This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kumano, H. Articles by Kuboki, T. Search for Related Content PubMed Articles by Kumano, H. Articles by Kuboki, T.

CONCERNS ABOUT DATA AND METHODOLOGY IN MULTIPLE CHEMICAL SENSITIVITY PAPER

RESPONSE

Graduate School of Medicine; University of Tokyo; Tokyo, Japan

We appreciate your letter to the editor concerning our paper because your comments prompted us to reexamine the results of our data analysis and to revise our paper accordingly. We acknowledge that there were some errors and insufficient statements in Tables 5 to 8 as well as in the body of text, although these problems did not affect the conclusions of our paper.

First, we will provide responses to your fourth and third comments, which are highly relevant to the necessary corrections. Then, we will proceed to reply to the points raised. Please refer to the revised tables and corrections made in the body of our paper along with the following replies.

As noted in your fourth comment, the values of the standard error of mean (SEM) were incorrectly cited, although all of the estimates of the mean values and the results of statistical comparisons were correct. In addition, the estimates of the mean values and SEM values were presented only for the female subjects because the original models included gender as a fixed effect. Space limitation prevented the presentation of results for both genders, whereas the numbers of observations in symptom and random prompts and the p values (df, t) of statistical comparisons were for all subjects. As the effect of gender was significant in none of the 21 measures in either model, we have revised our models to not include gender as a variable; we have computed the estimates for all subjects in Tables 5 to 8. Therefore, the results were changed, but this did not affect the significance of differences for each measure presented in these tables.

You mentioned sample size and degrees of freedom in Table 5 (and also in Tables 6 to 8) in your third comment. We agree that, to include the number of random prompts as n in the tables, is confusing, because it may give the impression that this was the number of subjects rather than observations; thus, we deleted the numbers from the tables. The two linear mixed models included a random effect for subjects and a fixed effect for the event of experiencing symptoms in the between-interview comparison (Tables 5 and 6) or a fixed effect for the groups in the between-group comparison (Tables 7 and 8). Statistical comparisons were performed for a fixed effect in each measure and the degrees of freedom were also calculated.

We revised the tables with regard to the points mentioned above and added the ranges of scores of physical symptoms, three mood measures other than depressive mood, and a depressive mood measure as 0 to 100, 0 to 300, and 0 to 600.

Replies to your other comments are as follows.

With regard to your first comment, 12 of 14 patients diagnosed as having agoraphobia had no history of panic disorder, which can hardly be distinguished from phobias regarding places where relevant chemicals may be present, because the cardinal feature of agoraphobia is "anxiety about being in places or situations from which escape might be difficult or in which help may not be available in the event of having an unexpected or situationally predisposed panic attack or panic-like symptoms" according to Diagnostic and Statistical Manual of Mental Disorders, Four Edition (DSM-IV) criteria. Therefore, agoraphobia without history of panic disorder diagnosed in the present study may be part of multiple chemical sensitivity (MCS) symptoms, and it does not necessarily follow that the sample was biased.

With regard to your second comment, we want to note that the relative standard deviation (RSD) of repeatability test was used as the degree of measurement accuracy, as we mentioned in Methods section: The concentrations of exposure were measured by five active and five passive samplers in each of several different places in indoor and outdoor environment for 8 hours and for 1 week, respectively, the RSD of five measurements was calculated for each chemical in each place, and the largest RSD was designated as the RSD of the measured substance. For patient A5 whom you mentioned, the repeatability measures (%) of toluene were 3.95 and 3.23 for active and passive sampling methods, respectively, as cited in a previous paper by Shinohara et al. (2004) that you referred in your later comment, and C_AS x (100 – RSD_AS) = 1085 is larger than C_PS x (100 + RSD_PS) = 939.

With regard to your fifth comment, we should remind you that we controlled for the effect of gender in our original analyses, of which magnitude was different between the between-interview comparisons including seven female subjects with MCS and the between-group comparisons including not only seven female subjects with MCS but also 10 female control subjects. After excluding the effect of gender, the differences in the mean and SEM values between Table 6 and Table 8 or between Table 5 and Table 7 have become much smaller but are still present because these were still estimates from the two different models.

With regard to your sixth comment, we used a nonparametric test for exposure levels measured by passive sampling because many data points were below the levels of detection and many of the distributions were far from normal. On the other hand, we regarded the symptom profile data as continuous, which were evaluated by the lengthening and shortening bar scale with 21 steps (converted into the range of 0 to 100 for display in the tables).

We regarded the 1999 consensus criteria as the minimum criteria and we also used other clinical examinations to obtain supportive findings. However, as the significance of the variations in these clinical examinations has not been widely approved and it was not the focus of our paper, we did not mention the use of these laboratory examinations.

Finally, we wish to mention the tendency for accumulation of MCS in the same families. Although the present study included only one couple (A12 and A13) that lived together, there have been many case reports of patients with MCS from the same families. However, this cannot be taken as decisive evidence supporting the significance of personal interactions as the etiology of MCS. Members of the same family have a greater probability of environmental exposure to the same chemicals, which could lead to the occurrence of MCS.

DOI:10.1097/01.psy.0000268322.92720.da

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kumano, H. Articles by Kuboki, T. Search for Related Content PubMed Articles by Kumano, H. Articles by Kuboki, T.