This study is the first to investigate the relation between objectively measured physical fitness and hospitalization due to HLDD in a prospective design. The hypothesis that low physical fitness is a risk factor for hospitalization due to HLDD was not supported.
As mentioned in the introduction, a beneficial effect on low back disorders should be expected with increasing leisure time physical activity (LTPA) sufficiently intense to increase cardio-respiratory fitness. However, in this study, neither self-reported LTPA nor the objectively measured physical fitness was predictive of later hospitalization due to HLDD. From a theoretical point of view, physical activity should increase the resistance of the intervertebral discs as discussed by Nachemson [19
]. So, it cannot be ruled out that the negative finding in our study might be caused by a lack of statistical power (a type 2 error). Still, not even a tendency of an association was observed with an incidence of HLDD of 1.7% among men with a rather low physical fitness (mean (SD) 27.5 (3.5)), and 1.7% among men with a quite high physical fitness level (mean (SD) 38.8 (5.4)). Also, age-adjusted mean baseline values of physical fitness were practically identical for those who were later hospitalized due to HLDD and those who were not, 32.74 and 32.75, respectively (not shown in table). Furthermore, to gain an even larger exposure contrast between men with a high and a low physical fitness, we compared the incidence of HLDD between the lowest and the highest quintiles of physical fitness (not shown in table). The incidence among men with a fitness level of 15–26 (lowest quintile) was 1.9%; among men in the highest quintile 39–78 (highest quintile) it was 2.1%. Based on the above reasoning and the results, we find it unlikely that a lack of statistical power explains our findings.
Since no previous studies have investigated the association between physical fitness and HLDD, we are not able to compare our results with those of others. Previous studies on low back pain may indicate that a high level of physical fitness may be associated with less LBP as demonstrated by Heneweer et al. [5
] in a cross-sectional case–control study. However, as mentioned, only a small fraction of people who suffer from low back pain do so due to HLDD [20
]. Therefore, it is plausible that high physical fitness may prevent low back pain, but not HLDD.
As previously shown and indicated in Table in this paper, strenuous occupational work is a significant predictor for risk of hospitalization due to HLDD. It is well known that strenuousness at work is a combined factor of both the external physical work demand and the physical fitness of the worker [21
]. However, strenuous work tasks in this study include lifting of heavy burdens and awkward body positions, which may rather call for high muscular capacity than aerobic capacity. Occupational exposure to strenuous work tasks may not necessarily induce a training effect in muscles and connective tissues in the low back region. This is supported by a former study among elite athletes [22
]. In contrast, strenuous exposures may be harmful because it may overload tissues of the lower back [19
]. Instead, physical training during leisure has shown to improve capacity and reduce risk of low back disorders [23
], but Table also shows that no effect is seen from leisure time physical activity on HLDD. This indicates that strenuousness at work is the most important modifiable risk factor for hospitalization due to HLDD among men. Men in sedentary jobs may have HLDD not leading to hospitalization, but being treated as out-patients more often than men in strenuous jobs, a possibility which cannot be evaluated in this study.
This study has the advantage of data from objectively measured physical fitness and an objectively evaluated low back disease. Furthermore, the study has the advantage of a prospective design among individuals without back disorders at baseline. In this study, physical fitness is given as aerobic capacity relative to body weight and therefore, naturally, fitness levels are highly associated with weight levels. The univariate analysis presents a high association between body weight and HLDD suggesting that weight might modify a possible association between high aerobic capacity and HLDD. Future studies should therefore investigate the role of directly measured aerobic capacity and the risk of HLDD and separately evaluate the possible modifying effect of body weight. Fitness varies considerably across age groups. For example, previously defined fitness categories suggest that a physical fitness of ~30 ml O2
can be defined as “very poor” among males below the age of 30 and “very good” among males above the age of 75 [24
]. In the current study, fitness above 32 ml O2
has been defined as high fitness both for younger as well as for older males as the stratification of fitness in low and high categories in the current study was also statistically derived (i.e. based on the median value of the whole population). This may introduce a skewed distribution of age in the two fitness groups – i.e. many young males will fall in the category of high physical fitness and many older males will fall in the category of low physical fitness. However, since the age in this study ranged from 40–58 years, the median fitness of 32 ml O2
only ranged from poor to fair according to previous categorisations. Thus, due to the relatively narrow age span, the skewness can be overcome by controlling for age in all analyses. The sufficiency of this action is supported by the fact that the current stratification of fitness previously has proven highly valid for the determination of other health outcomes [7
]. However, the lack of repeated measures of physical fitness during the relatively long follow-up period may contribute to misclassification of physical fitness. One further limitation is the relatively small number of end-points, narrowing the possibility of conducting interaction analyses and increasing risk of a type-2 statistical error. Moreover, the inclusion of only middle-aged Caucasian men may make our results less relevant for other age groups, ethnic groups, and women.