Appropriate selection of exercise intensity in patients with interstitial lung disease

Training intensities in pulmonary rehabilitation are commonly based on xed percentages of peak heart rate (HRpeak), heart rate reserve (HRR), or peak work load (Wpeak). For patients suffering from interstitial lung disease (ILD) it is unknown, whether those intensities are appropriate when compared to individual ventilatory thresholds (VT1 and VT2) derived from cardiopulmonary exercise testing (CPET). The aim of this study was to compare xed HR percentages with HRs at VT1 and VT2. tests Comparisons HRs VT1


Introduction
Interstitial lung diseases (ILDs) are characterized by exertional dyspnea, exercise-induced hypoxemia, and exercise intolerance [1]. Thus, exercise limitation is a common feature in patients with ILD and is closely associated with increased mortality, particularly in idiopathic pulmonary brosis (IPF) [2]. Major contributors to exercise limitation in ILD include alterations in pulmonary gas exchange, ventilatory and skeletal muscle dysfunction [2]. Reduced diffusion capacity and impaired pulmonary circulation due to capillary destruction and hypoxic pulmonary vasoconstriction result in insu cient oxygen-hemoglobin saturation during exercise [3,4]. Exertional hypoxemia was shown to attenuate cerebral oxygenation, potentially affecting exercise tolerance [5]. Beside hypoxemia, abnormal heart rate responses to exercise have been demonstrated, associated with low exercise capacity and poor prognosis [6]. Moreover, quadriceps muscle force (20-25%) was shown to be reduced in ILD compared with healthy controls, considerably contributing to exercise impairment regardless of the underlying type of ILD [7][8][9]. More pronounced muscle atrophy in skeletal muscles of the lower limbs compared to upper limbs suggests physical inactivity as an important cause of muscle dysfunction and exercise limitation in ILD patients [10].
Exercise training represents a key component of pulmonary rehabilitation for people suffering from chronic lung disease including ILD, associated with improvement of symptoms, physical function and quality of life [11][12][13]. Principles of exercise training in patients with chronic respiratory disease are comparable with those valid for healthy individuals [14,15], including personalized exercise prescription and progression of training load [11]. The exercise intensity applied is of utmost importance for training success and is commonly set at xed percentages of peak values of walking velocity, heart rates or workloads [16][17][18][19]. However, such xed percentages may not re ect optimal exercise intensities in patients suffering from various heart or lung diseases [20,21].
Incremental cardiopulmonary exercise testing (CPET) represents the tool of choice to assess exercise capacity, cardiovascular risk, and functional capacity, and thus, the most valuable basis for developing exercise prescription and assessing training effects on an individual basis [21]. CPET provides two important measures, the ventilatory threshold 1 (VT1) and 2 (VT2), that allows to differentiate between exercise intensity domains, i.e., moderate, high, severe, and extreme [21], which can be assessed reliably and reproducibly and performed safely even in patients with severe exercise intolerance [22].
Although a threshold-based training model may be superior to the relative percentage concept [23], it seems not to be widely applied in pulmonary rehabilitation including ILD [16][17][18]. Thus, the relationship between VT1 and VT2 derived from CPET and xed percentages of peak heart rate (HRpeak), heart rate reserve (HRR), and peak work load (Wpeak) remains to be evaluated, especially for ILD patients.
The aim of this study was to compare the individual heart rates at VT1 and VT2 with those at 70%Wpeak, 70%HRpeak and 60%HRR. Due to the speci c limitations in ILD we hypothesized that the relation between those intensity measures would differ within different types of ILD and from those of a sedentary healthy control population.

Subjects
A total of 120 patients, who were referred to the department of pulmonology, medical university of Vienna between 2018-2020, were included in this study, 80 patients with diagnosis of interstitial lung disease and 40 age-, weight-and height-matched control subjects (table 1). All patients included in this study had CPET assessment data available. The study was conducted in accordance with the ethical principles laid down in the declaration of Helsinki 1975 and the protocol was approved by the Ethics committee of the medical university of Vienna.

Cardiopulmonary exercise test (CPET)
All patients underwent a symptom limited CPET on an Ergoline 800 bicycle (Sensormedics, United States) with respiratory gas-exchange analysis, using a step protocol with progressive increase in workload every minute according to a total exercise time between 8-12 minutes. The increment was adapted to the expected maximum working capacity. Patients were encouraged to exercise until exhaustion. A cycling frequency of 60-80 revolutions per minute (rpm) had to be maintained. The test was ended when the subject failed to maintain a pedal frequency of at least 60 rpm. Blood pressure was measured every 2 minutes and continuous 12-lead electrocardiogram and oxygen saturation (SpO 2 ) were recorded. Breathby-breath minute ventilation (VE), carbon dioxide output (VCO 2 ) and oxygen uptake (VO 2 ) were measured using Sensormedics 2900 Metabolic Measurement Cart. The respiratory exchange ratio (RER) was de ned as VCO 2 /VO 2 , the oxygen pulse was calculated by VO 2 /heart rate and the ventilatory equivalent for oxygen uptake (VE/VO 2 ) and the ventilatory equivalent for carbon dioxide production (VE/VCO 2 ) were measured. VT1 was determined using the V-slope method, double-checked by establishing the nadir of VE/VO 2 versus work rate relationship. VT2 was determined using the point of increase of the VE versus VCO 2 , double-checked by establishing the nadir of VE/VCO 2 versus work rate relationship. Blood gas analysis was measured at rest, at VT1 and at peak exercise. Absolute values were measured and % of predictive values were assessed using reference values for CPET by Hansen and Jones.

Statistical analysis
Statistical analysis was performed by IBM SPSS version 27.0 (IBM SPSS Statistics for Windows, Chicago,IL, USA). Normal distribution of the data was veri ed by the Kolmogorov-Smirnov test and Shapiro-Wilk test. After a descriptive data analysis, between group differences in baseline characteristics were analysed using the Student´s t-test for normally distributed data. For non-normally distributed data, Mann-Whitney U test was used to assess the group differences. All tests were conducted as two-sided and p-value<0.05 was considered signi cant. Comparisons between HRs at VT1 and VT2 and HRs at 70%HRpeak, 70%Wpeak and 60%HRR were performed using descriptive statistics presenting numbers and corresponding percentages. To visualize the differences between HRs determined at VT1 and VT2 and HRs assessed as a percentage at 70%HRpeak, 70%Wpeak and 60%HRR, the data was scaled using the min-max normalization, so that for every individual the values of VT1 and VT2 would correspond to the numbers 0 and 1 and the rest of the formulas was rescaled accordingly, with the same linear transformation. Figures 1 and 2 show boxplots of the scaled HR values determined by the 3 formulas, for ILD patients and the control group, respectively.

Subjects characteristics
Characteristics of ILD patients and controls are shown in table 1.
A total of 120 subjects were included for analysis, 80 patients with diagnosed ILD and 40 matched controls. The mean age of the ILD patients was 54.6 ± 13 years, 70 women (58%), 50 men (42%). Anthropometric data did not differ between ILD patients and controls. Patients with IPF were older than those with CTD and had a higher body mass compared to patients with CTD and sarcoidosis. Compared to controls, resting HRs were higer in patients with sarcoidosis, and SpO 2 values were lower in those with CTD and IPF.
Included types of ILD and pulmonary function in ILD patients are shown in table 2. Out of the 80 ILD patients, 32 suffered from IPF, 37 from connective tissue disease (CTD) and 11 from sarcoidosis. Twenty eight (37.5%) ILD patients had restrictive lung function. In the ILD group the mean forced ventilatory capacity (FVC) was 85.8 ± 21.4%pred and the mean carbon monoxide transfer factor (TLCO) was 60.4 ± 20.8%pred. None of the patients were on long-term oxygen therapy.
Responses to maximal exercise are shown in table 3.
Physiological responses (VO 2 , W, SpO 2 , HR) determined at maximal exercise were all signi cantly lower in ILD patients compared to controls. This is true for all types of ILD with the exception of sarcoidosis patients, who had similar HRpeak values as controls. VO 2 peak (%pred) was also higher in patients with sarcoidosis compared to IPF.
Ventilatory thresholds and heart rates at xed percentages of peak heart, peak power output, and heart rate reserve Ventilatory thresholds were signi cantly higher in %VO 2 peak (p<0.001), in %Wpeak (p<0.040) and %HRpeak (p<0.001) in the patient group with ILD compared to controls, whereas both, VT1 and VT2 were signi cantly lower at %VO 2 peakpred (table 4). Mean HRs at VT1 did not differ between groups, but mean HRs at VT2 were signi cantly lower in ILD patients. HRs at xed percentages, i.e., at 70%HRpeak, at 70%Wpeak and at 60%HRR were signi cantly lower in the ILD patients compared to controls. In all patients except one the VT2 could be assessed.
HRs at 70%HRpeak were lower than the HRs at VT1 in 66% of the IPF patients, 54% of the CTD patients and 55% of the patients with sarcoidosis ( gure 1) compared to 18% in the control group ( gure 2).

Discussion
In the present study, HRs at VT1 and VT2 have been compared to xed HR percentages, i.e., of 70%HRpeak, 70%Wpeak, and 60%HRR in patients with ILD, ILD subgroups and an age-matched healthy control group. Our ndings demonstrate differences in performance characteristcs and the related scattering of xed HR percentages when compared to the individual VT1 and VT2. Patients with ILD had lower exercise capacity (VO 2 peak and Wpeak) and lower cardorespiratory responses (HRpeak and SpO 2 peak) to maximal exercise than controls. Comparisons between ILD types revealed higher VO 2 peak (%pred) and peak HRs in patients with sarcoidosis compared to those with CTD, which is in agreement with other studies (2,3,31). Scattering of xed HR percentages is rather small for HRs at 70%Wpeak and 60%HRR but comparatively large for HR at 70%HRpeak ( gure 1). In contrast to the control group, HR at 70%HRpeak in ILD is at or slightly below the HR at VT1. However, the scatter range is probably too large to generate optimal individual training effects, because exercise intensity may be below VT1 in some ILD patients or above VT1 in others.
Assessment of appropiate exercise intensities in patients with chronic diseases becomes more and more important. It has been suggested that people with ILD may need more careful planning and modi cation of their exercise prescription than healthy subjects or even patients with COPD [25]. Generally, VTs derived from CPET ensure individual physiological adaptations to exercise and can help to nd the optimal training "zones" [27]. VT1 and VT2 form boundaries for the determination of 3 training zones (from low to high) successfully applied in athletes and patients as well [21,28]. Whereas in athletes the largest proportion of the training volume is performed at intensities below VT1 [28], in patients suffering from lung diseases, including ILD, intensities above VT1 are preferentially applied in rehabilitation [12,16,29]. This is at least partly based on the early study by Casaburi et al., who evaluated effects of various training intensities in COPD patients. These authors found reduced ventilatory requirements and improved exercise tolerance after training at intensities above VT1, due to metabolic adaptations within the working muscles resulting in lower blood lactate concentration, diminished carbon dioxide production and associated lower exercise ventilation [29].
The individual application of training intensities based on CEPT is particularly needed by patients suffering from different diseases. For instance, several training studies in chronic heart failure patients implicated the VT1 as an useful and valid method for individual training prescription [27,30,31]. In those patients, the proper assessment of training intensity was emphasized because of the high inter-patient variance. Similarly, intensity prescription based on HR identi cation at the VT was also highlighted for patients with left-ventricular dysfunction (LVDF) [24]. Even in healthy subjects it was shown that exercising according to a xed HRR for 12 weeks, VO 2 peak was increased in only 42% of the total group when compared to a signi cantly improved VO 2 peak in all individuals exercising according to the range between VT1 and VT2 [23,32]. It was also suggested that due to the heterogenity of ILD patients, i.e., those suffering from sarcoidosis, modi cation and program adjustment of the standard pulmonary rehabilitation format, including individual prescription of training intensity, are required [33]. Our ndings con rm the large variability of heart rate responses to exercise (CPET) and the considerable scattering of xed HR percentages in comparison to HRs at the individual VTs in ILD patients. Thus, as claimed for cardiac rehabilitation [31], or even more important, the approach of xed HR percentages may be inaccurate in a large proportion of ILD patients undertaking rehabilitation and should be replaced by individual VTs determined by CPET.
To the best of our knowledge, this is the rst study comparing HRs at the individual VTs and xed HR percentages. Thus, the presented ndings derived from a relatively large cohort of ILD patients not only highlight the importance of CEPT but also provide valuable basis for training intensity prescription for those patients.
This study may be limited by the inter-observer variability in the determination of ventilatory threshold. In order to minimise the bias the ventilatory thresholds were determined and cross-checked by two different observers. The patients in our study were only mild to moderately limited, which explain on one side that the VT2 could be asessed in all but 2 patients and on the other hand the relatively mild impairement in exercise capacity, which was nevertheless signi cantly lower compared to the control group.

Conclusions
The presented ndings demonstrate large variability of HR responses to exercise (CPET) in ILD patients and a considerable scattering of xed HR percentages in comparison to HRs at the individual VTs. Thus, when compared to xed HR percentages, the use of individual VTs is more appropriate to prescribe exercise intensity in the rehabiltation of ILD patients.

Declarations
Ethics approval: The study was conducted in accordance with the ethical principles laid down in the declaration of Helsinki 1975 and the protocol was approved by the Ethics committee of the medical university of Vienna (EC 1462).

Consent for publication:
Not applicable Availability of data and materials Page 8/17 All data generated or analysed during this study are included in this published article Funding:

No funding
Competing interests: The authors declare that they have no competing interests.
Author contributions: KV designed the study, had full access to all of the data in the study, performed study examination, acquired data, analysed and interpreted data and wrote the manuscript draft.
AL, DB, MRG, SS performed study examination and acquired data. PK, RHZ, MB analysed and interpreted data and wrote manuscript draft.
All listed authors read, revised and nally approved the manuscript and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved     Figure 1 Boxplots of the scaled HR values determined by the 3 formulas for ILD patients. Black lines in the graphics indicate the range between VT1 and VT2.