Benefits of self-paced concurrent training on lung function, cardiopulmonary fitness and fatigue perception in patients with multiple sclerosis

Abstract

Aim: Studying the effects of self-paced concurrent high-intensity interval training and resistance training (HIIT-RT) on respiratory function, cardiopulmonary fitness and fatigue perception in patients with multiple sclerosis (PwMS).

Methods: Twenty-three PwMS were randomized into a 12-week training group (three times per week) (TG, n = 11) or a control group (CG, n = 12). Lung function (spirometry), aerobic capacity (graded cardiopulmonary-exercise-testing) and perceived fatigue (Fatigue Severity Scale (FSS)) were evaluated pre- and post-intervention.

Results: The forced vital capacity (p = 0.036, Hedges'g (g) = 0.93), forced expiratory time (p = 0.045, g = 0.88), peak expiratory flow (p = 0.043, g = 0.89) increased in TG compared with CG. The TG showed an increase in peak aerobic power (p = 0.004, g = 1.34) and peak oxygen uptake (p < 0.001, g = 2.58) compared with CG. There was a decrease in ventilatory equivalent for carbon dioxide (p = 0.02, g = 1.02) and FSS scores (p < 0.001, g = 1.72) in TG comparatively with CG.

Conclusion: 12-week self-paced HIIT-RT enhanced lung function as well as aerobic fitness, and alleviated fatigue perception in PwMS.

Graphical Abstract

Plain language summary

Article highlights.

Introduction

• Patients with multiple sclerosis could experience increased fatigue perception resulting, at least in part, from their respiratory dysfunction and poor aerobic capacity.

• Concurrent high-intensity interval training and resistance training (HIIT-RT) was effective in improving aerobic capacity in MS patients but no study has explored yet its effects on spirometric parameters and perceived fatigue in this population.

Aim & methods

• This parallel-group randomized controlled study aimed to investigate the effectiveness of 12-week self-paced HIIT-RT (3 sessions/week) on pulmonary function (spirometry), aerobic fitness and metabolic responses (graded cardiopulmonary-exercise-testing), as well as perceived fatigue (Fatigue Severity Scale) in patients with relapsing-remitting MS (RRMS).

Results

• The HIIT-RT group (n = 11) showed an improvement in respiratory function (i.e., an increase in forced vital capacity, forced expiratory time and peak expiratory flow) and cardiopulmonary fitness (i.e., an increase in peak aerobic power and peak oxygen uptake, as well as a decrease in ventilatory equivalent for carbon dioxide) compared with the control group (n = 12, sedentary behavior).

• The fatigue perception and peak lactate level were also reduced in the HIIT-RT group comparatively with the control group.

Discussion

• The aerobic fitness improvement could be due to the benefits of training (mainly HIIT) on mitochondrial biogenesis, which might increase lipid oxidation and reduce the involvement of anaerobic lactic pathway, thus explaining lactate decrease.

• The improved aerobic capacity induced by HIIT-RT could participate in enhancing spirometric parameters through its benefits on lung function, which might increase respiratory volumes.

• Fatigue alleviation could be attributable to the beneficial effects of self-paced HIIT-RT on aerobic fitness and deconditioning.

Conclusion

• Given the safety and the effectiveness of self-paced HIIT-RT found in this study, caregivers may involve this non-pharmacological intervention in the management of respiratory dysfunction, perceived fatigue, poor aerobic capacity and disturbed metabolic responses in RRMS patients.

1. Introduction

Multiple sclerosis (MS) is a chronic autoimmune and inflammatory disease of the central nervous system (CNS) that leads to demyelination and axonal degeneration [1]. Fatigue represents one of the most disabling and frequent symptoms of MS, affecting up to 57% of persons [2]. Indeed, in patients with MS (PwMS), fatigue could result either directly from CNS lesions (e.g., cortical or subcortical pathways) induced by inflammatory damage (primary fatigue) or indirectly from MS-related factors such as sedentary lifestyle, sleep disorders and pharmacotherapy (secondary fatigue) [3]. In this context, increased fatigue perception has been shown to be induced, at least in part, by poor cardiopulmonary fitness in PwMS [4]. In addition, it has been reported that some PwMS could experience a respiratory dysfunction [i.e., decrease in forced expiratory volume in the first second (FEV1) and forced vital capacity (FVC)], mainly at an advanced stage of the disease [5] or sometimes at an early MS stage but without pulmonary symptoms [6,7]. Indeed, this MS-related lung dysfunction might result from inspiratory and expiratory muscle weakness that could be one of the perceived fatigue causes [8]. Disturbed lung function has been shown to be involved in reducing aerobic capacity in PwMS, reflecting poor exercise tolerance [9].

As part of the non-pharmacologic MS treatment, interval training seemed to be the safest and the most effective exercise modality for PwMS, since it has been reported to be the least hyperthermic training [10]. This could avoid exacerbation of neurological signs and visual disorders resulting from body hyperthermia that might inhibit nerve conduction in demyelinated axons (i.e., Uhthoff's phenomenon) [11]. In this context, high-intensity interval training (HIIT) (i.e., aerobic exercises) has been shown to be more effective than moderate-intensity continuous aerobic training (MICT) in alleviating MS-related fatigue perception [12] and poor aerobic capacity [13]. However, according to a previous meta-analysis, aerobic training combined with resistance training (RT) showed more pronounced benefits than either intervention alone on MS-induced disorders such as perceived fatigue, muscle weakness and postural imbalance [14]. In this sense, the beneficial effects of MICT combined with RT on fatigue perception have been demonstrated in PwMS [15,16]. Nevertheless, the effects of concurrent HIIT and RT (HIIT-RT) (i.e., combined training involving performance of aerobic exercises (e.g. HIIT) and RT in the same workout session, but sequentially not simultaneously [17]) on perceived fatigue have not been studied yet in PwMS. In fact, HIIT-RT can be a better alternative than MICT-RT to alleviate fatigue because HIIT has been shown to have more pronounced anti-fatigue and anti-inflammatory effects compared with MICT [12,18].

The effectiveness of HIIT-RT in remedying impaired aerobic capacity, muscle weakness, poor quality of life and glucose intolerance has been demonstrated in PwMS [19,20]. In this context, Zeelmaekers and Rihon [21] reported that a 12-week HIIT-RT increased peak oxygen uptake (VO_2peak) and peak heart rate (HR_peak) in PwMS. Besides, it has been demonstrated that HIIT-RT was more effective than continuous training combined with RT in increasing VO_2peak and time to exhaustion in PwMS [22].

Nevertheless, for spirometric parameters, there are only a few studies regarding aerobic training or RT effects in PwMS [23–25], but MS studies concerning the effects of concurrent training and mainly HIIT-RT are still lacking. Moreover, given the mutual association between fatigue perception, aerobic capacity and spirometric lung function [4,8], concurrent training, particularly HIIT-RT, can be the best exercise modality for improving pulmonary function in PwMS.

Furthermore, it is increasingly recommended to involve self-regulation (e.g., self-selection of recovery time and exercise intensity based on the participant's perceived exertion) in training protocols, since the self-paced exercise has been shown to have benefits on perceptual [26], cardiorespiratory [27], psycho-cognitive [28–31] and physical [26,31–33] disorders compared with the imposed exercise in PwMS and other populations (e.g., elderly, obese persons). Additionally, performing exercises in a self-paced manner has been demonstrated to increase the participants' enjoyment and adherence to training sessions [29,30]. In this context, the effectiveness of self-paced HIIT compared with imposed HIIT in improving aerobic capacity has been attributed to the possibility for participants to choose to perform the exercises at a higher intensity during the self-paced training [27]. This could induce greater demands on their cardiorespiratory capacity and increase their glycogen depletion, thus promoting mitochondrial biogenesis via the stimulation of AMPK to respond to the exercise-related energy demand [27]. Moreover, it has been reported that the beneficial effects of self-paced RT on physical capacity could derive from its effectiveness in improving neuromotor function through the optimization of physiological processes that could promote neuronal and muscular adaptations [32]. Therefore, based on the above-mentioned arguments, the self-paced nature of exercise was chosen for the training program of the current study.

To the best of our knowledge, no study has explored yet the impact of self-paced HIIT-RT on spirometric parameters and perceived fatigue in PwMS. To provide an effective new therapy to alleviate MS-related cardiorespiratory dysfunction that could disrupt patients' social life and independence during daily activities [34], we aimed to investigate the effects of self-paced HIIT-RT on respiratory function, aerobic capacity, perceived fatigue and metabolic responses in PwMS. We hypothesize that all these parameters could be improved in the training group (HIIT-RT) compared with the control group (no training intervention).

2. Methods

2.1. Participants & ethical approval

This study was carried out among PwMS who were recruited from the Neurology department and evaluated in the department of Physical Medicine and Functional Rehabilitation of the Habib Bourguiba University Hospital. This research was performed after receiving a written informed consent from all participants. After approval of this study by the South Institutional Human Research Ethics Committee (N°0184/2019), its prospective registration was performed in the Pan African Clinical Trial Registry database (PACTR20200746530958). This study was conducted according to the World Medical Association Declaration of Helsinki.

2.2. Eligibility criteria

Included patients were aged under 60 years old and had a relapsing-remitting MS (RRMS) as well as an expanded disability status scale (EDSS) score <6 (i.e., ability to walk 100 m without assistance or rest) [35]. The non-inclusion criteria were as follows: pregnancy, obesity [body mass index (BMI) ≥30 kg/m²], MS relapses, training or physiotherapy program, cardiopulmonary and psychiatric diseases, uncorrected vestibular and visual disturbances as well as orthopedic disorders in the past year requiring medical supervision.

2.3. Study design & experimental protocol

For this parallel-group randomized controlled study, participants were assigned to either a training group (TG) or a control group (CG), and were assessed before (T0: pre-tests) and after (T1: post-tests) 12 weeks in the morning (8 a.m.). The pre- and post-tests included spirometric test and questionnaire to assess, respectively, lung function and perceived fatigue as primary outcomes, as well as cardiopulmonary exercise testing (CPET) to evaluate aerobic capacity and metabolic responses as secondary outcomes (Figure 1). Before beginning the protocol, collection of clinical and sociodemographic data was performed, and the impedancemetry, the Caen Chronotype Questionnaire [36], as well as the International Physical Activity Questionnaire [37] were used to determine anthropometric parameters, chronotype and fitness level, respectively.

Figure 1.

Experimental protocol.

Note: ETO₂: End-tidal oxygen pressure; ETCO₂: End-tidal carbon dioxide pressure; FET: Forced expiratory time; FVC: Forced vital capacity; FEV1: Forced expiratory volume in the first second, FIF_50%: Forced inspiratory flow at 50% FVC; FEF_50%: Forced expiratory flow at 50% FVC; FEF_25–75%: Forced expiratory flow at 25–75% of FVC; HR_peak: Peak heart rate; lactate_peak: Peak lactate level; PAP: Peak aerobic power; PEF: Peak expiratory flow; RPE: Rating of perceived exertion; RM: One-repetition maximum; T0: Pre-test; T1: Post-test; VO_2peak: Peak oxygen uptake; VeqO₂: Ventilatory equivalent for oxygen; VeqCO₂: Ventilatory equivalent for carbon dioxide; VE_peak: Peak pulmonary ventilation.

The CG participants were requested to maintain their usual lifestyle and complete the pre- and post-tests without receiving any exercise intervention throughout the study. The TG participants were asked to perform 12-week concurrent training (three sessions per week) in the afternoon (between 4 p.m. and 7 p.m.) in a gym. To familiarize the TG with the training equipment and exercises, four familiarization sessions were conducted prior to the training protocol (48 h after pre-tests) (Figure 1). The 1-RM was assessed in the last two sessions of familiarization (after pre-tests) and training program (before post-tests) for the following exercises: calf raise, squat, push down, arm curl, seated row and bench press. In these sessions, exercises were performed with the heaviest load that participants could lift a maximum of ten times without experiencing excessive discomfort. The 1-RM was estimated using this formula: “Estimated 1-RM = lifted load / [1.0278 - (0.0278 × number of repetitions (reps))]” [38]. The 1-RM was calculated only for the TG to partly control RT intensity and assess maximal dynamic muscle strength [39]. The assessment of 1-RM was not performed for the CG due to the unavailability of most participants in this group.

During each training session, TG performed, successively, 3–5 min of warm-up [on treadmill at 30% HR_peak, rate of perceived exertion (RPE) (Borg scale (6–20) = 6], two self-paced aerobic exercises using the treadmill and elliptical trainer or bike (HIIT), four self-paced strengthening exercises (upper and lower extremities, anterior and posterior trunk muscles) and finally self-stretching exercises (5–7 min). The HIIT was conceived according to the recommendations of the American College of Sports Medicine [40] and was performed in a self-paced mode. During weeks 1–6, participants performed 5 blocks of 1-min workout at 80–90% HR_peak (RPE = 16–18) interspersed with an active recuperation of 1-min workout at 40–50% HR_peak (RPE = 8–10), without leading to respiratory and perceptual disorders [27]. During weeks 7–12, a progressive increase was performed only in the high-intensity blocks duration which became 2-min workout at 80–90% HR_peak, so the HIIT total duration increased from 23 min in weeks 1–6 to 33 min in weeks 7–12, as presented in Figure 1.

Strengthening exercises were hypertrophic, designed based on the recommendations of Dalgas et al. [10], and performed in a self-paced manner. The lifted load, the exercises execution order, as well as the rest time between series and exercises were adjusted according to participants' RPE and capacities (e.g., rest time deemed sufficient to produce maximum effort during the next exercise or set) [41]. The exercises used in the present study were chosen based on previous MS studies [42–44]. In each session, participants had to select 4 exercises [upper (n = 1) and lower (n = 1) extremities, anterior (n = 1) and posterior (n = 1) trunk muscles]. The suggested exercises were as follows: (i) for upper extremity: shoulder exercises using pulley and cables machine or dumbbells, arm curl and push down; (ii) for lower extremity: leg extension, leg press, leg curl, squat, lunges and calf raise. For trunk, the following exercises were proposed: anterior muscles: abdominal crunch and sit-up, exercises using the converging chest press, bench press and Pec Deck machines, as well as posterior muscles: bipedal and unipedal bridge and exercises using the seated row and pull down machines.

As summarized in Figure 1, an augmentation in the training load was performed throughout weeks by increasing the number of sets and reps as well as the exercise intensity (% RM), as detailed below:

• Weeks 1–2: 3 sets × 10 reps with lightweight (35% RM; RPE = 7–10) (anatomical adaptation)

• Weeks 3–4: 3 sets × 12 reps with light to moderate weight (≤65% RM; RPE = 10–13)

• Weeks 5–6: 4 sets × 10 reps with light to moderate weight (<65% RM; RPE = 10–14).

• Weeks 7–8: 4 sets × 10 reps with moderate to heavy weight (65–70% RM; RPE = 14–16).

• Weeks 9–10: 4 sets × 8 reps with heavy weight (65–75% RM; RPE = 14–17).

• Weeks 11–12: 4 sets × 8 reps with a higher weight than weeks 9–10 (up to 80% RM; RPE = 18).

2.4. Testing

2.4.1. Aerobic capacity & metabolic responses: spiroergometric parameters

To assess cardiorespiratory capacity, participants performed a maximal CPET on an ergocycle (Ergoline, Ergoselect 100, France) under physician supervision using a validated protocol for PwMS [45,46]. During the first 3-min rest, diastolic (DBP) and systolic (SBP) blood pressure as well as HR were recorded. After 5-min warm-up (cycling at 0 watts), the test began at an initial power of 5 watts then a gradual increase of 15 watts/min was performed in the exercise intensity. During the test, participants were required to maintain their cycling rate at 60 rpm until they felt voluntarily fatigued [46]. HR was continuously recorded using a 12-lead electrocardiogram (ECG). Blood pressure (i.e., SBP + DBP) and pulse oxygen saturation (SpO₂) were measured every 3 min using manual blood pressure monitor and pulse oximeter, respectively. The RPE was evaluated at each level using the modified Borg scale (0–10) [47]. Gas exchanges were recorded throughout the test by a gas analyzer (Metasys TR-B, Brainware, Toulon, France) which was calibrated before each test with certified reference gases [16% of oxygen (O₂) and 4% of carbon dioxide (CO₂)]. The respiratory exchange ratio (RER = CO₂ production / O₂ consumption) was recorded every 15 seconds.

The CPET was considered complete and maximal if it met at least two of these criteria: VO₂ plateau, HR ≥90% HR_peak predicted by age (220-age), RPE >9 (Borg 0–10), RER ≥1.1 [46], lactatemia ≥8–10 mmol and cycling rate ≤45 rpm [48]. The test-discontinuation criteria included dyspnea, lower-extremity cramps, chest pain, pallor or cyanosis, SBP >250mmHg and/or DBP >115 mmHg, abnormalities on the ECG (arrhythmia and intraventricular conduction disorders) and SpO₂ <85% [49].

The analyzedparameters were as follows:

• VO_2peak (ml/min/kg), peak aerobic power (PAP) (watts) and HR_peak (bpm)

• Peak pulmonary ventilation (VE_peak) (l/min): reflects the variation in respiratory rate and tidal volume [VE = respiratory rate (cycle/min) × tidal volume (l)] [46]

• Ventilatory equivalent for oxygen (VeqO₂) and ventilatory equivalent for carbon dioxide (VeqCO₂): the quantity of air that must be ventilated to bring 1L of O₂ and eliminating 1L of CO₂, respectively. The decrease in VeqO₂ and VeqCO₂ reflects a better gas exchange and good respiratory function [50].

• Oxygen pulse (VO₂/HR) (ml/bpm): shows the amount of O₂ absorbed by the peripheral tissues in each heart beat (efficiency of cardiac pumping) [50].

• End-tidal oxygen pressure (ETO₂) (mmHg): its decrease reflects an hypoxemia resulting from exercise and its increase at the beginning of exercise indicates an hyperventilation [50].

• End-tidal carbon dioxide pressure (ETCO₂) (mmHg): its decrease in response to exercise indicates an alteration of the ventilation/perfusion ratio and hyperventilation. The increase in ETCO₂ during exercise reflects alveolar hypoventilation [50].

• Peak lactate level (lactate_peak) (mmol/l): was measured by the Lactate Pro 2 analyzer (Arkray, Shiga, Japan) at the last stage of CPET, as a biomarker of metabolic responses.

2.4.2. Lung function: spirometric parameters

Respiratory function was assessed using the Spirobank II spirometer (MIR, Rome, Italy). Assessment was performed in the sitting and lying positions to evaluate diaphragmatic weakness. In fact, a drop in FVC superior to 10% when moving from the sitting to the lying position indicates an unilateral diaphragm dysfunction (ΔFVC_position = (FVC_sitting - FVC_lying / FVC_sitting) >10%) [51]. A ΔFVC_position >30% reflects a bilateral diaphragmatic weakness [51]. The best (highest value) of the three trials performed in each position was taken for analysis. The FVC test was performed based on the recommendations of the spirometer user manual and the American Thoracic Society (ATS)/European Respiratory Society (ERS) [52]. Before starting the test, a nose clip was placed on the participants' nostrils to prevent the air passage. Then, participants were asked to (1) insert the turbine mouthpiece into their mouth with closing their lips around it, (2) perform two to three normal inhalations and exhalations, (3) realize one deep, fast and maximal inhalation followed by a 6-s brutal, rapid and maximal exhalation, and (4) finally perform a deep and maximal inhalation.

Participants were instructed not to smoke, practice exhaustive exercise for 60 min before the test, use intoxicants within 8 h before the test, and wear tight-fitting clothing that might prevent chest and abdominal distension [52].

For the acceptability criteria of the three trials, each expiratory phase was accepted only if participants did not (1) cough and close their glottis, especially during the first second of expiration, (2) stop the maneuver early, and (3) perform an incomplete maximal expiration [52]. The trial was also accepted if the extrapolated volume (i.e., volume of exhaled air before starting the test) was <0.15 l and there were no air leak, flow variability and mouthpiece obstruction [52]. The trial was also considered as acceptable if participants fulfilled one of these three criteria at the end of the forced expiration: expiratory time ≥15 s, plateau of expiration (i.e., volume ≤0.025 l in the last second of expiration), and reproducible FVC [52].

After accepting these three trials, their repeatability was checked by calculating the difference between the two highest FVC values and the two highest FEV1 values. If the difference was less than or equal to 0.15 l (for FVC and FEV1), the maneuvers were considered reproducible [52]. Otherwise, other trials were carried out until these two conditions were met (without exceeding eight trials) [52].

The following spirometric parameters were measured in this study:

• -The FVC (l), FEV1 (l), Tiffineau index: FEV1/FVC (%), peak expiratory flow (PEF) (l/s), forced expiratory time (FET) (s) and forced expiratory flow at 25–75% of FVC (FEF_25–75%) (l/s).
• -The FEV1/PEF: value >8 ml/l/min reflects an obstruction of the upper airways that may be extra or intra-thoracic [53].
• -To differentiate between extra and intra-thoracic obstructions, the forced inspiratory and expiratory flows at 50% FVC (FIF_50% and FEF_50%, respectively) were also measured in order to calculate the FIF_50%/FEF_50% ratio which is considered as normal if it is greater than 1. Indeed, this ratio is normal in the case of an intra-thoracic obstruction but it is less than 1 in the case of an extra-thoracic obstruction [53].

2.4.3. Fatigue perception

Perceived fatigue was assessed using the French version of the Fatigue Severity Scale (FSS) that represents a reliable and validated one-dimensional scale for PwMS [54]. This questionnaire is composed of 9 questions to which participants responded on a scale ranging from 1 “strongly disagree” to 7 “strongly agree” to indicate how much they agree with each item [54]. A total score ≥4 corresponds to a fatigue threshold score [54].

2.5. Power analysis calculation

Sample size calculation was performed a priory using the procedures suggested by Beck et al. [55] and the “G * POWER 3.1” software [56] where 0.05 and 0.95 were set as values of power and alpha, respectively. Given that this study is the first to investigate the effects of HIIT-RT on spirometric parameters and fatigue perception in PwMS, sample size was calculated based on a previous study that revealed the effectiveness of MICT-RT in alleviating MS-related perceived fatigue [43]. According to this latter study, estimation of effect size (ES) was set at 0.87 [43]. In total, to achieve the required power, data from around 20 participants (10 per group) would be enough to reduce the risk of type 2 statistical error. Given the necessity of taking into account the dropout rate, a revision of this sample (n = 20) was needed. In fact, the assumption of 10% for non-attendance in training program ≥7 sessions or for absenteeism in post-tests [20] resulted in a revised sample of 22 participants [(22.22) (N’) = N(20) / (1 - % dropout (0.10)].

2.6. Randomization & blinding

Procedure of block randomization (computerized random numbers) was used to randomize 30 adults with RRMS who were allocated to a TG (n = 15) or a CG (n = 15). This procedure was made by an external researcher who used an automatic web-based randomization program (http://www.jerrydallal.com/random/randomize.htm) (allocation ratio 1:1) to create the random allocation sequence that included the participants' identification code (ID) with the corresponding group. This sequence was generated using random block sizes of 2, 4 and 6 PwMS (block of 6 2 2 6 4 6 4) [57].

For blinding and masking, opaque and sealed envelopes containing participants' ID were used to make the investigators (i.e., who carried out the pre- and post-tests) blinded to the allocation. Opening of envelopes occurred prior to the first training session by the caregiver (Sonda Jallouli) who performed the training program and who was unaware of data collected during pre- and post-tests.

2.7. Statistical analysis

Data analysis was performed using the Statistica 12 software (StatSoft, Maisons-Alfort, France). Tables were used to present data as mean ± standard deviation (if normal distribution) or median (interquartile range [lower quartile; upper quartile]) (if non-normal distribution or discrete variables). When data were normally distributed (Shapiro-Wilk test) and variances were homogeneous (Levene test), parametric tests were conducted. In the absence of normality and homogeneity, we used non-parametric tests.

The qualitative data (i.e., chronotype, physical activity and educational levels and MS modifying treatments) were compared between TG and CG at T0 using the Chi-Square test of independence (χ²). Baseline comparisons between the two groups in quantitative data were performed using the independent samples t-test (weight, fat mass, BMI and age) or the Mann-Whitney U test (lean mass, total body water, MS duration and EDSS). The independent samples t-test was used to compare between TG and CG in baseline values of all spirometric and spiroergometric parameters except initial values of FVC in the sitting position, VE_peak, oxygen pulse, respiratory rate, ETCO₂ and FSS scores, which were analyzed using the Mann-Whitney U test.

The pre-post intervention change (Δ) was compared between both groups (Δ = values _after - values _before training or sedentary lifestyle) using the independent samples t-test for FSS scores as well as all spirometric and spiroergometric parameters except ETCO₂, oxygen pulse (VO₂/HR), respiratory rate, FVC in the lying position, FEV1, FIF_50%, FEF_50%, FEF_25–75%, FEV1/FVC, which were analyzed by the Mann-Whitney U test.

The two-way repeated measures ANOVA [group (TG vs CG) × time (T0 vs T1)] was used to compare ΔFVC position between the two groups. The changes from T0 to T1 in all 1-RM measures were analyzed within the TG using the dependent samples t-test.

For data analyzed by the independent samples t-test or Mann-Whitney U test, the determination of ES was performed by calculating the Hedges'g (g) (i.e., corrected form of Cohens'd (d)) given the sample size difference between both groups. For the paired samples t-test analyzed data, “d” was calculated as ES. The “g” and “d” were interpreted as follows: small (d <0.20; g = 0.15), medium (0.2 < d <0.5; g = 0.40) and large (d ≥0.8; g = 0.75) [58,59]. For assessing the ANOVA significance (for Δ FVC_position), the ES was calculated using the partial eta-squared (ηp²). The percentage change was calculated as follows: % Δ = (values _after - values _before / values _before) × 100. For statistical significance, a probability level of 0.05 was accepted.

3. Results

3.1. Recruitment & participant flow

PwMS were recruited between February and December 2021 until the required sample size was achieved. The study protocol started in January 2022 and the last follow-up occurred in June 2022. Among 90 PwMS screened for eligibility, 30 patients were included and randomized into two groups (n = 15 per group). Only 23 participants completed the whole study protocol and their data were taken for analysis [TG: n = 11 (6 women + 5 men) and CG: n = 12 (11 women + 1 man)]. The dropouts' reasons are displayed in Figure 2.

Figure 2.

CONSORT flow diagram illustrating the recruitment, randomization and follow-up of participants during the study.

3.2. Baseline data

At T0, there were no significant differences between the two groups in sociodemographic and anthropometric data, chronotype, physical fitness level and clinical data (Table 1). No significant differences were found at baseline between TG and CG in the primary outcomes (spirometric parameters and FSS scores) and secondary outcomes (spiroergometric parameters) (Table 2). Baseline clinical interpretation showed that the Δ FVC_position was inferior to 10% in all participants, thus indicating the lack of diaphragmatic weakness.

Table 1.

Baseline comparisons in participants' characteristics and clinical data between both groups.

	TG (n = 11)	CG (n = 12)	p-value
Sociodemographic and anthropometric data
Age (years)	33.09 ± 10.26	36.83 ± 8.07	0.34
Weight (kg)	67.03 ± 12.11	64.66 ± 13.30	0.85
BMI (kg/m2)	23.27 ± 3.35	22.96 ± 4.39	0.35
Lean mass (kg)	46.70 (18.60 [40.70; 59.30])	43.85 (7.90 [42.95; 50.85])	0.78
Total body water (kg)	34.40 (13.15 [30.25; 43.40])	31.85 (4.85 [31.10; 35.95])	0.44
Fat mass (kg)	17.05 ± 6.50	21.23 ± 11.86	0.29
Chronotype	Morning type (n = 3) Neither type (n = 8)	Morning type (n = 7) Neither type (n = 5)	0.13
Physical fitness	Weak (n = 3) Moderate (n = 8)	Weak (n = 5) Moderate (n = 7)	0.26
Educational level	Secondary school (n = 2) University degree (n = 7) Master degree (n = 2)	Middle school (n = 3) Secondary school (n = 3) University degree (n = 4) Doctoral degree (n = 2)	0.09

Clinical data
EDSS	1.50 (1.50 [1.00; 2.50])	2.75 (1.50 [2.00; 3.50])	0.14
MS duration (years)	2.00 (9.00 [1.00; 10.00])	4.50 (5.00 [3.00; 8.00])	0.83
Disease modifying treatments for MS	First-line treatment (n = 9) Second-line treatment (n = 2)	First-line treatment (n = 9) Second-line treatment (n = 3)	0.69

Note: Quantitative data are presented as mean ± standard deviation or median (interquartile range [lower quartile; upper quartile].

BMI: Body mass index; CG: Control group; EDSS: Expanded disability status score; MS: Multiple sclerosis; n: Participant number; TG: Training group.

Table 2.

Baseline comparisons in the scores of perceived fatigue as well as the spirometric and spiorergometric parameters between both groups.

Parameters	TG (n = 11)	CG (n = 12)	p-value	t or U
Fatigue perception
FSS scores (points)	4.30 (1.34 [4.10; 5.44])	4.32 (0.84 [4.10; 4.94])	0.71	U = 59

Spirometric parameters
FVCsitting position (l)	3.73 (1.53 [3.22; 4.75])	3.47 (0.77 [3.17; 3.93])	0.21	U = 45.00
FVClying position (l)	3.93 ± 0.84	3.66 ± 0.70	0.41	t = 0.84
Δ FVCposition (sitting-lying/sitting)	0.01 ± 0.05	-0.01 ± 0.09	0.43	t = 0.79
FET (s)	4.50 ± 1.09	4.60 ± 1.34	0.84	t = -0.19
PEF (l/s)	6.27 ± 1.88	5.48 ± 1. 09	0.33	t = 0.99
FEV1 (l)	3.37 ± 0.69	2.88 ± 0.65	0.09	t = 1.73
FEV1/FVC (%)	85.72 ± 14.33	80.60 ± 9.16	0.31	t = 1.03
FEV1/PEF (ml/l/min)	9.29 ± 1.84	9.35 ± 2.09	0.95	t = -0.07
FIF50% (l)	2.28 ± 0.96	2.05 ± 0.90	0.56	t = 0.59
FEF50% (l)	4.18 ± 1.39	3.26 ± 0.75	0.06	t = 1.98
FIF50%/FEF50%	0.57 ± 0.21	0.63 ± 0.22	0.55	t = -0.61
FEF25–75% (l/s)	3.67 ± 1.22	2.84 ± 0.97	0.08	t = 1.81

Spiroergometric parameters and metabolic responses at maximal CPET
Lactatepeak (mmol/l)	7.62 ± 1.69	7.75 ± 1.87	0.86	t = -0.18
VO2peak (ml/min/kg)	43.97 ± 7.53	44.54 ± 7.98	0.86	t = -0.18
PAP (watts)	114 ± 30.07	132.50 ± 22.61	0.11	t = -1.67
HRpeak (bpm)	146.00 ± 28.58	147.33 ± 29.25	0.91	t = -0.11
Tidal volume (l)	1.63 ± 0.65	1.69 ± 0.45	0.81	t = -0.25
VeqO2	47.64 ± 5.39	52.48 ± 7.10	0.08	t = -1.83
VeqCO2	50.39 ± 4.87	51.43 ± 6.01	0.65	t = - 0.45
ETO2 (mmHg)	118.73 ± 2.28	119.08 ± 2.91	0.75	t = -0.32
VEpeak (l/min)	62.00 (33.90 [44.90; 78.80])	56.78 (7.66 [54.62; 62.29])	0.69	U = 59.00
ETCO2 (mmHg)	42.00 (7.00 [38.00; 45.00])	43.00 (24.00 [35.00; 59.00])	0.85	U = 62.50
Oxygen pulse (ml/bpm)	40.21 (17.25 [32.55; 49.80])	48.27 (4.54 [45.42; 49.95])	0.16	U = 42.50
Respiratory rate (cycle/min)	40.00 (10.00 [36.00; 46.00])	36.00 (7.00 [33.00; 40.00])	0.13	U = 41.00

Note: Data are presented as mean ± standard deviation or median (interquartile range [lower quartile; upper quartile]).

CG: Control group; CPET: Cardiopulmonary exercise testing; ETCO2: End-tidal carbon dioxide pressure; ETO2: End-tidal oxygen pressure; FET: Forced expiratory time; FSS: Fatigue severity scale; FVC: Forced vital capacity; FEV1: Forced expiratory volume in the first second; FIF_50%: Forced inspiratory flow at 50% FVC; FEF_50%: Forced expiratory flow at 50% FVC; FEF_25–75%: Forced expiratory flow at 25–75% of FVC; HR_peak: Peak heart rate; Lactate_peak: Peak lactate level (at the last stage of CPET); PAP: Peak aerobic power, PEF: Peak expiratory flow; TG: Training group; t: t-test; U: Mann-Whitney U Test; VO_2peak: Peak oxygen uptake; VeqO₂: Ventilatory equivalent for oxygen; VeqCO₂: Ventilatory equivalent for carbon dioxide; VE_peak: Peak pulmonary ventilation.

3.3. Aerobic capacity & metabolic responses: spiroergometric parameters

Statistical analysis revealed that the Δ of PAP and VO_2peak were higher (i.e., a greater increase in these parameters) in TG than in CG (PAP: 30.65%, p = 0.004, g = 1.34; VO_2peak: 42.08%, p < 0.001, g = 2.58) (Table 3). T-tests also showed that TG presented lower Δ of lactate_peak and VeqCO₂ (i.e., a greater decrease in these parameters) compared with CG (lactate_peak: 64.05%, p < 0.001, g = 2.03; VeqCO₂: 15.92%, p = 0.02, g = 1.02) (Table 3). However, no significant differences were found between TG and CG for HR_peak, VeqO₂, VE_peak, ETO₂, ETCO₂, oxygen pulse, tidal volume and respiratory rate (Table 3).

Table 3.

Effects of training on fatigue perception as well as spiroergometric and spirometric parameters.

Parameters	TG (n = 11)	CG (n = 12)	p-value	t or U	g
Fatigue perception
Δ FSS scores (points)	-1.63 ± 1.24	0.44 ± 1.16	<0.001***	t = -4.12	1.72

Spirometric parameters
Δ FVCsitting position (l)	0.11 ± 0.29	-0.09 ± 0.16	0.036*	t = 2.23	0.93
Δ FET (s)	1.15 ± 2.09	-0.57 ± 1.77	0.045*	t = 2.13	0.88
Δ PEF (l/s)	1.18 ± 1.47	-0.24 ± 1.69	0.043*	t = 2.14	0.89
Δ FVClying position (l)	0.20 (0.34 [-0.03; 0.31])	0.04 (0.25 [-0.15; 0.10])	0.18	U = 44.00	-
Δ FEV1 (l)	-0.01 (0.42 [-0.3; 0.12])	-0.03 (0. 92 [-0.21; 0.08])	0.90	U = 63.50	-
Δ FEV1/FVC (%)	-2.90 (3.00 [-3.30; -0.30])	1.25 (5.15 [-1.60; 3.55])	0.17	U = 43.00	-
Δ FEV1/PEF (ml/l/min)	-1.39 ± 1.24	-0.95 ± 2.61	0.62	t = -0.51	-
Δ FIF50% (l)	-0.16 (0.93 [-0.85; 0.08])	-0.55 (0.90 [-0.75; 0.15])	0.97	U = 65.00	-
Δ FEF50% (l)	0.08 (1.11 [-0.54; 0.57])	-0.08 (1.07 [-0.61; 0.46])	0.98	U = 65.00	-
Δ FIF50%/FEF50%	-0.13 ± 0.27	-0.12 ± 0.24	0.93	t = -0.086	-
Δ FEF25–75% (l/s)	-0.17 (0.78 [-0.46; 0.32])	0.06 (0.68 [-0.36; 0.32])	0.56	U = 56.00	-

Spiroergometric parameters and metabolic responses at maximal CPET
Δ Lactatepeak (mmol/l)	-4.06 ± 2.84	1.05 ± 2.19	<0.001***	t = - 4.85	2.03
Δ VO2peak (ml/min/kg)	8.29 ± 7.52	-10.17 ± 6.79	<0.001***	t = 6.18	2.58
Δ PAP (watts)	12.73 ± 32.35	-25.83 ± 25.12	0.004**	t = 3.21	1.34
Δ HRpeak (bpm)	10.09 ± 22.60	5.33 ± 16.77	0.57	t = 0.58	-
Δ Tidal volume (l)	-0.12 ± 0.51	-0.07 ± 0.33	0.66	t = -0.45	-
Δ VeqCO2	-5.01 ± 8.64	3.07 ± 7.18	0.02*	t = -2.45	1.02
Δ VeqO2	1.22 ± 13.19	2.52 ± 10.10	0.79	t = -0.27	-
Δ ETO2 (mmHg)	-0.91 ± 2.47	-0.83 ± 2.72	0.95	t = -0.07	-
Δ VEpeak (l/min)	-4.19 ± 12.48	-2.96 ± 10.26	0.80	t = -0.26	-
Δ ETCO2 (mmHg)	-6.00 (3.00 [-7.00; -4.00])	-4.50 (13.50 [-10.50; 3.00])	0.93	U = 64.00	-
Δ Oxygen pulse (ml/bpm)	-5.98 (26.48 [-16.88; 9.60])	-4.44 (11.60 [-7.40; 4.20])	0.88	U = 63.00	-
Δ Respiratory rate (cycle/min)	-1.00 (11.00 [-7.00; 4.00])	-3.00 (13.00 [-7.00; 6.50])	0.95	U = 64.50	-

Note: Data are presented as mean ± standard deviation or median (interquartile range [lower quartile; upper quartile]).

Δ: Pre-post intervention change calculated as follows: Δ = values _after - values _before training or sedentary lifestyle; CG: Control group; CPET: Cardiopulmonary exercise testing; ETO₂: End-tidal oxygen pressure; ETCO₂: End-tidal carbon dioxide pressure; FET: Forced expiratory time; FSS: Fatigue severity scale; FVC: Forced vital capacity; FEV1: Forced expiratory volume in the first second; FIF_50%: Forced inspiratory flow at 50% FVC; FEF_50%: Forced expiratory flow at 50% FVC; FEF_25–75%: Forced expiratory flow at 25–75% of FVC; g: Hedges'g (effect size); HR_peak: Peak heart rate; Lactate_peak: Peak lactate level (at the last stage of CPET); PAP: Peak aerobic power, PEF: Peak expiratory flow; TG: Training group; t: t-test; U: Mann-Whitney U Test; VO_2peak: Peak oxygen uptake; VeqO₂: Ventilatory equivalent for oxygen; VeqCO₂: Ventilatory equivalent for carbon dioxide; VE_peak: Peak pulmonary ventilation. *, **, ***: Significant group effect at p < 0.05; p < 0.01, and p < 0.001, respectively.

3.4. Lung function: spirometric parameters

T-tests reported that the Δ of FVC_{sitting position}, FET and PEF were greater (i.e., a higher increase in these parameters) in TG than in CG (FVC_{sitting position}: 5.67%, p = 0.036, g = 0.93; FET: 37.88%, p = 0.045, g = 0.88; and PEF: 23.25%, p = 0.043, g = 0.89) (Table 3). However, there were no significant differences between TG and CG for FEV1, FEV1/FVC, FEV1/PEF, FIF_50%, FEF_50%, FIM_50%/FEM_50% and FEF_25–75% (Table 3).

For Δ FVC_position, ANOVA showed only a significant time effect (F_{(1, 21)} = 6.36, p = 0.019, ηp² = 0.23). No significant group effect (F_{(1, 21)} = 1.26, p = 0.27) or interaction (group × time) (F _(1,21) = 0.87, p = 0.36) was found for this parameter (Table 3).

3.5. Fatigue perception

T-test revealed that Δ of FSS scores were lower (i.e., a greater decrease in FSS scores) in TG than in CG (48.09%, p < 0.001, g = 1.72) (Table 3).

3.6. Muscle strength

The 1-RM measures (in TG) were higher at T1 than at T0 for calf raise (55.60%, p < 0.001, d = 2.70), squat (48.20%, p = 0.0013, d = 1.34), push down (23.14%, p = 0.0002, d = 1.72), arm curl (34.56%, p < 0.001, d = 2.02) and seated row (41.64%, p = 0.00011, d = 1.85). However, no significant time effect was found for bench press (p = 0.84).

3.7. Safety outcomes

No adverse effects related to the intervention were observed during this research.

4. Discussion

The current study aimed to evaluate the effects of 12-week self-paced HIIT-RT on pulmonary function, aerobic fitness, perceived fatigue and metabolic responses in PwMS. The main results revealed the effectiveness of this concurrent training in improving lung function (i.e., spirometric parameters), aerobic capacity and metabolic responses (i.e., spiroergometric parameters) as well as in alleviating fatigue perception.

For aerobic fitness, PAP and VO_2peak increased by 30.65% and 42.08%, respectively, in TG compared with CG. The TG showed a decrease in VeqCO₂ compared with CG (15.92%). Accordingly, previous studies revealed that 24-week aerobic training (cycling at 50–80% W_max) combined with whole body strengthening (1–3 sets of 8–12 reps at 15–30% of maximum load) improved VO_2peak in PwMS [60]. Similarly, Zaenker et al. [20] observed an increase in maximum tolerated power (MTP) following HIIT (5 × 1 min at 90–110% MTP interspersed with 3-min active recuperation) associated with RT (4–5 sets of 10–15 reps) for 3 months in PwMS. The increased PAP, found in our study, could be explained by the improvement of VO_2peak since it could lead to better muscle aerobic capacity and function through enhancing muscle blood O₂ uptake and muscle mitochondrial oxidative capacity including an increased oxidative enzyme activity [61]. This improvement could direct muscle energy metabolism toward lipid oxidation much more than carbohydrate oxidation, which could increase glycogen storage and delay acidosis related to lactic acid production, thus enabling participants to extend the exercise duration during CPET, leading to a better PAP [62,63]. The increased lower-extremity muscle strength found in TG (i.e., increased 1-RM: calf raise: 55.60%, squat: 48.20%) could also explain the PAP improvement given the involvement of muscle weakness in limiting exercise capacity through reducing the ability of muscles to withstand the resistance imposed on them during pedaling [64,65]. The increased aerobic capacity (PAP and VO_2peak) could be explained by the role of HIIT in increasing lipid utilization [66], which might be remarkable via the decrease of lactate_peak level in TG compared with CG (64.05%). Indeed, HIIT has been shown to increase the PGC-1α which might promote mitochondrial biogenesis [66,67]. This could lead to increased fatty acids oxidation which reduced the use of anaerobic lactic pathway, thus explaining the decreased lactate_peak level [66]. In contrast to our finding for lactate_peak, a previous MS study revealed an increase in this parameter following a 12-week HIIT (5 × 1 min at 90–110% MTP interspersed with active recuperation) associated with RT [20]. This disparity between results could be attributable to the differences in baseline level of lactate_peak (8.03 vs. 6.90 mmol/l), sociodemographic data (age: young adults: 35.17 vs. middle-aged adults: 44.60 years old) and clinical data [(MS form: RR (100%) vs. RR (84.61%) + secondary progressive MS (11.54%) + primary progressive MS (3.85%)] [20]. These authors attributed the increased level of lactate_peak found in their study to the augmentation of VO_2peak, MTP and HR_peak indicating an increased muscle activity that could lead to delayed exhaustion [20]. For the other spiroergometric parameters (HR_peak, VE_peak, VeqO₂, ETO₂ and ETCO₂, oxygen pulse, tidal volume and respiratory rate), no significant effects of HIIT-RT were found in the present study. Similarly, Hansen et al. [68] observed no effect of 6-month aerobic moderate training (1–3 × 6–10 min at RPE = 12–14) combined with RT (1–4 × 10–15 reps) on VeqO₂, ETO₂, ETCO₂ and oxygen pulse in PwMS who suffered from respiratory dysfunction. Our findings could be explained by the necessity of adding respiratory muscle training that aimed to improve respiratory function in particular, such as inspiratory exercises or expiratory training associated with upper-extremity exercises [68].

Regarding spirometric parameters, compared with CG, the TG showed an increase in FVC (5.67%), FET (37.88%) and PEF (23.25%). The lack of MS studies regarding the concurrent training impact on spirometric data led us to compare our findings with those of studies using aerobic training or RT, which revealed contradictory results. Indeed, our findings are similar to those of Rasova et al. [23] who observed an improvement in FVC following a 2-month aerobic training (60% VO_2peak) in PwMS. In contrast, Balavi and Ghanbarzade [24] reported a decrease in pulmonary function following a 12-week RT in PwMS. However, Rampello et al. [25] found no effect of 8-week aerobic training on lung function in PwMS. This discrepancy in results could be due to the difference in the training protocol (concurrent (HIIT-RT) vs. RT) [24] or duration (12 vs. 8 weeks) [25]. Our findings could be explained by the effectiveness of intense aerobic training in improving O₂ uptake, activating sleeping alveoli and raising respiratory rate, which could increase the alveolar air volume and alveoli elasticity leading to enhanced FVC and respiratory function [69]. Besides, the increased aerobic capacity could explain the enhanced spirometric parameters, as the training-induced improved aerobic fitness could participate in increasing lung volumes through enhancing pulmonary functioning [70]. For the other spirometric data, the HIIT-RT had no effects on FEF_50%, FIF_50%, FIF_50%/ FEF_50%, FEV1, FEV1/PEF, FEV1/FVC, FEF_25–75% and Δ FVC_position in the current study. Similar findings were found in previous studies which revealed no significant effects of 4-week HIIT (5 × 1 min at 90% VO_2peak interspersed with 3 min at 20W) on FEF_25–75%, FEV1 and FEV1/FVC in physically active healthy people [71]. The non-significant change in these expiratory flows and volumes following HIIT-RT could be explained by its insufficient intensity and duration to induce lung inflation that could allow a better airways stretching [71]. Indeed, this pulmonary stretching has been shown to be responsible for reducing airway resistance by decreasing smooth muscle tone leading to more dilated airways [71]. The lack of training effect on ΔFVC_position could be attributable to the absence of diaphragmatic weakness in our participants at baseline, which reduced the possibility of alleviating this disorder (ceiling effect) [43].

Regarding fatigue perception, the FSS scores decreased by 48.09% in TG compared with CG. This result is consistent with that of other researchers who found that PwMS perceived less fatigue following a 12-week concurrent training (cycling at 65–75% HR_peak + RT) [15,43]. This finding could be explained by the role of training in improving aerobic fitness since it has been shown that an increase in VO_2peak could predict the attenuation of MS-related perceived fatigue (R² = 13%) [4]. Furthermore, the anti-fatigue effect of training could be, at least in part, attributable to the fact that our participants were probably suffering from secondary fatigue induced by deconditioning, rather than from primary fatigue (caused by the disease per se) [3], since they presented a weak to moderate physical activity level at baseline.

This study presents three limitations. First, the findings of the current research cannot be generalized for patients with severe neurological incapacity (EDSS ≥6) or those living with progressive MS. Second, we cannot determine if the training benefits, observed in this study, depend on its duration because there were no intermediate data measurements (e.g., at the middle of the intervention). Third, this study is limited by the lack of assessment of other parameters that could explain the aerobic capacity improvement such as sleep quality [72].

5. Conclusion

Based on this study among patients with RRMS, the self-paced HIIT-RT may be an effective therapy for improving respiratory function and attenuating fatigue perception probably due to its benefits on cardiopulmonary fitness, metabolic responses and deconditioning. Indeed, the beneficial effects of HIIT-RT on aerobic capacity can be mediated by an increased mitochondrial biogenesis and a decreased anaerobic lactic pathway use. Therapists may consider this non-pharmacological intervention when treating MS-related pulmonary disorders and perceived fatigue. For future research, it would be relevant to investigate the effectiveness of self-paced HIIT-RT on respiratory function in PwMS according to their neurological disability, disease form, disease-modifying treatment, age and sex. Researchers are also encouraged to perform a deeper analysis of lung function based on investigating functional residual capacity and airway resistance using plethysmography in patients with advanced MS [73].

Author contributions

Sonda Jallouli, Omar Hammouda, Mariem Damak, Salma Sakka and Chokri Mhiri developed the study aim and design (conceptualization and methodology). Mariem Damak, Salma Sakka and Chokri Mhiri realized patients' recruitment and clinical data collection. Sonda Jallouli performed the training program; she is a physiotherapist and doctor in Biological Sciences Applied to Sports and Physical Activities (experimentation). Rami Maaloul, Rim Kammoun, Tarak Driss, Giovanni de Marco, Walid Feki, Sameh Ghroubi and Mohamed Habib Elleuch collected spirometric and spiroergometric data pre- and post-intervention (investigation and data acquisition) with providing the necessary equipment to these measurements (resources). Sonda Jallouli performed statistical analysis (formal analysis). Rim Kammoun helped Sonda Jallouli in interpreting results related to spirometric data (interpretation). Sonda Jallouli wrote the first manuscript draft (writing - original draft) and was the planner and the manager of this study (project administration). The supervisor of the current research was Omar Hammouda (supervision). The manuscript was critically reviewed by all authors who approved its versions submitted for revision, its final version and its version intended for publication (writing - review & editing). All listed authors 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 and to the journal to which the work was submitted and will be published.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, stock ownership or options and expert testimony.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The authors state that they have obtained appropriate institutional review board approval (by the south Institutional Human Research Ethics Committee (0184/2019)) and/or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations.

In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.

Clinical trial registration

This study was prospectively registered in accordance with WHO and ICMJE standards in the Pan African Clinical Trial Registry database (PACTR20200746530958).