Centenarians are a fascinating population that has increased in numbers since the middle of the last century. Factors such as aging of the general population, improvement in social and medical conditions, genetic and personal characteristics of each elderly individual as well as environmental aspects, may contribute towards the capacity to survive for more than 100 years. It has been estimated that between 2000 and 2050, the number of centenarians will increase more than 16-fold, with around 850,000 centenarians living in the US in 2050, for example [1]. In Portugal, the number of centenarians has almost tripled over the last 10 years from 589 in 2001 [2] to 1526 in 2011 [3] and recent projections from the National Institute of Statistics of Portugal indicate a substantial increase in these numbers in the future.

At present, the full mechanisms associated with reaching 100 years of life or longer are not completely understood. A genetic factor has been suggested in family studies, and many centenarians report very good health into their nineties, having been able to avoid critical periods in their lives. The living environment also seems to play an important part, and Sardinia (Italy), New England (US) and some regions in Japan are examples of places where longevity seems to be a rule for a significant number of people [4].

Health conditions associated with extreme aging are not simple: cognitive decline, muscle weakness and immobilization are present in a large number of very old individuals, frequently in association [5]. There is, however, great heterogeneity regarding these health parameters in near-centenarians and centenarians, with functional capacity ranging from being bedridden to full working capacity. One of the first centenarian studies involving German citizens showed that 20–30% were still completely active, 50% had some type of dementia and 20–30% were bedridden and severely impaired [6]. More recently a study of Chinese centenarians (369 male and 1751 female) showed good self-reported health in 56% of male and 50.6% of female centenarians [7]; in the same study, 51% of the men and 46.4% of the women were still active and 43.3% of men and 25.4% of women still had normal or high cognitive scores. Similar results were reported in the Ipsen survey in France, which showed that 65.2% of centenarian men still had good or very good health, compared with 56.5% of female centenarians. In this study significantly more men than women had normal or high cognitive test results (39.1% versus 17.9%, respectively [8]. In contrast, in the Tokyo centenarian study only 20% of centenarians enjoyed physical and cognitive well being, regardless of their sex [9].

Physical exercise, either aerobic or with an emphasis upon specific motor abilities, has been shown to improve cognition in elderly people [10]. Both types of physical exercise are important to an independent life and previous research showed an association between exercise and cognition in middle aged and elderly individuals [11]. Gait speed is an important factor in old age, and constitutes a predictive indicator of functional capacity in older adults [12]. Handgrip strength was related to cognition in very old subjects: those with the lower values in handgrip were at higher risk of subsequent functional decline and mortality [13]. Furthermore, handgrip strength and walking test results have been shown as prognostic indicators of subsequent mortality in hospitalized elderly people [14]. Several authors have shown an association between handgrip strength and physical and cognitive functionality in old and very old people and also that handgrip strength is an important predictor of functionality in very old people [13–16].

However, it is not known whether handgrip strength (and walking parameters) are related to cognitive functions in centenarians. Our hypothesis is that both parameters are still related even in very old subjects. Thus, this study aims to evaluate a group of centenarians from two regions of Portugal (Oporto and Beira Interior) and to look for any relationship between their cognitive status, their disease burden and some measures of their physical capacity, namely handgrip and walking parameters.

To the best of our knowledge, this is the first study on physical performance and cognition performed with a sample of Portuguese centenarians. Results showed a significant correlation between cognitive status and some motor function in this group of people older than 100 years (namely handgrip strenght and time spent in the three metre walk). In contrast, no correlation was found between handgrip strength and walking velocity or between physical performance and the presence/absence of severe disease.

The significant correlation that we found between some aspects of physical performance and cognitive scores in centenarians seems important [12, 15]. As individuals grow old, motor abilities are reduced as compared with younger adults. The loss of muscular mass correlates with loss of strength and can be assessed using the handgrip test, which is an indirect measure of health and muscular function [27, 28]. Cognitive changes are also identified in healthy elderly people but, if they are not related to disease, such changes can be compensated. Elderly people achieve cognitive responses which are similar to those seen in younger adults although this seems to occur through a more expanded use of cortical areas, as seen using neuroimaging studies [29, 30]. This expanded utilization of cortical areas can occur with simple motor tasks as well as with more complex motor activities involving coordination, either because old people need to aggregate more motor neurons in different areas or because they need to activate more networks in order to achieve the same result as that obtained by younger adults [31, 32]. In future studies, neuroimaging of centenarian brains with functional MRI may help to clarify this issue.

Handgrip is an useful single marker of frailty and an important predictor of functionality [15, 16]. The relationship found between handgrip and cognitive scores is important, since both items are related to the autonomous functionality of the subjects. A co-dependence of cognition and motor activity has been found in several research studies. For example, the performance in the Trail Making Test, when impaired, is associated with a reduction in walking velocity, particularly with obstacles, in older individuals [11]. Furthermore, walking performance and several neuropsychological tests, including memory, language and executive functions seem to be correlated [33]. It is possible that in centenarians, as in non-centenarian elderly people, motor function and cognition may share some neural networks due to a reduction in neural substrates, and also because cognition is necessary for motor function, or because more extended networks are necessary for both functions. We can also argue that this is probably what conveys the positive effect of physical exercise over neurodegenerative disorders. It has been demonstrated that the increase in complexity of movements and coordination tasks and other types of motor interventions in these patients are associated with an improvement in cognition [10, 32–35]. Different exercise programmes in healthy adults seem to lead to similar cognitive improvements. In a group of older adults both progressive resistance training and multi component training (balance, coordination and resistance) were similarly effective at improving cognition [35], and coordination training and cardiovascular training produced similar positive cognition results in a group of older adults although these types of training activate different areas of the brain [10, 35–37]. On the other hand, several longitudinal studies showed that old people with walking difficulties have a higher risk of developing dementia [38, 39]. Taken together with our results, these studies may suggest that the activation of motor networks of different types involves the activation of cognitive networks.

We could not find any correlation between handgrip strength and time to perform the three metre tests in our sample of centenarians. We could expect that both physical functions would be correlated, since they represent the motor capacity of the subject. However the neural subtract for both characteristics is different: strength seems to be a pure motor function related to the performance of the motor cortex and the peripheral nervous system and muscles, while walking implies a much more sophisticated situation, associating motor cortex, basal ganglia and cerebellum. The absence of correlation observed in our study may thus result from these different neural processes [40].

We would expect centenarians to show some cognitive decline, muscular weakness or immobilization or even a joint decline in these three parameters. Centenarians are survivors of several bouts of diseases but these may present clinically in various ways [7–9, 41]. In our study, about 70% of the individuals had some form of severe disease. However, the ones with severe disease did not show any statistical difference in motor function or cognition when compared with the ones with no disease. This result may be related to the small number of subjects we studied, but it may also be due to the fact that the questionnaire used to obtain information about chronic diseases was very vague, associating information about several disorders but without defining the degree of severity of each one. It is possible that the use of a more detailed inquiry with a gradient of severity could yield a more precise result. It is also possible that, due to the exceptional resilience that being a centenarian represents, disease may not affect the motor and cognitive function in the same way as with younger subjects. Further studies are needed to ascertain this.

Our study had some limitations. Firstly, our strategy was to collect data from the subjects with information on three tests – handgrip, walking and SMMSE; we did not use data from centenarians who, for some reason, did not have one of the three measures. In setting this exclusion criterion, it is possible that we had some form of selection of the more competent people in both motor and cognitive areas. Secondly, although the short version of the MMSE is the most frequent test used in centenarians, due to its simplicity, it is also its simplicity that warrants some degree of caution, because it gives a rough approach to cognitive status.

Our study found that there is a positive association between motor and cognitive function in centenarians. This suggests that even in very old individuals, an association between motor capacity and cognition is still present.