Blood contamination and the measurement of salivary progesterone and estradiol

The impact of blood leakage due to microinjury to the oral cavity on the measurement of salivary reproductive hormones was examined. Saliva samples were collected before, immediately after, and then every 15 min for 1 h following vigorous tooth brushing. Blood in saliva was quantified by visual inspection of discoloration and an immunoassay for transferrin. Levels of progesterone increased, and levels of estradiol decreased, in saliva after microinjury. These changes were present only immediately after microinjury. The findings have implications for the use of salivary assays in biobehavioral research, short-term dynamic investigations, pharmacokinetic analyses, and studies of chronobiological changes in progesterone and estradiol levels.     

Bacteria in the oral mucosa and its effects on the measurement of cortisol, dehydroepiandrosterone, and testosterone in saliva

Bacteria load in saliva was experimentally manipulated, and the consequences for the measurement of salivary testosterone (T), dehydroepiandrosterone (DHEA), and cortisol (C) were examined. Healthy adults (n = 19) donated the first saliva sample upon rising after which they rinsed their mouths with water, waited 10 min, and donated a second sample. Samples were either left untreated or passed through a 0.22-microm filter and then frozen at -80 degrees C or incubated at room temperature (RT) for 10 days. Aliquots of each sample were cultured on agar to determine baseline and post-incubation (or freezing) bacteria load. Bacteria counts were not significantly influenced by rinsing (with water), were substantially reduced by filtration, and increased by incubation at RT. Average levels of salivary T and C, but not DHEA, were significantly lower in samples stored at RT than samples frozen the day of collection. The change in bacteria count induced by storing samples at RT was associated with a change in testosterone but not cortisol or DHEA. When samples were passed through a 0.22-microm filter bacteria counts were reduced, and the association between bacteria and testosterone was reduced to non-significant. These findings contribute to a growing body of literature revealing that the process of sample collection, storage, and handling can dramatically influence the accuracy of information generated when salivary biomarkers are integrated into research and clinical diagnostics.     

Changes in pain perception and hormones pre- and post-kumdo competition

The psychological stress of competition is a powerful stimulus affecting numerous hormones, which in turn change how pain is perceived. This study investigated whether a kumdo (kendo) team competition may be related to changes in hormones and pain. Seventeen healthy male kumdo practitioners participated in this experiment. Pain experiments were conducted by applying noxious stimuli with a thermal stimulator 10 min before a kumdo competition and 30 min post-competition. Serum testosterone, cortisol, beta-endorphin levels, pain thresholds, pain ratings at 48 °C and during blood sampling (sampling pain), anxiety, blood pressure, and heart rate were measured pre- and post-competition. Anxiety, pain threshold, testosterone/cortisol ratio, and blood pressure were significantly higher pre-competition compared to post-competition, while cortisol and pain ratings were significantly lower pre-competition than post-competition. There were significant correlations between the number of previous competitions and testosterone levels both pre-competition and post-competition. In pre-competition measurements, sampling pain increased with an increase in systolic blood pressure, heart rate, and beta-endorphins, and a decrease in age. In post-competition measurements, sampling pain increased with an increase in diastolic blood pressure and a decrease in testosterone levels. These results indicate that severe psychological pre-competition stress was associated with reduced pain ratings, perhaps in order to improve athletic performance. This also suggests that competitors may be at risk of potential injury due to changes in pain perception, and careful consideration should be taken to avoid potential injury before and during competition.     

Hormones,behavior, and social network analysis: Exploring associations between cortisol,testosterone, and network structure

We used a new interdisciplinary paradigm of social network analysis (SNA) to investigate associations between hormones and social network structures. We examine these biobehavioral processes and test hypotheses about how hormones are associated with social network structures using exponential random graph modeling (ERGM) in a cohort of first-year students (n = 74; 93% female; M age = 27 years) from a highly competitive, accelerated nursing program. Participants completed friendship nominations and as a group simultaneously donated saliva (later assayed for cortisol and testosterone). ERGM analyses revealed that salivary cortisol levels were inversely associated with the number of outgoing ties (i.e., network activity). By contrast, testosterone was not related to friendship network structure. Integration of SNA and salivary bioscience creates a novel approach to understanding hormone–behavior relationships within the context of human social ecologies.     

The effects of saliva collection,handling and storage on salivary testosterone measurement

Several endocrine parameters commonly measured in plasma, such as steroid hormones, can be measured in the oral fluid. However, there are several technical aspects of saliva sampling and processing that can potentially bias the validity of salivary testosterone measurement. The aim of this study was to evaluate the effects caused by repeated sampling; 5 min centrifugation (at 2000, 6000 or 10,000g); the stimulation of saliva flow by a cotton swab soaked in 2% citric acid touching the tongue; different storage times and conditions as well as the impact of blood contamination on salivary testosterone concentration measured using a commercially available ELISA kit. Fresh, unprocessed, unstimulated saliva samples served as a control. Salivary testosterone concentrations were influenced neither by repeated sampling nor by stimulation of salivary flow. Testosterone levels determined in samples stored in various laboratory conditions for time periods up to 1 month did not differ in comparison with controls. For both genders, salivary testosterone levels were substantially reduced after centrifugation (men F = 29.1; women F = 56.17, p < 0.0001). Blood contamination decreased salivary testosterone levels in a dose-dependent manner (men F = 6.54, p < 0.01, F = 5.01, p < 0.05). Salivary testosterone can be considered A robust and stable marker. However, saliva processing and blood leakage can introduce bias into measurements of salivary testoterone using ELISA. Our observations should be considered in studies focusing on salivary testosterone.     

Changes in adrenal and testicular activity monitored by salivary sampling in males throughout marathon runs

Measurement of cortisol and testosterone in saliva samples provided by marathon runners at 6.4 km (4-mile) intervals has been used for monitoring acute changes in adrenal and testicular activity, and the changes compared with mean values in timed samples on five rest days. The collection of mixed whole saliva was well accepted; the missed sample rate in the 8 runners in the Cardiff marathon was less than 10%. On rest days, salivary cortisol and testosterone were within the normal male range and showed a circadian rhythm; mean values at 08.00 h (23.5 nmol L-1; 258 pmol L-1, p less than 0.001, p less than 0.001 respectively) were higher than at 22.00 h (2.8 nmol L-1; 130 pmol L-1). In samples collected at 09.00 h, immediately prior to the Cardiff marathon, cortisol (25.1 nmol L-1) and testosterone (304 pmol L-1) were higher than the mean values (14.9 nmol L-1; 209 pmol L-1) on non-run days. Concentrations of both steroids increased during the marathon; testosterone peaked (442 pmol L-1) at 21 miles, whereas cortisol continued to increase, being maximal (87.9 nmol L-1) at 30 min after completion of the run. Four of the runners in the Cardiff marathon also participated in the Bristol marathon and the changing patterns in salivary hormones were strictly comparable. Salivary sampling would appear to be of value in monitoring acute and rhythmic changes in endocrine function in marathon runners. The temporal relationship between changes in salivary cortisol and testosterone are consistent with direct inhibition of testicular secretion by high cortisol concentrations.     

Salivary concentration of progesterone and cortisol significantly differs across individuals after correcting for blood hormone values

Between‐individual variation of salivary progesterone (P4) and cortisol levels does not always closely reflect blood hormone concentrations. This may be partly a function of individual differences in salivary hormone excretion. We tested whether time of day at sampling and ethnicity contributed to individual variation in salivary hormones after adjusting for blood hormone levels. Forty‐three Caucasian and 15 Japanese women (18–34 years) collected four sets of matched dried blood spot (DBS) and saliva specimens across a menstrual cycle (N = 232 specimen sets). Linear fixed‐effects (LFE) models were used to estimate the effects of diurnal variation and ethnicity on salivary P4 and cortisol while adjusting for DBS levels. For each hormone, women with exclusively positive or negative residuals (unexplained variance) from the LFE models were categorized as high‐ or low‐saliva‐to‐DBS hormone ratio (SDR; high or low salivary secretors), respectively. We found that salivary P4 (P < 0.05) was significantly higher in early morning compared to the afternoon, after controlling for DBS levels, ethnicity, and BMI. After further adjusting for this diurnal effect, significant individual variation in salivary P4 and cortisol remained: sixteen and nine women, respectively were categorized as low or high salivary secretors for both hormones (P < 0.001), suggesting systematic individual‐specific variation of salivary hormonal concentration. We conclude that when saliva is used to quantify P4 or cortisol levels, time of day at sampling should be controlled. Even with this adjustment, salivary P4 and cortisol do not closely mirror between‐ individual variation of serum P4 and cortisol in a substantial proportion of individuals. Am J Phys Anthropol 149:231–241, 2012. © 2012 Wiley Periodicals, Inc.     

Rapid endocrine responses of young men to social interactions with young women

It is well-established that males of many nonhuman vertebrate species exhibit hormonal reactions to stimuli from potential mates. The present studies were designed to test replication of preliminary findings suggesting that human males may exhibit such reactions as well. In Experiment 1, young men (n=115) provided saliva samples before and after either conversing with a woman confederate or sitting alone for 15 min. Changes from baseline in salivary testosterone concentrations were significantly greater among the men exposed to women, but only among subjects tested in the afternoon. In Experiment 2, male subjects (n=99) interacted with either a male or a female confederate with saliva samples collected before and after these interactions and all experimental sessions conducted in the afternoon. Men who interacted with women exhibited significant elevations of testosterone relative to both their own baseline concentrations and to change scores among the men who interacted with other men. In addition, women confederates' ratings of men's extraversion and degree of self-disclosure were positively correlated with changes in testosterone. In both experiments, furthermore, changes in cortisol concentrations from baseline were significantly greater among men who spoke with women relative to men in the control conditions. The results provide evidence that social interactions with potential mates can in fact trigger specific patterns of endocrine responses in human males.     

Adiponectin: Serum-saliva associations and relations with oral and systemic markers of inflammation

This study addresses gaps in our understanding about the validity and utility of using salivary adiponectin to index serum adiponectin levels. Matched blood and saliva samples were collected on a single occasion from healthy adults (n = 99; age 18–36 years, 53% male). Serum and saliva was assayed for adiponectin and inflammatory cytokines (IL-1β, IL-6, IL-8, TNFα), and saliva was also assayed for markers of blood contamination (transferrin), total protein (salivary flow rate) and matrix metalloproteinase-8 (MMP-8). We examined the extent to which salivary adiponectin was associated with serum adiponectin, and the influence of potential confounders on the serum-saliva correlation, including age, sex, body mass index, and markers of inflammation, oral health, salivary blood contamination, and flow rate. Findings revealed a modest serum-saliva association for adiponectin, and strong positive associations between salivary adiponectin and salivary levels of inflammatory cytokines, MMP-8, transferrin, and total protein. By contrast, salivary adiponectin was not related to serum levels of inflammatory activity. The magnitude of the serum-saliva association was strengthened when controlling for total protein in saliva, blood leakage into oral fluid, salivary inflammatory cytokines, and MMP-8. The pattern of findings extends our understanding of salivary adiponectin and its potential use as an index of circulating adiponectin levels.     

Validation of a cortisol enzyme immunoassay and characterization of salivary cortisol circadian rhythm in chimpanzees (Pan troglodytes)

Monitoring concentrations of stress hormones is an important tool for behavioral research and conservation for animals both in the wild and captivity. Glucocorticoids can be measured in mammals as an indicator of stress by analyzing blood, feces, urine, hair, feathers, or saliva. The advantages of using saliva for measuring cortisol concentrations are three-fold: it is minimally invasive, multiple samples can be collected from the same individual in a short timeframe, and cortisol has a relatively short response time in saliva as compared with other materials. The purpose of this study was to: (1) conduct an adrenocorticotropic hormone (ACTH) challenge as a physiological validation for an enzyme immunoassay to measure salivary cortisol in chimpanzees and (2) characterize the circadian rhythm of salivary cortisol in chimpanzees. We determined that salivary cortisol concentrations peaked 45 min following the ACTH challenge, which is similar to humans. Also, salivary cortisol concentrations peaked early in the morning and decreased throughout the day. We recommend that saliva collection may be the most effective method of measuring stress reactivity and has the potential to complement behavioral, cognitive, physiological, and welfare studies.     

Differences in Salivary Alpha-Amylase and Cortisol Responsiveness following Exposure to Electrical Stimulation versus the Trier Social Stress Tests

Background

Cortisol is an essential hormone in the regulation of the stress response along the HPA axis, and salivary cortisol has been used as a measure of free circulating cortisol levels. Recently, salivary alpha-amylase (sAA) has also emerged as a novel biomarker for psychosocial stress responsiveness within the sympathetic adrenomedullary (SAM) system.

Principal Findings

We measured sAA and salivary cortisol in healthy volunteers after exposure to the Trier Social Stress Test (TSST) and electric stimulation stress. One hundred forty-nine healthy volunteers participated in this study. All subjects were exposed to both the TSST and electric stimulation stress on separate days. We measured sAA and salivary cortisol levels three times immediately before, immediately after, and 20 min after the stress challenge. The State (STAI-S) and Trait (STAI-T) versions of the Spielberger Anxiety Inventory test and the Profile of Mood State (POMS) tests were administered to participants before the electrical stimulation and TSST protocols. We also measured HF, LF and LF/HF Heart Rate Variability ratio immediately after electrical stimulation and TSST exposure. Following TSST exposure or electrical stimulation, sAA levels displayed a rapid increase and recovery, returning to baseline levels 20 min after the stress challenge. Salivary cortisol responses showed a delayed increase, which remained significantly elevated from baseline levels 20 min after the stress challenge. Analyses revealed no differences between men and women with regard to their sAA response to the challenges (TSST or electric stimulations), while we found significantly higher salivary cortisol responses to the TSST in females. We also found that younger subjects tended to display higher sAA activity. Salivary cortisol levels were significantly correlated with the strength of the applied electrical stimulation.

Conclusions

These preliminary results suggest that the HPA axis (but not the SAM system) may show differential response patterns to distinct kinds of stressors.     

Hormonal changes in satisfied and dissatisfied shift workers across a shift cycle.

Although the literature claims that shift work is harmful, it overlooks the fact that that many shift workers are satisfied and stay healthy. There is little knowledge of the biological mechanisms mediating the differences in susceptibility. The present study compared satisfied and dissatisfied shift workers with respect to major anabolic and catabolic hormones. Forty-two male shift workers, with an extremely rapidly rotating shift schedule, were divided into two groups according to their ratings of satisfaction with their work hours. Morning blood samples were taken during the first and last morning shift in the shift cycle. Serum was analyzed with respect to testosterone, cortisol, and prolactin. Dissatisfied shift workers had lower morning testosterone than satisfied ones, but they did not significantly differ with respect to cortisol or prolactin. Low testosterone levels were, in addition, associated with a greater sleep need, disturbed sleep/wakefulness, and an increased need for recovery after the work period, the latter being the best predictor of testosterone levels. The only change across the shift cycle concerned a significant decrease of morning cortisol at the end of the shift cycle. High morning cortisol was related to having a morning personality and fewer sleep problems before the morning shift. Dissatisfaction with the shift system seems related to lower testosterone levels, which in turn are related to disturbed sleep/wakefulness and increased need for sleep and recovery. Furthermore, morning cortisol was reduced across a shift cycle. It is suggested that reduced testosterone levels may be part of a mechanism of shift work maladjustment.     

Women's intercollegiate athletic competition: Cortisol,testosterone, and the dual-hormone hypothesis as it relates to status among teammates

Recent research suggests that testosterone and cortisol jointly regulate dominance motivation and, perhaps, the status relationships that are affected by it. For this article, the results of six different studies of women's intercollegiate athletic competition were combined to give a sample size of almost ninety women for whom we had before- and after-competition values for salivary cortisol and testosterone for at least one and sometimes two competitions. For many of these women, we had surveys that allowed us to assess their status with teammates. In no matter what sport (soccer, softball, volleyball, and tennis) levels of salivary cortisol and testosterone increased when women participated in athletic competition. Salivary levels of C and T appear to rise in parallel during competition and increases in levels of one hormone are significantly related to increases in the other. Salivary levels of these hormones typically decreased for teammates who did not play but watched the competition from the sidelines. For women who played in two competitions, individual differences in the positive effect of competition on cortisol and testosterone were conserved from one competition to the next, affirming the personal consistency of endocrine responses to competition. Status with teammates was positively related to before-competition levels of testosterone, but only for women with relatively low before-competition levels of cortisol. This result provides novel support for the “dual-hormone hypothesis” as it relates to predicting social status in women's athletic teams — natural social groups of individuals who know each other and whose social hierarchy has evolved over the course of practice and play for at least one and, in some cases, several years of intercollegiate athletic competition.     

Women's intercollegiate volleyball and tennis: Effects of warm-up, competition, and practice on saliva levels of cortisol and testosterone

In virtually all sports, participants “warm-up” prior to formal competition. Women athletes from a highly ranked varsity college volleyball team and, in a second study, a highly ranked varsity college tennis team gave saliva samples before warm-up, at mid-warm-up (volleyball) or after warm-up (tennis), and immediately after intercollegiate competition. For volleyball and tennis, warm-up was associated with a substantial elevation in saliva levels of testosterone which was carried over through the period of actual competition. Cortisol levels were relatively unchanged during warm-up, but typically rose during competition. Thus, as women prepare for athletic competition by warming up, testosterone levels rise in apparent anticipation of the coming contest and then remain high through the period of play. In volleyball and tennis, after-practice testosterone level was significantly higher than before-practice level, and practice session increases in testosterone (but not cortisol) were positively correlated with increases in testosterone during intercollegiate competition. When practice and competitive play share as yet undetermined key elements, individual differences in this endocrine response to “competition” appear stable across practice and intercollegiate competition.     

Interpretation of the discrepancy observed between plasma free and salivary testosterone levels in man

The differences observed between plasma free and saliva testosterone levels do not result from methodological artefacts. We postulate that androgens metabolism in salivary glands modulates saliva testosterone levels both in men and women. In women, the "excess" of saliva testosterone (compared to the calculated free testosterone concentration), could be due to conversion of androstenedione in the salivary glands. The rate of this transformation is exaggerated in hirsute patients resulting in a significantly higher excess of saliva testosterone. In men, on the contrary, a "deficit" in saliva testosterone is observed suggesting testosterone transformation in salivary glands. This deficit is more marked in impotent men and appears to be reversed by testosterone administration. Saliva testosterone levels do not merely reflect plasma free hormone concentrations but also afford potentially useful information on androgens metabolism.     

Body size and allometric variation in facial shape in children

Objectives

In group‐living primates, it has been reported that the alpha male exhibits high concentrations of cortisol and testosterone in the context of mating competition. We investigated how the presence of females affected salivary cortisol and testosterone levels in males from a small captive group of chimpanzees (Pan troglodytes). Specifically, we assessed whether the presence of females resulted in a rapid increase in salivary cortisol and testosterone levels in the alpha male.

Materials and methods

We compared the social behavior and salivary hormone concentrations of four males before and after the presentation of receptive females. Three times a day, we collected saliva samples, a useful matrix for investigating short‐term hormonal changes, and measured cortisol and testosterone concentration by liquid chromatography–tandem mass spectrometry (LC–MS/MS).

Results

The frequency of inter‐male aggression increased in the presence of females, indicating intense competition among males. Salivary cortisol levels increased in all males in the presence of females; however, the increase was significantly more pronounced in the alpha male. We found a complex three‐way interaction among the presence of females, sampling timings, and male dominance rank in the analysis of salivary testosterone. Contrary to our prediction, a post hoc analysis revealed that salivary testosterone levels decreased after female introduction and that the alpha male did not show a higher level of salivary testosterone.

Conclusions

Our study provides experimental evidence suggesting that the presence of females plays a significant role in the rank‐related variation in the cortisol levels in male chimpanzees. Furthermore, our findings demonstrate the usefulness of salivary hormones for detecting short‐term physiological changes in studies of socioendocrinology.

     

Changes in the diurnal rhythms of cortisol,melatonin, and testosterone after 2, 4, and 7 consecutive night shifts in male police officers

Night work is associated with a large range of acute health problems and possibly also health consequences in the long run. Yet, only very few field studies specifically investigate the effects of consecutive night shift on key physiological regulatory systems. In this field study, we investigated the effects of consecutive night shifts on three hormones, melatonin, cortisol, and testosterone, among police officers at work. More specifically, the aim was to investigate how the diurnal rhythms of melatonin, cortisol, and testosterone responded to two, four, and seven consecutive night shifts and a corresponding number of days for recovery. The study was part of the “In the Middle of the Night” project and included 73 male police officers from five different police districts. The participants were exposed to three intervention conditions: “2+2”: two consecutive night shifts followed by two consecutive day recovery days; “4+4”: four consecutive night shifts followed by four consecutive recovery days; “7+7”: seven consecutive night shifts followed by seven consecutive recovery days. On the last day with night shift and the last recovery day in each intervention, the participants collected saliva samples every 4th hour when awake. The diurnal rhythms of melatonin, cortisol, and testosterone were all affected differently by an increasing number of consecutive night shifts: the amplitude of the melatonin rhythm was suppressed by 4.9% per day (95% CI 1.4–8.2% per day; p = 0.006). The diurnal rhythm of cortisol phase was delayed with an increasing number of night shifts by 33 min/day (95% CI 18–48 min per day; p ≤ 0.001), but did not show any changes in amplitude. For the diurnal rhythm of testosterone, there was no effect of the number of consecutive night shifts and the diurnal rhythm completely followed the sleep/wake cycle. We found that there were no differences in the rhythms of melatonin, cortisol, and testosterone after 2, 4, and 7 recovery days, respectively. In conclusion, we found signs of desynchronization in terms of suppressed amplitude of melatonin and phase delay of salivary cortisol as a consequence of the increasing number of consecutive night shifts among police officers at work. Lack of synchronization has been suggested as a possible mechanism linking night work to disease, but this remains to be determined.     

Salivary testosterone determination in studies of child health and development

Measurement of hormones in children's saliva has excited interest because of numerous potential applications in developmental studies. Although assays of children's saliva for some hormones (e.g., cortisol) are widely available and used, the availability and use of assays of children's saliva testosterone is restricted. By adapting a commercially available serum testosterone kit, our laboratory has developed a reliable, efficient, and highly sensitive procedure for measuring testosterone in children's saliva that does not require separation or extraction. The minimum detection limit was 0.8 pg/mL. Intraassay coefficients of variation (CV) were between 3.66 and 6. 78% at concentrations 9.25 to 86.41 pg/mL, and interassay CVs were between 5.70 and 6.61% at concentrations of 7.3 to 118.51 pg/mL. The standard curve was highly reproducible (M slope = -0.70 and Mr = 0. 99). Method accuracy, determined by spike recovery, and linearity, determined by serial dilution, were 99.20 and 92.80%, respectively. Values from matched serum and saliva samples showed strong linear relationships. The assay captured near 99.09% of the range of individual differences in boys' (N = 90) and girls' (N = 85), ages 8-12, am and pm salivary testosterone levels. This assay can be easily applied to the investigation of testosterone-behavior relations in the context of studies on child health and development. It may help many child development researchers improve or expand their research activities.     

Cues to sex- and stress-hormones in the human male face: functions of glucocorticoids in the immunocompetence handicap hypothesis

The stress-linked version of the immunocompetence handicap hypothesis has been proposed to account for inconsistencies in relationships between testosterone and immune response. The model has received some support from studies demonstrating roles of stress hormones in relationships between testosterone, immune function and secondary sexual ornamentation. Such work, however, has relied on artificial elevation of testosterone so may not reflect relationships in natural populations. We created human male facial stimuli on the basis of naturally co-occurring levels of salivary testosterone and the stress hormone cortisol. In Study 1 we tested female preferences for male faces with cues to combinations of the hormones across the menstrual cycle, and in Study 2 we tested perceptions of health and dominance in a novel set of facial stimuli. Females preferred cues to low cortisol, a preference that was strongest during the fertile phase of the menstrual cycle. The effects of cortisol on attractiveness and perceived health and dominance were contingent upon level of testosterone: the effects of the stress hormone were reduced when testosterone was high. We propose explanations for our results, including low cortisol as a cue to a heritable component of health, attractiveness as a predictor of low social-evaluative threat (and, therefore, low baseline cortisol) and testosterone as a proxy of male ability to cope efficiently with stressors.     

Sexy thoughts: effects of sexual cognitions on testosterone, cortisol, and arousal in women

Previous research suggests that sexual stimuli increase testosterone (T) in women and shows inconsistent effects of sexual arousal on cortisol (C), but effects of cognitive aspects of arousal, rather than behaviors or sensory stimuli, are unclear. The present study examined whether sexual thoughts affect T or C and whether hormonal contraceptive (HC) use moderated this effect, given mixed findings of HC use confounding hormone responses. Participants (79 women) provided a baseline saliva sample for radioimmunoassay. We created the Imagined Social Situation Exercise (ISSE) to test effects of imagining social interactions on hormones, and participants were assigned to the experimental (sexual) or one of three control (positive, neutral, stressful) conditions. Participants provided a second saliva sample 15 min post-activity. Results indicated that for women not using HCs, the sexual condition increased T compared to the stressful or positive conditions. In contrast, HC using women in the sexual condition had decreased T relative to the stressful condition and similar T to the positive condition. The effect was specific to T, as sexual thoughts did not change C. For participants in the sexual condition, higher baseline T predicted larger increases in sexual arousal but smaller increases in T, likely due to ceiling effects on T. Our results suggest that sexual thoughts change T but not C, baseline T levels and HC use may contribute to variation in the T response to sexual thoughts, and cognitive aspects of sexual arousal affect physiology.