This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Küchler, T. Search for Related Content PubMed Articles by Küchler, T. Related Articles Related Correspondence

In Reply

Reference Center on Quality of Life in Oncology, Kiel, Germany

To start, we would like to thank Dimeo and Thiel for their interest in our study that was published in the Journal of Clinical Oncology on July 1, 2007.¹ Their comments focus on two possible mechanisms that are different from ours and that might provide explanations of the results. First, they critically review the data on the residual tumor classification (RTC); second, they address the issue of causes of death. We would like to respond to these issues separately.

First, as Dimeo and Thiel stated correctly, RTC is a basic predictor of survival. Accordingly, in our trial, we found a significant and substantial difference in survival between patients with R0 resection compared with those with R1 or R2 resections. This information was not included in our report because it was an expected result. However, the differences between the experimental group (EG) and control group (CG) regarding the R status had to be analyzed carefully. Just a look at the distribution, showing more R0 resections in the EG, seems to be a convincing explanation for the survival differences found. In fact, those differences in RTC between the groups were, at best, marginally significant (² = 10.058; P = .106), thus contributing to the overall survival differences.² In the next step, a multivariate analysis (backward regression) that took every factor into account that might contribute to those differences, psycho-oncological support remained an independent significant prognostic factor. This was true despite the proportions of variance explained by RTC (and TNM stage and tumor site). These results were presented comprehensively in Table 4 of the original article, and we hope that this additional information is satisfactory.

Furthermore, Dimeo and Thiel commented that "the proportion of patients with R0 resection is further diluted by the inclusion of patients with benign tumors." To briefly answer this concern, we will say that in a secondary analysis (not reported because in a randomized trial you do not exclude patients from an intent-to-treat analysis), we excluded the patients with no or benign tumors and found similar results.

Second, Dimeo and Thiel stated, "Since death causes are not evaluated, it is not clear if...". This is not correct. As we described,² we used (and combined) three different sources of information for the 10-year follow-up: the Cancer Registry, the family doctors, and the National Registry of Citizens. In this process, we regarded the information of "cancer as cause of death" from the cancer registry as being reliable. As a second approach, we contacted the family doctors of all patients who were not in the cancer registry to determine survival status and to get as much information as possible about the study patients during the time between the 2- and 10-year follow-up. The family doctors informed us about those patients who were definitely alive (ie, "...was here 4 weeks ago"), and we reported this information to the Cancer Registry. The doctors were able to verify, for all but three patients, that cancer was the cause of death. Of those three patients (two from the EG, one from the CG), two died as a result of cardiovascular disease and one died as a result of "alcohol abuse." Finally, there were 16 patients for whom the family doctors had no information about survival status. For those patients we had to use the National Registry of Citizens, which only certifies the death but not the cause of death. Given that we knew from the family doctors that these patients were cancer patients at last contact, we attributed cancer as the cause of death for the 13 patients who were dead according to the Registry. Of the remaining three patients, two were certainly alive (one from the EG, one from the CG), given that we finally reached them on the phone. Furthermore, we discovered (after an almost detective-like search), that one patient (from the CG) had remarried and left the country. In the analysis we treated her as being alive. Thus, we are quite confident that we not only determined the survival status of all 271 study patients, but in the cases of the nonsurvivors, we determined the causes of death.

The reason that we did not use the additional information from the family doctors (especially on those three patients who died as a result of reasons other than cancer) was simply that it would have been a bias in favor of our findings.

In summary, as already stated, this trial has several weaknesses. In our opinion these are addressed in the accompanying editorial.³ We do not think that either the statistical analyses or the determination of causes of death in the 10-year follow-up are part of these weaknesses.

AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

REFERENCES

1. Küchler Th, Bestmann B, Rappat S, et al: Impact of psychotherapeutic support on patients with gastrointestinal cancer undergoing surgery: 10-year survival results of a randomized trial. J Clin Oncol 25:2702-2708, 2007[Abstract/Free Full Text]

2. Sachs L: Angewandte Statistik. Berlin, Germany, Springer, 2006, pp 239-243

3. Andrykowski MA: Survival benefits associated with provision of psychotherapeutic support to patients with gastrointestinal cancer: Lots of bang for a few bucks? J Clin Oncol 25:2644-2645, 2007[Free Full Text]

Related Correspondence

• Does Psychotherapy Affect the Survival of Cancer Patients? More Questions Than Answers
  Fernando Dimeo and Eckhard Thiel
  JCO 2007 25: 5664-5665 [Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Küchler, T. Search for Related Content PubMed Articles by Küchler, T. Related Articles Related Correspondence