Some comments about the background of: Hemilä et al. (2006) Int j sports Medicine 27: 336-341 Physical activity and the risk of pneumonia

Download 127.37 Kb.
Size127.37 Kb.
  1   2   3   4   5   6   7   8

Harri Hemilä, PhD, MD 15 May 2006

Department of Public Health

University of Helsinki



This commentary is available at:

Some comments about the background of:
Hemilä et al. (2006) Int J Sports Medicine 27: 336-341

Physical activity and the risk of pneumonia
In 1994, Medicine and Science in Sports and Exercise (MSSE) published a review by David Nieman: Exercise, upper respiratory tract infection, and the immune system. MSSE 26:128-139, 1994

in which Nieman speculated that there would be a "J"-curve type of relationship between susceptibility to respiratory infections and physical activity.

The "J"-hypothesis was, however, purely speculative.

The right hand part of the curve, the speculated elevation of respiratory infection risk associated with heavy physical activity, was based mainly on studies with marathon runners. However, sore throat or coughing after several hours running is no evidence of viral or bacterial etiology, but can be caused by the long ventilatory exertion.

The proposed valley in the "J"-curve was not based on any studies with actual clinical outcomes, but only on laboratory studies of the immune system which are only surrogate markers for infection susceptibility. However, clinical trialists have frequently emphasized that there is wide spread overemphasis on surrogates and often the surrogates produce unreliable results and in some cases even results that are opposite to the effects on the clinically relevant outcome (see e.g. Fleming and DeMets: Surrogate end points in clinical trials: are we being misled? Ann Intern Med. 1996;125(7):605-13. ).
One study that has been cited as support to the valley of the "J"-curve is the trial by Nieman et al.

Physical activity and immune function in elderly women. MSSE 25:823-831, 1993

However, there were only 30 randomized participants in this very small trial (“32 sedentary healthy elderly women were recruited from the surrounding community…” “the 32 sedentary subjects were randomly assigned to either the experimental [walking] or control group” “2 subjects in the walking group dropped out” italics mine).

In the control group 50% (8/16) of participants had upper respiratory infection (URI), and in the experimental (walking) group 22% (3/14) had URI. Fisher exact test gives P(2-tail)=0.21 for the comparison of these two randomized groups. Obviously this randomized trial does not provide any evidence to support the concept that moderate exercise might improve resistance to URI.

In that 1993 study Nieman also had a third group that was not a randomized arm, but consisted of clearly different kinds of participants (“12 highly conditioned elderly women between ages of 65 and 84 were also recruited from several states”, italics mine). However, Nieman inappropriately included this non-randomized group of very different kinds of participants to his statistical analysis as if the three groups were comparable. It is surprising that the journal reviewers of that manuscript did not point out that the randomization covers only the two groups of 14-16 people and that it is misleading to include the non-randomized group in the same analysis as if the non-randomized group was comparable with the two randomized groups. Also, the participants were on average 73 years old and, even if the randomized trial would have found benefit of exercise at P<0.05 level, the findings could not be extrapolated to young or middle aged people.

Nieman's 1994 paper in MSSE has been cited 119 times overall, and 15 of these citations were in MSSE which originally published the "J"-curve hypothesis. These figures reflect the high importance of Nieman’s 1994 paper. Furthermore, the "J"-curve has been graphically described in about a dozen reviews that I have read. Therefore the 1994 paper and the "J"-curve hypothesis have been highly influential.

In 1999, Nieman's group wrote a new review in MSSE in which they again showed in a graph the "J"-curve and a mirror image curve showing the speculated “immune function” which would have its peak at moderate exercise

MSSE 1999;31:57-66

In this new figure there were dots and squares presented at the lines indicating that the curves would have experimental basis - as if the lines were drawn through those “experimental” points. However, the points were purely fictional and in my opinion highly misleading to ordinary reader who may not have enough time when glancing the paper to see that the "experimental points" had no factual basis. I am surprised that none of the manuscript reviewers paid any attention to such fictional points.

Download 127.37 Kb.

Share with your friends:
  1   2   3   4   5   6   7   8

The database is protected by copyright © 2023
send message

    Main page