Electronic Letters to:
|
|
eLetters: Electronic letters published:
-
![[Read eLetter]](/icons/misc/sectionDOWN.gif)
Re: Skinfold thickness
- Joseph Kim, Tom Meade, and Andy Haines (12 April 2006)
-
BMI and mortality risk
- Dr Rachel Gill, Hywell Dinsdale Epidemiological Analyst (21 March 2006)
|
|
|||
|
Joseph Kim, Lecturer London School of Hygiene and Tropical Medicine, Tom Meade, and Andy Haines
Send letter to journal:
joseph.kim{at}lshtm.ac.uk Joseph Kim, et al.
|
Dear Dr Gill, The decision to use fourths of BMI, rather than a more traditional five-group categorisation of BMI (i.e., < 18 kg/m2, 18-25 kg/m2, 25-30 kg/m2, 30-35 kg/m2, > 35 kg/m2) was based on the availability of data. Though we considered using the more traditional approach, there were insufficient numbers of subjects and fatal events in the extreme BMI categories to justify its use. There were only 3 all-cause deaths (1 CHD- related) among the 6 subjects with BMI < 18 kg/m2 and 10 deaths (5 CHD -related) among the 18 subjects with BMI > 35 kg/m2. Thus, we were unable to show the “U”-shaped curve found in previous studies. The all-cause mortality group included participants who had died from a variety of conditions. Only around one-third of deaths (even less in women) were due to CHD so it is unsurprising particularly given the limited sample size no significant relationship was observed. The relationship between obesity and total mortality may vary from study to study depending on the contribution of different causes of death to overall mortality. Further research involving competing risk analysis would certainly help clarify this point. |
|||
|
|
|||
|
Dr Rachel Gill, Public Health Doctor Directorate of Public Health, Adur, Arun and Worthing Teaching PCT, Hywell Dinsdale Epidemiological Analyst
Send letter to journal:
Rachel.gill{at}aaw.nhs.uk Dr Rachel Gill, et al.
|
Dear Editors, We were surprised by the conclusions drawn by Kim, Meade and Haines in their recent article ‘Skinfold thickness, body mass index, and fatal coronary heart disease: 30 year follow up of the Northwick Park heart study’, published in your journal. Whilst table one does show a significant difference between the distribution of BMI in the cohort who die of CHD and those who do not, the same data as presented in tables 2A and 2B suggest a linear relationship. This contradicts the U-shaped relationship reported by most studies of all cause mortality and BMI.[1][2] A linear relationship suggests that even an underweight male who increased their BMI by 3kg/m2 (3.5kg/m2 in women) to move into the ‘normal’ BMI range would have a 29% increase risk of fatal CHD (48% in women), which seems unlikely. An increase in CHD risk for individuals with low BMI may have been disguised by the groupings employed by the researchers in table three. The ‘standard’ group includes underweight individuals, which seems unwise even when forced by small numbers. The researchers claim to have replicated their findings when using the more standard groupings of BMI employed by Flegal and colleagues. This can only mean they found an increased risk of CHD death in the overweight and obese categories, yet like Flegal et al, they observed no increase in all cause mortality in the slightly overweight and moderately obese groups. Given the large proportion of deaths CHD accounts for (33% in this sample), it seems surprising that the increase due to CHD did not come through in the all cause analysis. The only way to reconcile these findings would be if slightly overweight or obese individuals had a slightly reduced mortality risk from causes of death other than CHD. |
|||
