Original Contribution

Effect of Caffeine on Intravenous Adenosine-Induced Hyperemia in Fractional Flow Reserve Measurement

Hidenari Matsumoto, MD, PhD1,*;  Kenji Nakatsuma, MD1,*;  Toshihiko Shimada, MD2;  Shunpei Ushimaru, MD1; Mikiko Mikuri, MD1;  Taketoshi Yamazaki, MD, PhD1;  Tetsuya Matsuda, MD, PhD3

Hidenari Matsumoto, MD, PhD1,*;  Kenji Nakatsuma, MD1,*;  Toshihiko Shimada, MD2;  Shunpei Ushimaru, MD1; Mikiko Mikuri, MD1;  Taketoshi Yamazaki, MD, PhD1;  Tetsuya Matsuda, MD, PhD3

Abstract: Background. The interaction between caffeine and adenosine is still a matter of debate. Aims. We examined whether caffeine attenuated intravenous adenosine-induced hyperemia in the measurement of fractional flow reserve (FFR) and whether an increased dose of adenosine overcame the caffeine antagonism. Methods. FFR was measured using different adenosine doses (140, 175, and 210 µg/kg/min) and papaverine as a reference standard in patients with intermediate coronary stenoses, who refrained from caffeine for >24 h (no-caffeine group; n = 14) and those who consumed caffeine (caffeine group; n = 28). Results. The median caffeine level in the caffeine group was 2.9 mg/L (interquartile range, 1.8-4.6 mg/L). In the no-caffeine group, FFR with adenosine did not decrease above the dose of 140 µg/kg/min (0.769, 0.771, and 0.770 at 140, 175, and 210 µg/kg/min, respectively) and was not significantly different from that with papaverine (0.765). In the caffeine group, adenosine overestimated FFR (140 µg/kg/min: 0.813, P<.001; 175 µg/kg/min: 0.806, P<.01; 210 µg/kg/min: 0.794, P=.01) compared with papaverine (0.779). The difference in FFR between papaverine and 140 µg/kg/min dose of adenosine was significantly greater in the caffeine group than in the no-caffeine group (0.034 vs 0.004; P<.05). Conclusion. Caffeine attenuates intravenous adenosine-induced hyperemia in FFR measurement. Even increased adenosine doses up to 210 µg/kg/min cannot fully surmount the antagonism.

J INVASIVE CARDIOL 2014;26(11):580-585

Key words: adenosine, caffeine, coronary hyperemia, fractional flow reserve


Induction of maximal coronary hyperemia is essential for a reliable fractional flow reserve (FFR) measurement. Adenosine produces hyperemia through activation of A2a receptors in vascular smooth muscles.1 Presently, continuous intravenous adenosine is used most often because of its advantages, including short duration of action, absence of QT-interval prolongation, and sustained hyperemia.2-4 It is generally thought that caffeine attenuates adenosine-induced hyperemia by blocking adenosine receptor activity.5,6 However, the effect of caffeine on adenosine-induced hyperemic response is still in dispute.7 Conflicting results have been reported in the literature concerning the effect of caffeine on adenosine stress single-photon emission computed tomography (SPECT).8,9 A previous report demonstrated that an increased dose of adenosine could overcome the caffeine antagonism on SPECT.9 Similar concerns apply to the FFR measurement in the catheterization laboratory. Regarding the effect of caffeine on the FFR measurement, there are limited data using intracoronary adenosine in a small number of subjects.10 Thus far, to our knowledge, the interaction between caffeine and intravenous adenosine in the FFR measurement is yet to be determined. 

The aims of this study were to examine whether caffeine attenuated intravenous adenosine-induced hyperemia and whether increased doses of adenosine overcame the caffeine antagonism in the FFR measurement. In this study, the FFR was measured using different adenosine doses and papaverine as a reference standard in patients who consumed and those who withheld from consuming caffeine-containing products. 


The study protocol was approved by the local ethics committee. Written informed consent was obtained from all patients before the examination. 

Study patients. This prospective, single-center study enrolled 47 patients who underwent clinically indicated coronary angiography and FFR assessment for intermediate coronary stenosis, which was defined as 40% to 70% on the basis of visual estimation during angiography. Exclusion criteria included acute myocardial infarction, second- or third- degree atrioventricular block, prior coronary artery bypass grafting, patients on theophylline-containing medications, and any contraindications for adenosine or papaverine. Of the 47 patients, 16 were asked to refrain from caffeine-containing products >24 h before the test (no-caffeine group) and 31 were not (caffeine group). In accordance with our institution’s routine protocol, patients were asked to abstain from food for >3 hours before the catheterization, but not from beverages. Because of a difference in dietary habits, the elderly Japanese population generally does not drink as much coffee as the Western population. In our preliminary study, serum caffeine levels in patients who underwent coronary angiography were lower than those in the previous studies.8-10 To produce a wide range of caffeine levels, 20 patients in the caffeine group were given orally 100 or 200 mg of caffeine (Estaron Mocha; SS Pharmaceutical Co, Ltd), which is comparable to 1-2 large shots of espresso,9 just before the catheterization.

FFR measurement. Diagnostic coronary angiography was performed through radial or femoral approach in multiple projections. After coronary angiography, distal coronary pressure and aortic pressure were simultaneously measured at baseline and during maximal hyperemia using commercially available FFR systems (RadiAnalyzer and PressureWire Certus; St. Jude Medical) and a 5 Fr guiding catheter without side holes, as described previously.11 The FFR was calculated on a beat-to-beat basis as the mean distal coronary pressure (Pd) divided by the mean aortic pressure (Pa) during maximal hyperemia. 

Hyperemic stimuli. Intracoronary isosorbide dinitrate (2 mg) was administered before each stimulus. Two different orders of hyperemic agents (adenosine or papaverine first) were used alternately in order to exclude the influence of the order of stimuli. The second stimulus was administered after hemodynamic variables returned to baseline with an interval of at least 5 minutes. Adenosine was infused continuously via the large brachial vein at a dose of 140 µg/kg/min. Then, the infusion rate was increased to 175 µg/kg/min and thereafter to 210 µg/kg/min. All infusions were continued for at least 2 minutes, and data were obtained during the last 30 seconds of adenosine infusion at each dose.2 If a beat-to-beat variation in Pd/Pa ratio was observed, the FFR was defined as the lowest Pd/Pa ratio. The existence of cyclic variation was noted when Pd/Pa fluctuated >0.05 during the last 30 seconds at each dose of adenosine. Intracoronary papaverine (10-12 mg in the right coronary artery or 15-20 mg in the left coronary artery)4 was given through the coronary catheter followed by 5 mL of saline.

Quantitative coronary angiography. Quantitative coronary angiography was performed using an auto-edge detection method with a commercially available system (CMS version 6.0; Medis) by an independent investigator who was blinded regarding the patients’ clinical data. Reference diameter, minimum lumen diameter, and percent diameter stenosis were measured using the external diameter of the catheter as a scaling device.

Serum caffeine. Just before the administration of adenosine, blood samples were taken from the guiding catheter. Serum caffeine concentrations were measured using an enzyme-multiplied immunoassay technique. 

Statistical analysis. Statistical analysis was performed with PASW Statistics 18 (SPSS, Inc). Unless otherwise indicated, data were expressed as mean ± standard deviation for quantitative variables and as frequency with percentage for categorical variables. Comparisons between different stimuli were done with the paired t-test or Wilcoxon signed-rank test for quantitative variables and with McNemar test for categorical variables. Hemodynamic parameters at increased doses of adenosine were compared at 140 µg/kg/min dose or after papaverine. Between-group comparisons were done with the unpaired-samples t-test or Mann-Whitney U-test for quantitative variables and with the χ2 test or Fisher’s exact test for categorical variables. Correlation between caffeine level and difference in Pd/Pa ratio between papaverine and 140 µg/kg/min dose of adenosine was assessed with the use of Spearman’s rank correlation coefficient. A P-value <.05 was considered a statistically significant difference.


Patient characteristics. In the no-caffeine group, adenosine was discontinued at a dose of 210 µg/kg/min due to hypotension in 1 patient and intolerable chest pain in another patient. In the caffeine group, early termination of adenosine infusion was required at a dose of 175 µg/kg/min due to the development of third-degree atrioventricular block in 1 patient and at a dose of 210 µg/kg/min due to intolerable chest pain in 2 patients. Accordingly, the study protocol was completed in the remaining 28 patients of the caffeine group and 14 of the no-caffeine group. Indications for diagnostic cardiac catheterization included an abnormal or equivocal test result on magnetic resonance imaging (n = 14), SPECT (n = 6), and computed tomography angiogram (n = 4), presence of typical angina (n = 8), and angiographic follow-up after stent implantation (n = 10). Patient characteristics are summarized in Table 1. All target vessels had Thrombolysis in Myocardial Infarction (TIMI) flow grade 3. The FFR was measured in vessels located in the vascular territory of a previous infarction in 4 patients in the caffeine group and 3 patients in the no-caffeine group. Successful cannulation of the pressure wire distal to the stenosis was achieved in all patients. There were no procedure-related complications.

Caffeine concentration. In the caffeine group, the median consumptions of coffee and tea within 24 hours before the test were 0.5 cups (interquartile range [IQR], 0-1 cups) and 2 cups (IQR, 1-3 cups), respectively. The serum caffeine levels in the caffeine group ranged from 1.1 to 9.6 mg/L (median, 2.9 mg/L; IQR, 1.8-4.6 mg/L). In the no-caffeine group, serum caffeine levels were undetectable (<1 mg/L) in all patients.

Hemodynamic responses. The hemodynamic responses to hyperemic stimuli are shown in Table 2. At baseline and after papaverine, there were no significant differences in the hemodynamic parameters between the two groups. In the no-caffeine group, no further decrease in Pd/Pa ratio was observed during adenosine infusion above a dosage of 140 µg/kg/min (P=.53 [140 µg/kg/min vs 175 µg/kg/min] and P=.67 [140 µg/kg/min vs 210 µg/kg/min]). The mean Pd/Pa ratios induced by adenosine were not significantly different from those induced by papaverine (P=.56, P=.49, and P=.52 at doses of 140, 175, and 210 µg/kg/min, respectively). In contrast, in the caffeine group, the dose of 210 µg/kg/min produced a further decrease in Pd/Pa ratio compared with 140 µg/kg/min (P=.20 [140 µg/kg/min vs 175 µg/kg/min] and P<.01 [140 µg/kg/min vs 210 µg/kg/min]). Compared with papaverine, adenosine significantly overestimated Pd/Pa ratio by 0.034 (P<.001) at a dose of 140 µg/kg/min, 0.028 (P<.01) at 175 µg/kg/min, and 0.016 (P=.01) at 210 µg/kg/min. In the subgroup with caffeine level of ≤ 2.0 mg/L (n = 10; mean caffeine level, 1.6 ± 0.3 mg/L), the differences in Pd/Pa ratio between adenosine (140 µg/kg/min, 0.797 [P=.07]; 175 µg/kg/min, 0.795 [P=.15]; and 210 µg/kg/min, 0.793 [P=.19]) and papaverine (0.778) were small and not significant.

Figure 1 shows the individual values of Pd/Pa ratio at baseline and at peak action of each stimulus. Functionally significant lesions, defined as papaverine-induced FFR of ≤0.80, were present in 16 and 9 patients in the caffeine and no-caffeine groups, respectively. In the no-caffeine group, none had a false-negative result. On the other hand, in the caffeine group, false-negative results were demonstrated in 5 (31%), 4 (25%), and 2 (13%) patients at doses of 140, 175, and 210 µg/kg/min, respectively. Notably, Pd/Pa ratio during adenosine remained above the currently accepted upper limit of the gray zone of FFR (0.75 to 0.80) in 2 patients (22%) and 1 patient (11%) at doses of 140 and 175 µg/kg/min, respectively. In the caffeine group, cyclic variations during adenosine infusion were observed in 11 patients (39%), 8 patients (29%), and 8 patients (29%) at doses of 140, 175, and 210 µg/kg/min, respectively. In the no-caffeine group, cyclic variation occurred in only 1 patient (7%) at a dose of 140 µg/kg/min. The frequency of cyclic variation in the caffeine group was greater at all doses of adenosine (P<.05) than in the no-caffeine group. Caffeine level in the caffeine group was not significantly correlated with the difference in Pd/Pa ratio between papaverine and 140 µg/kg/min dose of adenosine (P=.76). Pertinent findings for each patient in the caffeine group are listed in Supplementary Table 1 (available online at www.invasivecardiology.com).

Side effects. The adverse effects of hyperemic stimuli are summarized in Table 3. Although prolongation of the QT-interval was observed after papaverine injection in most patients, none had ventricular fibrillation or tachycardia. No other side effects were observed. During administration of adenosine, the frequency of symptoms increased in a dose-dependent manner. Aminophylline was not required to reverse the side effects of adenosine. 


The major findings in the present study are summarized as follows: (1) intravenous adenosine at a dose of 140 µg/kg/min resulted in overestimation of the FFR in patients who consumed caffeine; (2) increasing the adenosine dose up to 210 µg/kg/min partially surmounted the caffeine antagonism; and (3) concerns about tolerability and safety of increased doses of adenosine remained.

Effect of caffeine on FFR measurement. Previous studies showed that intravenous adenosine at a dose of 140 µg/kg/min could induce hyperemia comparable to intracoronary papaverine,3,12 and no further change in Pd/Pa was observed at doses of >140 µg/kg/min.3 In accordance with the previous studies, the FFR was not significantly different between papaverine and 140 µg/kg/min dose of adenosine when patients abstained from caffeine before the test. In contrast, the same dose of adenosine did not achieve maximal hyperemia in patients who consumed caffeine. These findings suggest that caffeine interferes with adenosine-induced hyperemia. 

Aqel et al demonstrated that intravenous administration of caffeine at 4 mg/kg did not alter FFR induced by 30-50 µg dose of intracoronary adenosine.10 There are several factors that potentially affect the discrepancy from our findings. First, the caffeine antagonism could differ according to the route of administration of adenosine. Intravenous adenosine is metabolized before it reaches the coronary artery because of its short half-life,2 whereas intracoronary administration achieves high levels of adenosine in the coronary artery without being metabolized. Second, recent reports suggested that 300 or 600 µg dose of intracoronary adenosine is required to induce hyperemia comparable to intravenous adenosine.13,14 Submaximal hyperemia due to insufficient doses of intracoronary adenosine might be associated with the lack of change in FFR. Additionally, the small sample size (n = 10) might be inadequate to prove a statistically significant change.

The present and previous studies9 did not find the acceptable caffeine level or concentration-response effect of caffeine levels on hyperemia. The results might be attributed to peripheral intravenous infusion of adenosine. In the presence of caffeine, adenosine concentration in the coronary artery could also affect the hyperemic response. Because of its short half-life, adenosine concentration in the coronary artery could vary between individuals and could be influenced by several factors, such as circulation time and presence of anemia.2 Although a dose-dependent relationship was not found in the present study, FFR was not significantly different between adenosine and papaverine in the subgroup with low caffeine level (≤2.0 mg/L). In addition, the mean caffeine concentration in the study that demonstrated the interaction between adenosine and caffeine was much higher (6.2 mg/L)9 than in the other (3.1 mg/L).8 Further studies are required to investigate the acceptable caffeine level for both intravenous and intracoronary adenosine at varying doses and the required time for abstention from caffeine on a large sample size. 

Increased doses of adenosine. Reyes E et al demonstrated that 210 µg/kg/min dose of adenosine surmounted the inhibitory effect of caffeine on adenosine stress SPECT.9 Although additional decreases in Pd/Pa ratio were observed at increased doses of adenosine in patients who consumed caffeine in the present study, even 210 µg/kg/min dose of adenosine did not achieve maximal hyperemia comparable to papaverine. A limited accuracy of SPECT compared with FFR15 might account for the discrepant result because the difference in Pd/Pa between papaverine and 210 µg/kg/min dose of adenosine was statistically significant, albeit modest. In addition, our results raise concerns about tolerability and safety for its widespread use.

Cyclic change in FFR. We found a higher frequency of cyclic variation in the caffeine group than in the no-caffeine group. It was presumed that cyclic hyperemia resulted from a variation in adenosine concentration in the coronary artery and an insufficient trough level of adenosine for induction of maximal hyperemia.2 In the presence of variation in adenosine concentration in the coronary artery, a variation in coronary vasodilator effect could be augmented by caffeine. This phenomenon is clinically relevant in both the catheterization laboratory and imaging center. Cyclic changes in Pd/Pa during the pressure pullback curve recording hamper accurate assessment in diffuse or multiple lesions. In non-invasive stress testing, continuous systemic and coronary hemodynamics cannot be monitored. Under the condition of cyclic change, injection of contrast agents or tracers at the low ebb of hyperemia can result in false-negative results. 

Clinical implications. Our results support the recommendations in protocols for non-invasive adenosine stress testing that patients should refrain from caffeine before the test.16,17 In clinical practice, it is not uncommon that an intermediate stenosis with undetermined functional significance is incidentally diagnosed by coronary angiography. As it now stands, it is preferable that caffeine is withheld before coronary angiography when possible. Otherwise, alternative drugs that are unaffected by caffeine, such as papaverine and nicorandil,18 would be appropriate. Although intracoronary adenosine10 or intravenous bolus injection of regadenoson19 might overcome the caffeine antagonism, further investigation would be required. Recently, a promising vasodilator-free pressure-derived index of coronary stenosis severity (instantaneous wave-free ratio)20 has been introduced. A hybrid instantaneous wave-free ratio-FFR approach21 with the aforementioned vasodilators might be feasible if clinically validated.

Study limitations. This study has possible limitations. First, a direct comparison was not performed between before and after administration of caffeine in the same patient for ethical reasons. Second, the present study was conducted on a relatively small number of patients. Lastly, adenosine was not administered through a central vein because we performed transradial coronary catheterization in most of our study patients. However, a recent study has reported that peripheral adenosine infusion induces hyperemia comparable to central infusion.22 


The intake of caffeine before FFR measurement attenuates peripheral intravenous adenosine-induced coronary hyperemia. Even though an increase in adenosine dose up to 210 µg/kg/min might partially surmount the caffeine antagonism, concerns about tolerability and safety remain. Our results suggest that abstention from caffeine is necessary before adenosine stress testing to avoid submaximal hyperemia.


  1. Olsson RA, Davis CJ, Khouri EM, Patterson RE. Evidence for an adenosine receptor on the surface of dog coronary myocytes. Circ Res. 1976;39(1):93-98.
  2. Wilson RF, Wyche K, Christensen BV, Zimmer S, Laxson DD. Effects of adenosine on human coronary arterial circulation. Circulation. 1990;82(5):1595-1606.
  3. De Bruyne B, Pijls NH, Barbato E, et al. Intracoronary and intravenous adenosine 5’-triphosphate, adenosine, papaverine, and contrast medium to assess fractional flow reserve in humans. Circulation. 2003;107(14):1877-1883.
  4. Pijls NH, Sels JW. Functional measurement of coronary stenosis. J Am Coll Cardiol. 2012;59(12):1045-1057.
  5. Smits P, Schouten J, Thien T. Cardiovascular effects of two xanthines and the relation to adenosine antagonism. Clin Pharmacol Ther. 1989;45(6):593-599.
  6. Fredholm BB, AP IJ, Jacobson KA, Klotz KN, Linden J. International union of pharmacology. Xxv. Nomenclature and classification of adenosine receptors. Pharmacol Rev. 2001;53(4):527-552.
  7. Salcedo J, Kern MJ. Effects of caffeine and theophylline on coronary hyperemia induced by adenosine or dipyridamole. Catheter Cardiovasc Interv. 2009;74(4):598-605.
  8. Zoghbi GJ, Htay T, Aqel R, Blackmon L, Heo J, Iskandrian AE. Effect of caffeine on ischemia detection by adenosine single-photon emission computed tomography perfusion imaging. J Am Coll Cardiol. 2006;47(11):2296-2302.
  9. Reyes E, Loong CY, Harbinson M, Donovan J, Anagnostopoulos C, Underwood SR. High-dose adenosine overcomes the attenuation of myocardial perfusion reserve caused by caffeine. J Am Coll Cardiol. 2008;52(24):2008-2016.
  10. Aqel RA, Zoghbi GJ, Trimm JR, Baldwin SA, Iskandrian AE. Effect of caffeine administered intravenously on intracoronary-administered adenosine-induced coronary hemodynamics in patients with coronary artery disease. Am J Cardiol. 2004;93(3):343-346.
  11. Vranckx P, Cutlip DE, McFadden EP, Kern MJ, Mehran R, Muller O. Coronary pressure-derived fractional flow reserve measurements: recommendations for standardization, recording, and reporting as a core laboratory technique. Proposals for integration in clinical trials. Circ Cardiovasc Interv. 2012;5(2):312-317.
  12. van der Voort PH, van Hagen E, Hendrix G, van Gelder B, Bech JW, Pijls NH. Comparison of intravenous adenosine to intracoronary papaverine for calculation of pressure-derived fractional flow reserve. Cathet Cardiovasc Diagn. 1996;39(2):120-125.
  13. Lopez-Palop R, Carrillo P, Frutos A, et al. Comparison of effectiveness of high-dose intracoronary adenosine versus intravenous administration on the assessment of fractional flow reserve in patients with coronary heart disease. Am J Cardiol. 2013;111(9):1277-1283.
  14. Leone AM, Porto I, De Caterina AR, et al. Maximal hyperemia in the assessment of fractional flow reserve: Intracoronary adenosine versus intracoronary sodium nitroprusside versus intravenous adenosine: the NASCI (Nitroprussiato versus Adenosina nelle Stenosi Coronariche Intermedie) study. JACC Cardiovasc Interv. 2012;5(4):402-408.
  15. Melikian N, De Bondt P, Tonino P, et al. Fractional flow reserve and myocardial perfusion imaging in patients with angiographic multivessel coronary artery disease. JACC Cardiovasc Interv. 2010;3(3):307-314.
  16. Henzlova MJ, Cerqueira MD, Mahmarian JJ, Yao SS, Quality assurance committee of the American Society of Nuclear C. Stress protocols and tracers. J Nucl Cardiol.  2006;13(6):e80-e90.
  17. Kramer CM, Barkhausen J, Flamm SD, Kim RJ, Nagel E; Society for Cardiovascular Magnetic Resonance Board of Trustees Task Force on Standardized Protocols. Standardized cardiovascular magnetic resonance imaging (CMR) protocols, society for cardiovascular magnetic resonance: board of trustees task force on standardized protocols. J Cardiovasc Magn Reson. 2008;10:35.
  18. Jang HJ, Koo BK, Lee HS, et al. Safety and efficacy of a novel hyperaemic agent, intracoronary nicorandil, for invasive physiological assessments in the cardiac catheterization laboratory. Eur Heart J. 2013;34(27):2055-2062.
  19. Tejani FH, Thompson RC, Iskandrian AE, McNutt BE, Franks B. Effect of caffeine on spect myocardial perfusion imaging during regadenoson pharmacologic stress: rationale and design of a prospective, randomized, multicenter study. J Nucl Cardiol. 2011;18(1):73-81. Epub 2010 Nov 17.
  20. Sen S, Escaned J, Malik IS, et al. Development and validation of a new adenosine-independent index of stenosis severity from coronary wave-intensity analysis: results of the ADVISE (Adenosine Vasodilator Independent Stenosis Evaluation) study. J Am Coll Cardiol. 2012;59(15):1392-1402.
  21. Petraco R, Park JJ, Sen S, et al. Hybrid iFR-FFR decision-making strategy: implications for enhancing universal adoption of physiology-guided coronary revascularization. EuroIntervention. 2013;8(10):1157-1165.
  22. Seo MK, Koo BK, Kim JH, et al. Comparison of hyperemic efficacy between central and peripheral venous adenosine infusion for fractional flow reserve measurement. Circ Cardiovasc Interv. 2012;5(3):401-405.


*Joint first authors.

From the 1Cardiovascular Center, Rakuwakai Otowa Hospital, Kyoto, Japan;  2Department of General Medicine, Nara City Hospital, Nara, Japan; and 3Department of Systems Science, Kyoto University Graduate School of Informatics, Kyoto, Japan. 

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted January 17, 2014, provisional acceptance given April 7, 2014, final version accepted May 15, 2014.

Address for correspondence: Hidenari Matsumoto, MD, PhD, Cardiovascular Center, Rakuwakai Otowa Hospital, Kyoto, 2, Otowachinji-cho, Yamashina-ku, Kyoto, 607-8062, Japan. Email: matsumoto.hidenari@gmail.com