0.039). Thus,among men with longer sitting time, compared to those who did notspend time in household activities, the HR for mortality was 0.80 (95%CI: 0.60–1.08) and 0.43 (95% CI: 0.27–0.69) for those men who spentN0 to 2 h/d and N2 h/d in HPA, respectively (P for trend b 0.001).In women, HPA was associated with reduced risk of death bothin those with longer and in those with shorter sitting times (P forinteraction = 0.671).

Additional analyses were performed to rule out preexisting subclin-ical disease (e.g., cardiovascular disease or cancer) as a factor leading tochanges in HPA and its association with mortality; specifically, werepeated the analyses after excluding the 120 deaths occurring in thefirst year of follow-up. The main results did not materially change;results in the same direction were obtained after additionally excluding

Table 2Mortality risk according to time spent in household physical activity among older adults, Spain

Time spent in household physical activity

Men

0 h/d N0 to 2 h/d N2 h/d

N/deaths 550/262 471/140 225/65Model 1: HR (95% CI) 1 (Ref.) 0.64 (0.50–0.82) 0.60 (0.43–0.82)Model 2: HR (95% CI) 1 (Ref.) 0.67 (0.52–0.86) 0.63 (0.45–0.87)Model 3: HR (95% CI) 1 (Ref.) 0.77 (0.59–1.00) 0.69 (0.50–0.96)Model 4: HR (95% CI) 1 (Ref.) 0.79 (0.61–1.03) 0.73 (0.53–1.01)

HR: Hazard ratio. CI: Confidence interval.Model 1: adjusted for age (years) and educational level (no education, primary, secondary or hModel 2: adjusted as inmodel 1 and formarital status (single,married, divorced, widowed), livi(yes, no), and history of coronary heart disease (yes, no), stroke (yes, no), diabetes mellitus (yModel 3: adjusted as in model 2 and for smoking (never, former, current smoker), alcohol drinModel 4: adjusted as in model 3 and for sitting time (h/d).

deaths in the second and the third year of follow-up, though statisticalprecision was substantially reduced (data not shown).

Discussion

In this long-term cohort study in Spanish older adults, HPAwas asso-ciated with reduced mortality in women regardless of time spentseated, obesity, activity limitations and somephysical health conditions.HPA was also linked to lower mortality in sedentary men.

The importance of HPA in older adults has usually been under-estimated because, while it is commonly performed at light-intensity,the PA guidelines mostly focus on moderate-to vigorous PA (USDHHS,2008b; World Health Organization, 2010). In fact, populations withlower levels of moderate-to-vigorous PA such as older adults, womenand people with low socio-economic status usually engage in higherlevels of HPA (Bicalho et al., 2010; Csizmadi et al., 2011; Duncan et al.,2009; Hallal et al., 2011; Khaing Nang et al., 2010; Mann et al., 2013;Marquez et al., 2011; Ng et al., 2009). Hence, it is of public and clinicalrelevance to establish whether these subpopulations could benefitfrom HPA.

, 2003.

P for trend Women P for trend

≤2 h/d N2 to 4 h/d N4 h/d

713/337 523/116 392/500.060 1 (Ref.) 0.59 (0.46–0.76) 0.44 (0.33–0.59) b0.0010.129 1 (Ref.) 0.68 (0.53–0.88) 0.49 (0.36–0.66) b0.0010.249 1 (Ref.) 0.71 (0.55–0.91) 0.51 (0.38–0.68) b0.0010.398 1 (Ref.) 0.72 (0.56–0.93) 0.52 (0.39–0.70) b0.001

igher education).ng alone (yes, no), bodymass index (kg/m2), agility limitation (yes, no),mobility limitationes, no), hip fracture (yes, no), and cancer (yes, no).king (never, former, current drinker), and leisure-time physical activity (min/d).

Table 3Mortality according to household physical activity among older adults, by socio-demographic and health status variables, Spain, 2003.

Time spent in household physical activity

Men P for trend Women P for trend

0 h/d N0 to 2 h/d N2 h/d ≤2 h/d N2 to 4 h/d N4 h/d

b75 years, N/deaths 305/97 327/66 164/37 278/59 340/37 305/23HR (95% CI) 1 (Ref.) 0.69 (0.46–1.03) 0.69 (0.41–1.16) 0.972 1 (Ref.) 0.56 (0.36–0.86) 0.46 (0.28–0.76) b0.001

≥75 years, N/deaths 244/165 144/73 62/27 435/278 182/79 87/27Adjusted HR (95% CI) 1 (Ref.) 0.90 (0.65–1.26) 0.79 (0.51–1.22) 0.257 1 (Ref.) 0.77 (0.57–1.05) 0.56 (0.39–0.81) b0.001

No formal education, N/deaths 261/134 190/64 101/35 421/220 291/66 209/30HR (95% CI) 1 (Ref.) 0.69 (0.49–0.98) 0.65 (0.41–1.03) 0.738 1 (Ref.) 0.61 (0.43–0.88) 0.51 (0.34–0.76) b0.001

Primary or higher education, N/deaths 289/128 281/76 124/29 292/117 231/50 183/19HR (95% CI) 1 (Ref.) 0.68 (0.45–1.01) 0.59 (0.37–0.96) 0.121 1 (Ref.) 0.88 (0.60–1.30) 0.54 (0.33–0.87) 0.021

BMI b 30 kg/m2, N/deaths 402/203 349/108 174/55 445/219 339/79 224/31HR (95% CI) 1 (Ref.) 0.78 (0.59–1.04) 0.73 (0.51–1.04) 0.437 1 (Ref.) 0.64 (0.48–0.86) 0.50 (0.35–0.73) b0.001

BMI ≥ 30 kg/m2, N/deaths 148/59 121/32 52/10 268/118 183/38 168/19HR (95% CI) 1 (Ref.) 0.79 (0.40–1.60) 0.55 (0.23–1.32) 0.490 1 (Ref.) 0.68 (0.44–1.03) 0.49 (0.28–0.85) 0.002

No agility limitation, N/deaths 232/82 235/55 116/29 154/55 178/42 119/14HR (95% CI) 1 (Ref.) 0.88 (0.58–1.33) 0.74 (0.45–1.22) 0.828 1 (Ref.) 0.70 (0.44–1.12) 0.51 (0.27–0.98) 0.008

With agility limitation, N/deaths 317/180 236/85 109/36 559/282 345/74 273/35HR (95% CI) 1 (Ref.) 0.71 (0.51–1.00) 0.70 (0.44–1.10) 0.333 1 (Ref.) 0.67 (0.51–0.88) 0.52 (0.37–0.74) b0.001

No mobility limitation, N/deaths 287/103 323/77 161/39 187/51 234/46 168/22HR (95% CI) 1 (Ref.) 0.77 (0.51–1.16) 0.68 (0.44–1.06) 0.717 1 (Ref.) 0.75 (0.49–1.14) 0.78 (0.48–1.28) 0.234

With mobility limitation, N/deaths 262/160 148/62 64/26 526/286 289/70 224/27HR (95% CI) 1 (Ref.) 0.80 (0.56–1.16) 0.78 (0.44–1.27) 0.394 1 (Ref.) 0.67 (0.50–0.88) 0.40 (0.28–0.59) b0.001

No comorbidities, N/deaths 338/132 341/104 172/43 458/190 399/81 287/31HR (95% CI) 1 (Ref.) 0.95 (0.69–1.31) 0.80 (0.52–1.21) 0.987 1 (Ref.) 0.77 (0.55–1.07) 0.57 (0.39–0.84) 0.003

With comorbidities, N/deaths 211/131 130/36 54/21 255/147 123/35 106/18HR (95% CI) 1 (Ref.) 0.46 (0.28–0.75) 0.53 (0.32–0.87) 0.080 1 (Ref.) 0.62 (0.42–0.89) 0.44 (0.27–0.71) b0.001

HR: Hazard ratio. CI: Confidence interval. Analyses adjusted as in model 4 in Table 2.

17D. Martínez-Gómez et al. / Preventive Medicine 61 (2014) 14–19

The results of this study concur with two recent studies in olderadults reporting that HPA might reduce all-cause mortality (Chenet al., 2012; Ottenbacher et al., 2012). Other studies in middle-agedpeople have also obtained similar results. Autenrieth et al. found thatlight-intensity HPA was related to 18% lower all-cause mortality in4672 men and women aged 25–74 years (Autenrieth et al., 2011).Also, Stamatakis et al. (2009) reported that HPA might protect againstall-cause mortality in 13,726 individuals aged 35 years or older partici-pating in the Scottish Health Survey. Finally, Besson et al. (2008) found aprotective effect of HPA onmortality in 14,903 Englishmen andwomen.

A

>2 h/d(103/43)

>0-2 h/d(177/43)

0 h/d(158/59)

>2 h/d(122/22)

>0-2 h/d(294/97)

0 h/d(392/204)

Haz

ard

ratio

for

all-c

ause

mor

talit

y

1.50

1.25

1

0.75

0.50

0.25

0

≥8 h/d in sitting time <8 h/d in sitting time

Reference

0.80

0.43

0.97

0.64

0.90

(N / deaths)Household PA

P for interaction=0.039

Fig. 1. Mortality risk across categories of household physical activity (PA) and sitting time ameducational level (no education, primary, secondary or higher education), marital status (singlimitation (yes, no), mobility limitation (yes, no), coronary heart disease (yes, no), stroke (y(never, former, current smoker), alcohol drinking (never, former, current drinker), and leisure-by sitting time.

In contrast, no association was observed between HPA and mortality in7456 individuals aged 35–55 years from the Whitehall II Cohort Study(Sabia et al., 2012). This discrepancy with previous studies might bedue to the specific characteristics of the Whitehall II Cohort Study;participants were mainly men, employed, who spent only a short timein HPA (i.e., mean duration ≈ 24 min/d). However, none of thesestudies examined whether the study association was independent ofsitting time, which is a unique contribution of the present study.

Recent data using objectively measured PA show that older adultsspend most of their waking time in sedentary behaviors or in light-

B

>4 h/d(109/9)

>2-4 h/d(195/35)

≤2 h/d(201/67)

>4 h/d(186/32)

>2-4 h/d(327/81)

≤2 h/d(512/270)

1.50

1.25

1

0.75

0.50

0.25

0

≥8 h/d in sitting time <8 h/d in sitting time

Reference

0.68

0.57

0.35

0.64

0.82

P for interaction=0.671

ong older adults, Spain, 2003. A: men. B: women. Analyses were adjusted for age (years),le, married, divorced, widowed), living alone (yes, no), body mass index (kg/m2), agilityes, no), diabetes mellitus (yes, no), hip fracture (yes, no), and cancer (yes, no), smokingtime PA (min/d). P for interaction indicates whether the effect of HPA on mortality varies

18 D. Martínez-Gómez et al. / Preventive Medicine 61 (2014) 14–19

intensity PA (Evenson et al., 2012; Troiano et al., 2008). Moreover, olderindividuals usually replace sedentary time by light PA. Since there isevidence of the detrimental effect of sedentary behavior on mortalityin the elderly (Koster et al., 2012), which is independent of moderate-to-vigorous PA, it was unclear whether the beneficial effect of HPA –

performed mainly at light-intensity – on mortality was mediated bysedentariness. This study extends previous research by showing thatthe HPA is associated with reduced risk of death in (i) older women,regardless of sedentariness, and (ii) sedentary men.

Being physically active contributes to longer life expectancy(USDHHS, 2008a), and the time that women spend in household choresmay be important in this relationship since it is a way to get the PA theyneed. Older women engage in lower levels ofmoderate and vigorous PAthan older men (Evenson et al., 2012; Troiano et al., 2008). Yet, olderwomen are less sedentary than men (Evenson et al., 2012; Troianoet al., 2008). Probably, this apparent contradiction (i.e., both less PAand less sedentariness) means that women tend to spend more timethan men in light-intensity PA. A substantial part of this light-intensityPA might be obtained in household, which is the single domain of PA(i.e., transport, leisure-time, work and home) where women devotemore time than men, although levels of household activity in womenmay depend on cultural norms (Chen et al., 2012; Ottenbacher et al.,2012).

In the fully-adjusted analyses, HPA did not show an association withreducedmortality in oldermen,which contrastswith previous research(Chen et al., 2012; Ottenbacher et al., 2012). In fact, although thepartially-adjusted models showed a certain association with mortality,it disappeared after further adjustment. Similar results were alsoobserved in both apparently healthy and unhealthy (e.g., obese, withactivity limitations) men or after excluding deaths in the first years offollow-up. One possible explanation is that the time spent in HPA bymen, which was substantially smaller than that by women, was notsufficient to translate into reduced mortality because it representsonly a small part of their daily total PA. However, among sedentarymen,we did find a survival benefit related to HPA. Results of the presentstudy suggest that in this subgroup of men (who are usually very inac-tive in terms of moderate-to-vigorous PA), N2 h/d spent in HPA couldcounteract the adverse effect of too much sitting time on mortality.

Several mechanisms could account for the reducedmortality associ-ated with HPA in older adults (Löllgen et al., 2009). In cross-sectionalstudies, light-intensity PA has shown a favorable associationwith sever-al cardiometabolic risk factors, including waist circumference, glucose,triglycerides, HDL-cholesterol, and clustered metabolic risk (Camhiet al., 2011; Gardiner et al., 2011; Hawkins et al., 2012; Healy et al.,2007, 2008). Moreover, longer time spent in light-intensity PA mea-sured by accelerometry has been associated with reduced arterial stiff-ness (Gando et al., 2010). Recently, an intervention study found thatinterrupting sitting time with short bouts of light-intensity PA reducedpostprandial glucose and insulin levels in overweight and obese adults(Dunstan et al., 2012). Therefore, some researchers have suggestedthat light-intensity PA (also called the non-exercise part of the activitycontinuum) may be a more realistic and achievable approach to doingPA in individualswith activity limitations orwith very sedentary behav-ior (e.g., most older adults), because sitting time is not usually replacedby PA at moderate or vigorous intensity (Manns et al., 2012; Powellet al., 2011). Preliminary evidence suggests that a reduction of seden-tary time by increasing bouts of light-intensity PA is feasible in theolder population (Gardiner et al., 2011).

This study has some limitations. First, HPA was self-reported. Thus,despite the fact that questionnaire-based data on household choreshas been shown to be reasonably reliable and valid in older adults(Forsén et al., 2010), our results should be interpretedwith caution. Sec-ond, HPAwas obtained only at baseline. Although the analyses assumedthat this lifestyle is stable over time, some changes are still possible andwould likely have led to an underestimation of the protective effect ofHPA on mortality. Third, information on cause-specific mortality was

not available, so the specific pathways (cardiovascular, respiratory,etc.) for the study association are unknown. Fourth, information oncognitive function was obtained in 2001 with the Mini-Mental StateExamination (Faubel et al., 2009), but unfortunately it was not done inthe 2003 assessment, so it could not be included as a covariate in theanalyses. Lastly, neither was income considered, despite the fact that itcould be a confounder of the study association. This study also hadseveral strengths. Specifically, the analyses were adjusted for numerousconfounders, and the relatively large sample size allowed examiningwhether the study association varied according to socio-demographiccharacteristics (e.g., age) andhealth conditions (e.g., obesity and activitylimitations). These analyses suggest, for example, thatmost older adults(whether or not they have activity limitations) may benefit from HPA,although further testing in experimental studies is needed to supportthis assertion and for providing more precise estimates of benefit.

In conclusion, HPA was associated with a substantial reduction ofmortality regardless of sitting time in older women, and in sedentarymen. While current PA guidelines emphasize moderate-to-vigorousPA for maintaining or increasing health, they also recommend thatolder adults adapt their PA to their own capacity (USDHHS, 2008b;World Health Organization, 2010). Thus, the overarching recommenda-tion in health care settings for this population is to avoid being physical-ly inactive. This study provides further support for the claim that “somePA (even of light intensity) is better than none” (USDHHS, 2008b;World Health Organization, 2010). Future research should includeobjective measures of HPA (e.g., direct observation and SenseCam) toprovide new insights regarding its role on survival among older adults.

Contributors

DMG and FRA had full access to all of the data in the study and takeresponsibility for the integrity of the data and the accuracy of the dataanalysis. Study concept and design: DMG, PGC, and FRA. Acquisition ofdata: PGC, and LLM. Analysis and interpretation of data: DMG, PGC,and FRA. Drafting of the manuscript: DMG and FRA. Critical revision ofthe manuscript for important intellectual content: DMG, PGC, LLM,and FRA. Statistical analysis: DMG, PGC, and FRA. Obtained funding:PGC and FRA. Administrative, technical, or material support: PGC andLLM. Study supervision: FRA. All authors have read and approved thefinal manuscript.

Funding

This work was funded by FIS grant 12/1166, FP7-HEALTH-2012-Proposal No: 305483-2 (FRAILOMIC Initiative), and by the ‘CátedraUAM de Epidemiología y Control del Riesgo Cardiovascular’. DMG had a“Juan de la Cierva” contract from the Spanish Ministry of Science.

Conflict of interest statement

The authors declare that there are no conflicts of interests.

References

Autenrieth, C.S., Baumert, J., Baumeister, S.E., et al., 2011. Association between domains ofphysical activity and all-cause, cardiovascular and cancer mortality. Eur. J. Epidemiol.26, 91–99.

Balboa-Castillo, T., Guallar-Castillón, P., León-Muñoz, L.M., Graciani, A., López-García, E.,Rodríguez-Artalejo, F., 2011a. Physical activity and mortality related to obesity andfunctional status in older adults in Spain. Am. J. Prev. Med. 40, 39–46.

Balboa-Castillo, T., León-Muñoz, L.M., Graciani, A., Rodríguez-Artalejo, F., Guallar-Castillón, P., 2011b. Longitudinal association of physical activity and sedentary behav-ior during leisure time with health-related quality of life in community-dwellingolder adults. Health Qual. Life Outcomes 9, 47.

Besson, H., Ekelund, U., Brage, S., et al., 2008. Relationship between subdomains of totalphysical activity and mortality. Med. Sci. Sports Exerc. 40, 1909–1915.

Bicalho, P.G., Hallal, P.C., Gazzinelli, A., Knuth, A.G., Velásquez-Meléndez, G., 2010. Adultphysical activity levels and associated factors in rural communities of Minas GeraisState, Brazil. Rev. Saude Publica 44, 884–893.

19D. Martínez-Gómez et al. / Preventive Medicine 61 (2014) 14–19

Camhi, S.M., Sisson, S.B., Johnson, W.D., Katzmarzyk, P.T., Tudor-Locke, C., 2011.Accelerometer-determined moderate intensity lifestyle activity and cardiometabolichealth. Prev. Med. 52, 358–360.

Chen, L.J., Fox, K.R., Ku, P.W., Sun, W.J., Chou, P., 2012. Prospective associations betweenhousehold-, work-, and leisure-based physical activity and all-cause mortalityamong older Taiwanese adults. Asia Pac. J. Public Health 24, 795–805.

Csizmadi, I., Lo Siou, G., Friedenreich, C.M., Owen, N., Robson, P.J., 2011. Hours spent andenergy expended in physical activity domains: results from the Tomorrow Project co-hort in Alberta, Canada. Int. J. Behav. Nutr. Phys. Act. 8, 110.

Duncan, G.E., Goldberg, J., Buchwald, D., Wen, Y., Henderson, J.A., 2009. Epidemiology ofphysical activity in American Indians in the Education and Research Towards Healthcohort. Am. J. Prev. Med. 37, 488–494.

Dunstan, D.W., Kingwell, B.A., Larsen, R., et al., 2012. Breaking up prolonged sitting re-duces postprandial glucose and insulin responses. Diabetes Care 35, 976–983.

Evenson, K.R., Buchner, D.M., Morland, K.B., 2012. Objective measurement of physical ac-tivity and sedentary behavior among US adults aged 60 years or older. Prev. ChronicDis. 9, E26.

Faubel, R., López-García, E., Guallar-Castillón, P., Graciani, A., Banegas, J.R., Rodríguez-Artalejo, F., 2009. Usual sleep duration and cognitive function in older adults inSpain. J. Sleep Res. 18, 427–435.

Forsén, L., Loland, N.W., Vuillemin, A., et al., 2010. Self-administered physical activityquestionnaires for the elderly: a systematic review of measurement properties.Sports Med. 40, 601–623.

Gando, Y., Yamamoto, K., Murakami, H., et al., 2010. Longer time spent in light physical activ-ity is associatedwith reduced arterial stiffness in older adults. Hypertension56, 540–546.

Gardiner, P.A., Eakin, E.G., Healy, G.N., Owen, N., 2011. Feasibility of reducing older adults'sedentary time. Am. J. Prev. Med. 41, 174–177.

Guallar-Castillón, P., Rodríguez-Artalejo, F., Díez Gañán, L.D., Banegas Banegas, J.R.,Lafuente Urdinguio, P.L., Herruzo Cabrera, R.H., 2001. Consumption of alcoholic bev-erages and subjective health in Spain. J. Epidemiol. Community Health 55, 648–652.

Gutiérrez-Fisac, J.L., López, E., Banegas, J.R., Graciani, A., Rodríguez-Artalejo, F., 2004. Prev-alence of overweight and obesity in elderly people in Spain. Obes. Res. 12, 710–715.

Hallal, P.C., Knuth, A.G., Reis, R.S., et al., 2011. Time trends of physical activity in Brazil(2006–2009). Rev. Bras. Epidemiol. 14, S53–S60.

Hawkins, M., Belalcazar, L.M., Schelbert, K.B., Richardson, C., Ballantyne, C.M., Kriska, A.,2012. The effect of various intensities of physical activity and chronic inflammationin men and women by diabetes status in a national sample. Diabetes Res. Clin.Pract. 97, e6–e8.

Healy, G.N., Dunstan, D.W., Salmon, J., et al., 2007. Objectively measured light-intensityphysical activity is independently associated with 2-h plasma glucose. DiabetesCare 30, 1384–1389.

Healy, G.N., Wijndaele, K., Dunstan, D.W., et al., 2008. Objectively measured sedentarytime, physical activity, and metabolic risk: the Australian Diabetes, Obesity and Life-style Study (AusDiab). Diabetes Care 31, 369–371.

Khaing Nang, E.E., Khoo, E.Y., Salim, A., Tai, E.S., Lee, J., Van Dam, R.M., 2010. Patterns ofphysical activity in different domains and implications for intervention in a multi-ethnic Asian population: a cross-sectional study. BMC Public Health 10, 644.

Koster, A., Caserotti, P., Patel, K.V., et al., 2012. Association of sedentary time with mortal-ity independent of moderate to vigorous physical activity. PLoS One 7, e37696.

Löllgen, H., Böckenhoff, A., Knapp, G., 2009. Physical activity and all-cause mortality: anupdated meta-analysis with different intensity categories. Int. J. Sports Med. 30,213–224.

López-García, E., Faubel, R., León-Muñoz, L., Zuluaga, M.C., Banegas, J.R., Rodríguez-Artalejo, F., 2008. Sleep duration, general and abdominal obesity, and weight changeamong the older adult population of Spain. Am. J. Clin. Nutr. 87, 310–316.

Mann, K.D., Hayes, L., Basterfield, L., Parker, L., Pearce, M.S., 2013. Differing lifecourse as-sociations with sport-, occupational- and household-based physical activity at age49–51 years: the Newcastle Thousand Families Study. Int. J. Public Health 58, 79–88.

Manns, P.J., Dunstan, D.W., Owen, N., Healy, G.N., 2012. Addressing the nonexercise partof the activity continuum: a more realistic and achievable approach to activity pro-gramming for adults with mobility disability? Phys. Ther. 92, 614–625.

Marquez, D.X., Hoyem, R., Fogg, L., Bustamante, E.E., Staffileno, B., Wilbur, J., 2011. Physicalactivity of urban community-dwelling older Latino adults. J. Phys. Act. Health 8,S161–S170.

Martinez-Gonzalez, M.A., Lopez-Fontana, C., Varo, J.J., Sanchez-Villegas, A., Martinez, J.A.,2005. Validation of the Spanish version of the physical activity questionnaire usedin the Nurses' Health Study and the Health Professionals' Follow-up Study. PublicHealth Nutr. 8, 920–927.

Matthews, C.E., Chen, K.Y., Freedson, P.S., et al., 2008. Amount of time spent in sedentarybehaviors in the United States, 2003–2004. Am. J. Epidemiol. 167, 875–881.

Ng, S.W., Norton, E.C., Popkin, B.M., 2009. Why have physical activity levels declinedamong Chinese adults? Findings from the 1991–2006 China Health and NutritionSurveys. Soc. Sci. Med. 68, 1305–1314.

Ottenbacher, A.J., Snih, S.A., Karmarkar, A., et al., 2012. Routine physical activity and mor-tality in Mexican Americans aged 75 and older. J. Am. Geriatr. Soc. 60, 1085–1091.

Owen, N., Sparling, P.B., Healy, G.N., Dunstan, D.W., Matthews, C.E., 2010. Sedentary be-havior: emerging evidence for a new health risk. Mayo Clin. Proc. 85, 1138–1141.

Powell, K.E., Paluch, A.E., Blair, S.N., 2011. Physical activity for health: what kind? Howmuch? How intense? On top of what? Annu. Rev. Public Health 32, 349–365.

Sabia, S., Dugravot, A., Kivimaki, M., Brunner, E., Shipley, M.J., Singh-Manoux, A., 2012. Ef-fect of intensity and type of physical activity on mortality: results from theWhitehallII Cohort Study. Am. J. Public Health 102, 698–704.

Stamatakis, E., Hamer, M., Lawlor, D.A., 2009. Physical activity, mortality, and cardiovascu-lar disease: is domestic physical activity beneficial? The Scottish Health Survey —1995, 1998, and 2003. Am. J. Epidemiol. 169, 1191–1200.

Troiano, R.P., Berrigan, D., Dodd, K.W., Mâsse, L.C., Tilert, T., McDowell, M., 2008. Physicalactivity in the United States measured by accelerometer. Med. Sci. Sports Exerc. 40,181–188.

United StatesDepartment ofHealth andHumanServices, 2008a. Physical ActivityGuidelinesAdvisory Committee Report. http://www.health.gov/paguidelines/committeereport.aspx.

United States Department of Health and Human Services, 2008b. Physical Activity Guide-lines for Americans. http://www.health.gov/paguidelines/guidelines/default.aspx.

van der Ploeg, H.P., Chey, T., Korda, R.J., Banks, E., Bauman, A., 2012. Sitting time and all-cause mortality risk in 222 497 Australian adults. Arch. Intern. Med. 172, 494–500.

World Health Organization, 2010. Global recommendations on physical activity for health.Geneva. http://www.who.int/dietphysicalactivity/factsheet_recommendations/en/index.html.