Childhood Body Weight and CVD, Cancer in Older Adulthood
Childhood Body Weight and CVD, Cancer in Older Adulthood
Taking the results from the present study together, we found that there was little evidence of a clear relationship between BMI measured at 11 years of age and 9 independent morbidities assessed up to 67 years later. The occasional positive associations that were apparent—with smoking-related cancers and peripheral vascular disease—might have been generated by chance alone, given the large number of models necessarily conducted in the course of our analyses.
Our results agree with some existing findings. In a systematic review in which the authors stratified studies into age at BMI assessment (<7, 7–17, or 18–30 years), investigators found that the positive relation with coronary heart disease was strongest in the older group (7 studies) and null in the youngest (3 studies), whereas the aggregated result in the intermediate age group (7 studies) for a 1–standard deviation increase in BMI (hazard ratio = 1.09, 95% confidence interval: 1.00, 1.07) somewhat resembled our own (hazard ratio = 1.00, 95% confidence interval: 0.94, 1.07). In a separate review of 8 studies that featured stroke as the outcome of interest, 3 studies had null results, which is in keeping with the finding herein. To the best of our knowledge, ours is the first examination of the link between childhood BMI and other presentations of CVD, particularly heart failure and peripheral vascular disease. Results from a few studies have suggested a positive association between childhood BMI and colorectal cancers, but null results have also been reported, as they have for lung cancer. The mean and standard deviations for BMI in the present study approximate those from other studies in similar era and with similar ages at BMI assessment, including samples drawn from the United Kingdom. These studies also revealed positive associations with coronary heart disease; therefore, it is perhaps unlikely to be the narrow distribution of BMI in our cohort members, one that is very lean by contemporary standards, that is responsible for the negative results.
Although the present study has a series of strengths, including the unusually comprehensive range of health endpoints, the use of direct measurements of childhood BMI rather than adult recall, the high proportion of the original study members traced, and the utilization of a nationally representative sample, it is of course not without its weaknesses. The previously described tracking of BMI between childhood and adulthood means that the apparently increased rates of peripheral vascular disease in study members with higher weight in early life may instead be due to adult overweight. We had no repeated measurement of BMI with which to test this hypothesis. Although we were able to analyze the relationships of BMI with an array of health outcomes, there were too few cancer events to facilitate analyses for selected malignancies, such as head and neck cancer and pancreatic cancer. A further potential explanation for our generally negative results is that, unlike in other published analyses, we used nonfatal endpoints only. However, there is evidence that, at least for BMI measured in middle-aged adults, similar findings are apparent for both CVD and cancer irrespective of whether the outcome is incidence (nonfatal events) or mortality. Consistent with our various ethical agreements, the morbidity and mortality data have each been merged separately with the early-life data, creating 2 distinct data sets. Because of this, it was not possible for us to ascertain whether there was a different relationship of pre-adult BMI with the onset of a first event via hospitalization (etiology) or survival from it (prognosis). It is also the case that if an obese adult who was also obese in childhood died suddenly without any recorded morbidities, he or she might contribute to the finding of no association between childhood obesity and later morbidity risk. Again, however, the absence of linked morbidity and mortality data does not facilitate scrutiny of this hypothesis.
In conclusion, a link between pre-adult body weight and later morbidity was largely lacking in the present study. The general paucity of studies in this field should be addressed.
Discussion
Taking the results from the present study together, we found that there was little evidence of a clear relationship between BMI measured at 11 years of age and 9 independent morbidities assessed up to 67 years later. The occasional positive associations that were apparent—with smoking-related cancers and peripheral vascular disease—might have been generated by chance alone, given the large number of models necessarily conducted in the course of our analyses.
Comparison With Other Studies
Our results agree with some existing findings. In a systematic review in which the authors stratified studies into age at BMI assessment (<7, 7–17, or 18–30 years), investigators found that the positive relation with coronary heart disease was strongest in the older group (7 studies) and null in the youngest (3 studies), whereas the aggregated result in the intermediate age group (7 studies) for a 1–standard deviation increase in BMI (hazard ratio = 1.09, 95% confidence interval: 1.00, 1.07) somewhat resembled our own (hazard ratio = 1.00, 95% confidence interval: 0.94, 1.07). In a separate review of 8 studies that featured stroke as the outcome of interest, 3 studies had null results, which is in keeping with the finding herein. To the best of our knowledge, ours is the first examination of the link between childhood BMI and other presentations of CVD, particularly heart failure and peripheral vascular disease. Results from a few studies have suggested a positive association between childhood BMI and colorectal cancers, but null results have also been reported, as they have for lung cancer. The mean and standard deviations for BMI in the present study approximate those from other studies in similar era and with similar ages at BMI assessment, including samples drawn from the United Kingdom. These studies also revealed positive associations with coronary heart disease; therefore, it is perhaps unlikely to be the narrow distribution of BMI in our cohort members, one that is very lean by contemporary standards, that is responsible for the negative results.
Strengths and Limitations
Although the present study has a series of strengths, including the unusually comprehensive range of health endpoints, the use of direct measurements of childhood BMI rather than adult recall, the high proportion of the original study members traced, and the utilization of a nationally representative sample, it is of course not without its weaknesses. The previously described tracking of BMI between childhood and adulthood means that the apparently increased rates of peripheral vascular disease in study members with higher weight in early life may instead be due to adult overweight. We had no repeated measurement of BMI with which to test this hypothesis. Although we were able to analyze the relationships of BMI with an array of health outcomes, there were too few cancer events to facilitate analyses for selected malignancies, such as head and neck cancer and pancreatic cancer. A further potential explanation for our generally negative results is that, unlike in other published analyses, we used nonfatal endpoints only. However, there is evidence that, at least for BMI measured in middle-aged adults, similar findings are apparent for both CVD and cancer irrespective of whether the outcome is incidence (nonfatal events) or mortality. Consistent with our various ethical agreements, the morbidity and mortality data have each been merged separately with the early-life data, creating 2 distinct data sets. Because of this, it was not possible for us to ascertain whether there was a different relationship of pre-adult BMI with the onset of a first event via hospitalization (etiology) or survival from it (prognosis). It is also the case that if an obese adult who was also obese in childhood died suddenly without any recorded morbidities, he or she might contribute to the finding of no association between childhood obesity and later morbidity risk. Again, however, the absence of linked morbidity and mortality data does not facilitate scrutiny of this hypothesis.
In conclusion, a link between pre-adult body weight and later morbidity was largely lacking in the present study. The general paucity of studies in this field should be addressed.
Source...