RT Journal
A1 Weaver MA
T1 DIetary composition during weight-loss maintenance
JF JAMA
JO JAMA
YR 2012
FD September 19
VO 308
IS 11
SP 1087
OP 1087
DO 10.1001/2012.jama.11611
UL http://dx.doi.org/10.1001/2012.jama.11611
AB 
I believe the authors overinterpreted 1 key result in their study. They commented that “C-reactive protein [CRP] also tended to be higher with the very low-carbohydrate diet” and concluded that the very low-carbohydrate diet “may increase cortisol excretion and CRP.” The authors implied that the low–glycemic index diet may be preferable to the very low-carbohydrate diet despite the fact that the latter outperformed the former on most key outcome measures. However, the overall null hypothesis of equal mean CRP values across the 3 diets could not be rejected at the specified .05 significance level, with a P value of .13. Nevertheless, the authors reported the results of a test for linear trend across the 3 diets, arranged from highest to lowest glycemic load and assuming equal spacing, with a corresponding P value of .05. It is this test that the authors used to support the conclusion.