TITLE: Hot Flashes Among Prostate Cancer Patients Undergoing Androgen Deprivation Therapy: Psychosocial and Quality of Life Issues PRINCIPAL INVESTIGATOR:

The gold standard for objectively measuring hot flashes in women is an increased sternal skinconductance level (SCL), but validation studies in prostate cancer patients are lacking. In the laboratory,an SCL increase of ≥1.78 micro-mho in 45 s had a sensitivity of 68% and a positive predictive value of100% in detecting self-reported hot flashes among prostate cancer patients. Outside the laboratory, 71%of the objective markers of hot flashes were accompanied by a subjective report of a hot flash, and 65%of subjective reports occurred in the absence of an objective criterion. This study demonstrates thatsternal skin conductance can be used to detect hot flashes in men in a manner analogous to its utilizationamong women. Such use would improve outcome analysis of treatment studies.Full ManuscriptHot flashes are prevalent and troublesome in prostate cancer survivors (PCS) undergoingandrogen deprivation therapy. A hot flash is a transient sensation of heat or flushing with rapid onset andcan be accompanied by sweating, shortness of breath, and dizziness (Quella, Loprinzi, & Dose, 1994).As many as 80% of men undergoing androgen ablation report hot flashes (Karling, Hammar, &Varenhorst, 1994; Schow, Renfer, Rozanski, & Thompson, 1998; Spetz, Hammar, Lindberg, Spangberg,& Varenhorst, 2001), which have been associated with poorer physical well-being (Nishiyama,Kanazawa, Watanabe, Terunuma, & Takahashi, 2004).The etiology and basic biobehavioral mechanisms of hot flashes remain unresolved and thesearch for safe and effective treatments continues. Few, if any, placebo-controlled clinical trials oftreatments for hot flashes among PCS have been published, but a large placebo effect of up to 66% hasbeen observed in treatment studies of hot flashes among women (Nelson, 2004; Nelson et al., 2006).These effects, however, have been based on self-reported hot flashes, rather than objectivemeasurement. Objective assessment of hot flashes would allow for improved outcome analysis as wellas aid studies of the pathophysiology of hot flashes.The gold standard for objective measurement of hot flashes in women is sternal skinconductance monitoring. Skin conductance is primarily a measure of sweat gland activity and ispositively correlated with the number of active sweat glands and their rate of secretion (Dawson, Schell,& Filion, 2000). Skin conductance levels (SCLs) are measured in micro-mho (μmho), a unit of electricalconductance.Two laboratory studies with menopausal women showed that 64% and 100% of subjectivereports of spontaneous hot flashes were accompanied by a SCL magnitude increase of ≥2 μmho within 30 s; whereas, 90% and 97% of such SCL increases were accompanied by a subjective report (de Bakker& Everaerd, 1996; Freedman, 1989). During hot flashes triggered through application of heat, 100% ofthe SCL increases of ≥2 μmho within 30 s were accompanied by an event mark (Freedman, 1989).These studies also demonstrated that sternal skin conductance was a better measure of hot flashes thanother physiological indicators. Sternal skin conductance as a measurement of hot flashes has not been validated for men, anduse of criteria developed for women may not be appropriate as studies of sweat glands in the sternalregion have shown sex differences. The density of functioning sweat glands on the chest is suggested to be greater in women than men (Knip, 1969). Moreover, one study found that men had significantlygreater sweat secretion rates on the chest during passive heat exposure than women despite similar skinblood flow (Inoue et al., 2005), and another has demonstrated similar results for men and womenmatched on aerobic capability (VO2max) and surface area-to-weight ratio (Frye & Kamon, 1981). Youngmen had significantly greater sweat rates on the chest than preovulatory, postovulatory, and amenorrhealyoung women in the first 30-min of exercise as ambient temperature was increasing. Such sex differences in sweat gland functioning on the chest suggest that the SCL magnitude fordetecting hot flashes may be different between men and women. Due to greater sweating rates amongmales, the base magnitude criterion of 2 μmho for women may be too low for men and result insignificant measurement error. This is suggested by one study, which found a significant mean SCLincrease of 8.7 μmho during hot flashes among castrated prostate cancer survivors (Spetz, Pettersson, etal., 2001).However, some research suggests otherwise. The SCL magnitude during hot flashes might becomparable between males and females since sweating declines with age (Armstrong & Kenney, 1993;Inoue, Shibasaki, Hirata, & Araki, 1998) and prostate cancer patients have been older thanpostmenopausal women within studies of hot flashes (de Bakker & Everaerd, 1996; Hanisch, Palmer, &Coyne, 2006; Spetz, Petterson, et al., 2001). Furthermore, postmenopausal women with hot flashessweat more on the chest than asymptomatic postmenopausal women and menstruating women(Freedman & Subramanian, 2005). Thus, despite the large SCL increases during hot flashes in prostatecancer patients, it is unclear if the SCL magnitude indicative of hot flashes in menopausal women isvalid for PCS (Carpenter, 2005b). Due to a lack of validation studies, we aimed to determine the bestSCL indicator of a hot flash in PCS during a laboratory session and to test the laboratory results in realworld settings.Methods ParticipantsEight PCS participating in an ambulatory study of hot flashes and who reported experiencing anaverage of at least 6 hot flashes a day participated in a controlled laboratory study. Eligibility criteria forthe ambulatory study included ongoing androgen deprivation therapy, ECOG criteria of 0-3, and nocurrent radiation, chemotherapy, or myelosuppressive medications. The laboratory participants wererecruited from prostate cancer support groups and through fliers. This study was approved by theUniversity of Pennsylvania’s Institutional Review Board, the Clinical Trials Committee of theAbramson Cancer Center, and the General Clinical Research Center (GCRC).The men gave informed consent, and were paid $100 in compensation. Participants wereprimarily Caucasian (75%) and most had earned at least a college degree (75%). Their ages ranged from54 to 83 years and averaged at 68.0 years. All participants were receiving leuprolide, a gonadotropin-releasing hormone agonist. Although one participant ate a diet rich in soy and another drank green tea tocontrol hot flashes, no other medications or therapies intended to reduce hot flashes were used. MeasuresSternal skin conductance. Skin conductance levels were recorded using a 0.5 constant voltagecircuit (Lykken & Venabless, 1971) built in to the front end of single channel of a Biolog® recorder(UFI Model 3992/1 SCL, UFI, Morro Bay, CA) and Meditrace® silver/silver chloride electrodes(Graphic Controls, Buffalo, NY) or Model 1081-HFD silver/silver chloride electrodes (UFI, Morro Bay,CA). Electrodes were 1.5 cm in diameter and filled with .05M KCI Unibase/glycol paste (Scheider &Fowles, 1978). The Biolog monitor is a solid state device containing a microprocessor and 2 MBmemory. It is powered by a standard 9 volt battery and was programmed to sample 12 bit skinconductance data at 1 Hz (once per second).Event marking. Participants were instructed to depress the event-mark buttons on the Biolog®when they felt a hot flash occurring. The data was time stamped when the event-mark buttons werepressed. The Biolog® emitted an auditory signal and displayed a visual message on the LCD to alertparticipants that their subjective hot flash had been recorded. Hot flash questionnaire. In addition to the event marker, participants recorded the time as well asthe severity, bother, duration, and the physical and mental symptoms of the hot flashes. Hot flashseverity and bother were measured on a 5-point scale (0=not at all, 4=extreme). The duration of the hot flash was scored as the total number of minutes. ProceduresParticipants were tested individually at the GCRC within the Hospital of the University ofPennsylvania. They did not consume caffeine or alcoholic products 4 hours before testing and did notconsume food for 2 hours before and during the entire testing period. During the testing session, participants were supine on a bed and wore only a light cotton hospital gown. Across testing sessions,the ambient temperature did not drop below 21°C or exceed 26°C.All participants were connected to the Biolog® monitor by 1030 h. After a 30 min rest forstabilization, monitoring of hot flashes continued until 1500 h. Electrodes were placed two inches belowthe collar bone and four inches apart centered from the sternal midline. Skin sites were cleaned withalcohol, and any chest hair was trimmed before electrode placement. During the laboratory study, theUFI electrodes became available for testing. Five participants wore the Meditrace electrodes for the firsthalf of the testing session then were fitted with UFI electrodes. The research assistant regularly checkedthe skin conductance monitor, the participant’s well-being, and ensured that the participant was notsleeping. If a spontaneous hot flash did not occur within three hours, a heating test was administered.Following Sturdee and colleagues (1978), participants were covered with multiple blankets to increasebody temperature. If a hot flash did not occur within 30 min, 8 oz of decaffeinated hot tea was ingestedto further increase body temperature (Wurster, McCook, & Randall, 1966).Hot flash data was also collected for 24 hours outside the laboratory. Research assistants metparticipants at their homes and connected a Biolog® monitor to begin recording at 1100 hour.Participants were instructed to participate in their regular activities with the exception of body-in-wateractivities (e.g., showering, bathing, or swimming) until the assistant returned the following day anddisconnected the monitoring equipment. During the ambulatory monitoring, participants pressed theevent marker when they felt a hot flash was occurring. Data AnalysisFirst, we examined the concordance between the men’s self-reported hot flashes and theobjective criteria of hot flashes previously validated for women. The SCL data and participants’ eventmarks were recorded on a RAM card in the Biolog® during the monitoring session. Afterward, datawere downloaded into a PC via customized software (DPS v.2.1®, UFI, Morro Bay, CA) andgraphically displayed on screen. The DPS automatically and sequentially scanned SCL data for an SCLmagnitude of ≥2 μmho within 30 s and flagged such magnitude increases and event marks.In accord with previous laboratory studies, the occurrence of a hot flash was determined by theparticipant’s subjective report. A true-positive hot flash was defined as the co-occurrence within a 5-minperiod of a subjective report and the SCL criterion, and a false-negative hot flash was the occurrence ofthe subjective report without the SCL criterion. A false-positive hot flash was the SCL criterion lackingsubjective corroboration. Sensitivity was calculated as the number of true-positives divided by the sumof true-positives and false-negatives. The positive predictive value (PPV) was determined by the numberof true-positives divided by the sum of true-positives and false-positives.Secondly, the SCL data was analyzed by the Receiver Operation Characteristic (ROC) curvestatistic (Green & Swets, 1966) to determine the optimal SCL cut-off point for identification of a hotflash in men. The SCLs for 5 min preceding and 15 min following the self-report of a hot flash werevisually scanned for maximum SCL increases in 30, 45, 60, and 75 s. Likewise, maximum increases in 5 randomly-selected 20-min SCL periods during non-hot flash times from each participant were identifiedso that true-negatives, the absence of both an event mark and specific SCL magnitudes, could bedetermined. These data were used in the ROC analysis to compute the sensitivity, specificity, and PPVof various SCL magnitudes. Specificity was calculated by the number of true-negatives divided by the sum of true-negatives and false-positives.To help identify artifact in ambulatory monitoring, descriptive statistics of various parameters ofthe SCLs were calculated to describe the SCL signature accompanying a subjective report by men in thelaboratory. The SCL signature of hot flashes in women is a discrete event characterized by a rapid SCLincrease followed by a gradual SCL decline (Carpenter, 2005a). Four minutes of SCLs preceding the peak of the SCL increase during hot flashes were identified and used to calculate two baseline periods.Baseline 1 is the average SCL of the first 30 s of the fourth minute preceding the peak, and baseline 2 isthe average of the first 30 s of the third minute preceding the peak. In addition, SCL magnitude changes and the SCL decrease following a hot flash were reviewed.Lastly, we examined the concordance between objective measurement and subjective report ofhot flashes in ambulatory conditions. Trained data analysts reviewed possible hot flash events flagged by the DPS to determine a valid hot flash according to the SCL profile (i.e., SCL magnitude andsignature) validated for women and identified in this paper for men. Objective hot flashes werecompared to subjective reports during waking hours. Waking hours were determined by diary entries ofwhen the men got out of bed for the day and went to bed for the night.Results Laboratory monitoringNo technical difficulties were encountered during the laboratory study. Seven men experienced21 spontaneous hot flashes as indicated by self report. The heating test was administered to one man,who reported one triggered hot flash. Twelve hot flashes were accompanied by the SCL magnitude of≥2 μmho within 30 s including the heat-induced hot flash. Sensitivity of the SCL magnitude was 55%.The PPV was 92%. In the ROC analysis, the area under the curve for the 30-, 45-, 60-, and 75-s periodswas .960, .970, .964, and .966, respectively. Table 1 presents the sensitivity, specificity, and PPV ofvarious SCL magnitude increases in 45 s.Signature of laboratory hot flashes. The average SCL for baseline 1 and 2 was 2.93 μmho (SD =1.39) and 2.99 μmho (SD = 1.47), respectively. All subjective reports of hot flashes were accompaniedby a SCL increase. A SCL magnitude of ≥1.78 μmho in 45 s occurred in 68% of subjective hot flashes,and the SCL increase peaked at a range of 1.55 to 22.37 μmho. Except for 1 subjective hot flash, thepeak SCL increase occurred after the participants event marked the onset of a hot flash. The timebetween subjective report and SCL peak ranged from 1 s to 198 s. The SCL at 1, 5, and 10 min after theSCL peak had decreased an average of 2.09 μmho (SD = 2.62), 3.44 μmho (SD = 3.26), 4.46 μmho (SD= 3.91), respectively. See Figure 1 for comparison of sternal conductance increases during hot flashes inthe laboratory and ambulatory settings.Subjective experiences of laboratory hot flashes. All but one subjective hot flash wasexperienced as a feeling of warmth. Other descriptors of hot flashes included perspiration/sweating(68%), clammy skin (50%), and flushing (32%). No participants indicated that they experienceddizziness, shortness of breath, muscle tension, nausea, dry mouth, headache, heart palpitations, ornegative emotions during hot flashes. Hot flashes were not considered very severe (M = 1.59, SD = .80)or bothersome (M = 1.36, SD = .73). Participants reported the duration of hot flashes to be 4 minuteslong (SD = 2.02) on average. Ambulatory monitoringNo technical difficulties were encountered during the ambulatory study. The men averagedbeing awake for 15.7 (SD = 1.4) hours, and during this time, reported multiple hot flashes (M = 7.88; SD= 3.18; R = 5 12). When using the SCL profile developed for women (Freedman, 1989), 24 objectivehot flashes were detected. Seventy-five percent of the objective hot flashes were accompanied by anevent mark and 45 event marks occurred in the absence of an objective hot flash. Similarly, 31 objectivehot flashes were detected when using the criteria, ≥1.78 μmho increase in 45 s, identified in this paper. Seventy-one percent of these objective hot flashes were accompanied by an event mark, and 41 eventmarks occurred in the absence of an objective hot flash. However, a 45-s magnitude of ≥1.03 μmhoreturned 82 objective hot flashes, 61% of which were accompanied by an event mark, and only 13 eventmarks occurred in the absence of an objective hot flash.Discussion This is the first laboratory study to determine an objective SCL profile for identification of hotflashes among prostate cancer survivors undergoing androgen deprivation therapy. Using the criteriaestablished for women of a SCL magnitude of ≥2 μmho within 30 s, sensitivity to detect subjective hot flashes in the laboratory was 55%, with a PPV of 91%. This compares to a sensitivity of 64% and a PPVof 90% found among menopausal women (de Bakker & Everaerd, 1996). However, analyses of thelaboratory data suggest that a better indicator of hot flashes in men consists of a SCL magnitude with alonger duration of 45 s and smaller μmho increase. A magnitude of ≥1.78 μmho in 45 s increasedsensitivity to 68% and provided a PPV of 100%. The SCL signature of a hot flash was similar between men and women. The SCL increase duringa hot flash was a distinct change from the relatively stable SCL preceding the hot flash. Most subjectivehot flashes were accompanied by a rapid SCL increase. The SCL decline following the peak was not a sharp drop but gradual. A comparison of the subjective characterization of hot flashes between men andwomen is not possible due to lack of or difference in data collection. Among men, hot flashes wereprimarily experienced as sensations of heat and sweat and were not severe or bothersome. None of thehot flashes lasted longer than 10 minutes; rather, 73% were less than 5 minutes in length.The SCL magnitude and signature were used conjointly as a profile to distinguish a hot flashevent from artifact in the ambulatory study. When the laboratory-based criteria were translated toambulatory settings, there was some loss of sensitivity to self-reported hot flashes, but the ≥1.78 μmhoin 45 s criteria continued to perform at higher level than the ≥2.0 μmho in 30 s criteria developed amongwomen. The sensitivity of the respective objective markers to detect subjective reports of hot flashes inambulatory settings was 35% and 29%. The SCL profile for hot flashes established for men alsoreturned fewer false alarms (i.e., subjective report without objective marker) than the women’s SCLprofile. This has implications for studies of hot flashes using objective measurement. If a greater SCLmagnitude is used to identify hot flashes, results would suggest that the men are not experiencing hotflashes when they report they are, and in addition, treatment might appear more efficacious than it reallyis, in terms of changes of objectively measured events.A weakness of this study was the inability to control ambient temperature. Higher temperaturesmight have impacted SCL increases during hot flashes. This is suggested by increasing sweating ratesduring heat exposure (Armstrong & Kenney, 1993; Inoue, Shibasaki, Hirata, & Araki, 1998). If this isthe case, ambient temperature might need to be a control variable in determining hot flashes inambulatory studies.More studies can be conducted to improve the accuracy of skin conductance in detecting hotflashes. It is notable that all subjective hot flashes were accompanied by an SCL increase but that themagnitude ranged from 0.32 to 15.75 μmho in 45 s. Likewise, one laboratory study of young Caucasianmales showed individual differences in spontaneous SCL activation and SCL activation followingphysical exertion (Rickles & Day, 1968). The SCL differences may be a result of dissimilarities inskinfold thickness, VO2max, sweat gland output, or sweat gland density. In addition, one study ofmenopausal women suggests that emotional distress might affect the SCL magnitude during self-reported hot flashes (Thurston, Blumenthal, Babyak, & Sherwood, 2005). More studies are needed todetermine what factors determine the degree of SCL increases during hot flashes. Additionally, futureresearch using psychophysiological stimuli with participants and matched controls would determine thespecificity of the SCL signature during hot flashes.The SCL profile established in the present study can be used for assessment of hot flashes inambulatory studies until a more accurate method for detecting hot flashes is developed. The results ofobjective measures of hot flashes in treatment studies, which have relied to-date on subjective report, could be important towards uncovering the mechanism behind the placebo effect, and in particular,whether this effect is reflected in changes in objectively measured events. Participants may report a decline in hot flashes due to their adaptation to and thus misperception or re-appraisal of the event,rather than the reduced occurrence of hot flashes. On the other hand, it is possible that a placeboresponse, demonstrated in subjective report but in the absence of changes in objectively recorded events, is nonetheless reflected in improvement in other measures of well-being. If that is the case, thenunderstanding the nature of this placebo response might aid in the development of cognitive-behavioralstrategies for the management of the significant discomfort associated with hot flashes.ReferencesArmstrong, C. G., & Kenney, W. L. (1993). Effects of Age and Acclimation on Responses to Passive Heat Exposure. Journal of Applied Physiology, 75(5), 2162-2167.Carpenter, J. S. (2005a). Physiological Monitor for Assessing Hot Flashes. Clinical NurseSpecialist, 19(1), 8-10. Carpenter, J. S. (2005b). State of the science: Hot flashes and cancer, part 1: Definition, scope,impact, physiology, and measurement. Oncology Nursing Forum, 32(5), 959-968.Dawson, M. E., Schell, A. M., & Filion, D. L. (2000). The electrodermal system. In J. T.Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of Psychophysiology (2 ed.): CambridgeUniversity Press. de Bakker, I. P., & Everaerd, W. (1996). Measurement of menopausal hot flushes: validation andcross-validation. Maturitas, 25(2), 87-98.Freedman, R. R. (1989). Laboratory and ambulatory monitoring of menopausal hot flashes.Psychophysiology, 26(5), 573-579.Freedman, R. R., & Subramanian, M. (2005). Effects of symptomatic status and the menstrualcycle on hot flash-related thermoregulatory parameters. Menopause-The Journal of the North AmericanMenopause Society, 12(2), 156-159.Frye, A. J., & Kamon, E. (1981). Responses to dry heat of men and women with similar aerobiccapacities. Journal of Applied Physiology, 50(1), 65-70.Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. New York,John Wiley & Sons, Inc.Hanisch, L. J., Palmer, S. C., & Coyne, J. C. (2006). Distress and hot flashes among prostatecancer patients. Poster presented at the Society of Behavioral Medicine 27th Annual Meeting, SanFrancisco, CA.Inoue, Y., Shibasaki, M., Hirata, K., & Araki, T. (1998). Relationship between skin blood flowand sweating rate, and age related regional differences. European Journal of Applied Physiology andOccupational Physiology, 79(1), 17-23.Inoue, Y., Tanaka, Y., Omori, K., Kuwahara, T., Ogura, Y., & Ueda, H. (2005). Sexandmenstrual cycle-related differences in sweating and cutaneous blood flow in response to passive heatexposure. European Journal of Applied Physiology, 94(3), 323-332.Karling, P., Hammar, M., & Varenhorst, E. (1994). Prevalence and duration of hot flushes aftersurgical or medical castration in men with prostatic-carcinoma. Journal of Urology, 152(4), 1170-1173.Knip, A. S. (1969). Measurement and regional distribution of functioning eccrine sweat glandsin male and female caucasians. Human Biology, 41(3), 380-387.Lykken, D.T., & Venables, P. H. (1971). Direct measurement of skin conductance: a proposalfor standardization, Psychophysiology, 8, 656-672.Nelson, H. D. (2004). Commonly used types of postmenopausal estrogen for treatment of hotflashes Scientific review. JAMA-Journal of the American Medical Association, 291(13), 1610-1620.Nelson, H. D., Vesco, K. K., Haney, E., Fu, R. W., Nedrow, A., Miller, J., et al. (2006).Nonhormonal therapies for menopausal hot flashes Systematic review and meta-analysis. JAMA-Journal of the American Medical Association, 295(17), 2057-2071.Nishiyama, T., Kanazawa, S., Watanabe, R., Terunuma, M., & Takahashi, K. (2004). Influenceof hot flashes on quality of life in patients with prostate cancer treated with androgen deprivationtherapy. International Journal of Urology, 11(9), 735-741.Quella, S., Loprinzi, C. L., & Dose, A. M. (1994). A qualitative approach to defining "hotflashes" in men. Urologic Nursing, 14(4), 155-158.Rickles, W. H., Jr., & Day, J. L. (1968). Electrodermal activity in non-palmar skin sites. Psychophysiology, 4(4), 421-435.Schneider, R., & Fowles, D. (1978). A convenient, non-hydrating electrolyte medium for themeasurement of electrodermal activity. Psychophysiology, 15, 483-486.Schow, D. A., Renfer, L. G., Rozanski, T. A., & Thompson, I. M. (1998). Prevalence of hotflushes during and after neoadjuvant hormonal therapy for localized prostate cancer. Southern MedicalJournal, 91(9), 855-857.Spetz, A. C., Hammar, M., Lindberg, B., Spangberg, A., & Varenhorst, E. (2001). Prospective evaluation of hot flashes during treatment with parenteral estrogen or complete androgen ablation formetastatic carcinoma of the prostate. Journal of Urology, 166(2), 517-520.Spetz, A. C., Pettersson, B., Varenhorst, E., Theodorsson, E., Thorell, L. H., & Hammar, M. (2001). Momentary increase in plasma calcitonin gene-related peptide is involved in hot flashes in mentreated with castration for carcinoma of the prostate. Journal of Urology, 166(5), 1720-1723.Sturdee, D. W., Wilson, K. A., Pipili, E., & Crocker, A. D. (1978). Physiological aspects ofmenopausal hot flush. British Medical Journal, 2(6130), 79-80. Thurston, R. C., Blumenthal, J. A., Babyak, M. A., & Sherwood, A. (2005). Emotionalantecedents of hot flashes during daily life. Psychosomatic Medicine, 67(1), 137-146.Wurster, R. D., McCook, R. D., & Randall, W. C. (1966). Cutaneous vascular and sweatingresponses to tympanic and skin temperatures. Journal of Applied Physiology, 21, 617-622.Author’s NoteThis research was supported by the US Department of Defense, Grant #DAMD17-02-1-0125 and theHospital of University of Pennsylvania’s General Clinical Research Center, Grant #RR00040. Thecontent of this publication does not necessarily reflect the views or policies of the Department ofDefense or the General Clinical Research Center.Note1. Five men provided data using both electrode types. To assess similarity of functioning, hot flashsignatures were generated across participants within electrode type. Signatures did not differ inshape or average magnitude of change across 30(4.8 μmho vs. 4.7 μmho) or 45-s epochs (5.4μmho vs. 5.4 μmho). Table 1: Accuracy of sternal skin conductance increases within 45 s for detecting hot flashes reported byprostate cancer patients in the laboratory Magnitude (μmho) Sensitivity (%)Specificity (%)PPV (%) ≥0.315100.0080.0075.86≥0.57590.9092.5086.96≥1.03068.2095.0088.24 ≥1.78068.20100.00100.00≥2.11059.10100.00100.00 Note: PPV = positive predictive value Figure 1. Mean sternal skin conductance levels during hot flashes reported in the laboratory and inambulatory settings +5+4+3+2+10-1-2-3-4