13 November 2002

Dear Editor,

We have some questions regarding the recent article by Dengate and Ruben describing the apparent effect of calcium propionate on attention deficit hyperactivity disorder related disorders.1

1. Although only 27 patients were studied in a double-blind crossover study, the authors chose not to give us the individual data. Instead, the authors cited a summary of the statistical analyses with a final P-value of 0.05. The statistical analysis used was quite complex so it would have been useful to have access to more of the raw data. Perhaps belatedly, would the authors send us the computer printout of the data summary and the analysis?

The statistical analysis of cross-over trials needs careful consideration due to the fact that there are potential confounding factors inherent in cross-over designs. Cross-over designs distort results in the presence of outliers. The authors do not report on this. Nor is there mention of the estimate of a ‘carry-over’ effect (where the effect of the first treatment influences the action of the subsequent treatment). If carry-over effect is significant then this has serious implications on the conclusion of the study.2

2. We noted that the scores from school teachers were omitted from analysis because compliance was difficult and not enough scores were properly completed. Even though the authors state that other papers have demonstrated a tight correlation between parents’ and teachers’ hyperactivity scores, we would like to see the data collected from the teachers. In particular, when teacher data was available, how often did it agree with the parent data? How often did it disagree?

3. We praise the authors for honestly mentioning that most parents did correctly ‘guess’ which bread they were using. However, the authors state that in some families, two children were enrolled in the study. Even though the control and test bread were supposedly indistinguishable, perhaps when eaten together side by side an expert mother could find a difference. Also, since test bread contains the calcium propionate preservative, it might be expected to last longer and be less mouldy in the Darwin climate. If so, was the correlation between bread type and behaviour better, worse, or the same, in the families with more than one child in the study?

4. Finally, we wonder why calcium propionate in bread is regarded as an additive chemical? Calcium is ubiquitous, as is propionic acid. The authors have not proposed a mechanism for the detected effect. According to our reading, calcium propionate, a 3 carbon fatty acid that replaces circulating glucose in ruminants, is a by-product of cheese making and is produced by bacteria taking part in normal fermentation reactions. As such, it seems to be itself a ‘food’ rather than a chemical. The only article we could find relating propionate consumption in bread to altered physiology was the observation in one paper that it seemed to delay gastric emptying slightly, perhaps endearing itself to diabetics who like to eat bread.3


  1. Top of page
  2. References
  • 1
    Dengate S, Ruben A. Controlled trial of cumulative behavioural effects of a common bread preservative. J. Paediatr. Child Health 2002; 38: 3736.
  • 2
    Woods JR, Williams JG, Tavel M. The two-period crossover design in medical research. Ann. Intern. Med. 1989; 10: 5605.
  • 3
    Darwiche G, Ostman EM, Liljeberg HG et al. Measurements of the gastric emptying rate by use of ultrasonography: studies in humans using bread with added sodium propionate. Am. J. Clin. Nutr. 2001; 74: 2548.