Effects of High-Intensity Interval Training (HIIT) on Patients with Musculoskeletal Disorders: A Systematic Review and Meta-Analysis with a Meta-Regression and Mapping Report

The aim was to assess the impact of high-intensity interval training (HIIT) on patients with musculoskeletal disorders. We conducted a search of Medline, Embase, PEDro, and Google Scholar. We conducted a meta-analysis to determine the effectiveness of HIIT on pain intensity, maximal oxygen consumption (VO2 max), disability, and quality of life (QoL). We employed the GRADE and PEDro scales to rate the quality, certainty, and applicability of the evidence. Results showed significant differences in pain intensity, with a moderate clinical-effect (SMD = −0.73; 95% CI: −1.40–−0.06), and in VO2 max, with a moderate clinical-effect (SMD = 0.69; 95% CI: 0.42–0.97). However, the meta-analysis showed no statistically significant results for disability (SMD = −0.34; 95% CI: −0.92–0.24) and QoL (SMD = 0.40; 95% CI: −0.80–1.60). We compared HIIT against other exercise models for reducing pain intensity and increasing VO2 max. The meta-analysis showed no significant differences in favour of HIIT. Meta-regression analysis revealed that pain intensity scores were negatively associated with VO2 max (R2 = 82.99%, p = 0.003). There is low-moderate evidence that the HIIT intervention for patients with musculoskeletal disorders can reduce pain intensity and increase VO2 max but has no effect on disability and QoL. Results also showed that HIIT was not superior to other exercise models in reducing pain intensity and increasing VO2 max.


Introduction
Musculoskeletal pain is an important public health issue because of its impact on quality of life (QoL) and the disability it can represent [1]. More than 20% of the world's population is affected by painful conditions, contributing to the high consumption of healthcare resources [2]. Pain management can be approached from several perspectives, both pharmacological and non-pharmacological, the latter of which includes physical agents, manual therapy, psychosocial interventions, patient education, and exercise training [3,4].
Exercise therapy has been reported to be highly effective in managing patients with musculoskeletal pain [5] and has been shown to produce hypoalgesia by releasing betaendorphins or endocannabinoids [6][7][8]. Exercise therapy also interacts with the autonomic,

Selection Criteria and Data Extraction
First, two independent reviewers (F.C.M. and L.S.M.), who assessed the relevance of the RCTs regarding the study questions and aims, performed a data analysis, which was performed based on information from the title, abstract, and keywords of each study. If there was no consensus or the abstracts did not contain sufficient information, the full text was reviewed. In the second phase of the analysis, the full text was used to assess whether the studies met all the inclusion criteria. Differences between the two independent reviewers were resolved by a consensus process moderated by a third reviewer [29]. Data described in the results were extracted by means of a structured protocol that ensured that the most relevant information was obtained from each study [30].

Methodological Quality Assessment
We used the Cochrane Handbook for Systematic Reviews of Interventions version 5.1.0 to assess the risk of bias in the included studies [30]. The assessment tool covers a total of 7 domains: (1) random sequence generation (selection bias), (2) allocation concealment (selection bias), (3) blinding of participants and personnel (performance bias), (4) blinding of outcome assessments (detection bias), (5) incomplete outcome data (attrition bias), Diagnostics 2022, 12, 2532 4 of 31 (6) selective reporting (reporting bias), and (7) other biases. Bias was assessed as low risk, high risk, or unclear risk.
The studies' methodological quality was assessed using the PEDro scale [31], which assesses the internal and external validity of a study and consists of 11 criteria: (1) specified study eligibility criteria, (2) random allocation of patients, (3) concealed allocation, (4) measure of similarity between groups at baseline, (5) patient blinding, (6) therapist blinding, (7) assessor blinding, (8) fewer than 15% dropouts, (9) intention-to-treat analysis, 10) intergroup statistical comparisons, and 11) point measures and variability data. The methodological criteria were scored as follows: yes (1 point), no (0 points), or do not know (0 points). The PEDro score for each selected study provided an indicator of the methodological quality (9-10 = excellent; 6-8 = good; 4-5 = fair; 3-0 = poor) [32]. We used the data obtained from the PEDro scale to map the results of the quantitative analyses.
Two independent reviewers (F.C.-M. and L.S.-M.) examined the quality of all the selected studies using the same methodology. Disagreements between the reviewers were resolved by consensus with a third reviewer. The concordance between the results (interrater reliability) was measured using Cohen's kappa coefficient (κ) as follows: (1) κ > 0.7 indicated a high level of agreement between assessors; (2) κ = 0.5-0.7 indicated a moderate level of agreement; and (3) κ < 0.5 indicated a low level of agreement) [33].

Evidence Map
We created a visual map of the scientific evidence for each article to visually display the information as a bubble plot. The review information is based on 3 dimensions:
Effect (y-axis): Each of the reviews was classified according to its methodological quality using the PEDro scale. 4. Statistically significant differences: Articles with statistically significant differences were marked with white dots.

Certainty of Evidence
The certainty of evidence analysis was based on classifying the results into levels of evidence according to the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework, which is based on five domains: study design, imprecision, indirectness, inconsistency, and publication bias [34]. The assessment of the five domains was conducted according to GRADE criteria [35,36]. Evidence was categorised into the following four levels accordingly: (a) High quality. Further research is very unlikely to change our confidence in the effect estimate. All five domains are also met; (b) Moderate quality. Further research is likely to have an important impact on our confidence in the effect estimate and might change the effect estimate. One of the five domains is not met; (c) Low quality. Further research is very likely to have a significant impact on our confidence in the effect estimate and is likely to change the estimate. Two of the five domains are not met; and, finally, (d) Very low quality. Any effect estimates are highly uncertain. Three of the five domains are not met [35,36].
For the study design domain, the recommendations were downgraded one level in the event there was an uncertain or high risk of bias and serious limitations in the effect estimate (more than 25% of the participants were from studies with fair or poor methodological quality, as measured by the PEDro scale). In terms of inconsistency, the recommendations were downgraded one level when the point estimates varied widely among studies, the confidence intervals showed minimal overlap, or when the I 2 was substantial or large (greater than 50%). At indirectness domain recommendations were downgraded when severe differences in interventions, study populations or outcomes were found (the recommendations were downgraded in the absence of direct comparisons between the interventions of interest or when there are no key outcomes, and the recommendation is based only on intermediate outcomes or if more than 50% of the participants were outside the target group). For the imprecision domain, the recommendations were downgraded by one level if there were fewer than 300 participants for the continuous data [37].

Data Synthesis and Analysis
The statistical analysis was conducted using MetaXL software (version 5.3 (EpiGear International, Sunrise Beach, Queensland, Australia) [38]. To compare the outcomes reported by the studies, we calculated the standardised mean difference (SMD) over time and the corresponding 95% confidence interval (CI) for the continuous variables. The statistical significance of the pooled SMD was examined as Hedges' g to account for a possible overestimation of the true population effect size in the small studies [39].
We used the same inclusion criteria for the systematic review and the meta-analysis and included three additional criteria: (1) In the results, there was detailed information regarding the comparative statistical data of the exposure factors, therapeutic interventions, and treatment responses; (2) the intervention was compared with a similar control group; and (3) data on the analysed variables were represented in at least three studies.
The estimated SMDs were interpreted as described by Hopkins et al. [40], that is, we considered that an SMD of 4.0 represented an extremely large clinical effect, 2.0-4.0 represented a very large effect, 1.2-2.0 represented a large effect, 0.6-1.2 represented a moderate effect, 0.2-0.6 represented a small effect, and 0.0-0.2 represented a trivial effect. We estimated the degree of heterogeneity among the studies using Cochran's Q statistic test (a p-value < 0.05 was considered significant) and the inconsistency index (I 2 ) [40]. We considered that an I 2 > 25% represented small heterogeneity, I 2 > 50% represented medium heterogeneity, and I 2 > 75% represented large heterogeneity [41]. The I 2 index is a complement to the Q test, although it has the same problems of power with a small number of studies [41]. When the Q-test was significant (p < 0.1) and/or the result of I 2 was >75%, there was heterogeneity among the studies, and the random-effects model was conducted in the meta-analysis. To detect publication bias and to test the influence of each individual study, we performed a visual evaluation of the Doi plot [42], seeking asymmetry. We also performed a quantitative measure of the Luis Furuya-Kanamori (LFK) index, which has been shown to be more sensitive than the Egger test in detecting publication bias in a metaanalysis of a low number of studies [43]. An LFK index within ±1 represents no asymmetry, exceeding ±1 but within ±2 represents minor asymmetry, and exceeding ±2 involves major asymmetry. To test each study's influence, we visually examined the forest plot and performed an exclusion sensitivity analysis. Lastly, we applied a meta-regression analysis to analyse the relationship between pain intensity and VO 2 max variables using a random effects model employing the effect size statistic (Hedges' g) of the pain intensity scores to correlate with the VO 2 max scores [44].

Results
The study search strategy was presented in the form of a flow diagram (Figure 1).

Characteristics of the Included Studies
The patients were diagnosed with a persistent musculoskeletal pain condition [2 knee osteoarthritis studies [45,46], two axial spondylarthritis studies [16,47], three studies on chronic nonspecific low back pain [17,48,49], one study on episodic migraineurs [50], one study on fibromyalgia [15], one study on subacromial pain syndrome [51], one study on rheumatoid arthritis and adult-juvenile idiopathic arthritis [52], and one study on general persistent pain condition with previous trauma [53], and all of them evaluated pain intensity, VO 2 max, disability, and health-related QoL. Table 1 lists the descriptive characteristics of the included studies.

Interventions
In all groups, HIIT was compared to other types of training or interventions (including controls and no interventions), with the exception of Bressel et al. [45], which studied a single HIIT and balance training group, and Sveaas et al. (2014 [16,47], which included an HIIT and moderate-intensity continuous training (MICT) group and another no exercise group. Of the studies referred to above, three had two groups: one HIIT group and one MICT group [15,17,46]. Atan and Karavelioglu [15] included a third standard care group. Two other studies had only one HIIT and one standard care group [48,51]. Two studies had an HIIT group and another group that maintained the activities of daily living [52] and their usual physical activity [54]. Flehr et al. [53] had one HIIT group and one yoga group, while Verbrugghe et al. [49] studied four groups with different types of HIIT. The total duration of the intervention ranged from 6 to 12 weeks, with most studies having a frequency of two to three times per week, except for Keogh et al. [46] and Atan and Karavelioglu [15], which had frequencies of four and five times per week, respectively. Table 2 presents extensive details on the intervention characteristics of the included studies.

Characteristics of the Included Studies
The patients were diagnosed with a persistent musculoskeletal pain condition [2 knee osteoarthritis studies [45,46], two axial spondylarthritis studies [16,47], three studies

Control Group
No intervention (n = 12) VO 2 max (mL/kg/min) No group × time interaction between the three groups (p = 0.14). HIIT showed no clear effect on pain intensity at the end of the intervention and at 9 months of follow-up.

HIIT:
Interval: 110 rpm with a resistance similar or slightly higher than the rest. Intensity was defined as "an intensity at which you felt it was quite difficult to complete sentences during the exercise". Rest: ∼70 rpm To avoid pain, progressive increase in initial sessions. MICT: 60-80 rpm. Intensity was defined as "An intensity at which you are able to speak in complete sentences during the exercise". To avoid pain, progressive increase in initial sessions  HIIT (AerT) + High-intensity Global and Core StrT HIIT protocol: -Warmup: 5 min cycling -HIIT Training: 5 × 1 min high-intensity cycling interval alternating with a 1 min cycling recovery period. Weekly increase of interval duration of 10 s until week 6.
Work/rest ratio:
Work/rest ratio:

Methodological Quality Results
We evaluated the studies' quality with the Cochrane assessment tool. Most of the studies had a low risk of selective reporting bias. The domain with the highest percentage of studies with a high risk of bias was the blinding of participants and personnel (performance bias). Figure 2 shows the risk of bias summary and risk of bias graph. The inter-rater reliability of the methodological quality assessment was high (κ = 0.787). All of the studies had an excellent or good methodological quality, except the one by Bressel et al. [45] Due to the nature of the interventions, none of the studies performed blinding of the patients or evaluators. Table 3 lists the PEDro scores for each study. The inter-rater reliability of the methodological quality assessment between assessors was high (κ = 0.815).

Methodological Quality Results
We evaluated the studies' quality with the Cochrane assessment tool. Most of the studies had a low risk of selective reporting bias. The domain with the highest percentage of studies with a high risk of bias was the blinding of participants and personnel (performance bias). Figure 2 shows the risk of bias summary and risk of bias graph. The interrater reliability of the methodological quality assessment was high (κ = 0.787). All of the studies had an excellent or good methodological quality, except the one by Bressel et al. [45] Due to the nature of the interventions, none of the studies performed blinding of the patients or evaluators. Table 3 lists the PEDro scores for each study. The inter-rater reliability of the methodological quality assessment between assessors was high (κ = 0.815).    1, patient choice criteria are specified; 2, random assignment of patients to groups; 3, hidden assignment; 4, groups were similar at baseline; 5, all patients were blinded; 6, all therapists were blinded; 7, all evaluators were blinded; 8, measures of at least one of the key outcomes were obtained from more than 85% of baseline patients; 9, intentionto-treat analysis was performed; 10, results from statistical intergroup comparisons were reported for at least one key outcome; 11, the study provides point and variability measures for at least one key outcome. 1, patient choice criteria are specified; 2, random assignment of patients to groups; 3, hidden assignment; 4, groups were similar at baseline; 5, all patients were blinded; 6, all therapists were blinded; 7, all evaluators were blinded; 8, measures of at least one of the key outcomes were obtained from more than 85% of baseline patients; 9, intention-to-treat analysis was performed; 10, results from statistical intergroup comparisons were reported for at least one key outcome; 11, the study provides point and variability measures for at least one key outcome. Figure 3 presents the results of the evidence map for the included studies.

Pain Intensity
The meta-analysis showed statistically significant differences for the HIIT intervention, with a moderate clinical effect in seven studies (SMD: −0.73; 95% CI −1.40-−0.06; p < 0.05) but with evidence of significant heterogeneity (Q = 32.57, p < 0.001, I 2 = 82%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, −1.73) indicating a low risk of publication bias (Figures 4A and

Pain Intensity
The meta-analysis showed statistically significant differences for the HIIT intervention, with a moderate clinical effect in seven studies (SMD: −0.73; 95% CI −1.40-−0.06; p < 0.05) but with evidence of significant heterogeneity (Q = 32.57, p < 0.001, I 2 = 82%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, −1.73) indicating a low risk of publication bias (Figures 4A and A1). The certainty of the evidence was low, showing that HIIT likely decreases pain intensity, having been downgraded due to imprecision (sample size < 300) and inconsistency (I 2 = 82%) ( Table 4).
Regarding the sub-analysis comparing HIIT against other therapeutic exercise models, the meta-analysis showed no significant differences for the HIIT intervention in 3 studies (SMD: −0.35; 95% CI −0.76-0.06, p ≥ 0.05) with no evidence of significant heterogeneity (Q = 1.37, p = 0.5, I 2 = 0%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed no asymmetry (LFK, 0.67) indicating a very low risk of publication bias (Figures 4B and A2).

VO 2 max
The meta-analysis showed significant differences for the HIIT intervention, with a moderate clinical effect in six studies (SMD: 0.69; 95% CI 0.42-0.97, p < 0.05), with no evidence of significant heterogeneity (Q = 4.06, p = 0.54, I 2 = 0%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, 1.33) indicating a low risk of publication bias (Figures 5A and A2). The certainty of the evidence was moderate, showing that HIIT probably increases VO 2 max, having been downgraded due to imprecision (sample size < 300) ( Table 4). dence of significant heterogeneity (Q = 4.06, p = 0.54, I 2 = 0%). The shape of the funnel a DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LF 1.33) indicating a low risk of publication bias (Figures 5A and A2). The certainty of evidence was moderate, showing that HIIT probably increases VO2 max, having be downgraded due to imprecision (sample size < 300) ( Table 4). Regarding the subanalysis comparing HIIT against other therapeutic exercise mo els, the meta-analysis showed no statistically significant differences for the HIIT interv tion in three studies (SMD: 0.28; 95% CI −0.31-0.87, p ≥ 0.05), with no evidence of sign cant heterogeneity (Q = 4.16, p = 0.13, I 2 = 52%). The shape of the funnel and DOI plot d not present asymmetry, and the LFK index showed no asymmetry (LFK, −0.31) indicati a very low risk of publication bias (Figures 5B and A2). Regarding the subanalysis comparing HIIT against other therapeutic exercise models, the meta-analysis showed no statistically significant differences for the HIIT intervention in three studies (SMD: 0.28; 95% CI −0.31-0.87, p ≥ 0.05), with no evidence of significant heterogeneity (Q = 4.16, p = 0.13, I 2 = 52%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed no asymmetry (LFK, −0.31) indicating a very low risk of publication bias (Figures 5B and A2).

Disability
The meta-analysis showed no statistically significant differences for the HIIT intervention in three studies (SMD: −0.34; 95% CI −0.92-0.24, p ≥ 0.05), with no evidence of significant heterogeneity (Q = 4.55, p = 0.21, I 2 = 34%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, −1.68) indicating a low risk of publication bias (Figures 6A and A3). The certainty of the evidence was moderate, showing that HIIT probably does not decrease disability, being downgraded due to imprecision (sample size <300) ( Table 4). significant heterogeneity (Q = 4.55, p = 0.21, I 2 = 34%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, −1.68) indicating a low risk of publication bias (Figures 6A and A3). The certainty of the evidence was moderate, showing that HIIT probably does not decrease disability, being downgraded due to imprecision (sample size <300) ( Table 4).

Quality of Life
The meta-analysis showed no significant differences for the HIIT intervention in 4 studies (SMD: 0.40; 95% CI −0.80-1.60, p ≥ 0.05), with evidence of significant heterogeneity (Q = 24.01, p < 0.001, I 2 = 88%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, 1.43), indicating a low risk of publication bias (Figures 6B and A3). The certainty of the evidence was low, showing that HIIT likely does not increase QoL, being downgraded due to imprecision (sample size < 300) and inconsistency (I 2 = 88%) ( Table 4).

Meta-Regression Analysis
In the meta-regression analysis, we explored the role of pain intensity scores in improving VO2 max function. The results showed that pain intensity was significantly and negatively correlated with VO2 max (β = −0.91; Z = −3.02; p = 0.003 and R 2 = 82.99%) ( Figure  7).

Quality of Life
The meta-analysis showed no significant differences for the HIIT intervention in 4 studies (SMD: 0.40; 95% CI −0.80-1.60, p ≥ 0.05), with evidence of significant heterogeneity (Q = 24.01, p < 0.001, I 2 = 88%). The shape of the funnel and DOI plot did not present asymmetry, and the LFK index showed minor asymmetry (LFK, 1.43), indicating a low risk of publication bias (Figures 6B and A3). The certainty of the evidence was low, showing that HIIT likely does not increase QoL, being downgraded due to imprecision (sample size < 300) and inconsistency (I 2 = 88%) ( Table 4).

Meta-Regression Analysis
In the meta-regression analysis, we explored the role of pain intensity scores in improving VO 2 max function. The results showed that pain intensity was significantly and negatively correlated with VO 2 max (β = −0.91; Z = −3.02; p = 0.003 and R 2 = 82.99%) (Figure 7).

Discussion
Our main goal was to analyse the effect of HIIT on the VO2 max, pain intensity, disability, and QoL of patients with musculoskeletal disorders. Our results suggest that HIIT has a significant moderate effect size on VO2 max and pain intensity but does not seem to improve the disability and QoL of patients with musculoskeletal disorders. We also found that pain intensity was negatively associated with VO2 max.
We found a moderate certainty of evidence of a moderate effect size of HIIT on VO2 max when compared with no intervention. Several authors also found that HIIT was superior to usual care or no intervention in improving VO2 max among patients with cardiovascular disorders or cancer [18,19,55]. We did not find that HIIT was superior to another exercise intervention on VO2 max; however, the results across systematic reviews differ [19,56,57]. It has been previously reported that HIIT induces muscular adaptations, such as mitochondrial biogenesis and increased intramuscular capillarisation [58,59] vascular adaptations, such as increased blood cell volume [60], and cardiac adaptations, such as increased cardiac output and contractility [59,61]. All of these mechanisms have been shown to play a role in VO2 max [62]. We found that the patients' pain intensity scores were negatively associated with VO2 max, which is an important predictor of all-cause mortality and cardiovascular disease [63,64]. It should be noted that patients with chronic pain and musculoskeletal disorders have shown an increased risk of cardiovascular and chronic disease and an increased risk of mortality due to cardiac disease [65,66]. An improvement in cardiorespiratory capacity has been shown to decrease the mortality risk by up to 16% [67,68]. HIIT appears to be an effective solution for improving patients' cardiorespiratory capacity. We found a low certainty of evidence of a moderate effect size of HIIT on pain intensity compared with no intervention. Geneen et al. found that physical activity appears to

Discussion
Our main goal was to analyse the effect of HIIT on the VO 2 max , pain intensity, disability, and QoL of patients with musculoskeletal disorders. Our results suggest that HIIT has a significant moderate effect size on VO 2 max and pain intensity but does not seem to improve the disability and QoL of patients with musculoskeletal disorders. We also found that pain intensity was negatively associated with VO 2 max.
We found a moderate certainty of evidence of a moderate effect size of HIIT on VO 2 max when compared with no intervention. Several authors also found that HIIT was superior to usual care or no intervention in improving VO 2 max among patients with cardiovascular disorders or cancer [18,19,55]. We did not find that HIIT was superior to another exercise intervention on VO 2 max; however, the results across systematic reviews differ [19,56,57]. It has been previously reported that HIIT induces muscular adaptations, such as mitochondrial biogenesis and increased intramuscular capillarisation [58,59] vascular adaptations, such as increased blood cell volume [60], and cardiac adaptations, such as increased cardiac output and contractility [59,61]. All of these mechanisms have been shown to play a role in VO 2 max [62].
We found that the patients' pain intensity scores were negatively associated with VO 2 max , which is an important predictor of all-cause mortality and cardiovascular disease [63,64]. It should be noted that patients with chronic pain and musculoskeletal disorders have shown an increased risk of cardiovascular and chronic disease and an increased risk of mortality due to cardiac disease [65,66]. An improvement in cardiorespiratory capacity has been shown to decrease the mortality risk by up to 16% [67,68]. HIIT appears to be an effective solution for improving patients' cardiorespiratory capacity.
We found a low certainty of evidence of a moderate effect size of HIIT on pain intensity compared with no intervention. Geneen et al. found that physical activity appears to induce exercise-induced hypoalgesia in patients with chronic pain; however, the results were inconsistent across the various exercise modalities [69]. When compared with another exercise intervention, HIIT did not show a greater effect. It has been shown that exerciseinduced hypoalgesia acts through the activation of nociceptive inhibitory pathways that release endogenous opioids and endocannabinoids [70]; however, populations with chronic pain often have exercise-induced hypoalgesia dysfunction [70,71]. Nonetheless, we found that HIIT appeared to be an effective modality for decreasing pain intensity. Patients with musculoskeletal disorders often present central sensitisation, a facilitation of the nociceptive signal in the central nervous system [72]. Quantitative sensory testing is employed to evaluate central nervous system nociceptive modulation [72]. HIIT has shown an intensitydependent [12,13] positive effect on pain tolerance [13] and pain thresholds [12,73]. In certain conditions, the presence of an inflammatory state can increase nociceptor activity and has been associated with pain intensity [71,[74][75][76]. After performing HIIT, a number of authors have found a decrease in inflammatory markers [77][78][79], such as C-reactive protein, tumour necrosis factor-alpha and interleukin-6 (IL-6), and a release of anti-inflammatory cytokines, such as IL-10 [79]. In contrast, other authors have found that HIIT induced an acute increase in IL-6 levels [80,81]; however, Pedersen proposed that this acute liberation will then induce an anti-inflammatory response [82]. Shanaki et al. observed a decrease in pro-inflammatory M1-macrophage markers and an increase in anti-inflammatory M2macrophage markers in mice after HIIT [83]. However, not all musculoskeletal conditions show reduced pain intensity in parallel with a decrease in pro-nociceptive or inflammatory serum markers [76,84], and not all musculoskeletal conditions progress with an increased inflammatory state [76].
We found a low level of evidence of no significant effect of HIIT on QoL compared with no intervention or usual care. Mugele et al. systematically reviewed the effect of HIIT on QoL, compared with usual care, and found unclear results [19]. QoL appears to be more closely related to interpretation and catastrophising than pain intensity [85], which might explain why we observed a decrease in pain intensity with no improvement in QoL. Monticone et al. found that a multidisciplinary treatment involving cognitive-behavioural therapy and exercise results in a significant improvement in QoL, while exercise alone resulted in little change [86]. We also found moderate certainty evidence of no significant effect of HIIT on disability compared with no intervention or usual care. Kamper et al. found that a treatment involving a physical and a psychological or social component had a greater effect on disability than physical therapy alone for patients with chronic low back pain. HIIT alone might be insufficient for improving disability or QoL in musculoskeletal disorders [87].
Time constraints and pain are two of the main barriers to physical activity for patients with musculoskeletal disorders [88][89][90]. Despite similar effects on VO 2 max and pain intensity with other exercise types, HIIT requires less training volume to achieve similar effects in the included studies that provide the control group's training duration [15,50]. Wewege et al. found that the most common adverse effects in patients with cardiovascular disease were musculoskeletal complaints; however, we observed that HIIT presented similar or almost no additional major or minor adverse events or pain flare-ups than no intervention or other exercise modalities [91]. Major cardiac adverse events during HIIT appear at a rate of 1 per 11,333 HIIT h in patients with cardiovascular disease [91] but with no significant difference in the overall adverse events rate between HIIT and MICT [91]. As recommended by Weston et al. if health professionals want to implement HIIT, they should evaluate patients on a case-by-case basis depending on their cardiac history [20]. Heisz et al. found that participants rated HIIT more enjoyable than MICT and that enjoyment increased with repeated HIIT when it remained constant with repeated MICT [92]. Health professionals should include HIIT in the management of musculoskeletal disorders, given that HIIT is a time-efficient, enjoyable, effective, and safe form of exercise. Finally, it is relevant to stress that it is important to prescribe exercise specifically for each patient and for each clinical condition, although in this work it has been grouped by variables, rather than by populations.

Limitations
We found low-to-moderate quality evidence for our results. Further studies are needed on the effects of HIIT on musculoskeletal disorders to confirm our results. The sample sizes of the included studies were often very small. Future studies should include larger sample sizes to improve the quality of the evidence. Due to the lack of sufficient data and the heterogeneity among the interventions (e.g., frequency, intervention duration), we could not establish the specific effect on each musculoskeletal disorder and the optimal HIIT parameters. Due to the small number of trials, we pooled the aerobic and anaerobic HIIT training studies; future systematic reviews should evaluate them separately. Only a few studies compared the effect of HIIT against high-intensity continuous training or other types of exercise; future studies should include this type of high-intensity training.
As recommended by the American Thoracic Society/American College of Chest Physicians Statement on Cardiopulmonary Exercise Testing, we included VO 2 peak and VO 2 max and used them interchangeably [93]. Quantitative sensory testing (e.g., pain pressure or thermal threshold, conditioned pain modulation, and temporal summation) is essential in pain research; future studies evaluating the effects of HIIT on musculoskeletal disorders should include these variables. In addition, no further meta-regression analysis could be performed due to the small number of articles sharing the outcomes of interest. Lastly, it is important to stress that there were 3 studies where HIIT was embedded in other exercise interventions such as balance exercise and continuous exercise. This is a clear limitation that should be considered when extrapolating the results [16,45,47].

Conclusions
There is low to moderate quality evidence that the HIIT intervention for patients with musculoskeletal disorders can improve pain intensity and VO 2 max but not disability and QoL. The results of the subanalyses showed that HIIT was not superior to other exercise models in improving pain intensity and VO 2 max . Clinically, this tells us that we can implement high-intensity interval exercise models if our goal is to improve pain intensity or increase cardiorespiratory fitness through maximal oxygen consumption. However, it is important to keep in mind two aspects: changes in pain intensity may not be accompanied by improvements in the subjective perception of quality of life or disability, at least, based on the data we currently have, and second, that this exercise model was not superior to other exercise models with respect to eliciting these clinical changes. This should be considered clinically. Low sample sizes and lack of prescription parameters emphasise the need for further research on HIIT in musculoskeletal disorders for its implementation in a clinical context.