Response to ‘Letter to the Editor’ regarding the article ‘Factors associated with reaching 90 years of age: a study of men born in 1913 in Gothenburg, Sweden’


Lars Wilhelmsen. Fregattgatan 16, SE-42674, V.Frölunda, Sweden.
(fax: +46-31-295182; e-mail:

Dear Sir,

We thank for the comments and respond for the four items below.

  • 1 We have been careful in not claiming that the results ‘provide a support for public health policies’, mainly because there are no clinical trial results that support such a policy. The areas under the ROC curves are never higher than 0.62304, and for the combination 0.682, which we agree, is not very high. However, it should be considered that that is for a very long follow-up – often up to 30–40 years, and higher ROC values are not to be expected. Half of those dying after 50 but before 90 years dye after 77 years of age. Thus, the predictive ability of a variable has to last for a long period to contribute significantly, and that is one of the messages of the study.
  • 2 It should be observed that most of the factors found in this study [1] agree with known risk factors for several important diseases, and we, therefore, do not believe that the findings are ‘chance findings’. We are well aware of the risk of chance findings.
  • 3 There has been a consistent finding in this cohort of men born in 1913 that obesity has not been a risk factor for coronary disease or total mortality, which we agree is at variance with many other studies. The reason may be that obesity has not been common so far among these men. Therefore, there is no reason to adjust for BMI. The area under the curve for BMI at the age of 54 years was 0.504.
  • 4 The study cohort was a third of all men born in 1913 in Gothenburg. There were only a few men (7) who died after child age and before age 50 when the study started.

Conflict of interest statement

No conflict of interest was declared.