- Email: [email protected]
The Development of Anticipatory Nausea in Patients Receiving Adjuvant Chemotherapy for Breast Cancer
Physiology & Behavior, Vol. 61, No. 5, pp. 737–741, 1997 Copyright q 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0031-9384/97 $17.00 / .00
PII S0031-9384(96)00528-8
The Development of Anticipatory Nausea in Patients Receiving Adjuvant Chemotherapy for Breast Cancer GUY H. MONTGOMERY* AND DANA H. BOVBJERG*† 1 *Psychiatry Department, Memorial Sloan-Kettering Cancer Center, New York, NY 10021 and †Department of Neurology and Neuroscience, Cornell University Medical College, New York, NY 10021 Received 18 June 1996; Accepted 14 October 1996 MONTGOMERY, G. H. AND D. H. BOVBJERG. The development of anticipatory nausea in patients receiving adjuvant chemotherapy for breast cancer. PHYSIOL BEHAV 61(5) 737–741, 1997.—Previous research on anticipatory nausea in cancer patients has focused on its occurrence in the clinic before chemotherapy infusions. The present study is the first, to our knowledge, to examine the development of anticipatory nausea across eight chemotherapy infusions for three time periods (night before, morning of, and immediately prior to each infusion). Based on classical conditioning experiments conducted with animal subjects, we hypothesized that the severity of anticipatory nausea would increase as the time for scheduled infusions approached. Eightytwo women diagnosed with Stage I or II breast cancer were assessed for the intensity of anticipatory nausea at three time periods prior to eight scheduled infusions of outpatient adjuvant chemotherapy. Analyses indicated a significant interaction between number of infusions experienced and temporal proximity to the infusion, supporting the hypothesis. Changes in the severity of anticipatory nausea across infusions were consistent with conditioned learning predictions. These results contribute to a growing recognition of the importance of conditioning principles for understanding side effects of chemotherapy for cancer and may have implications for the management of side effects secondary to a variety of pharmacotherapies in clinical practice. q 1997 Elsevier Science Inc. Classical conditioning
Human
Anticipatory nausea
Breast cancer
ONE of the most compelling examples of the relevance of classical conditioning to clinical medicine is the phenomenon of anticipatory nausea in patients receiving emetogenic chemotherapy treatment for cancer (4,30). During the course of repeated outpatient infusions of chemotherapy, typically administered in cycles with an infusion every 1–4 weeks for several months (12), some patients begin to experience nausea not only following treatment but also before they receive their infusion, when they return to the clinic for scheduled treatment (4,8,26,33). In the early 1980s, clinical investigators (7,25,27,31) recognized that this anticipatory nausea could be parsimoniously explained by conditioning processes. That is, outpatient chemotherapy can be viewed as a series of conditioning trials in which environmental cues (conditioned stimuli) become associated with administration of emetogenic chemotherapy infusions (unconditioned stimuli), such that reexposure to those stimuli begins to trigger nausea (conditioned response) (6,26,33). Support for the conditioning mechanism of anticipatory nausea comes from four lines of evidence. First, experimental studies with animal (10,29) and human (5) subjects have demonstrated the ability of an originally neutral cue to become a conditioned
Chemotherapy side effects
stimulus for nausea and vomiting after pairing with emetogenic drugs. Second, studies have found that many patients recall particular cues (e.g., the smell of the waiting room, the cancer center logo, the sight of the oncologist) that triggered their nausea prior to treatment infusions (1,14). Although the stimuli tend to be highly idiosyncratic, particular cues (putative conditioned stimuli) have been reported to elicit nausea (putative conditioned responses) in patients when they returned to the chemotherapy clinic even years after completing all treatment (9). Third, the prevalence of anticipatory nausea is higher among patients receiving more emetogenic regimens of chemotherapy (higher intensity of the putative unconditioned stimulus) (15,25,42). Fourth, the prevalence of anticipatory nausea is higher among patients receiving more infusions of emetogenic chemotherapy (increased number of ‘‘conditioning trials’’) (2,39). One of the more remarkable aspects of anticipatory nausea from a conditioning perspective are reports that some patients complain of being nauseated outside the clinic several hours before (or even the night before) their scheduled treatment infusions (4). This aspect of anticipatory nausea has received little research attention. Based on the extensive animal conditioning
1 Requests for reprints should be addressed to Dr. Dana H. Bovbjerg, Psychiatry Department, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. E-mail: [email protected]
737 / eh0c 2597 Mp 737 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
738
MONTGOMERY AND BOVBJERG
literature, there are at least two reasons to expect that conditioned reactions to proximal cues (cues closer in time to unconditioned stimuli) will be stronger than those to more distal cues (cues further in time from unconditioned stimuli). First, it is known that reducing the interstimulus interval (the time between conditioned stimuli and unconditioned stimuli) increases the probability that a conditioned response will develop (3,19). Second, given the importance of contingent relations between conditioned stimuli and unconditioned stimuli (35), conditioned nausea may be more likely to develop to proximal cues (e.g., the presence of the nurse) because these cues are more predictive of the occurrence of the unconditioned stimulus than are distal cues (e.g., looking at the appointment schedule). Based on these considerations, we hypothesized that the development of anticipatory nausea across repeated cycles of chemotherapy would be greater for patients’ reports of nausea in the clinic than for their reports of nausea earlier in the day or the night before scheduled treatment infusions. The development of anticipatory nausea, assessed for three time periods prior to each of eight infusions, was examined in the present study using data collected as part of a prospective longitudinal investigation of psychobehavioral side effects of chemotherapy. Previous studies involving some of the same patients focused on the development of anticipatory distress (13) and the development of conditioned reactions to a beverage experimentally paired with chemotherapy (5,20,37). To our knowledge, the present study is the first in the literature to examine the temporal patterning of anticipatory nausea during the acquisition of this putative conditioned response. METHODS
Participants Participants in the study included 82 women diagnosed with Stage I or II breast cancer, status post radical, modified radical, or segmental breast surgery, and scheduled to receive outpatient adjuvant infusions of chemotherapy (cyclophosphamide, methotrexate, 5-fluorouracil, and/or adriamycin) on a 3-week cycle as previously detailed (5). At each infusion, participants also received a regimen of standard antiemetic medications (5). The women were consecutively recruited to the study prior to their first infusion, and all provided written informed consent at that time. As part of the larger research project in which they were participating, some of the women in the present study were randomly assigned to receive a distinctive beverage immediately before treatment infusions. Preliminary analyses revealed no significant effects of this manipulation on the variables of interest in the present study, therefore data were collapsed across beverage conditions. Women included in the present study were older than 18 years of age, were not pregnant, had never previously received chemotherapy, were not scheduled to receive radiotherapy or oral chemotherapy during their chemotherapy treatment, and had complete data. This sample had a mean age of 48 years (range, 28–74; SE, 1.15), 73% were white, 68% married, 56% held college degrees, and 80% were working or on leave from work.
placing a slash across a line that was anchored by ‘‘Not at all nauseated,’’ and ‘‘As nauseated as I could be.’’ The distance in millimeters from the left end of the line provided the nausea score (range 0–100). At each assessment, patients completed separate ratings of the intensity of their nausea for: ‘‘Last night at home,’’ ‘‘This morning at home,’’ and ‘‘Right now.’’ On one occasion, all participants completed a standard demographic questionnaire and the Marlowe–Crowne Social Desirability Scale (11) in order to evaluate possible associations between social desirability and ratings of anticipatory nausea. Statistics Our primary analytic strategy for the nausea data was to utilize the raw VAS scores (0–100) in a repeated measures analysis of variance (rANOVA), using the SAS general linear model program (36). In these analyses there were two within-subject effects: infusion (infusions 1–8) and time period (last night, this morning, and right now). Of particular interest was the interaction term (infusion 1 time period), which represented a critical test of the study hypothesis that the development (across infusions) of anticipatory nausea differs for the three time periods. Although the F distribution of rANOVA is considered robust to skewed distributions (21), it was important to confirm the results of the above analyses using the more conservative approach of trichotomizing the VAS data. To this end, data were recoded as follows: VAS scores of 0 remained 0, VAS scores of 1 through 49 were recoded as 1, and VAS scores of 50 or greater were recoded as 2. To explore possible contributions of demographic or treatment variables (age, ethnic group, marital status, education, employment history, or specific chemotherapy regimen), each of these variables was dichotomized and entered separately as a predictor of anticipatory nausea in preliminary analyses (rANOVAs as described above). None of these variables was found to be related to anticipatory nausea (i.e., no main effects or interactions). RESULTS
Results revealed a pattern of anticipatory nausea consistent with the study hypothesis that the development of anticipatory nausea across repeated cycles of chemotherapy would be stronger for time periods more proximate to the chemotherapy infusion (see Fig. 1). Statistical support for this pattern of responses was
Measures and Procedures Following our established methodology for prospective longitudinal assessment of anticipatory nausea (1,5,32), all patients completed study materials in the clinic prior to each of eight consecutively scheduled chemotherapy infusions. Nausea was assessed using 10-cm visual analogue scales (VASs) on which patients were asked to indicate the severity of their nausea by
FIG. 1. Intensity of anticipatory nausea over the course of adjuvant chemotherapy (infusion number) for three different assessment time periods.
/ eh0c 2597 Mp 738 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
THE DEVELOPMENT OF ANTICIPATORY NAUSEA indicated by a significant infusion 1 time period interaction [F (14, 1134) Å 2.90, p õ 0.0002]. In addition, there were main effects of infusion [F(7, 567) Å 3.01, p õ 0.0041] and time period [F(2, 162) Å 10.13, p õ 0.0001]. The same pattern of results was seen when nausea scores were trichotomized prior to the rANOVA analyses. Again, we found an infusion 1 time period interaction [F(14, 1134) Å 2.33, p õ 0.0004], as well as main effects for infusion [F(7, 567) Å 2.74, p õ 0.008] and time period [F (2, 162) Å 6.52, p õ 0.002]. To determine the source of the significant interaction between infusion and time period, separate one-way rANOVAs were performed to examine the pattern of nausea across infusions for each of the three time periods (i.e., nausea reported for last night at home, this morning at home, and right now). These analyses revealed significant increases in intensity across the eight infusions for reports describing nausea right now [F (7, 567) Å 4.04, p õ 0.0002] and nausea this morning at home [F(7, 567) Å 3.07, p õ 0.0035]. On the other hand, reports of nausea last night at home did not differ across infusions [F (7, 567) Å 1.25, p õ 0.27]. One-way rANOVA analyses of the trichotomized data showed the same pattern of significant changes across infusions for nausea right now [F (7, 567) Å 3.15, p õ 0.0029], this morning [F (7, 567) Å 3.21, p õ 0.0024], and last night [F (7, 567) Å 1.53, p õ 0.155]. To provide a further description of the magnitude of changes in intensity of anticipatory nausea over the course of adjuvant chemotherapy, we conducted an effect size (e.s.) analysis (38) to compare nausea intensity prior to conditioning (infusion 1) to nausea seen at infusion 8 (e.g., after seven ‘‘conditioning trials’’). Effect sizes were calculated for each of the nausea assessment time periods based on differences in the means between infusion 1 and infusion 8. Consistent with the results of the rANOVA analyses above, the strongest effect size was revealed for reports of right now (e.s. Å 1.00), a weaker effect was seen for this morning at home (e.s. Å 0.43), and the weakest effect was demonstrated for last night at home (e.s. Å 00.07). It should be noted that following conditioning (at infusion 8), the significant differences in intensity of nausea reported for the three time periods [F (2, 162) Å 9.23, p õ 0.0002] were consistent with the study hypothesis (proximal conditioned responses greater than distal). Prior to conditioning (at infusion 1), however, the opposite pattern of responses was seen. Nausea was least intense in the clinic and highest at home the night before treatment [ F(2, 162) Å 3.10, p õ 0.05]. As a check on the possible influences of response bias in patients’ reports of nausea, we examined the relations between their scores on the Marlowe–Crowne Social Desirability Scale (11) and their reported nausea at each of the 24 assessments (8 infusions 1 3 time periods). As none of the Pearson correlation coefficients was found to be significant (p ú 0.05), we found no support for the possibility that nausea reports were biased by patients’ desires to present themselves in a positive manner. DISCUSSION
Consistent with the study hypothesis, we found that patients had greater increases in the severity of anticipatory nausea across eight chemotherapy infusions for times closer to scheduled infusions relative to more distal times. Nausea in the clinic before treatment infusions showed the greatest increases in severity across the eight chemotherapy infusions, while nausea at home the morning of chemotherapy infusions showed moderate increases, and nausea experienced the night before chemotherapy infusions did not increase significantly. To our knowledge, this study is the first to indicate that anticipatory nausea shows a
739 pattern of severity that depends upon both the number of infusions experienced by the patient and the temporal focus of the assessment relative to the patient’s scheduled infusion. Several factors should be kept in mind when considering the interpretation of these results. First, the present study relied on self-reported nausea data. This approach, however, is standard in the literature on anticipatory nausea, as there is no well-accepted objective measure (8,26,33,41). It is hoped that the technology to assess physiological indicators of the severity of nausea in humans will soon be widely available. Second, such self-report data may be responsive to demand characteristics of the study (28). However, patients were not informed of study hypotheses concerning conditioned learning (patients were told that the study was investigating side effects of chemotherapy), making it unlikely that their reports of nausea were biased toward a conditioned learning explanation of their symptoms. Moreover, we found no association between the severity of nausea reported by patients and their scores on a questionnaire that assessed the desire to present one’s self in a positive manner. Third, the potential confounding effects of unmeasured patient characteristics cannot be formally excluded in a correlational study such as the present one. The relatively homogeneous patient sample in the present study, however, reduces the likelihood of such confounding effects. It also should be noted that statistical analyses revealed that demographic or medical variables that could account for the significant interaction between infusion and time period did not. Fourth, patients’ reports of anticipatory nausea for the last night and this morning at home time periods were retrospective. Given the significant findings of the present study, one can now justify more comprehensive real-time assessments to confirm the changes in nausea during time periods prior to treatment infusions. Such studies could make use of recently developed frequent assessment methodologies (17). In the present study, one argument against any systematic bias based solely on the retrospective nature of the nausea reports for last night and this morning comes from the infusion 1 results, which demonstrated an opposite temporal gradient from that seen in subsequent infusions. Fifth, the mean levels of anticipatory nausea reported in the present study were modest. Although this is consistent with previous research, it would be of interest to investigate the development of anticipatory nausea with more emetogenic protocols (e.g., cisplatin regimens). Consistent with the study hypothesis, the results of the present study revealed a specific pattern of changes in the severity of anticipatory nausea across the chemotherapy infusions. Assessment of nausea for time periods closer to scheduled infusions demonstrated larger increases in severity of anticipatory nausea over the eight infusions than more distal time periods. These results are consistent with an extensive literature on classical conditioning processes in animals. Based on this literature, modern conceptualizations of conditioned learning describe conditioning processes as the formation of associations between events (35,40). These theoretical perspectives emphasize the importance of information that conditioned stimuli provide about unconditioned stimuli, rather than the importance of temporal relations between conditioned and unconditioned stimuli. That is, contingency is viewed as more important to the development of conditioned responses than contiguity (35,40). Contingent relations could have played a role in the pattern of nausea seen in the present study. For example, patients may have consistently interacted with the same nurse (who always wore a white lab coat), waited in the same waiting room, and received infusions in the same treatment room. Thus, clinic cues potentially could have been more consistently paired with infusions than nonclinic cues.
/ eh0c 2597 Mp 739 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
740
MONTGOMERY AND BOVBJERG
In addition, a venerable conditioning literature indicates that long interstimulus intervals (between conditioned and unconditioned stimuli) reduce the likelihood that a conditioned response will develop. If the interstimulus interval is too long, an organism may be unable to perceive the environmental relations, even when the conditioned stimulus reliably predicts the unconditioned stimulus (3,19). The results of the present study could be due to the longer time between cues experienced at home and the infusion compared with cues experienced in the clinic. Future studies will have to determine the relative contributions of these two conditioning mechanisms to the pattern of nausea reported in the present study. Another conditioning phenomenon relevant to the present results is the shape of acquisition curves seen in animal models. These typically show an initial rapid increase in response strength, after which additional conditioning trials have smaller incremental effects on the intensity of conditioned reactions (i.e., an asymptote is reached) (16,29). To our knowledge, the results of the present study are the first in the literature to demonstrate an analogous asymptotic curve in the development of anticipatory nausea in cancer chemotherapy patients. It must be noted that in the animal literature on classical conditioning, the conditioned stimuli are typically under experimental control. Implicit to the interpretation of the results of the present study is the existence of conditioned stimuli that triggered the nausea reactions reported for the different time periods. In this initial study, however, we did not explicitly pair any stimuli with chemotherapy or, indeed, identify any particular cues that served as conditioned stimuli. Previous studies of patients undergoing chemotherapy have confirmed that such cues exist but have found that specific cues are highly idiosyncratic (2,26). Given the highly individualized nature of patients’ experiences prior to treatment, and prior to individual chemotherapy infusions, it is not surprising that a wide range of environmental cues could serve as conditioned stimuli and that these could be very different for different people. For example, prior to treatment for cancer some patients may have had experience with intravenous lines, which could result in latent inhibition of conditioned reactions to those cues (18). Obviously, patients are also likely to differ in their experiences of specific cues over the day before individual chemotherapy infusions, as they all lead different lives (e.g., some may routinely drive to the clinic and park in the clinic lot, while others may routinely walk to the clinic and pass the same flower shop, etc.). Further complicating the assessment of conditioned stimuli for nausea, in some patients nausea can be triggered by mental images and thoughts about chemotherapy (34). The approach used in the present study, assessing the putative conditioned reaction across different time intervals before chemotherapy infusions, allowed us to avoid the difficulties posed by the idiosyncratic nature of conditioned stimuli under naturalistic conditions. The results of the present study suggest the relevance of experimental conditioning models for our understanding of the development of anticipatory nausea in chemotherapy patients. The anticipatory nausea reported by patients prior to their first infusion of chemotherapy is likely to involve additional mechanisms, which are as yet not well understood. Although not a focus of the present study, there are several possible mechanisms that could hypothetically account for the data. The phenomenon of anticipatory nausea prior to the first infusion could conceivably be due to low levels of nausea present in any random sample of 82 individuals. This hypothesis can not formally be ruled out due to a lack of a healthy volunteer control group for statistical comparison. However, attributing preinfusion 1 nausea to ‘‘back-
ground noise’’ does not explain differences between levels of anticipatory nausea last night, this morning, and right now, and some patients in the sample experienced far greater levels of anticipatory nausea prior to the first infusion than the mean data revealed (scores ranged from 0 to 86 for the night before the first infusion, from 0 to 54 for the morning of the first infusion, and from 0 to 55 just prior to the first infusion). Together, these two points detract from the background noise hypothesis and indicate the likely existence of other causal variables. For example, nervousness or anxiety about the upcoming chemotherapy could directly produce feelings of nausea. Relations between anxiety and nausea have been reported in the literature (8,26). In the present study, however, patients’ levels of emotional distress for the same three time periods prior to the first infusion were not found to be correlated with their reports of nausea. A second possible mechanism for anticipatory nausea prior to the first infusion is that patients’ beliefs or expectancies about chemotherapy may lead to feelings of nausea. Expectancies for nonvolitional outcomes have been labeled ‘‘response expectancies,’’ and are assumed to function directly in the production of nonvolitional outcomes (22,23). There is some evidence that patients’ expectancies of posttreatment nausea are related to their subsequent experience of this side effect of chemotherapy (1), but little research has investigated the impact of expectancies on anticipatory nausea (4). Future studies should consider explicitly examining the relations between patients’ expectations of anticipatory nausea and their later experience of anticipatory nausea both at the first and at subsequent infusions, as the construct of response expectancies may play a role in conditioned responses as well as in nonconditioned responses (24). The present study may have implications for clinical practice. First, failure to assess patients’ levels of anticipatory nausea prior to their arrival in the clinic may substantially underestimate their actual experience of nausea and its possible negative impact on quality of life. Second, health care professionals may wish to modify their interventions according to the temporal proximity of the chemotherapy infusion. Patients may benefit from prophylactic interventions they could apply at home the morning of (or even the night before) scheduled infusions. Additional therapeutic measures might need to be taken in the clinic as the time for an infusion approaches. In summary, the results of the present study indicate that the development of anticipatory nausea across repeated cycles of chemotherapy was greater for patients’ reports of nausea in the clinic than for their reports of nausea earlier in the day or the night before scheduled treatment infusions. This pattern of anticipatory nausea responses is consistent with predictions based on conditioning theory. These results contribute to a growing recognition of the importance of conditioning principles for our understanding of side effects of chemotherapy for cancer and may have implications for the management of side effects secondary to a variety pharmacotherapies in clinical practice. ACKNOWLEDGEMENTS
This research was supported by a training grant from The National Cancer Institute (CA09461) and research grants from The National Cancer Institute (CA66227) and the Department of the Army (DAMD1794-J-4141). We are required to indicate that ‘‘The content of the information does not necessarily reflect the position or policy of the government.’’ We acknowledge the assistance of Drs. William Redd, Terry DiLorenzo, and Paul Jacobsen on this project, the technical assistance of Ms. Noel Millea and Dorothy Parks, as well as the full cooperation of the Breast Service of Memorial Sloan-Kettering Cancer Center (Dr. Larry Norton, Chief).
/ eh0c 2597 Mp 740 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
THE DEVELOPMENT OF ANTICIPATORY NAUSEA
741
REFERENCES 1. Andrykowski, M. A.; Jacobsen, P. B.; Marks, E.; Gorfinkle, K.; Hakes, T. B.; Kaufman, R. J.; Currie, V. E.; Holland, J. C.; Redd, W. H. Prevalence, predictors, and course of anticipatory nausea in women receiving adjuvant chemotherapy for breast cancer. Cancer 62:2607–2613; 1988. 2. Andrykowski, M. A.; Redd, W. H. Longitudinal analysis of the development of anticipatory nausea. J. Consult. Clin. Psychol. 55:36– 41; 1987. 3. Baker, T. W. Properties of compound conditioned stimuli and their components. Psychol. Bull. 70:611–625; 1968. 4. Blasco, T. Anticipatory nausea and vomiting: Are psychological factors adequately investigated? Br. J. Clin. Psychol. 33:85–100; 1994. 5. Bovbjerg, D. H.; Redd, W. H.; Jacobsen, P. B.; Manne, S. L.; Taylor, K. L.; Surbone, A.; Norton, L.; Gilewski, T. A.; Hudis, C. A.; et al. An experimental analysis of classically conditioned nausea during cancer chemotherapy. Psychosom. Med. 54:623–637; 1992. 6. Burish, T. G.; Carey, M. P. Conditioned aversive responses in cancer chemotherapy patients: Theoretical and developmental analysis. J. Consult. Clin. Psychol. 54:593–600; 1986. 7. Burish, T. G.; Lyles, J. N. Effectiveness of relaxation training in reducing adverse reactions to cancer chemotherapy. J. Behav. Med. 4:65–78; 1981. 8. Carey, M. P.; Burish, T. G. Etiology and treatment of the psychological side effects associated with cancer chemotherapy. A critical review and discussion. Psychol. Bull. 104:307–325; 1988. 9. Cella, D. F.; Pratt, A.; Holland, J. C. Persistent anticipatory nausea, vomiting, and anxiety in cured Hodgkin’s disease patients after completion of chemotherapy. Am. J. Psychiatry 143:641–643; 1986. 10. Collins, K. H.; Tatum, A. L. A conditioned salivary reflex established by chronic morphine poisoning. Am. J. Physiol. 74:14–15; 1925. 11. Crowne, D. P.; Marlowe, D. The approval motive: Studies in evaluative dependence. New York, London, Sydney: John Wiley & Sons, Inc.; 1964. 12. DeVita, V. T.; Hellman, S.; Rosenberg, S. A. Cancer: Principles & practice of oncology. Philadelphia: Lippincott; 1989. 13. DiLorenzo, T. A.; Jacobsen, P. B.; Bovbjerg, D. H.; Cheng, H.; Hudis, C.; Sklarin, N. Sources of anticipatory emotional distress in women receiving chemotherapy for breast cancer. Ann. Oncol. 6:705–711; 1995. 14. Divgi, A. B. Oncologist-induced vomiting: The IGVID syndrome? N. Engl. J. Med. 320:189–190; 1989. 15. Dolgin, M. J.; Katz, E. R.; McGinty, K.; Siegel, S. E. Anticipatory nausea and vomiting in pediatric cancer patients. Pediatrics 75:547– 552; 1985. 16. Dworkin, B. R. Learning and physiological regulation. Chicago: The University of Chicago Press; 1993. 17. Eckenrode, J.; Bolger, N. Daily and within-day event measurement. In: Cohen, S.; Kessler, R. C.; Gordon, L. U., eds. Measuring stress. New York: Oxford University Press; 1995. 18. Fantino, E. The concept of latent inhibition and its application to psychosomatic medicine [editorial comment]. Psychosom. Med. 54:638–640; 1992. 19. Grings, W. W. Compound stimulus transfer in human classical conditioning. In: Black, A. H.; Prokasy, W. F., eds. Classical conditioning II: Current research and theory. New York: Appleton-CenturyCrofts; 1972:248. 20. Jacobsen, P. B.; Bovbjerg, D. H.; Schwartz, M. D.; Hudis, C. A.; Gilewski, T. A.; Norton, L. Conditioned emotional distress in women receiving chemotherapy for breast cancer. J. Consult. Clin. Psychol. 63:108–114; 1995.
21. Keppel, G. Design and analysis: A researcher’s handbook. Englewood Cliffs, NJ: Prentice Hall, Inc.; 1982. 22. Kirsch, I. Response expectancy as a determinant of experience and behavior. Am. Psychol. 40:1189–1202; 1985. 23. Kirsch, I. Changing expectations: A key to effective psychotherapy. Pacific Grove, CA: Brooks/Cole; 1990. 24. Montogmery, G. H.; Kirsch, I. Mechanisms of placebo analgesia: An empirical investigation. Psychol. Sci. 7:174–176; 1996. 25. Morrow, G. R. Prevalence and correlates of anticipatory nausea and vomiting in chemotherapy patients. J. Natl. Cancer Inst. 68:484– 488; 1982. 26. Morrow, G. R.; Dobkin, P. L. Anticipatory nausea and vomiting in cancer patients undergoing chemotherapy treatment: Prevalence, etiology and behavioral interventions. Clin. Psychol. Rev. 8:517–556; 1988. 27. Nesse, R. M.; Carli, T.; Curtis, G. G.; Kleinman, P. D. Pre-treatment nausea in cancer chemotherapy: A conditioned response? Psychosom. Med. 42:33–36; 1980. 28. Orne, M. T. Demand characteristics and the concept of quasi-controls. In: Rosenthal, R.; Rosnow, R. I., eds. Artifact in behavioral research. New York: Academic Press; 1969. 29. Pavlov, I. P. Conditioned refexes: An investigation of physiological activity of the cerebral cortex (Lecture III). Oxford, UK: Oxford University Press; 1927. 30. Redd, W. H. Management of anticipatory nausea and vomiting with cancer chemotherapy. In: Foley, K. M.; Bonica, J. J.; Ventafridda, V., eds. Advances in pain research and therapy: Second International Congress on Cancer Pain. New York: Raven Press, Ltd.; 1990:345. 31. Redd, W. H.; Andresen, G. V. Conditioned aversion in cancer patients. Behav. Therapist 4:3–4; 1981. 32. Redd, W. H.; Andresen, G. V.; Minagawa, R. Y. Hypnotic control of anticipatory emesis in patients receiving cancer chemotherapy. J. Consult. Clin. Psychol. 50:14–19; 1982. 33. Redd, W. H.; Andrykowski, M. A. Behavioral intervention in cancer treatment: Controlling aversion reactions to chemotherapy. J. Consult. Clin. Psychol. 50:1018–1029; 1982. 34. Redd, W. H.; Dadds, M. R.; Futterman, A. D.; Taylor, K. L.; Bovbjerg, D. H. Nausea induced by mental images of chemotherapy. Cancer 72:629–636; 1993. 35. Rescorla, R. A. Pavlovian conditioning. Am. Psychol. 43:151–160; 1988. 36. SAS Institute Inc. SAS procedures guide, version 6, 3rd ed. Cary, NC: SAS Institute Inc.; 1990. 37. Schwartz, M. D.; Jacobsen, P. B.; Bovbjerg, D. H. The role of nausea in the development of taste aversions among cancer chemotherapy patients. Physiol. Behav. 59:659–663; 1996. 38. Smith, M. L.; Glass, G. V.; Miller, T. I. The benefits of psychotherapy. Baltimore: Johns Hopkins University Press; 1980. 39. Tomoyasu, N.; Bovbjerg, D. H.; Jacobsen, P. B. Conditioned reactions to cancer chemotherapy: Percent reinforcement predicts anticipatory nausea. Physiol. Behav. 59:273–276; 1996. 40. Turkkan, J. S. Classical conditioning: The new hegemony. Behav. Brain Sci. 12:121–179; 1989. 41. Watson, M.; McCarron, J.; Law, M. Anticipatory nausea and emesis, and psychological morbidity: Assessment of prevalence among outpatients on mild to moderate chemotherapy regimens. Br. J. Cancer 66:862–866; 1992. 42. Wilcox, P. M.; Fetting, J. H.; Nettescheim, K. M.; Abeloff, M. D. Anticipatory vomiting in women receiving cyclophosphamide, methotrexate, and 5-FU (CMF) adjuvant chemotherapy for breast carcinoma. Cancer Treat. Rep. 66:1601–1604; 1982.
/ eh0c 2597 Mp 741 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
PII S0031-9384(96)00528-8
The Development of Anticipatory Nausea in Patients Receiving Adjuvant Chemotherapy for Breast Cancer GUY H. MONTGOMERY* AND DANA H. BOVBJERG*† 1 *Psychiatry Department, Memorial Sloan-Kettering Cancer Center, New York, NY 10021 and †Department of Neurology and Neuroscience, Cornell University Medical College, New York, NY 10021 Received 18 June 1996; Accepted 14 October 1996 MONTGOMERY, G. H. AND D. H. BOVBJERG. The development of anticipatory nausea in patients receiving adjuvant chemotherapy for breast cancer. PHYSIOL BEHAV 61(5) 737–741, 1997.—Previous research on anticipatory nausea in cancer patients has focused on its occurrence in the clinic before chemotherapy infusions. The present study is the first, to our knowledge, to examine the development of anticipatory nausea across eight chemotherapy infusions for three time periods (night before, morning of, and immediately prior to each infusion). Based on classical conditioning experiments conducted with animal subjects, we hypothesized that the severity of anticipatory nausea would increase as the time for scheduled infusions approached. Eightytwo women diagnosed with Stage I or II breast cancer were assessed for the intensity of anticipatory nausea at three time periods prior to eight scheduled infusions of outpatient adjuvant chemotherapy. Analyses indicated a significant interaction between number of infusions experienced and temporal proximity to the infusion, supporting the hypothesis. Changes in the severity of anticipatory nausea across infusions were consistent with conditioned learning predictions. These results contribute to a growing recognition of the importance of conditioning principles for understanding side effects of chemotherapy for cancer and may have implications for the management of side effects secondary to a variety of pharmacotherapies in clinical practice. q 1997 Elsevier Science Inc. Classical conditioning
Human
Anticipatory nausea
Breast cancer
ONE of the most compelling examples of the relevance of classical conditioning to clinical medicine is the phenomenon of anticipatory nausea in patients receiving emetogenic chemotherapy treatment for cancer (4,30). During the course of repeated outpatient infusions of chemotherapy, typically administered in cycles with an infusion every 1–4 weeks for several months (12), some patients begin to experience nausea not only following treatment but also before they receive their infusion, when they return to the clinic for scheduled treatment (4,8,26,33). In the early 1980s, clinical investigators (7,25,27,31) recognized that this anticipatory nausea could be parsimoniously explained by conditioning processes. That is, outpatient chemotherapy can be viewed as a series of conditioning trials in which environmental cues (conditioned stimuli) become associated with administration of emetogenic chemotherapy infusions (unconditioned stimuli), such that reexposure to those stimuli begins to trigger nausea (conditioned response) (6,26,33). Support for the conditioning mechanism of anticipatory nausea comes from four lines of evidence. First, experimental studies with animal (10,29) and human (5) subjects have demonstrated the ability of an originally neutral cue to become a conditioned
Chemotherapy side effects
stimulus for nausea and vomiting after pairing with emetogenic drugs. Second, studies have found that many patients recall particular cues (e.g., the smell of the waiting room, the cancer center logo, the sight of the oncologist) that triggered their nausea prior to treatment infusions (1,14). Although the stimuli tend to be highly idiosyncratic, particular cues (putative conditioned stimuli) have been reported to elicit nausea (putative conditioned responses) in patients when they returned to the chemotherapy clinic even years after completing all treatment (9). Third, the prevalence of anticipatory nausea is higher among patients receiving more emetogenic regimens of chemotherapy (higher intensity of the putative unconditioned stimulus) (15,25,42). Fourth, the prevalence of anticipatory nausea is higher among patients receiving more infusions of emetogenic chemotherapy (increased number of ‘‘conditioning trials’’) (2,39). One of the more remarkable aspects of anticipatory nausea from a conditioning perspective are reports that some patients complain of being nauseated outside the clinic several hours before (or even the night before) their scheduled treatment infusions (4). This aspect of anticipatory nausea has received little research attention. Based on the extensive animal conditioning
1 Requests for reprints should be addressed to Dr. Dana H. Bovbjerg, Psychiatry Department, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. E-mail: [email protected]
737 / eh0c 2597 Mp 737 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
738
MONTGOMERY AND BOVBJERG
literature, there are at least two reasons to expect that conditioned reactions to proximal cues (cues closer in time to unconditioned stimuli) will be stronger than those to more distal cues (cues further in time from unconditioned stimuli). First, it is known that reducing the interstimulus interval (the time between conditioned stimuli and unconditioned stimuli) increases the probability that a conditioned response will develop (3,19). Second, given the importance of contingent relations between conditioned stimuli and unconditioned stimuli (35), conditioned nausea may be more likely to develop to proximal cues (e.g., the presence of the nurse) because these cues are more predictive of the occurrence of the unconditioned stimulus than are distal cues (e.g., looking at the appointment schedule). Based on these considerations, we hypothesized that the development of anticipatory nausea across repeated cycles of chemotherapy would be greater for patients’ reports of nausea in the clinic than for their reports of nausea earlier in the day or the night before scheduled treatment infusions. The development of anticipatory nausea, assessed for three time periods prior to each of eight infusions, was examined in the present study using data collected as part of a prospective longitudinal investigation of psychobehavioral side effects of chemotherapy. Previous studies involving some of the same patients focused on the development of anticipatory distress (13) and the development of conditioned reactions to a beverage experimentally paired with chemotherapy (5,20,37). To our knowledge, the present study is the first in the literature to examine the temporal patterning of anticipatory nausea during the acquisition of this putative conditioned response. METHODS
Participants Participants in the study included 82 women diagnosed with Stage I or II breast cancer, status post radical, modified radical, or segmental breast surgery, and scheduled to receive outpatient adjuvant infusions of chemotherapy (cyclophosphamide, methotrexate, 5-fluorouracil, and/or adriamycin) on a 3-week cycle as previously detailed (5). At each infusion, participants also received a regimen of standard antiemetic medications (5). The women were consecutively recruited to the study prior to their first infusion, and all provided written informed consent at that time. As part of the larger research project in which they were participating, some of the women in the present study were randomly assigned to receive a distinctive beverage immediately before treatment infusions. Preliminary analyses revealed no significant effects of this manipulation on the variables of interest in the present study, therefore data were collapsed across beverage conditions. Women included in the present study were older than 18 years of age, were not pregnant, had never previously received chemotherapy, were not scheduled to receive radiotherapy or oral chemotherapy during their chemotherapy treatment, and had complete data. This sample had a mean age of 48 years (range, 28–74; SE, 1.15), 73% were white, 68% married, 56% held college degrees, and 80% were working or on leave from work.
placing a slash across a line that was anchored by ‘‘Not at all nauseated,’’ and ‘‘As nauseated as I could be.’’ The distance in millimeters from the left end of the line provided the nausea score (range 0–100). At each assessment, patients completed separate ratings of the intensity of their nausea for: ‘‘Last night at home,’’ ‘‘This morning at home,’’ and ‘‘Right now.’’ On one occasion, all participants completed a standard demographic questionnaire and the Marlowe–Crowne Social Desirability Scale (11) in order to evaluate possible associations between social desirability and ratings of anticipatory nausea. Statistics Our primary analytic strategy for the nausea data was to utilize the raw VAS scores (0–100) in a repeated measures analysis of variance (rANOVA), using the SAS general linear model program (36). In these analyses there were two within-subject effects: infusion (infusions 1–8) and time period (last night, this morning, and right now). Of particular interest was the interaction term (infusion 1 time period), which represented a critical test of the study hypothesis that the development (across infusions) of anticipatory nausea differs for the three time periods. Although the F distribution of rANOVA is considered robust to skewed distributions (21), it was important to confirm the results of the above analyses using the more conservative approach of trichotomizing the VAS data. To this end, data were recoded as follows: VAS scores of 0 remained 0, VAS scores of 1 through 49 were recoded as 1, and VAS scores of 50 or greater were recoded as 2. To explore possible contributions of demographic or treatment variables (age, ethnic group, marital status, education, employment history, or specific chemotherapy regimen), each of these variables was dichotomized and entered separately as a predictor of anticipatory nausea in preliminary analyses (rANOVAs as described above). None of these variables was found to be related to anticipatory nausea (i.e., no main effects or interactions). RESULTS
Results revealed a pattern of anticipatory nausea consistent with the study hypothesis that the development of anticipatory nausea across repeated cycles of chemotherapy would be stronger for time periods more proximate to the chemotherapy infusion (see Fig. 1). Statistical support for this pattern of responses was
Measures and Procedures Following our established methodology for prospective longitudinal assessment of anticipatory nausea (1,5,32), all patients completed study materials in the clinic prior to each of eight consecutively scheduled chemotherapy infusions. Nausea was assessed using 10-cm visual analogue scales (VASs) on which patients were asked to indicate the severity of their nausea by
FIG. 1. Intensity of anticipatory nausea over the course of adjuvant chemotherapy (infusion number) for three different assessment time periods.
/ eh0c 2597 Mp 738 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
THE DEVELOPMENT OF ANTICIPATORY NAUSEA indicated by a significant infusion 1 time period interaction [F (14, 1134) Å 2.90, p õ 0.0002]. In addition, there were main effects of infusion [F(7, 567) Å 3.01, p õ 0.0041] and time period [F(2, 162) Å 10.13, p õ 0.0001]. The same pattern of results was seen when nausea scores were trichotomized prior to the rANOVA analyses. Again, we found an infusion 1 time period interaction [F(14, 1134) Å 2.33, p õ 0.0004], as well as main effects for infusion [F(7, 567) Å 2.74, p õ 0.008] and time period [F (2, 162) Å 6.52, p õ 0.002]. To determine the source of the significant interaction between infusion and time period, separate one-way rANOVAs were performed to examine the pattern of nausea across infusions for each of the three time periods (i.e., nausea reported for last night at home, this morning at home, and right now). These analyses revealed significant increases in intensity across the eight infusions for reports describing nausea right now [F (7, 567) Å 4.04, p õ 0.0002] and nausea this morning at home [F(7, 567) Å 3.07, p õ 0.0035]. On the other hand, reports of nausea last night at home did not differ across infusions [F (7, 567) Å 1.25, p õ 0.27]. One-way rANOVA analyses of the trichotomized data showed the same pattern of significant changes across infusions for nausea right now [F (7, 567) Å 3.15, p õ 0.0029], this morning [F (7, 567) Å 3.21, p õ 0.0024], and last night [F (7, 567) Å 1.53, p õ 0.155]. To provide a further description of the magnitude of changes in intensity of anticipatory nausea over the course of adjuvant chemotherapy, we conducted an effect size (e.s.) analysis (38) to compare nausea intensity prior to conditioning (infusion 1) to nausea seen at infusion 8 (e.g., after seven ‘‘conditioning trials’’). Effect sizes were calculated for each of the nausea assessment time periods based on differences in the means between infusion 1 and infusion 8. Consistent with the results of the rANOVA analyses above, the strongest effect size was revealed for reports of right now (e.s. Å 1.00), a weaker effect was seen for this morning at home (e.s. Å 0.43), and the weakest effect was demonstrated for last night at home (e.s. Å 00.07). It should be noted that following conditioning (at infusion 8), the significant differences in intensity of nausea reported for the three time periods [F (2, 162) Å 9.23, p õ 0.0002] were consistent with the study hypothesis (proximal conditioned responses greater than distal). Prior to conditioning (at infusion 1), however, the opposite pattern of responses was seen. Nausea was least intense in the clinic and highest at home the night before treatment [ F(2, 162) Å 3.10, p õ 0.05]. As a check on the possible influences of response bias in patients’ reports of nausea, we examined the relations between their scores on the Marlowe–Crowne Social Desirability Scale (11) and their reported nausea at each of the 24 assessments (8 infusions 1 3 time periods). As none of the Pearson correlation coefficients was found to be significant (p ú 0.05), we found no support for the possibility that nausea reports were biased by patients’ desires to present themselves in a positive manner. DISCUSSION
Consistent with the study hypothesis, we found that patients had greater increases in the severity of anticipatory nausea across eight chemotherapy infusions for times closer to scheduled infusions relative to more distal times. Nausea in the clinic before treatment infusions showed the greatest increases in severity across the eight chemotherapy infusions, while nausea at home the morning of chemotherapy infusions showed moderate increases, and nausea experienced the night before chemotherapy infusions did not increase significantly. To our knowledge, this study is the first to indicate that anticipatory nausea shows a
739 pattern of severity that depends upon both the number of infusions experienced by the patient and the temporal focus of the assessment relative to the patient’s scheduled infusion. Several factors should be kept in mind when considering the interpretation of these results. First, the present study relied on self-reported nausea data. This approach, however, is standard in the literature on anticipatory nausea, as there is no well-accepted objective measure (8,26,33,41). It is hoped that the technology to assess physiological indicators of the severity of nausea in humans will soon be widely available. Second, such self-report data may be responsive to demand characteristics of the study (28). However, patients were not informed of study hypotheses concerning conditioned learning (patients were told that the study was investigating side effects of chemotherapy), making it unlikely that their reports of nausea were biased toward a conditioned learning explanation of their symptoms. Moreover, we found no association between the severity of nausea reported by patients and their scores on a questionnaire that assessed the desire to present one’s self in a positive manner. Third, the potential confounding effects of unmeasured patient characteristics cannot be formally excluded in a correlational study such as the present one. The relatively homogeneous patient sample in the present study, however, reduces the likelihood of such confounding effects. It also should be noted that statistical analyses revealed that demographic or medical variables that could account for the significant interaction between infusion and time period did not. Fourth, patients’ reports of anticipatory nausea for the last night and this morning at home time periods were retrospective. Given the significant findings of the present study, one can now justify more comprehensive real-time assessments to confirm the changes in nausea during time periods prior to treatment infusions. Such studies could make use of recently developed frequent assessment methodologies (17). In the present study, one argument against any systematic bias based solely on the retrospective nature of the nausea reports for last night and this morning comes from the infusion 1 results, which demonstrated an opposite temporal gradient from that seen in subsequent infusions. Fifth, the mean levels of anticipatory nausea reported in the present study were modest. Although this is consistent with previous research, it would be of interest to investigate the development of anticipatory nausea with more emetogenic protocols (e.g., cisplatin regimens). Consistent with the study hypothesis, the results of the present study revealed a specific pattern of changes in the severity of anticipatory nausea across the chemotherapy infusions. Assessment of nausea for time periods closer to scheduled infusions demonstrated larger increases in severity of anticipatory nausea over the eight infusions than more distal time periods. These results are consistent with an extensive literature on classical conditioning processes in animals. Based on this literature, modern conceptualizations of conditioned learning describe conditioning processes as the formation of associations between events (35,40). These theoretical perspectives emphasize the importance of information that conditioned stimuli provide about unconditioned stimuli, rather than the importance of temporal relations between conditioned and unconditioned stimuli. That is, contingency is viewed as more important to the development of conditioned responses than contiguity (35,40). Contingent relations could have played a role in the pattern of nausea seen in the present study. For example, patients may have consistently interacted with the same nurse (who always wore a white lab coat), waited in the same waiting room, and received infusions in the same treatment room. Thus, clinic cues potentially could have been more consistently paired with infusions than nonclinic cues.
/ eh0c 2597 Mp 739 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
740
MONTGOMERY AND BOVBJERG
In addition, a venerable conditioning literature indicates that long interstimulus intervals (between conditioned and unconditioned stimuli) reduce the likelihood that a conditioned response will develop. If the interstimulus interval is too long, an organism may be unable to perceive the environmental relations, even when the conditioned stimulus reliably predicts the unconditioned stimulus (3,19). The results of the present study could be due to the longer time between cues experienced at home and the infusion compared with cues experienced in the clinic. Future studies will have to determine the relative contributions of these two conditioning mechanisms to the pattern of nausea reported in the present study. Another conditioning phenomenon relevant to the present results is the shape of acquisition curves seen in animal models. These typically show an initial rapid increase in response strength, after which additional conditioning trials have smaller incremental effects on the intensity of conditioned reactions (i.e., an asymptote is reached) (16,29). To our knowledge, the results of the present study are the first in the literature to demonstrate an analogous asymptotic curve in the development of anticipatory nausea in cancer chemotherapy patients. It must be noted that in the animal literature on classical conditioning, the conditioned stimuli are typically under experimental control. Implicit to the interpretation of the results of the present study is the existence of conditioned stimuli that triggered the nausea reactions reported for the different time periods. In this initial study, however, we did not explicitly pair any stimuli with chemotherapy or, indeed, identify any particular cues that served as conditioned stimuli. Previous studies of patients undergoing chemotherapy have confirmed that such cues exist but have found that specific cues are highly idiosyncratic (2,26). Given the highly individualized nature of patients’ experiences prior to treatment, and prior to individual chemotherapy infusions, it is not surprising that a wide range of environmental cues could serve as conditioned stimuli and that these could be very different for different people. For example, prior to treatment for cancer some patients may have had experience with intravenous lines, which could result in latent inhibition of conditioned reactions to those cues (18). Obviously, patients are also likely to differ in their experiences of specific cues over the day before individual chemotherapy infusions, as they all lead different lives (e.g., some may routinely drive to the clinic and park in the clinic lot, while others may routinely walk to the clinic and pass the same flower shop, etc.). Further complicating the assessment of conditioned stimuli for nausea, in some patients nausea can be triggered by mental images and thoughts about chemotherapy (34). The approach used in the present study, assessing the putative conditioned reaction across different time intervals before chemotherapy infusions, allowed us to avoid the difficulties posed by the idiosyncratic nature of conditioned stimuli under naturalistic conditions. The results of the present study suggest the relevance of experimental conditioning models for our understanding of the development of anticipatory nausea in chemotherapy patients. The anticipatory nausea reported by patients prior to their first infusion of chemotherapy is likely to involve additional mechanisms, which are as yet not well understood. Although not a focus of the present study, there are several possible mechanisms that could hypothetically account for the data. The phenomenon of anticipatory nausea prior to the first infusion could conceivably be due to low levels of nausea present in any random sample of 82 individuals. This hypothesis can not formally be ruled out due to a lack of a healthy volunteer control group for statistical comparison. However, attributing preinfusion 1 nausea to ‘‘back-
ground noise’’ does not explain differences between levels of anticipatory nausea last night, this morning, and right now, and some patients in the sample experienced far greater levels of anticipatory nausea prior to the first infusion than the mean data revealed (scores ranged from 0 to 86 for the night before the first infusion, from 0 to 54 for the morning of the first infusion, and from 0 to 55 just prior to the first infusion). Together, these two points detract from the background noise hypothesis and indicate the likely existence of other causal variables. For example, nervousness or anxiety about the upcoming chemotherapy could directly produce feelings of nausea. Relations between anxiety and nausea have been reported in the literature (8,26). In the present study, however, patients’ levels of emotional distress for the same three time periods prior to the first infusion were not found to be correlated with their reports of nausea. A second possible mechanism for anticipatory nausea prior to the first infusion is that patients’ beliefs or expectancies about chemotherapy may lead to feelings of nausea. Expectancies for nonvolitional outcomes have been labeled ‘‘response expectancies,’’ and are assumed to function directly in the production of nonvolitional outcomes (22,23). There is some evidence that patients’ expectancies of posttreatment nausea are related to their subsequent experience of this side effect of chemotherapy (1), but little research has investigated the impact of expectancies on anticipatory nausea (4). Future studies should consider explicitly examining the relations between patients’ expectations of anticipatory nausea and their later experience of anticipatory nausea both at the first and at subsequent infusions, as the construct of response expectancies may play a role in conditioned responses as well as in nonconditioned responses (24). The present study may have implications for clinical practice. First, failure to assess patients’ levels of anticipatory nausea prior to their arrival in the clinic may substantially underestimate their actual experience of nausea and its possible negative impact on quality of life. Second, health care professionals may wish to modify their interventions according to the temporal proximity of the chemotherapy infusion. Patients may benefit from prophylactic interventions they could apply at home the morning of (or even the night before) scheduled infusions. Additional therapeutic measures might need to be taken in the clinic as the time for an infusion approaches. In summary, the results of the present study indicate that the development of anticipatory nausea across repeated cycles of chemotherapy was greater for patients’ reports of nausea in the clinic than for their reports of nausea earlier in the day or the night before scheduled treatment infusions. This pattern of anticipatory nausea responses is consistent with predictions based on conditioning theory. These results contribute to a growing recognition of the importance of conditioning principles for our understanding of side effects of chemotherapy for cancer and may have implications for the management of side effects secondary to a variety pharmacotherapies in clinical practice. ACKNOWLEDGEMENTS
This research was supported by a training grant from The National Cancer Institute (CA09461) and research grants from The National Cancer Institute (CA66227) and the Department of the Army (DAMD1794-J-4141). We are required to indicate that ‘‘The content of the information does not necessarily reflect the position or policy of the government.’’ We acknowledge the assistance of Drs. William Redd, Terry DiLorenzo, and Paul Jacobsen on this project, the technical assistance of Ms. Noel Millea and Dorothy Parks, as well as the full cooperation of the Breast Service of Memorial Sloan-Kettering Cancer Center (Dr. Larry Norton, Chief).
/ eh0c 2597 Mp 740 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597
THE DEVELOPMENT OF ANTICIPATORY NAUSEA
741
REFERENCES 1. Andrykowski, M. A.; Jacobsen, P. B.; Marks, E.; Gorfinkle, K.; Hakes, T. B.; Kaufman, R. J.; Currie, V. E.; Holland, J. C.; Redd, W. H. Prevalence, predictors, and course of anticipatory nausea in women receiving adjuvant chemotherapy for breast cancer. Cancer 62:2607–2613; 1988. 2. Andrykowski, M. A.; Redd, W. H. Longitudinal analysis of the development of anticipatory nausea. J. Consult. Clin. Psychol. 55:36– 41; 1987. 3. Baker, T. W. Properties of compound conditioned stimuli and their components. Psychol. Bull. 70:611–625; 1968. 4. Blasco, T. Anticipatory nausea and vomiting: Are psychological factors adequately investigated? Br. J. Clin. Psychol. 33:85–100; 1994. 5. Bovbjerg, D. H.; Redd, W. H.; Jacobsen, P. B.; Manne, S. L.; Taylor, K. L.; Surbone, A.; Norton, L.; Gilewski, T. A.; Hudis, C. A.; et al. An experimental analysis of classically conditioned nausea during cancer chemotherapy. Psychosom. Med. 54:623–637; 1992. 6. Burish, T. G.; Carey, M. P. Conditioned aversive responses in cancer chemotherapy patients: Theoretical and developmental analysis. J. Consult. Clin. Psychol. 54:593–600; 1986. 7. Burish, T. G.; Lyles, J. N. Effectiveness of relaxation training in reducing adverse reactions to cancer chemotherapy. J. Behav. Med. 4:65–78; 1981. 8. Carey, M. P.; Burish, T. G. Etiology and treatment of the psychological side effects associated with cancer chemotherapy. A critical review and discussion. Psychol. Bull. 104:307–325; 1988. 9. Cella, D. F.; Pratt, A.; Holland, J. C. Persistent anticipatory nausea, vomiting, and anxiety in cured Hodgkin’s disease patients after completion of chemotherapy. Am. J. Psychiatry 143:641–643; 1986. 10. Collins, K. H.; Tatum, A. L. A conditioned salivary reflex established by chronic morphine poisoning. Am. J. Physiol. 74:14–15; 1925. 11. Crowne, D. P.; Marlowe, D. The approval motive: Studies in evaluative dependence. New York, London, Sydney: John Wiley & Sons, Inc.; 1964. 12. DeVita, V. T.; Hellman, S.; Rosenberg, S. A. Cancer: Principles & practice of oncology. Philadelphia: Lippincott; 1989. 13. DiLorenzo, T. A.; Jacobsen, P. B.; Bovbjerg, D. H.; Cheng, H.; Hudis, C.; Sklarin, N. Sources of anticipatory emotional distress in women receiving chemotherapy for breast cancer. Ann. Oncol. 6:705–711; 1995. 14. Divgi, A. B. Oncologist-induced vomiting: The IGVID syndrome? N. Engl. J. Med. 320:189–190; 1989. 15. Dolgin, M. J.; Katz, E. R.; McGinty, K.; Siegel, S. E. Anticipatory nausea and vomiting in pediatric cancer patients. Pediatrics 75:547– 552; 1985. 16. Dworkin, B. R. Learning and physiological regulation. Chicago: The University of Chicago Press; 1993. 17. Eckenrode, J.; Bolger, N. Daily and within-day event measurement. In: Cohen, S.; Kessler, R. C.; Gordon, L. U., eds. Measuring stress. New York: Oxford University Press; 1995. 18. Fantino, E. The concept of latent inhibition and its application to psychosomatic medicine [editorial comment]. Psychosom. Med. 54:638–640; 1992. 19. Grings, W. W. Compound stimulus transfer in human classical conditioning. In: Black, A. H.; Prokasy, W. F., eds. Classical conditioning II: Current research and theory. New York: Appleton-CenturyCrofts; 1972:248. 20. Jacobsen, P. B.; Bovbjerg, D. H.; Schwartz, M. D.; Hudis, C. A.; Gilewski, T. A.; Norton, L. Conditioned emotional distress in women receiving chemotherapy for breast cancer. J. Consult. Clin. Psychol. 63:108–114; 1995.
21. Keppel, G. Design and analysis: A researcher’s handbook. Englewood Cliffs, NJ: Prentice Hall, Inc.; 1982. 22. Kirsch, I. Response expectancy as a determinant of experience and behavior. Am. Psychol. 40:1189–1202; 1985. 23. Kirsch, I. Changing expectations: A key to effective psychotherapy. Pacific Grove, CA: Brooks/Cole; 1990. 24. Montogmery, G. H.; Kirsch, I. Mechanisms of placebo analgesia: An empirical investigation. Psychol. Sci. 7:174–176; 1996. 25. Morrow, G. R. Prevalence and correlates of anticipatory nausea and vomiting in chemotherapy patients. J. Natl. Cancer Inst. 68:484– 488; 1982. 26. Morrow, G. R.; Dobkin, P. L. Anticipatory nausea and vomiting in cancer patients undergoing chemotherapy treatment: Prevalence, etiology and behavioral interventions. Clin. Psychol. Rev. 8:517–556; 1988. 27. Nesse, R. M.; Carli, T.; Curtis, G. G.; Kleinman, P. D. Pre-treatment nausea in cancer chemotherapy: A conditioned response? Psychosom. Med. 42:33–36; 1980. 28. Orne, M. T. Demand characteristics and the concept of quasi-controls. In: Rosenthal, R.; Rosnow, R. I., eds. Artifact in behavioral research. New York: Academic Press; 1969. 29. Pavlov, I. P. Conditioned refexes: An investigation of physiological activity of the cerebral cortex (Lecture III). Oxford, UK: Oxford University Press; 1927. 30. Redd, W. H. Management of anticipatory nausea and vomiting with cancer chemotherapy. In: Foley, K. M.; Bonica, J. J.; Ventafridda, V., eds. Advances in pain research and therapy: Second International Congress on Cancer Pain. New York: Raven Press, Ltd.; 1990:345. 31. Redd, W. H.; Andresen, G. V. Conditioned aversion in cancer patients. Behav. Therapist 4:3–4; 1981. 32. Redd, W. H.; Andresen, G. V.; Minagawa, R. Y. Hypnotic control of anticipatory emesis in patients receiving cancer chemotherapy. J. Consult. Clin. Psychol. 50:14–19; 1982. 33. Redd, W. H.; Andrykowski, M. A. Behavioral intervention in cancer treatment: Controlling aversion reactions to chemotherapy. J. Consult. Clin. Psychol. 50:1018–1029; 1982. 34. Redd, W. H.; Dadds, M. R.; Futterman, A. D.; Taylor, K. L.; Bovbjerg, D. H. Nausea induced by mental images of chemotherapy. Cancer 72:629–636; 1993. 35. Rescorla, R. A. Pavlovian conditioning. Am. Psychol. 43:151–160; 1988. 36. SAS Institute Inc. SAS procedures guide, version 6, 3rd ed. Cary, NC: SAS Institute Inc.; 1990. 37. Schwartz, M. D.; Jacobsen, P. B.; Bovbjerg, D. H. The role of nausea in the development of taste aversions among cancer chemotherapy patients. Physiol. Behav. 59:659–663; 1996. 38. Smith, M. L.; Glass, G. V.; Miller, T. I. The benefits of psychotherapy. Baltimore: Johns Hopkins University Press; 1980. 39. Tomoyasu, N.; Bovbjerg, D. H.; Jacobsen, P. B. Conditioned reactions to cancer chemotherapy: Percent reinforcement predicts anticipatory nausea. Physiol. Behav. 59:273–276; 1996. 40. Turkkan, J. S. Classical conditioning: The new hegemony. Behav. Brain Sci. 12:121–179; 1989. 41. Watson, M.; McCarron, J.; Law, M. Anticipatory nausea and emesis, and psychological morbidity: Assessment of prevalence among outpatients on mild to moderate chemotherapy regimens. Br. J. Cancer 66:862–866; 1992. 42. Wilcox, P. M.; Fetting, J. H.; Nettescheim, K. M.; Abeloff, M. D. Anticipatory vomiting in women receiving cyclophosphamide, methotrexate, and 5-FU (CMF) adjuvant chemotherapy for breast carcinoma. Cancer Treat. Rep. 66:1601–1604; 1982.
/ eh0c 2597 Mp 741 Thursday Apr 03 11:12 PM EL–PB (v. 61, no. 5) 2597