Rapid Responses to:

PAPERS: Shah Ebrahim, David Montaner, and Debbie A Lawlor Clustering of risk factors and social class in childhood and adulthood in British women's heart and health study: cross sectional analysis BMJ 2004; 328: 861 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Perm any 3 from 10; but who's the winner?

GH Hall   (11 April 2004)

Why are there no Framingham scores, and why the high threshold for hypertension?

Michael J White   (26 April 2004)

Perm any 3 from 10; but who's the winner? 11 April 2004
GH Hall, Retired EX1 2HW Send response to journal: Re: Perm any 3 from 10; but who's the winner?	It’s always refreshing to find a paper which starts with the authors saying “we set out to prove” and then admits they didn’t. Criticisms seem a bit churlish after that, but it’s a bit confusing to use separate terms “co-occurrence” and “clustering” when there’s a good deal of overlap in the meanings. If the number of women with, say, 3 risk factors is within the “expected” range, then it’s co-occurrence: if the number is greater than expected, then it’s clustering. Or, in the case of no associations, it’s neither- but still exceptional! Association, correlation, and collinearity are perfectly good words in this context. According to my calculator, there are 120 different arrangements of 3 selected from 10 variables. I wonder what the different frequencies of these possibilities were in the women with 3 abnormal findings? Lumping all the “3’s” together as a homogeneous group may hide- and dilute- a lot of interesting arrangements. The question of the meaning of non-independence of some predictors is indeed of great theoretical interest and could give clues about causes antecedent to as well as common to them. Also, a factor analysis might enable a reduction of the number of items sought before risk scoring. And who decided that Ethical Approval was "Not required"? Learning about unexpected risk factors is hardly likely to be a matter of indifference. Competing interests: None declared
Why are there no Framingham scores, and why the high threshold for hypertension? 26 April 2004
Michael J White, General Practitioner Stakes Lodge Surgery, Waterlooville, PO7 8NS Send response to journal: Re: Why are there no Framingham scores, and why the high threshold for hypertension?	I read this article with interest as it adds a new insight into the link between cardiovascular risk factors and social class. There are two areas that need clarification. Firstly, the authors define women as having hypertension if they had a systolic blood pressure greater than or equal to 160 mm Hg or diastolic blood pressure greater than or equal to95 mm Hg or were taking blood pressure medication. Current modern guidelines define people as having hypertension if they have a systolic blood pressure greater than or equal to140 mm Hg or diastolic blood pressure greater than or equal to 90 mm Hg (References 1 & 2 & 3). The authors only used two spot readings, and appeared to overlook any previous history of hypertension or high blood pressures unless the patient was taking blood pressure medication. Does this increase the risk of “white coat hypertension”? Why did the authors use such a high threshold to define hypertension, and why were previous blood pressure readings not taken into consideration? Would a different hypertensive threshold have made any difference to the results? Secondly, this is an article about the Framingham equation risk factors. Why no risk scores? I am disappointed that there are no ten year coronary risk scores derived from the data for those women aged 60 to 74 years. For those women aged 75 to 79 years, a “quasi” score calculated as if they were 74 years old may have to suffice, as the Framingham Score is only intended for people aged 35 to 74 years. Other advantages of calculating the Framingham score include the use of the actual measurements of the blood pressures and HDL:total cholesterol ratios, thus avoiding the issue of whether or not the participant has hypertension or significant hypercholesterolaemia. The disadvantage is that it is not always clear which blood pressure or cholesterol figures should be used: the pre-treatment or post-treatment readings? (Reference 4) In my experience, many patients with unexpectedly high Framingham scores do not have significantly obvious high individual risk factors. Rather they have a subtle combination of all round slightly increased risk factors that a physician taking a purely intuitive approach would overlook. Yours faithfully Michael J White MB BS MRCGP DCH DRCOG Reference 1 Bryan Williams, Neil R Poulter, Morris J Brown, Mark Davis, Gordon T McInnes, John F Potter, Peter S Sever, Simon McG Thom British Hypertension Society guidelines for hypertension management 2004 (BHS-IV): summary BMJ 2004;328:634-640, doi:10.1136/bmj.328.7440.634 Reference 2 Williams B, Poulter NR, Brown MJ, Davis M, McInnes GT, Potter JF, et al. Guidelines for management of hypertension: report of the fourth working party of the British Hypertension Society, 2004—BHS IV. J Hum Hypertens 2004;18: 139-85 Reference 3 Ramsay LE, Williams B, Johnston G, MacGregor G, Poston L, Potter J, et al. Guidelines for management of hypertension: report of the third working party of the British Hypertension Society. J Human Hypertens 1999;13: 569- 92 Reference 4 Michael J White, Ville Pettilä, Jörgen Vesti-Nielsen, Stuart J Pocock, Valerie McCormack, Francois Gueyffier, Robert H Fagard . Predicting risk of death from cardiovascular disease BMJ 2001;323:999, doi:10.1136/bmj.323.7319.999 Competing interests: None declared