Sex Differences and Caffeine Impact in Adenosine-Induced Hyperemia

Visual Abstract

Myocar dial blood flow (MBF) and myocardial flow reserve (MFR) assessments have become central in the clinical assessment of cardiac 82 Rb in PET. The valid measures of stress MBF and MFR require a full hemodynamic response during the acquisition. Adenosine is a frequently used pharmacologic stressor that attaches to the A 2A -adenosine receptor that mediates coronary vasodilation (1,2). When using adenosine as a pharmacologic stressor, patients are recommended to refrain from beverages, food, and analgesics containing caffeine for at least 12 h before the imaging session to avoid subpar stressing of the patients as caffeine nonselectively blocks the A 2A -adenosine receptor (1)(2)(3). Consumption of caffeine has been shown to reduce the hemodynamic response even at plasma concentrations as low as 1 mg/L, which might introduce falsepositive findings after the consequential reductions in the stress MBF and MFR (4)(5)(6). Although elevated caffeine plasma concentrations are known to affect the hemodynamic response when using adenosine as a stressing agent (4,5,7), it is unknown whether any sex-specific differences in the hemodynamic response exist (5,8).
This study aimed to evaluate the potential influence of sex on the association between plasma caffeine concentrations and stress MBF and MFR, respectively.

Study Population
This study comprised 40 young healthy volunteers ( , acquired with and without controlled caffeine consumption before the imaging session. Inclusion criteria were age .18 y, no participation in studies testing drugs, no regular consumption of medicine, no known medical condition, and no use of tobacco and euphoric substances (except alcohol) within 3 mo before study participation. Exclusion criteria were pregnancy, allergy, intolerance to theophylline or adenosine, any prior medical history of asthma, or inability to adhere to the study protocol. The Scientific Ethics Committee of the Capital Region of Denmark (protocol no. H-15009293) and the Danish Data Protection Agency approved this study, and all volunteers provided informed oral and written consent.
Imaging Protocol PET Acquisition. The 40 healthy volunteers were divided into 2 groups, both undergoing 4 82 Rb PET/CT imaging sessions, each consisting of an 82 Rb rest-stress protocol (Fig. 1A). All PET acquisitions, targeting injection doses of 1,100 MBq (30 mCi) 82 Rb, were obtained in 3-dimensional mode on a 128-slice Biograph mCT PET/CT system (Siemens Healthineers) and stored in list-mode format (9). Pharmacologic stressing was obtained using adenosine infused at 140 mg/kg/ min for 6 min with PET emission acquisition starting 2.5 min into the infusion (Fig. 1A). Before the rest scans, the volunteers underwent a low-dose CT for attenuation-correction purposes (120kVp; effective tube current, 26 mA [11 mAs quality reference]) acquired using a freebreathing protocol (10). The volunteers were instructed to abstain from caffeine at least 24 h before each of the 4 imaging sessions. All study volunteers underwent 2 imaging sessions without ingestion of caffeine and 2 imaging sessions after the ingestion of caffeine in a controlled setting using caffeine tablets (Fig. 1B). The caffeine tablets were dissolved in hot water and orally ingested 60 min before the rest perfusion scans. One-half of the volunteers had 2 serial imaging sessions with ingestion of 100 and 300 mg of caffeine, whereas the other half of the study volunteers underwent 2 serial imaging sessions after the ingestion of 200 and 400 mg of caffeine (Fig. 1B). The 4 imaging sessions were obtained in a randomized fashion. For this study, the plasma caffeine concentration is reported at the time of the stress myocardial perfusion imaging (MPI), thus, as an average of the measurements obtained at 75 and 90 min. All plasma caffeine concentrations were measured using high-performance liquid chromatography-mass spectrometry.
Repeatability. Test-retest repeatability of the rest and stress MBF and the MFR were calculated for the baseline scans (0 mg of caffeine ingested) with plasma caffeine concentrations , 1 mg/L using the coefficient of variance (15).
Regression plots for correlation between ingested caffeine and blood plasma concentration caffeine were calculated for all scans, using averaged plasma caffeine concentrations obtained at 75 and 90 min.
Concentration-response curves were obtained using a nonlinear curve-fit model, and the half-maximal relaxation (EC50, half-maximal effective concentration) was determined:

Statistical Analysis
Differences in the MBF and MFR were quantified using multivariable assessments (ANOVA) in R (GNU project). Descriptive analyses of continuous values were reported as mean 6 SD and range or median and IQR. Two-tailed P values of less than 0.05 were considered statistically significant. All data were checked for normality using the Sharpio-Wilk test. Dependency of MFR and caffeine ingestion was determined using a 2-way ANOVA, with P values less than 0.05 being considered significant. Regressional analyses between plasma caffeine concentration and ingested caffeine were obtained for both sexes.

Study Population
A total of 40 volunteers underwent 4 serial PET/CT scans within 27 d (IQR5 17; 36). The volunteers had a median age of 23 y  Of the 40 volunteers, 1 volunteer (man) was excluded due to failure to comply with the protocol. In addition, 4 scans were excluded because of significant motion during the scans (0-mg scans: n 5 2; 200-mg scan: n 5 1, 400-mg scan: n 5 1). In total, 152 of the 160 acquired imaging sessions (95%) were used for the subsequent analyses.

Rest MBF
Rest MBF assessments are shown in Figure 2. Significant differences in the rest MBF assessments for the 2 sexes were observed for MPI sessions with plasma caffeine concentrations # 1 mg/L (P , 0.001) (no caffeine ingested) and $ 5 mg/L (P 5 0.027) (caffeine ingested). Reduced heart rates were observed in the women after ingestion of caffeine whereas men had stable heart rates ( Table 1). Similar test-retest repeatability coefficients were reported for the men and women (16.1% and 13.5% [P 5 0.42], respectively) ( Table 2). Women had higher diastolic pressure than men (62.3 6 6.4 vs. 59.7 6 7.8 mm Hg, P 5 0.02) whereas lower systolic blood pressure was reported for women (104. 6 (Table 5). At rest, men had increased CVR when compared with women (Table 6).

Stress MBF
Women had higher heart rate and diastolic blood pressure, whereas systolic blood pressure was decreased when compared with men (heart rate: 85. 1 (Fig. 3). Similar test-retest repeatability coefficients were reported for the men and women (18.4% and 10.9% (P 5 0.29), respectively) ( Table 2). Significant changes in the hemodynamic response were observed for the men. In contrast, consistent stress MBF assessments were observed for women at plasma caffeine concentrations below 5 mg/L (Fig. 3). Corresponding to the change in the hemodynamic response, CVR was observed to increase for men even at low plasma caffeine concentrations, whereas no changes in CVR were observed for women (Table 6). Linear correlations between ingested caffeine and plasma concentrations of caffeine in both men and women were observed (Fig. 4). Increased plasma caffeine concentrations were measured in women compared with men for the same doses of ingested caffeine. Furthermore, an almost 3-fold increase in the sensitivity to caffeine was observed in men, with consequential reductions in the stress MBF assessments (EC50 [mg/L]: men % 3, women % 8) (Fig. 5). Multivariable analyses including stress MBF, sex, and plasma caffeine concentrations revealed that both caffeine plasma concentration (P , 0.001), sex (P , 0.001), and  No differences in test-retest repeatability was observed between the 2 sexes. the interaction of the 2 (P 5 0.049) had a significant impact on stress MBF assessments.

MFR
Comparable MFR assessments were observed for women and men (Fig. 6), with similar repeatability coefficients for the 2 sexes (women5 12.9%, men 5 20.6% [P 5 0.21]) ( Table 2). However, MFR was reduced in men when plasma caffeine concentrations exceeded 5 mg/L, whereas women had stable MFR at all plasma caffeine concentrations (Fig. 6). On an individual basis, MFR might be reduced even at plasma concentrations as low as 1.2 mg/L in men and 7.1 mg/L in women. Multivariable analyses including MFR, sex, and plasma caffeine concentrations revealed that sex and caffeine plasma concentration (both P , 0.001), as well as the interaction of sex and plasma caffeine concentration (P 5 0.049), had a significant impact on the MFR assessments.

DISCUSSION
In this randomized controlled crossover trial, we evaluated the impact of plasma caffeine concentrations on stress MBF and MFR assessments in healthy volunteers. The main finding was a sexspecific response in stress MBF and MFR assessments when the volunteers had stress perfusion imaging after caffeine consumption. Multivariable analyses of the data revealed that sex, caffeine concentration, and the interaction of the 2 were strongly associated with changes in stress MBF and MFR. Hence, in healthy volunteers, caffeine intake was associated with a different hemodynamic response in men and women with consequential differences in the stress MBF and MFR assessments. Men were found more sensitive to caffeine concentration, where plasma caffeine concentrations as low as 1.2 mg/L might affect the perfusion estimates compared with 7.1 mg/L in women.
Baseline rest and stress MBF (plasma caffeine concentrations , 1 mg/L) were increased in women, which might be explained by   the estrogenic effect on the vascular tone (17). Several studies have shown an evident trend for higher plasma caffeine and lower plasma paraxanthine (the most active caffeine metabolite) concentrations in women as compared with men, suggesting that women metabolize caffeine slower than men. Women with higher estrogen levels have been reported to have decreased cytochrome P450 1A2 enzyme activity and decreased caffeine clearance (18). Furthermore, a study has reported that women have 25% higher adenosine A 1A (at messenger RNA level) receptor and 40% lower A 2A receptor expression (at messenger RNA level) than men and ovariectomized women (19). Thus, one can speculate that the differences in antagonistic potency of caffeine observed between men and women in our study is partially attributed to sex-related expression, adenosine receptor reserve, and intracellular signaling. However, the mechanism(s) involved are likely to be more complex and need to be scrutinized in future studies. Stress MBF and MFR were observed to fluctuate for the women at different plasma caffeine concentrations. These effects may be caused by the low number of measurements obtained for the plasma caffeine concentrations ranging between 1 and 5 mg/L and, thus, statistical noise. The general reduction in stress MBF for scans obtained after caffeine consumption reported for the male volunteers was in concurrence with previous studies (5,7,16).

Rest
Test-retest repeatability was tested for the baseline scans for all subjects with concurrent plasma caffeine concentrations , 1 mg/L. In this context, test-retest repeatability was found in concordance with previous studies evaluating the short-term variation in rest MBF (15). The concurrent test-retest repeatability measures, therefore, suggest that reductions in the MBF and MFR assessments observed for high plasma caffeine concentrations were introduced by the ingestion of caffeine. This finding is supported by the observed physiologic responses to caffeine ingestion. In concordance with previous studies, women had greater increases in diastolic blood pressure after caffeine administration than men (Table 3) (20,21). One explanation for this effect may be men have a more sensitive baroreflex than women (20,21). Further, sex differences in hemodynamic response to caffeine may be related to sex steroid hormone concentrations (22)(23)(24). When  taken together, these studies suggest that men are more responsive to caffeine than women, a hypothesis supported by the findings in this study since caffeine had a greater potency in reducing stress MBF in men than women (Fig. 5 (26). In this study, CVR was found to increase in men when the plasma caffeine concentration increased, whereas women had no significant changes in the CVR after ingestion of caffeine (Table 6). Sex differences in the clearance of plasma-protein-binding and differences in the volume of distribution may also explain some of the differences in the perfusion response to caffeine between sexes (27). Regional conditions favor regadenoson and dipyridamole over adenosine as the drugs of choice. Although only adenosine was evaluated in this study, it is likely that our findings can also apply to regadenoson and dipyridamole.

Study Limitations
This study was conducted in healthy young volunteers with normal cardiac perfusion without perfusion defects, and we cannot rule out that results would have differed in the typical patient population undergoing cardiac 82 Rb PET testing. Patient populations are often elderly and have an intermediate likelihood of ischemic heart disease in addition to lower sex hormone concentrations, caffeine pharmacokinetics, and pharmacodynamics. Combined, these differences might cause different hemodynamic responses than those reported in this study. However, we find that an establishment of sex-specific data in young normal volunteers is important before investigating the impact of caffeine consumption in an elderly cohort in whom comorbidities and coronary artery disease are common. Another limitation is that our study was conducted in a single center using only 1 imaging system, which might affect the overall MBF and MFR assessments.

CONCLUSION
In healthy volunteers, we found that caffeine consumption before MPI affected men and women differently. Associations between plasma caffeine and significant reductions in stress MBF and MFR were found at lower plasma caffeine concentrations in men than in women.