Correlation of stress factors with sustained depression of natural killer cell activity and predicted prognosis in patients with breast cancer.

Natural killer (NK) cell activity and psychological status were measured at baseline and at 3 months into treatment, as part of the National Cancer Institute (NCI) Protocol 79-C-111, randomizing breast cancer patients to lumpectomy/radiation v mastectomy. Patients who were found to have positive axillary lymph nodes also received combination chemotherapy (Adriamycin [Adria Laboratories, Columbus, OH], plus Cytoxan [Mead Johnson Pharmaceuticals, Evansville, IN] or methotrexate, plus 5-fluorouracil [5-FU]). Seventy-five patients were entered onto this behavioral immunology protocol at the time of data analysis. We reported in an earlier publication that NK activity was an important predictor of patient baseline prognosis relevant to nodal status. In that study, by using multiple regression analyses, 51% of the baseline NK activity variance could be accounted for by entering three distress indicators into the equation (patient "adjustment," lack of social support, and fatigue/depression symptoms). On reassessment of NK activity after 3 months, it was found that NK activity was not affected by the interim administration of chemotherapy and/or radiotherapy. However, consistent with our earlier findings, NK activity levels remained markedly lower in patients with positive nodes than in patients with negative nodes (at 60 to 1 effector to target cell [E:T] ratio, mean of 18% lytic activity v mean of 31% lytic activity [t = 1.87, P less than .05]). Even though average levels of NK activity were lower for patients with more tumor burden, there was still a substantial range of NK activity levels within the node positive patient group, as well as within the patient group as a whole. We hypothesized that differences in levels of NK activity could be predicted on the basis of baseline distress factors found to be significant in our earlier report. In fact, we found that we could account for 30% of NK activity level variance at 3 months follow-up on the basis of baseline NK activity, fatigue/depression, and lack of social support. Therefore, although neither radiation nor chemotherapy appeared to affect NK activity, tumor burden was again clearly associated with NK activity levels, and a significant amount of baseline and 3-month NK activity could be predicted on the basis of CNS-mediated effects. At the least, such factors provide a psychological marker of host biological status.