Singh-Manoux A, Kivimaki M, Glymour MM, Elbaz A, Berr C, Ebmeier KP, Ferrie JE, Dugravot A.
Timing of onset of cognitive decline: results from Whitehall II prospective cohort study.
BMJ. 2011;344:d7622.
PubMed.
This study by Singh-Manoux and colleagues is an important addition to the characterization of cognitive change during middle age and older adulthood. It stands out because of the large sample size (7,390 subjects) and the length of time over which the subjects were followed (10 years).
Rarely do we have an opportunity to examine cognitive change over a decade across so many individuals. It is important to point out that these results confirm what much of the literature on cognitive aging has already shown. In the 2004 review that they cite, our statement that there was little evidence for age-related decline before the age of 60 specifically referred to the known results from the few longitudinal studies existing at that time. We went on to say that many cross-sectional studies did find age-related declines across the entire adult lifespan (from age 20 to 90).
In particular, we noted a lack of knowledge regarding individuals in midlife (from 30 to 60). We pointed out the mismatch between longitudinal and cross-sectional data during midlife, and noted that both types of designs have their own pitfalls; the truth is likely to reside somewhere between them. One cross-sectional study in which I was involved showed that cognitive differences from age 30-40 occurred to nearly the same extent as those from age 60-70 (Park et al., 2002). Many other studies have shown similar results. It is important, but not surprising, that the authors have been able to confirm such results with longitudinal data. By filling in the gap between 45 and 60, this new study helps us see that longitudinal change in the same people can be detected with the power of a large sample, and that perhaps the cross-sectional estimates were not far off the mark.
One other important point is that, although the cognitive changes detected in this study were statistically significant, this was due to the large sample size having immense power to detect quite small effects. For all categories of cognitive skills except reasoning, the age-related decline in scores over 10 years amounted to about one item or less. For example, on the memory test, the average 45- to 49-year-old was able to remember 0.58 of an item less when tested again 10 years later. Such small changes may not be detectable within a person in the clinic. For example, if people coming into a clinic have a “true” memory score of 8 out of 20, they might score a 9 when tested on Monday, but a 7 on Friday, even though the test could still be considered highly reliable. If those people come in again in 10 years and score a 7, it would be almost impossible to determine if that is because they’re being tested on a Friday, or because their “true” score has declined to 7 and the current score is reflecting that change.
This study can detect these small changes because it averages over thousands of people, but knowing that such small changes occur will have limited usefulness in practice. This may also be why previous longitudinal studies with smaller samples failed to detect cognitive change at younger ages. This study does provide an important indication of how much cognitive change over a decade might be considered typical during middle adulthood, so that individuals exhibiting higher levels of change over time could be targeted for potential interventions.
References:
Park DC, Lautenschlager G, Hedden T, Davidson NS, Smith AD, Smith PK.
Models of visuospatial and verbal memory across the adult life span.
Psychol Aging. 2002 Jun;17(2):299-320.
PubMed.
Comments
Massachusetts General Hospital
This study by Singh-Manoux and colleagues is an important addition to the characterization of cognitive change during middle age and older adulthood. It stands out because of the large sample size (7,390 subjects) and the length of time over which the subjects were followed (10 years).
Rarely do we have an opportunity to examine cognitive change over a decade across so many individuals. It is important to point out that these results confirm what much of the literature on cognitive aging has already shown. In the 2004 review that they cite, our statement that there was little evidence for age-related decline before the age of 60 specifically referred to the known results from the few longitudinal studies existing at that time. We went on to say that many cross-sectional studies did find age-related declines across the entire adult lifespan (from age 20 to 90).
In particular, we noted a lack of knowledge regarding individuals in midlife (from 30 to 60). We pointed out the mismatch between longitudinal and cross-sectional data during midlife, and noted that both types of designs have their own pitfalls; the truth is likely to reside somewhere between them. One cross-sectional study in which I was involved showed that cognitive differences from age 30-40 occurred to nearly the same extent as those from age 60-70 (Park et al., 2002). Many other studies have shown similar results. It is important, but not surprising, that the authors have been able to confirm such results with longitudinal data. By filling in the gap between 45 and 60, this new study helps us see that longitudinal change in the same people can be detected with the power of a large sample, and that perhaps the cross-sectional estimates were not far off the mark.
One other important point is that, although the cognitive changes detected in this study were statistically significant, this was due to the large sample size having immense power to detect quite small effects. For all categories of cognitive skills except reasoning, the age-related decline in scores over 10 years amounted to about one item or less. For example, on the memory test, the average 45- to 49-year-old was able to remember 0.58 of an item less when tested again 10 years later. Such small changes may not be detectable within a person in the clinic. For example, if people coming into a clinic have a “true” memory score of 8 out of 20, they might score a 9 when tested on Monday, but a 7 on Friday, even though the test could still be considered highly reliable. If those people come in again in 10 years and score a 7, it would be almost impossible to determine if that is because they’re being tested on a Friday, or because their “true” score has declined to 7 and the current score is reflecting that change.
This study can detect these small changes because it averages over thousands of people, but knowing that such small changes occur will have limited usefulness in practice. This may also be why previous longitudinal studies with smaller samples failed to detect cognitive change at younger ages. This study does provide an important indication of how much cognitive change over a decade might be considered typical during middle adulthood, so that individuals exhibiting higher levels of change over time could be targeted for potential interventions.
References:
Park DC, Lautenschlager G, Hedden T, Davidson NS, Smith AD, Smith PK. Models of visuospatial and verbal memory across the adult life span. Psychol Aging. 2002 Jun;17(2):299-320. PubMed.
Make a Comment
To make a comment you must login or register.