Nature 470, 492–497 (2011)

Half of the data points were inadvertently omitted from the published version of Fig. 4a; the statistical analyses in the text and figure legend, however, do refer to the complete data set. The corrected figure is shown here and has been corrected in the online versions of the paper.

Figure 1
figure 1

PACAP associations with PTSD symptoms

Full size image

In addition, we present additional information to clarify two results reported in the Article regarding plasma pituitary adenylate cyclase-activating polypeptide (PACAP) levels and post-traumatic stress disorder (PTSD) symptom associations. In the Article, we reported replication of the association between PACAP levels and hyperarousal subscale, because this was the most robust association in the initial cohort. We now present the analyses separately for initial, replication and combined cohorts in Table 1. All associations but one are significant in the replication cohort. The second issue concerns potential medical confounds that could underlie the reported association. Although we do not have medical chart data on most patients, we do have responses from a health questionnaire administered during collection of trauma history and other data. We have now reanalysed the associations for the PTSD symptom scale (PSS) hyperarousal and total symptoms using subjective reports of health condition from the questionnaires as covariates. These data are presented in Table 2 and do not show any effect of health- and illness-related questions on the relationship between PACAP and PTSD symptoms. None of these additions affect the results and conclusions of the original Article.

Table 1 PACAP associations with PTSD symptoms
Full size table
Table 2 Health-adjusted PACAP associations
Full size table