Exercise-Induced Desaturation During a Six-Minute Walk Test in Patients with Pulmonary Arterial Hypertension : A Retrospective, Observational Study

Background: The study aimed to evaluate the practical implication of exercise-induced oxygen desaturation (EID) in pulmonary arterial hypertension (PAH) patients. Methods: We conducted a single-center, retrospective, observational study from April 2016 to March 2018. Twenty patients diagnosed with PAH after right-sided heart catheterization (RHC) were assessed using trans-thoracic echocardiogram (TTE), pulmonary function test (PFT) , 6-minute walking test (6MWT) with gas analysis, body composition test, and muscle power tests. The occurrence of adverse events was assessed. Results: Ten patients showed EID during the 6MWT. Patients were divided into an EID group and a non-EID group. Epidemiological characteristics, TTE data, and RHC data did not differ between the two groups. Forced expiratory volume in the rst second (FEV 1 ), 6-minute walking distance, and peak oxygen consumption were signicantly lower in the EID group (all p<0.05). The EID group showed higher risk of emergency room visits, readmission, lung transplantation surgery, and death (OR = 13.5, 95% CI 1.20– 152.21). We developed a predictive scale for exercise-induced desaturation (PSEID) in PAH (AUC=0.91, 95% CI 0.75–1.00). A PSEID score of 4 or more predicted EID with 90% sensitivity and 90% specicity. Conclusions: EID occurred in half of the PAH patients during the 6MWT, and this group showed poor prognosis with more events, such as emergency room visits, re-admission, lung transplantation, and death. Our PSEID using 6MWT and PFT can assist physicians in the early identication of patients at risk of adverse events.

and the occurrence of exercise-induced desaturation (EID) are correlated with mortality. 6, 7 Some patients with PAH show EID; however, no previous studies have revealed a relationship between the severity of PAH and EID. The aim of this study was to evaluate the occurrence of EID and elucidate the contributing factors in patients with PAH.

Study design and population
This was a single-center, retrospective, observational study. The enrolled patients had been diagnosed with PAH using trans-thoracic echocardiogram (TTE) and right-sided heart catheterization (RHC) between April 2016 and March 2018. We categorized the types of PAH into four groups according to etiology: idiopathic PAH = group I, congenital heart disease = group II, connective tissue disease (CTD) = group III, and portopulmonary hypertension = group IV. We excluded patients with any uncorrected congenital heart disease, including patent foramen ovale. The patients who were in a decompensated state requiring advanced or intravenous medical therapy, patients who were not able to walk without oxygen support, and patients with other physical problems which interfered with exercise were excluded. Finally, 20 patients (male = 3, female = 17) were included with a mean age of 46 years ( Table 1). The study protocol was approved by the institutional review board (approval number: H-1903-018-077) of Pusan National University Hospital, Busan, Korea. The study was conducted according to the principles of the Declaration of Helsinki, and the requirement for informed consent was waived due to the retrospective nature of the research. Six-minute walk test with gas analysis Exercise performance was evaluated using the 6-minute walk test (6MWT) with percutaneous oxygen saturation (SpO 2 ) monitoring and gas analysis. The exercise tests were performed when the patient was considered to be in a medically stable state just before discharge or at the outpatient clinic. 6MWT was performed on a 30-metre walking track according to the ATS guidelines and observed by one skilled physical therapist. 9 SpO 2 was observed using the wrist Ox 3150 pulse oximeter (Nonin Medical Inc., Plymouth, MN, USA) during the 6MWT. Using the K4b2 system (Cosmed, Rome, Italy), peak oxygen consumption (VO 2peak ) and the minute ventilation/carbon dioxide production (VE/VCO 2 ) slope were measured simultaneously. 10 Muscle function test Grip strength (GS) was tested using a hand-held dynamometer (Jamar® Hydraulic Hand Dynamometer; Sammons Preston Patterson Medical Products Inc., Bolingbrook, IL, USA). 11 In a seated position with elbow exed at 90 degrees, patients were asked to grip the dynamometer as strongly as possible three times. Knee extensor strength (KES) was measured using a digital hand-held dynamometer (Jtech Medical, Salt Lake City, UT, USA). 12 Patients sat in a chair with folded arms and knees bent at a 35-degree angle. The head of the dynamometer and the leg were wrapped with a belt and KES was measured three times.

De nition of sarcopenia
Based on the consensus de nition of the Asian Working Group for sarcopenia, [13][14][15] we diagnosed sarcopenia if the following conditions were met: (1) a gait speed < 1.0 m/sec or a maximal gait speed < 28 kg for men or < 18 kg for women and (2) Appendicular skeletal muscle mass (ASM)/height 2 < 7.0 kg/m 2 for men or < 5.7 kg/m 2 for women. Bioimpedance analysis was used to measure the muscle mass (Inbody; Biospace, Seoul, Korea).

De nition of event
We de ned an event as any of the following events that occurred during the rst year after the exercise test: (1) emergency room (ER) visit, (2) re-admission due to cardiopulmonary problem, (3) lung transplantation surgery, and (4) death.

Statistical analysis
Data were analyzed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA). We used the Wilcoxon ranksum test for evaluating continuous data. Qualitative data were analyzed with the Fisher's exact test. To evaluate the relationship between EID and cardiac, pulmonary, and exercise function, a Pearson correlation analysis was used. Statistical signi cance was de ned by a p-value of less than 0.05. The receiver operating characteristic (ROC) curve analysis was performed using SPSS for Windows version 21.0 (SPSS Inc., Chicago, IL, USA).

Cardiac and pulmonary function
Using TTE, systolic function, evaluated with LV ejection fraction and LV GLS, and diastolic function, evaluated with the E/E' ratio, did not differ between the two groups. Mean PAP, PVR, PCWP, CI, and RV GLS were not signi cantly different. FEV 1 was signi cantly lower in the EID group (70% vs. 83%, p = 0.037). FVC and DLCO did not differ between the two groups (Tables 2 and 3).   Table 3). Eight patients met the criteria for sarcopenia, and seven were in the EID group (p = 0.020, odds ratio (OR) = 21.00, 95% con dence interval (CI) = 1.780-248.10) ( Table 1).

One-year prognosis
In the EIP group, six patients required an ER visit and readmission. Among these, one patient died, and one patient underwent lung transplantation surgery. In the non-EIP group, only one patient required an ER visit and readmission (p = 0.057, OR = 13.50, 95% CI = 1.20-152.21).
When PSEID was same or above 4, the occurrence of EID was predicted with 90% sensitivity and 90% speci city (Table 5 and Fig. 1).  17 reported that patients with PAH desaturated during the 6MWT, we aimed to investigate the exerciseinduced desaturation during exercise and the severe COPD that may explain this phenomenon.
In advanced COPD patients, ventilation capacity decreases for the following reasons: reduced elastic lung recoil, increased airway narrowing, increased airway resistance, expiratory limitation, and dynamic hyperin ation. 18 Furthermore, decreased FEV 1 , increased reserve volume, sputum accumulation, and large bullae disturb effective gas exchange. Breathlessness and desaturation eventuate from the imbalance between increased ventilatory demand and decreased ventilatory capacity. 5  revealed that EID was a predictive factor for the decline in the functional capacity of patients with COPD. 22 However, there are some studies that reported the existence of EID in patients with PAH, while there were no studies on the relationship between EID and outcomes. Therefore, the risk assessments suggested by the 2015 ESC/ERS guidelines do not include the occurrence of EID. 5 In this study, half of the PAH patients showed EID during the 6MWT, and they showed lower 6MWD and FEV 1 . Interestingly, functional class-categorized by subjective symptoms-did not in uence the occurrence of EID. There was no statistical difference in the distribution of WHO classi cations in the EID and non-EID groups. Further, Manes et al. demonstrated that patients in different clinical subgroups showed hemodynamic and survival differences but no WHO functional class differences. 16 Therefore, subjective symptoms during daily living or exercise are insu cient to evaluate PAH patients, and PFT and exercise testing with monitoring of oxygen saturation should be evaluated.
Andrianopoulos et al. reported that the sensitivity of a baseline SpO 2 ≤ 95% for the prediction of EID was 81.0%, speci city was 49.2%, and positive and negative predictive values were 50.8% and 80.0%, respectively. 7 Additionally, they suggested that a DLCO < 50%, an FEV 1 < 45%, and a PaO 2 < 10 kPa could predict the occurrence of EID. In this study, 6MWD, percentage of predicted 6MWD, percentage of predicted DLCO, percentage of predicted FEV 1 , and VO 2peak were correlated with EID. We suggest that a PSEID predictive scale include these factors, as well as the presence of ILD and sarcopenia. These tests are commonly used for PAH patients. Further, as mentioned in the above results, the predicted accuracy of the PSEID was higher than expected, with 90% sensitivity and speci city. Thus, it would be applicable for clinical use. However, as there is a limit to conducting research on PAH patients in a single center, it is believed that such a "PSEID" needs to be extended to a multi-center study in order to be useful in clinical use.
There was no signi cant difference regarding the comorbidities in the two groups in this study. However, this could be a statistical error given the small sample size. Four of 10 patients in the EID group were diagnosed with interstitial lung disease (ILD) on chest CT scan. In contrast, none of the 10 non-EID patients had ILD. For the reasons mentioned above, we assume that PAH etiology would not be different between the two groups. PAH with connective tissue disease (CTD) was found in one of 10 patients in the non-EID group and four of 10 patients in the EID group. Four CTD patients showed ILD patterns on chest CT, and they also showed signi cantly lower DLCO. Exercise based pulmonary rehabilitation and breathing retraining are effective therapeutic interventions that improve physical performance, shortness of breath, and the quality of life in COPD patients. 23 However, the effect of rehabilitation on the improvement of desaturation or FEV1 is still controversial because exercise could not restore the destroyed lung parenchyma and airways. In PAH, the safety and effectiveness of exercise training has now been well established. A meta-analysis of 16 trials found the overall risk of adverse events during exercise to be 4.7%. 24 Another meta-analysis found that exercise training led to improvements in the 6MWD, VO 2peak, and peak workload. 25

Conclusion
EID occurred frequently during the 6MWT. It occurred in half of the PAH patients, and this group had a worse prognosis with more events, such as emergency room visits, re-admissions, lung transplantations, and deaths, in our study. We conclude that pulmonary function, exercise capacity, and functional status are more closely related to the occurrence of EID, rather than cardiac function. Our PSEID can assist physicians in the early identi cation of patients at risk of adverse events.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.