Sildenafil improves microvascular O2 delivery-to-utilization matching and accelerates exercise O2 uptake kinetics in chronic heart failure

Priscila A. Sperandio, Mayron F. Oliveira, Miguel K. Rodrigues, Danilo C. Berton, Erika Treptow, Luiz E. Nery, Dirceu R. Almeida, J. Alberto Neder


Nitric oxide (NO) can temporally and spatially match microvascular oxygen (O2) delivery (Q̇o2mv) to O2 uptake (V̇o2) in the skeletal muscle, a crucial adjustment-to-exercise tolerance that is impaired in chronic heart failure (CHF). To investigate the effects of NO bioavailability induced by sildenafil intake on muscle Q̇o2mv-to-O2 utilization matching and V̇o2 kinetics, 10 males with CHF (ejection fraction = 27 ± 6%) undertook constant work-rate exercise (70–80% peak). Breath-by-breath V̇o2, fractional O2 extraction in the vastus lateralis {∼deoxygenated hemoglobin + myoglobin ([deoxy-Hb + Mb]) by near-infrared spectroscopy}, and cardiac output (CO) were evaluated after sildenafil (50 mg) or placebo. Sildenafil increased exercise tolerance compared with placebo by ∼20%, an effect that was related to faster on- and off-exercise V̇o2 kinetics (P < 0.05). Active treatment, however, failed to accelerate CO dynamics (P > 0.05). On-exercise [deoxy-Hb + Mb] kinetics were slowed by sildenafil (∼25%), and a subsequent response “overshoot” (n = 8) was significantly lessened or even abolished. In contrast, [deoxy-Hb + Mb] recovery was faster with sildenafil (∼15%). Improvements in muscle oxygenation with sildenafil were related to faster on-exercise V̇o2 kinetics, blunted oscillations in ventilation (n = 9), and greater exercise capacity (P < 0.05). Sildenafil intake enhanced intramuscular Q̇o2mv-to-V̇o2 matching with beneficial effects on V̇o2 kinetics and exercise tolerance in CHF. The lack of effect on CO suggests that improvement in blood flow to and within skeletal muscles underlies these effects.

  • sildenafil
  • blood flow
  • heart failure
  • hemodynamics
  • near-infrared spectroscopy
  • oxygen consumption
  • kinetics

the inability to maintain an adequate driving pressure for blood-myocite oxygen (O2) diffusion [i.e., microvascular partial pressure of O2 (Po2mv)] is paramount to explain the slowness of exercise O2 uptake (V̇o2) kinetics in patients with chronic heart failure [CHF; as recently reviewed by Poole and colleagues (34)]. To keep a sufficiently high Po2mv, however, O2 delivery should be spatially and temporally matched to V̇o2 of individual fibers. In this context, seminal studies found that intramuscular Po2mv in rodents with CHF was critically low either at rest-to-contractions transition (7, 14) or during early recovery (12), i.e., when V̇o2 should be increasing or decreasing most rapidly, respectively. Importantly, it was demonstrated that reduced nitric oxide (NO) bioavailability exerted a key mechanistic role on on- and off-exercise O2 delivery-to-utilization uncoupling in these animal preparations (24, 25).

In intact humans with CHF, previous studies have concomitantly assessed the rate of change in phase II (“muscle”) V̇o2 (39) and intramuscular fractional O2 extraction {∼deoxygenated hemoglobin + myoglobin ([deoxy-Hb + Mb]) by near-infrared spectroscopy (NIRS)} (18) to gain insight into the dynamic matching of O2 delivery to utilization (9, 10, 37). This analysis is based on the widely accepted concept that Po2mv and changes thereof reflect the delivery/utilization ratio, i.e., V̇o2/microvascular O2 delivery (Q̇o2mv) as deoxygenation (18). In fact, decreases in O2 delivery relative to O2 needs speeded (and heightened) on-exercise [deoxy-Hb + Mb] kinetics but slowed postexercise [deoxy-Hb + Mb] recovery in patients with CHF (9, 10, 37). The therapeutic potential of increasing NO bioavailability to improve the dynamic coupling of Q̇o2mv to V̇o2, thereby accelerating V̇o2 kinetics and enhancing exercise tolerance, however, remains unexplored in these patients. Moreover, considering that stimulation of muscle metaboreceptors by hypoxia-related by-products might further increase the ventilatory drive, thereby predisposing to oscillatory breathing, it could be hypothesized that better O2 delivery-to-utilization matching induced by NO could contribute to breathing stability in CHF (23).

In the present study, therefore, we aimed to investigate the effects of increased NO bioavailability through acute pharmacological inhibition of muscle cGMP-specific phosphodiesterase-5 (PDE5) (32) by sildenafil intake (20) on peripheral muscle Q̇o2mv-to-V̇o2 matching and V̇o2 kinetics at the transition to and from constant work-rate exercise in patients with stable CHF. We hypothesized that compared with placebo, sildenafil would improve muscle oxygenation and accelerate V̇o2 kinetics with positive consequences on exercise tolerance in these patients.



This was a prospective study involving 10 nonsmoking men recruited from the CHF Outpatients Clinic of The São Paulo Hospital, Federal University of São Paulo. Patients had an established diagnosis of CHF (ischemic or idiopathic cardiomyopathy) for at least 4 yr, three-dimensional echodopplercardiography showing left-ventricle ejection fraction <35%, and New York Heart Association functional classes IIIII (Table 1). No patient had cardiac resynchronization therapy or a left-ventricle assist device. Patients were excluded from the study if they had clinical and/or functional evidences of chronic expiratory flow limitation (forced expiratory volume in 1 s/forced vital capacity ratio <0.7), anemia (actual Hb values as 15.5 ± 1.4 g%), unstable angina or significant cardiac arrhythmias, and myocardial infarction within the previous 12 mo. Renal function parameters were: serum creatinine = 1.2 ± 0.2 g/dl, estimated glomerular filtration rate (Cockroft-Gault) = 67.1 ± 8.7 ml·min−1 ·1.73 m−2, and urea = 48.4 ± 17.2 mg/dl. Study participants gave a written, informed consent, and the study protocol was approved by the Institutional Medical Ethics Committee.

View this table:
Table 1.

Resting and peak exercise data (n = 10)

Study protocol.

After familiarization, subjects performed an individualized ramp-incremental exercise test (5–10 W/min) to determine the difference between V̇o2 at the gas exchange threshold (GET; by the V-slope method) and V̇o2 at peak exercise (ΔV̇o2peak-GET). On different days, subjects underwent a constant work-rate exercise test to the limit of tolerance (Tlim; s) at a V̇o2 equivalent to 40–50% of the ΔV̇o2peak-GET (∼70–80% peak work rate if the GET was not identified) 1 h after sildenafil (50 mg) or placebo intake.

Cardiopulmonary exercise testing.

The tests were performed on an electronically braked cycle ergometer (Corival 400, Lode, The Netherlands) at 50 ± 5 rpm, and they were preceded and followed by unloaded pedaling at 0 W for 3 min. V̇o2 (ml/min), carbon dioxide (CO2) output (V̇co2; ml/min), minute ventilation (V̇e; l/min), and end-tidal partial pressures for CO2 (mmHg) were measured breath by breath (CardiO2 System, Medical Graphics, St. Paul, MN). During the incremental test, values were averaged as arithmetic mean of 20 s, and peak V̇o2 was the highest mean value recorded. The relationship between the rates of changes in V̇e and V̇co2 (ΔV̇e/ΔV̇co2) was established from unloaded exercise to the respiratory compensation point. Exercise oscillatory breathing, either during incremental or constant work-rate tests, was assumed as present when oscillations occurred ≥60% of exercise data at an amplitude >15% of resting values with minimal average amplitude of 5 l/min, and there was a regular oscillation (SD of three consecutive cycle lengths within 20% of the average) (31).

Skeletal muscle oxygenation.

Skeletal muscle oxygenation profiles of the left vastus lateralis were evaluated with a commercially available NIRS system (Hamamatsu NIRO-200, Hamamatsu Photonics, Japan). The methodology of continuous-wave NIRS has been described elsewhere (16). Among the NIRS variables, [deoxy-Hb + Mb] has been used as a proxy of fractional O2 extraction in the microcirculation, reflecting the balance between O2 delivery and utilization (18, 37). To improve intra- and intersubject comparability, values (μM/cm) were recorded as a delta (Δ) and expressed relative (%) to the amplitude of variation from baseline to the steady state (within 2 SD of the local mean) with placebo.

Central hemodynamics.

Cardiac output (CO) was measured throughout the constant work-rate tests using a calibrated signal-morphology impedance cardiography device (PhysioFlow PF-05, Manatec Biomedical, France). Values were recorded as mean of seven beats, as indicated by the manufacturer. The PhysioFlow device and its methodology have also been described in detail elsewhere (11). In preliminary experiments, the coefficient of variation (CV) for changes in CO during exercise was 3.3%, and stepwise changes in CO were consistent with those predicted from V̇o2 values, as described previously in CHF patients (27). Values (l/min) were recorded as a delta (Δ) from baseline and expressed relative (%) (27) to amplitude of variation from baseline to the steady state (within 2 SD of the local mean) after placebo.

Data analysis.

The breath-by-breath V̇o2, [deoxy-Hb + Mb], and CO data were time aligned to the start of exercise to 180 s after exercise cessation and interpolated second by second. Breaths outside 4 SD of the local V̇e mean were deleted, and V̇o2 was averaged into 5-s bins to further improve signal stability (26). The kinetics of these responses were determined by nonlinear regression using a least-square technique (Marquardt-Levenberg; SigmaPlot 10.0, Systat Software, San Jose, CA).

Exercise onset.

o2 was fitted from 30 s of baseline pedaling to 180 s after the onset of exercise. For [deoxy-Hb + Mb] kinetics, the analyses were conducted on data from 30-s baseline cycling to the steady-state response. The model used for fitting the kinetic response of V̇o2 and Δ[deoxy-Hb + Mb] without an “overshoot” (see below) was [Y](t)=[Y](b)+A[1e(tTD)/τ]6 (1) where b refers to baseline-unloaded cycling, and A, TD, and τ are the amplitude, time delay, and time constant of the exponential response (i.e., time to reach 63% of the final value), respectively. For V̇o2 analysis, we deleted the data relative to the cardiodynamic phase (9). The overall kinetics of V̇o2 and [deoxy-Hb + Mb] were determined by the mean response time (MRT) = τ + TD. For CO, we calculated the one-half time (t1/2; s), and an estimate of MRT was obtained as t1/2 × 1.44. A two-component, monoexponential model was applied to [deoxy-Hb + Mb] data with an overshoot (4, 10, 18) [Y](t)=[Y](b)+A1[1e(tTD1)/τ1]A2[1e(tTD2)/τ2](2) where the subscripts 1 and 2 correspond to the two sequential components (upward and downward, respectively). The area under the [deoxy-Hb + Mb] overshoot (AUC; a.u.) and a longer τ2 (Eq. 2) were used as additional indicators of impaired Q̇o2mv-to-V̇o2 matching (4) (Fig. 1). On preliminary trials, the CV for the kinetic parameters of the [deoxy-Hb + Mb] response ranged between 5% and 11% [first test − second test mean bias and range: τ = 0.3 s (−0.3 s to 0.8 s), and TD = 0.1 s (−0.4 s to 0.5 s)].

Fig. 1.
Fig. 1.

Skeletal muscle deoxygenated hemoglobin + myoglobin ([deoxy-Hb + Mb]) profiles at the onset of constant work-rate exercise after placebo (○) and sildenafil (●) intake in a representative patient with chronic heart failure (CHF). Values are expressed relative to the amplitude of the coefficient of variation from baseline to the steady state with placebo. TD, time delay; τ1, time constant of the upward component; τ2, time constant of the downward component; OS, overshoot.

Exercise recovery.

We used data from the last 30 s of exercise to 180 s of recovery to calculate the kinetics of the primary component of response. The model used for fitting the kinetic response of V̇o2 and [deoxy-Hb + Mb] was [Y](t) = [Y](ss) − A·[1 − e−(t − TD)/τ].

MRT of [deoxy-Hb + Mb] and CO was determined as described above. On preliminary trials, the CV for the kinetic parameters of the [deoxy-Hb + Mb] response ranged between 3% and 6% [first test − second test mean bias and range: τ = 3.4 s (−1.9 s to 7.9 s), and TD = 1.2 s (−1.4 s to 2.7 s)].

Statistical analysis.

Results were summarized as mean ± SD. The primary end point of the study was changes in on-exercise MRT-[deoxy-Hb + Mb] after sildenafil when compared with placebo. Secondary end points included Tlim, AUC (Eq. 2), τ-V̇o2/MRT-Δ[deoxy-Hb + Mb], V̇o2 kinetics, and amplitude of oscillatory breathing. To contrast within-subject resting and exercise responses, paired t- or Wlicoxon tests were used as appropriate. Pearson's product moment correlation was used to assess the level of association between continuous variables. The level of statistical significance was set at P < 0.05 for all tests.


Peak exercise capacity.

Peak work rate and V̇o2 of all patients were below the age- and gender-corrected lower limits of normality. Eight patients were on Weber's class C and two on class B. The GET was identified in all but two subjects. As anticipated by the long-term β-blocker therapy, patients showed impaired peak heart rate (Table 1). Exercise oscillatory breathing (31) was present in all but one patient during the incremental and constant work-rate exercise tests.

On- and off-exercise response dynamics after placebo.

All fitted data were included in the kinetics analysis, as coefficient of determination values ranged from 0.90 to 0.99. On- and off-exercise V̇o2 and CO kinetics with placebo were slower than reported previously in healthy males of similar age (10) (Table 2). On-exercise CO kinetics were consistently slower than V̇o2 dynamics, whereas off-exercise rates of change were remarkably similar (Fig. 2). There were significant relationships between these variables both at the onset of and during recovery from exercise (R = 0.69, and R = 0.81, respectively; P < 0.05).

View this table:
Table 2.

On- and off-exercise kinetic parameters for V̇o2, [deoxy-Hb/Mb], and cardiac output after placebo or sildenafil intake (n = 10)

Fig. 2.

Mean response time (MRT) of oxygen (O2) uptake (V̇o2), cardiac output (CO), and [deoxy-Hb + Mb] at the onset of and recovery from exercise after placebo (open bars) and sildenafil (solid bars) intake in patients with CHF. Data are mean (SD). *P < 0.05 for between-treatment comparisons on a given time point; †P < 0.05 for onset vs. recovery on a given treatment; ‡P < 0.05 for CO vs. V̇o2 and [deoxy-Hb + Mb] at a given time point; §P < 0.05 for V̇o2 vs. [deoxy-Hb + Mb] at a given time point.

At the onset of exercise, [deoxy-Hb + Mb] increased rapidly with a rate of change that was faster than the V̇o2 and CO responses. In contrast, MRT-[deoxy-Hb + Mb] during recovery did not differ from MRT-V̇o2 and MRT-CO (Fig. 2). An overshoot in [deoxy-Hb + Mb] was identified in eight of 10 patients (see Fig. 1 for a representative patient). AUC and the kinetics of its downward component were inversely related to on-exercise τ-V̇o2 (R = 0.80, and R = 0.73, respectively; P < 0.05).

Effects of sildenafil on exercise tolerance and physiological responses.

Sildenafil increased Tlim compared with placebo by ∼20% (P < 0.05; Table 3). Despite a lack of effect on the cardiopulmonary responses at exercise cessation (Table 3), sildenafil significantly decreased the amplitude of V̇e oscillations (% of mean) and their cycle length (21.4 ± 15.2% vs. 50.8 ± 10.4% and 19.9 ± 10.7% vs. 48.3 ± 12.1% for sildenafil and placebo, respectively). In addition, the number of cycles was reduced in two patients, and oscillatory breathing disappeared in other two patients.

View this table:
Table 3.

Main physiological responses just prior to exercise cessation after placebo or sildenafil intake (n = 10)

On- and off-exercise V̇o2 kinetics were accelerated by sildenafil intake (Table 2 and Figs. 24). There was a significant relationship between Δ (sildenafil-placebo) Tlim with treatment-induced decrements in τ-V̇o2 (R = −0.68; P < 0.05). In contrast, sildenafil failed to accelerate CO dynamics either at the onset of or recovery from exercise (Table 2 and Fig. 2). Consequently, V̇o2 and CO kinetics were no longer correlated after sildenafil treatment (P > 0.05).

Fig. 3.

Time course of V̇o2 at the onset of exercise after placebo (○) and sildenafil intake (●) in patients with CHF. Values were averaged in 10-s bins [mean (SE)] and expressed relative (%) to the amplitude of variation from start of loaded exercise to the 3rd min.

Fig. 4.

Individual effects of placebo and sildenafil intake in τ of V̇o2 at the onset (On-exercise; left) and recovery (Off-exercise; right) from exercise in patients with CHF. *P < 0.05.

On-exercise Δ[deoxy-Hb + Mb] kinetics were slowed by sildenafil (+25%; Table 2 and Fig. 2), and the AUC was lessened [median (range) = 2,085 (857–3,115) a.u. vs. 428 (125–628) a.u.; P < 0.05] or even abolished (n = 2; Fig. 1). Consequently, τ-V̇o2/MRT-[deoxy-Hb + Mb] after sildenafil was lower at the onset of exercise compared with placebo (1.77 ± 0.89 s vs. 3.31 ± 1.0 s, respectively; P < 0.05). In addition, [deoxy-Hb + Mb] recovery was faster with sildenafil (−15%; Table 2 and Fig. 2). Interestingly, improvement in on-exercise muscle oxygenation with active treatment {i.e., higher Δ (sildenafil-placebo) MRT-[deoxy-Hb + Mb]} was related to larger decrements in the amplitude of V̇e oscillations (R = −0.66; P = 0.045), faster V̇o2 kinetics (Fig. 5, top), and greater increases in exercise tolerance (Fig. 5, bottom).

Fig. 5.

Correlations between sildenafil-related changes in on-exercise MRT-[deoxy-Hb + Mb] with τ-V̇o2 kinetics (top) and time-to-exercise intolerance (Tlim; bottom) in patients with CHF.


This study is the first to demonstrate that compared with placebo, acute PDE5 inhibition with sildenafil led to a closer matching between intramuscular O2 delivery and utilization during the transition to and from constant work-rate exercise in patients with CHF. These beneficial consequences of active treatment speeded pulmonary V̇o2 kinetics, which suggests that peripheral muscle V̇o2 kinetics were faster, reduced the amplitude of oscillatory breathing, and increased exercise capacity. Improvements in muscle oxygenation and aerobic metabolism were not related to changes in CO dynamics, thereby suggesting that blood flow redistribution to and within the working muscles underlies these effects.

CHF-related O2 delivery-to-utilization mismatching.

Disease-induced decreases in O2 delivery relative to muscle O2 demands have been found to accelerate and amplify rest-to-exercise decrements in Po2mv and to slow its recovery after exercise (8, 12, 24, 25). This is physiologically equivalent to faster and heightened on-exercise increases in microvascular fractional O2 extraction (∼[deoxy-Hb + Mb], the mirror image of Po2mv) and slower postexercise [deoxy-Hb + Mb] recovery (9, 10, 18, 37). Moreover, fractional O2 extraction transiently overshoots the subsequent steady-state value (Fig. 1) when O2 delivery is markedly delayed relative to O2 needs (4, 9, 10, 18). All of these derangements are expected to increase O2 deficit and promote metabolic abnormalities associated with muscle fatigability and poor exercise tolerance (34). The effects of sildenafil on these abnormal patterns of intramuscular (de)oxygenation have not yet been determined in patients with CHF.

Sildenafil on muscle Q̇o2mv-to-V̇o2 matching.

The present study describes a novel mechanism (improved Q̇o2mv-to-V̇o2 matching) to explain the ergogenic properties of sildenafil (5, 9a, 22) in patients with CHF. The first hypothetical explanation for this finding is an increase in bulk muscle blood flow due to faster CO dynamics. However, our data indicate that this was not the case (Tables 2 and 3 and Fig. 2). In fact, Guazzi et al. (22) were the first to show that increased V̇o2/work-rate slope and faster postexercise V̇o2 kinetics with sildenafil were loosely related to improvements in central hemodynamics. These findings led the authors to hypothesize that CO had been redistributed to skeletal muscles after sildenafil administration, probably due to enhanced shear stress vasodilation in local conduit vessels. This hypothesis makes sense under the view that CHF is associated with impaired endothelial function and reduced NO bioavailability (15, 28), and sildenafil can improve flow-mediated vasodilation and local NO levels (21). Although we did not measure blood flow through these vessels, this might help to explain our findings of improved on- and off-exercise muscle oxygenation despite unaltered CO dynamics. However, vasodilation of feeding arteries is not an obligatory requisite for a better dynamic coupling between Q̇o2 and V̇o2 at the capillary level; in fact, the time course of skeletal muscle capillary hyperemia at the onset of exercise might differ markedly from the increased blood flow through larger conduit arteries (18, 35). Moreover, based on the stability of mean arterial pressure (Table 3) and CO, it is unlikely that there was a substantial decline in total peripheral vascular resistance after sildenafil. It is conceivable, therefore, that even if bulk blood flow increased with sildenafil (5, 21), it had the additional effect of more precisely matching intramuscular Q̇o2mv to V̇o2; i.e., active treatment might have redistributed blood flow within the working muscles (34).

Role of NO on Q̇o2mv-to-V̇o2 matching.

Diffusion of NO to surrounding vascular smooth muscle cells activates local cGMP, which induces vasodilation by inhibition of calcium influx into the cell, activation of calcium-ATPase pumps, and opening of potassium channels, thereby leading to hyperpolarization and relaxation (15, 38). Interestingly, however, NO can modulate not only Q̇o2mv but also, through inhibition of cellular respiration, muscle V̇o2 (36). Therefore, by acting on both determinants of tissue oxygenation (delivery and utilization), NO is the “ideal” candidate to exert a commanding role in muscle Q̇o2mv-to-V̇o2 matching (34). In line with the extensive evidence derived from animal studies (17, 24, 25), increased muscle NO bioavailability might have not only increased Q̇o2mv but also concomitantly decreased local O2 needs, thereby allowing better temporal and spatial Q̇o2mv-to-V̇o2 matching and a faster rate of transcapillary O2 flux.

Study implications and directions for future research.

The novel mechanistic explanation for the effects of sildenafil on exercise tolerance in CHF provided in the present study (Q̇o2mv-to-V̇o2 matching) may also help explain why measurements of central hemodynamics are poorly predictive of the ergogenic effects of sildenafil in these patients (21, 22). Moreover, our results open the perspective to combine sildenafil with other strategies known to increase muscle NO bioavailability [e.g., exercise training (2, 13) and l-arginine (30) or dietary nitrate supplementation (3)] in future trials in CHF. Considering that the speeding effect of sildenafil on postexercise muscle deoxygenation might have accelerated the rate of oxidative rephosphorylation (39), it seems interesting to investigate its value in improving patients' ability to perform repetitive activities, which are probably relevant to patients' daily functioning. Our sample comprised middle age subjects not older than 65 yr, and senescence is associated with further reductions in muscle NO bioavailability in CHF (7, 34). Therefore, we may have underestimated the beneficial effects of sildenafil on Q̇o2mv-to-V̇o2 matching, an issue to be investigated in older patients.

Confirming the reports of Guazzi and coworkers (22), we found a marked decrease in exercise-related oscillatory breathing with sildenafil, which was largely independent of changes in CO. Although we are uncertain whether changes in pulmonary vascular resistance and/or facilitation of intrapulmonary gas diffusion after sildenafil played a mechanistic role in this regard, improvements in oscillatory breathing and intramuscular oxygenation were significantly correlated. It is possible, therefore, that less accumulation of metabolic by-products due to better muscular O2 delivery-to-utilization matching reduced activation of groups III and IV afferents and the ventilatory drive, thereby contributing to breathing stability (9a, 19, 23, 33). Larger trials specifically powered to this issue are needed to further investigate the potential relationships among improved intramuscular O2 delivery-to-utilization coupling, enhanced muscle bioenergetics, decreased ergoreflex activation, and lower exercise ventilatory stimuli in CHF.

Study limitations.

We recognize that the small sample size and the acute nature of the present study indicate that it remains to be experimentally demonstrated whether the observed beneficial effects of sildenafil could be extrapolated to less-severe patients and whether they would be long lasting. These positive results, albeit obtained in a selected group of stable patients, now justify larger longitudinal investigations with a sizeable number of patients with different degrees of disease severity to look specifically at these physiological outcomes. It also remains to be tested whether these effects of sildenafil intake can also be reproduced in normal subjects.

As a noninvasive method, NIRS has a number of limitations (as discussed extensively elsewhere) (16); however, it remains the only available approach to interrogate the microcirculation (small arterioles, capillary, and venules) during whole-body exercise in intact humans. We restrained our analysis to [deoxy-Hb + Mb], as this variable is insensitive to blood volume changes, and its time course has been found to be remarkably equivalent to fractional O2 extraction when muscle venous outflow is carefully isolated in animal preparations [as discussed at length in Barbosa et al. (4) and Ferreira et al. (17)]. A caveat particularly pertinent to the present study, however, is its inability to differentiate Hb from Mb with regard to light absorption. However, by comparing muscle deoxygenation with Po2mv, determined by phosphorescence quenching within the same muscle region in rat preparations, Koga et al. (29) confirmed that the deoxygenation signal did provide a valid index of local fractional O2 extraction kinetics during exercise transients.

In the present study, we relied on a single exercise transition to extract the parameters of V̇o2 kinetics to avoid repetitive tests (and drug intake) in a very disabled population. Although we recognize that the signal-to-noise ratio would have been improved by averaging multiple transitions (1, 26, 39), the supra-GET bout provided a response amplitude from baseline (∼0.5 l/min of V̇o2 and ∼30 l/min of V̇e), which was large enough for fidelity of parameters estimation while avoiding the confounding effects of the ventilatory oscillations. Also, importantly, the effects of sildenafil on on-exercise τ-V̇o2 were commensurate with those found in response to interventions in CHF (1) and above the test-to-test variation in our laboratory during supra-anaerobic threshold exercise (10.2%).


Oversignaling of the NO pathway by PDE5 inhibition through a single dose of sildenafil enhanced on- and off-exercise O2 delivery-to-utilization matching at the microcirculatory level, reduced the amplitude of oscillatory breathing, and accelerated V̇o2 kinetics with positive effects on tolerance to endurance exercise in CHF. The lack of effect of sildenafil on CO suggests that improvement in blood flow to and within skeletal muscles is mechanistically linked to these actions. Our data indicate that sildenafil is a drug particularly suited to pathophysiology of CHF, as it enhances muscle oxygenation and pulmonary V̇o2 kinetics during exercise without the need of major functional improvements in the failing heart. Results of this cross-sectional investigation, however, should be confirmed in larger longitudinal studies to assess the long-term effects of sildenafil intake on these outcomes in patients with different degrees of disease severity.


L. E. Nery and J. A. Neder are Established Investigators of CNPq (Brazilian National Research Council). The authors have no relationship with industry.


Author contributions: L.E.N., D.R.A., and J.A.N. conception and design of research; P.A.S., M.F.O., M.K.R., D.C.B., and E.T. performed experiments; P.A.S., M.F.O., M.K.R., D.C.B., E.T., L.E.N., D.R.A., and J.A.N. analyzed data; P.A.S., M.F.O., E.T., L.E.N., D.R.A., and J.A.N. interpreted results of experiments; P.A.S., M.K.R., and J.A.N. prepared figures; J.A.N. drafted manuscript; P.A.S., D.C.B., E.T., L.E.N., D.R.A., and J.A.N. edited and revised manuscript; P.A.S., M.F.O., M.K.R., D.C.B., E.T., L.E.N., D.R.A., and J.A.N. approved final version of manuscript.


The authors are grateful to all colleagues from the Pulmonary Function and Clinical Exercise Physiology Unit [Federal University of São Paulo (UNIFESP), Brazil] for their friendly collaboration.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 7.
  7. 8.
  8. 9.
  9. 9a.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
View Abstract