Lifetime Physical Activity and the Risk of ALS
In this population based study, 636 (84%) of the 760 patients who gave informed consent to participate in the study between 1 January 2006 and 31 December 2010 returned the questionnaire. Of the 2332 population based controls who gave informed consent, 93% returned their questionnaires (2166 controls). Table 1 shows the characteristics of the 636 patients and 2166 controls. The patient characteristics of the responders and non-responders were similar. Of the 2802 participants, 2281 (81.4%) had completed the questionnaires on physical activities without any missing values for duration in years or hours per week. The distributions for gender, age at onset and site of onset in ALS patients were similar to those previously reported in population based studies.
A greater amount of leisure time physical activity was associated with an increased risk of ALS in the present study (adjusted OR 1.08, p=0.008) (Table 2). This is also illustrated in figure 1, showing the mean cumulative scores for leisure time activity (patient mean = 1.51, 95% CI 1.30 to 1.72; control mean = 1.25, 95% CI 1.18 to 1.32; p=0.004). Occupational and total physical activity were not associated with the risk of ALS (Table 2); no dose–response relationship was seen with physical activity (figure 2) and none of the vigorous physical activities showed a significant association with ALS (Table 3).
(Enlarge Image)
Figure 1.
Mean leisure time activity for amyotrophic lateral sclerosis (ALS) patients and controls. Patient mean=1.51, 95% CI 1.30 to 1.72; control mean=1.25, 95% CI 1.18 to 1.32.
(Enlarge Image)
Figure 2.
ORs with 95% CIs for the relationship between quartiles of leisure time, occupational and total activity, and the risk of amyotrophic lateral sclerosis. ORs were adjusted for gender, age at onset, body mass index, current smoking, current alcohol consumption and level of education. The physical activity score was categorised into quartiles (Q) based on the data of controls. Q1, 1st quartile; Q2, 2nd quartile; Q3, 3rd quartile; Q4, 4th quartile.
Maximum MET scores did not differ significantly between ALS patients and controls, implying that there was no difference in the maximum intensity of activities (all p values >0.35, not shown).
Survival analyses showed that none of the cumulative measures of physical activity was associated with survival (all p values >0.10). Of 636 patients, 63% died before the censoring date of 8 August 2011. The cumulative measures of leisure time, occupational and total activity did, however, show a significant relation with age at onset (all HR 0.94–0.95, p≤0.009). In order to show whether this effect was specific for patients or valid for age at the time of the questionnaire for controls, two additional analyses were performed: (1) an interaction term of diagnosis and physical activity was introduced into the model (all p values >0.45) and (2) a multivariate Cox regression was performed in controls using questionnaire completion as the event (p≤0.002). Both indicated that the relationship between physical activity and age at onset was an age related effect and thus not disease related. Kaplan–Meier curves of total activity of both survival and age at onset are shown in the online supplementary figure S1.
Results
In this population based study, 636 (84%) of the 760 patients who gave informed consent to participate in the study between 1 January 2006 and 31 December 2010 returned the questionnaire. Of the 2332 population based controls who gave informed consent, 93% returned their questionnaires (2166 controls). Table 1 shows the characteristics of the 636 patients and 2166 controls. The patient characteristics of the responders and non-responders were similar. Of the 2802 participants, 2281 (81.4%) had completed the questionnaires on physical activities without any missing values for duration in years or hours per week. The distributions for gender, age at onset and site of onset in ALS patients were similar to those previously reported in population based studies.
A greater amount of leisure time physical activity was associated with an increased risk of ALS in the present study (adjusted OR 1.08, p=0.008) (Table 2). This is also illustrated in figure 1, showing the mean cumulative scores for leisure time activity (patient mean = 1.51, 95% CI 1.30 to 1.72; control mean = 1.25, 95% CI 1.18 to 1.32; p=0.004). Occupational and total physical activity were not associated with the risk of ALS (Table 2); no dose–response relationship was seen with physical activity (figure 2) and none of the vigorous physical activities showed a significant association with ALS (Table 3).
(Enlarge Image)
Figure 1.
Mean leisure time activity for amyotrophic lateral sclerosis (ALS) patients and controls. Patient mean=1.51, 95% CI 1.30 to 1.72; control mean=1.25, 95% CI 1.18 to 1.32.
(Enlarge Image)
Figure 2.
ORs with 95% CIs for the relationship between quartiles of leisure time, occupational and total activity, and the risk of amyotrophic lateral sclerosis. ORs were adjusted for gender, age at onset, body mass index, current smoking, current alcohol consumption and level of education. The physical activity score was categorised into quartiles (Q) based on the data of controls. Q1, 1st quartile; Q2, 2nd quartile; Q3, 3rd quartile; Q4, 4th quartile.
Maximum MET scores did not differ significantly between ALS patients and controls, implying that there was no difference in the maximum intensity of activities (all p values >0.35, not shown).
Survival analyses showed that none of the cumulative measures of physical activity was associated with survival (all p values >0.10). Of 636 patients, 63% died before the censoring date of 8 August 2011. The cumulative measures of leisure time, occupational and total activity did, however, show a significant relation with age at onset (all HR 0.94–0.95, p≤0.009). In order to show whether this effect was specific for patients or valid for age at the time of the questionnaire for controls, two additional analyses were performed: (1) an interaction term of diagnosis and physical activity was introduced into the model (all p values >0.45) and (2) a multivariate Cox regression was performed in controls using questionnaire completion as the event (p≤0.002). Both indicated that the relationship between physical activity and age at onset was an age related effect and thus not disease related. Kaplan–Meier curves of total activity of both survival and age at onset are shown in the online supplementary figure S1.
Source...