INTRODUCTION
In 2015, the European Society of Cardiology (ESC) and the European Respiratory Society (ERS) guidelines proposed a multiparametric tool of risk stratification to estimate the prognosis of patients with pulmonary arterial hypertension (PAH) and guide treatment strategies. These recommendations suggest classifying patientsas low (mortality risk estimation <5% at 1 year), intermediate (mortality risk estimation between 5-10% at 1 year) and high (>10% mortality risk estimation at 1 year) risk, using clinical, imaging, and hemodynamic variables and estimation of aerobic capacity, which are known to be associated with patient prognosis. Among the latter, guidelines suggest using the World Health Organization (WHO) functional capacity (FC), the sixminute walking test (6MWT) and oxygen consumption (VO2) assessment in cardiopulmonary exercise testing (CPET). 1 This risk evaluation tool in PAH was retrospectively validated with different methods by three groups of investigators. The three validation models demonstrated an adequate capacity to discriminate patient prognosis. However, regardless of the method used, no group considered the use of CPET or VO2 assessment, employing only the WHO FC and 6MWT, 2-4 which represents a limitation, as both the FC and 6MWT present a prognostic performance inferior to CPET. 5,6 Firstly, there is no standardized way of assessing FC, a subjective parameter which reflects in the high interobserver variability in its evaluation, and the poor correlation between FC and maximum VO2. 7,8 The 6MWT also presents several limitations: it is influenced by age, body mass index, presence of comorbidities, 9 has little correlation with hemodynamic variables, 10 the changes in the distance walked are not translated into modifications in the prognosis, 11,12 there is a “learning” effect, 13 and above all, it has a “ceiling” effect which determines its poor sensitivity in low-risk patients. 14
The objective of our study was to establish what proportion of low-risk patients with PAH in FC I or II and who walk >440 meters in the 6MWT present moderate or high-risk parameters in CPET.
METHODS
Variable design and definition
A multicenter cross-sectional study was carried out including patients >18 years with PAH, in agreement with the 2015 ESC/ERS guideline definition, 1 who were at low-risk acaccording to the simplified risk evaluation method of the French registry: FC I-II, NT-proBNP <300 pg/mL and 6MWT >440meters. 4 Patients with congenital diseases were excluded from the study.
The CPET was performed in a treadmill (H/P Cosmos, Mercury Med, Germany) within 90 days following risk stratification and the respiratory gas exchange was continuously analyzed (Cosmed Quark CPET OMNIA 1.6/7 software). A Bruce or modified Bruce protocol was used, selected by the operator according to the clinical conditions of the patient.
Dynamic electrocardiographic changes (ST depression >1 mm) and arrhythmia occurrence were reported. The recorded variables included heart rate (HR), blood pressure, peripheral arterial oxygen saturation (SaO2), VO2, rate of carbon dioxide production (VCO2) and minute ventilation (VE). The respiratory quotient or respiratory exchange ratio (RER=VCO2/ VO2) was used as indicator of maximum exercise. Maximum exertion was considered for RER >1.1. Peak VO2 was definedas average VO2 during the last minute of exercise and was expressed as milliliters/minute/kilogram of body weight (ml/min/kg), and was also reported as percentage of the predicted VO2 value (according to prespecified tables which take into account sex, age and body surface area). Functional capacity was defined as normal when peak VO2/predicted VO2 was ≥85%. The three variables considered for PAH risk evaluation in the ESC guideline were: 1 peak VO2, its percentage with respect to predicted VO2 and the VE/VCO2 slope. The proportion of patients presenting these parameters in a risk category above low was established (peak VO2 ≤15 ml/min/ kg, percentage with respect to predicted VO2 ≤65% and a VE/ VCO2 slope ≥36.
Statistical analysis
Categorical variables were expressed as frequency and percentage, and were analyzed using the chi-square test or Fisher´s exact test, as appropriate. Continuous variables were expressed as mean and standard deviation, or median an interquartile range (IQR), according to their distribution. Student´s t test or the Wilkinson rank sum test was used to compare two groups, as appropriate. Statistical significance was considered for p <0.05. Stata 13.0 software package was used for statistical analyses.
RESULTS
A total of 18 patients were included in the study and 16 (89%) were women. Median age was 43.5 years (IQR 33-51 years) and median time from diagnosis to evaluation was 4.7 years (IQR 1.8-8.6 years). In half of the cases (n=9) the etiology of HAP was idiopathic, in 6 cases it was associated with connective tissue disease, in 2 cases it was secondary to human immunodeficiency virus (HIV) and in one case to portal hypertension. Baseline population characteristics are described in Table 1.
All patients evaluated presented an exercise functional capacity below 85% of the predicted value (Table 2). Only one female patient presented the three parameters assessed in the low-risk level. Eight patients (44%) had at least one abnormal parameter, 7 patients (39%) presented 2 altered parameters and in 2 patients all the parameters were anomalous. The most frequently altered parameters were the percent- age with respect to predicted VO2 and the VE/VCO2 slope in 67% of cases. Only 4 patients presented VO2 <15 ml/kg/min. No patient evidenced peak VO2 or percentage of predicted VO2 values in the high risk category. However, 6 patients (33%) presented a VE/ VCO2 slope considered as high risk (Figure 1).
DISCUSSION
Despite having included a population of patients who at the time of CPET were at a low-risk level and with preserved functional capacity, all the patients evaluated presented a peak VO2 <85% of the predicted value, which indicates at least mild impairment of functional capacity. 15 Only one patient presented all the parameters evaluated within the low-risk profile. Surprisingly, this patient was a woman treated with three drugs: phosphodiesterase-5 inhibitor, endothelin receptor antagonist, and intravenous prostanoid, which were started simultaneously in the context of cardiogenic shock 30 months prior to this evaluation. Although this evolution would seem unexpected, it is consistent with the results recently published by Boucly et al., who report a 5-year survival of 91% in 76 high-risk patients who were treated with initial triple therapy, including a parenterally administered prostanoid. 16
The rest of the patients (92%) presented at least one altered parameter in CPET. This may be due to the inherent limitations of FC assessment and the diagnostic sensitivity of the 6MWT. The population of our study was significantly younger than the low-risk population of the European registries in which the risk stratification proposed by the ESC/ERS guidelines was validated: median age of 43.5 years vs. mean of 52±17 years in the German registry; mean of 57±17 years in the French registry and median of 57 (IQR 39-68) years in the Swedish registry. 2-4 The fact that the patients are younger may affect the sensitivity of the 6MWT to detect FC impairment. 14 In the same sense, the fact that patients have several years of disease evolution and therefore have performed the 6MWT on repeated occasions can generate a "learning" effect that conditions its diagnostic capacity. 13) Recent publications have sought to subdivide the moderate-risk population into two strata: moderate- low and moderate-high risk. 17 Even Badagliacca et al. demonstrated and validated the value of CPET in the substratification of moderate-risk patients. 18 This substratification has important implications when making therapeutic decisions, both pharmacological as in time of inclusion on the lung transplantation list. 19 However, and despite the fact that low- risk patients present a non-negligible rate of events, as the 5-year mortality in the Swedish, French, and German registries was 8%, 9% and 31.9% respectively, 2-4 little has been done to identify which patients considered to be at low risk are more likely of disease progression and even of suffering fatal events. Similarly, the lowest mortality, 8% at 5 years, reported by the Swedish registry, does not seem at all auspicious if we consider that our population had an average age of 43 years. The CPET is then seen as a very useful tool to better identify those patients at higher risk who require a more aggressive treatment strategy with triple therapy and/or parenteral drug administration.
A recent study analyzed the mortality of incident patients diagnosed with PAH according to whether they started treatment before or after the publication of the 2015 ESC/ERS guidelines, which advocate the initiation of treatment with two drugs. 20 Although the investigators found that as of 2015 the percentage of patients who started combination treatments had significantly increased, this did not translate into a reduction of mortality; nevertheless, what was associated with lower mortality was reaching a low-risk stratum. In this same sense, although our work does not allow us to draw definitive conclusions, it can hypothesize about the importance of reaching a state of "absolute low risk". We understand as such, an exhaustive risk assessment strategy based on a method that allows FC to be estimated in an objective, measurable, reproducible manner that is correlated with hemodynamic parameters and whose variations correlate with the prognosis of the disease, such as CPET. (5,6) Another clinical scenario of potential value for CPET concerns those patients who, thanks to treatment, have reached a low-risk state and for some reason the possibility of de-escalating treatment is being considered. Although, in principle, we do not believe that there is any valid situation that warrants de-escalating treatment when the objective, that is so difficult to achieve, has been reached, it is not unusual for this possibility to be raised, both in the literature and in daily practice. 21,22. In this sense, CPET could establish if the patient really has a preserved FC prior to the start of medication reduction and could detect a possible deterioration of the FC before it is clinically manifest. However, and mindful of the presence of studies that show that patients who worsened their condition by reducing the intensity of treatment do not recover the lost well-being when a rescue treatment is administered, we strongly advise against venturing to reduce the intensity of treatment if no serious adverse effects occur. 23.
Our work has multiple limitations. In the first place, it is a small sample due to the fact that we are analyzing a disease considered rare, where most of the patients present at moderate or high-risk levels, and less than 30% are in FC I or II. 2-4 Secondly, although these are all patients who have been diagnosed a while ago, there is a great dispersion regarding the time of disease development. Another point that deserves to be mentioned is the heterogeneity that exists in terms of treatment. Although 12 patients (67%) are being treated with 2 drugs (an endothelin receptor antagonist and a 5-phosphodiesterase inhibitor), 2 are treated with 1 alone, and 4 with 3 drugs: 2 with selexipag, 1 with intravenous epoprostenol and 1 with subcutaneous treprostinil. This treatment heterogeneity suggests that, although they are patients with different evolutionary stages, it does not rule out the hypothesis underlying this study, which does not intend to compare one patient with another directly, but rather the possibility of disease progression with the consequent need to eventually require a more aggressive treatment, either pharmacological or a transplant, according to the baseline treatment.
CONCLUSIONS
All except one of the patients considered as low risk and who walk more than 440 meters in 6 minutes, presented at least one altered variable in CPET. The VE/VCO2 slope and the percentage of predicted O2 consumption were the most frequently altered variables. The VE/VCO2 slope was the only one that showed values considered as high risk. The CPET could have a place in the substratification of low-risk patients and thus allow the identification of a stratum of "absolute low risk". The importance of reaching this level of "absolute low risk" in terms of prognosis and patient survival should be evaluated in prospective studies.