Rethinking aerobic exercise intensity prescription in adults with spinal cord injury: time to end the use of "moderate to vigorous" intensity?

The paper addresses how to prescribe aerobic exercise intensity for adults with spinal cord injury (SCI), a critical element of exercise guidelines that currently recommend "moderate to vigorous" intensity without clear physiological thresholds. Given the physiological consequences of SCI on cardiovascular and respiratory responses, thresholds derived from non-disabled populations may not transfer. The study examines whether fixed relative intensities (e.g., %VO2peak, %HRpeak) align across individuals with SCI to the same exercise intensity domains (moderate, heavy, severe) and evaluates alternative domain-based prescriptions grounded in lactate thresholds, aiming to improve clarity, accuracy, and applicability of SCI exercise guidance.

- Existing SCI exercise guidelines (2018) recommend aerobic exercise at "moderate to vigorous" intensity but lack physiological definitions for SCI. - Non-disabled guidelines define intensity zones via %VO2max/%HRmax, reserves, and RPE, but these may not apply to SCI due to altered cardiovascular and respiratory control. - Domain-based prescription (moderate, heavy, severe) is supported by consistent VO2 and lactate kinetics within domains. - In non-disabled adults, no fixed %VO2max or %HRmax ensures uniform domain placement; %VO2R and %HRR relationships diverge from assumed linearity, questioning their use for individualized prescription. - There is limited SCI-specific evidence on domain transitions (e.g., CP/CS) and VO2 kinetics across domains, underscoring the need for SCI-tailored methods.

Design: Retrospective analysis of athlete data collected in a university lab; ethical approval obtained; informed consent provided. Participants: 134 competitive athletes (98 male, 36 female) categorized into paraplegia (PARA; n=47), tetraplegia (TETRA; n=20), and alternate health conditions (NON-SCI; n=67). PARA injuries T4–L2; TETRA C3–C7. Sports included handcycling, para-alpine ski, paratriathlon, wheelchair basketball, wheelchair rugby, wheelchair tennis. Exercise testing: Submaximal step test (3-min stages; target 6–8 stages) followed by graded exercise test (GXT) to exhaustion. Modes matched to sport: handcycle on Cyclus 2, arm crank ergometry (Lode Angio), or wheelchair propulsion on a motorized treadmill using athletes’ wheelchairs. - Submax protocols: HC/ACE started 15–60 W with 10–20 W increments every 3 min; WCP started 0.7–2.8 m·s−1, increased 0.2–0.4 m·s−1 every 3 min. WCP was discontinuous (45–60 s between stages) for blood sampling. Measurements: Breath-by-breath VO2 (Metalyzer 3B), HR (Polar RS400), RPE (Borg 6–20; final minute of each stage), ear-lobe capillary [BLa] at end of each stage. Calibration performed per manufacturer guidelines. Termination criteria (submax): [BLa] > 4 mmol·L−1 or RPE ≥ 17 (RPE criterion emphasized in TETRA due to potential blunted lactate responses). GXT: 15 min recovery then start at workload where [BLa] rose 0.5 mmol·L−1 above rest; increments of 10–20 W·min−1 (HC/ACE) or 0.1 m·s−1·min−1 (WCP) to volitional exhaustion; VO2, HR monitored continuously; RPE and [BLa] at test end. Data processing: 30-s rolling averages used; peak VO2 and HR defined as highest 30-s values during GXT. Submax stage-end VO2 and HR expressed as %VO2peak and %HRpeak. Lactate thresholds: LT1 as intersection of horizontal and ascending segments in log-[BLa] vs log-VO2; LT2 defined as [BLa] equal to LT1 + 1.5 mmol·L−1. VO2 and corresponding HR at LT1 and LT2 determined (HR via interpolated VO2:HR linear relationship). RPE at LT1 and LT2 derived from participant-specific quadratic RPE–[BLa] models. Domains defined: moderate (<LT1), heavy (LT1–LT2), severe (>LT2). Statistical analyses: Shapiro–Wilk for normality; dynamic multilevel models with lagged independent variables for RPE vs %VO2peak and RPE vs %HRpeak, accounting for repeated measures (stage nested within participant), initial condition (i=1), and lagged effects (i−1). Fixed/random effects considered for sex, group (PARA/TETRA/NON-SCI), and exercise mode (ACE/HC/WCP). One-way ANOVA with Bonferroni post hoc tested group differences at LT1 and LT2 for VO2 (absolute, relative, %VO2peak), HR (beats·min−1, %HRpeak), and RPE; effect sizes categorized per standard thresholds. Domain distributions computed at 5% increments from 35–95% for %VO2peak and %HRpeak.

- RPE modeling: - %VO2peak model: RPE significantly influenced by initial %VO2peak (i=1), current (i>1), and lagged (i−1) %VO2peak (all P<0.01). Group effect present for i>1 (P=0.01), with TETRA showing greater within-individual variability; PARA and NON-SCI grouped together. - %HRpeak model: RPE significantly influenced by initial, current, and lagged %HRpeak (all P<0.01); fixed group effect showed PARA differed from TETRA/NON-SCI. - Derived correspondences (Table 2): For a given RPE, TETRA required higher %VO2peak and similar-to-higher %HRpeak than PARA. Examples: RPE 12 ≈ 52% VO2peak and 68% HRpeak (PARA) vs 58% VO2peak and 71% HRpeak (TETRA); RPE 15 ≈ 70% VO2peak, 82% HRpeak (PARA) vs 76% VO2peak, 85% HRpeak (TETRA). - LT1 responses: - Significant group effects for absolute VO2, relative VO2, %VO2peak, HR, and %HRpeak (all P<0.01). - TETRA had lower absolute and relative VO2 at LT1 than PARA (ES 0.86, 1.00) and NON-SCI (ES 1.00, 1.62), but higher %VO2peak at LT1 than PARA (ES 0.96) and NON-SCI (ES 0.94). - HR at LT1 was lower in TETRA than PARA (ES 2.33) and NON-SCI (ES 2.74), yet %HRpeak at LT1 was higher in TETRA than PARA (ES 1.00) and NON-SCI (ES 0.69). - RPE at LT1 did not differ between groups (P=0.62). - LT2 responses: - Significant group effects for absolute VO2, relative VO2, %VO2peak, HR, and %HRpeak (all P<0.01). - TETRA had lower absolute and relative VO2 at LT2 than PARA (ES 0.83, 0.97) and NON-SCI (ES ~1.03, 1.72), but higher %VO2peak at LT2 than PARA (ES 1.26) and NON-SCI (ES 1.06). - HR at LT2 was lower in TETRA than PARA (ES 3.00) and NON-SCI (ES 3.55), while %HRpeak at LT2 was higher in TETRA than PARA (ES 0.93) and NON-SCI (ES 0.62). - RPE at LT2 did not differ between groups (P=0.19). - Domain distribution at fixed % thresholds: - No fixed %VO2peak or %HRpeak commonly used for exercise prescription placed all participants within the same domain. Many fixed percentages showed participants spread across moderate, heavy, and severe domains. - Intensity classification translation: - Thresholds mapping non-disabled intensity bands to SCI differed for PARA and TETRA (Table 3), indicating that adopting non-disabled cut-points misclassifies intensity in SCI. Combining these with domain distributions further shows heterogeneity at "moderate" and "vigorous" intensities. - Overall: Fixed %VO2peak and %HRpeak are unsuitable for achieving uniform domain-specific intensity in adults with SCI; RPE at LT1 (~11) and LT2 (~15) appeared similar across groups, suggesting potential utility but with inter-individual variability (SD ~1–2).

The study demonstrates that in adults with SCI, fixed relative intensities (%VO2peak, %HRpeak) do not ensure homogenous placement within exercise intensity domains, undermining the scientific validity of prescribing "moderate to vigorous" intensity using such metrics. Tetraplegia alters the relationship between physiological responses and perceived exertion, yielding higher %VO2peak and %HRpeak at lactate thresholds despite lower absolute workloads, reflecting autonomic and cardiorespiratory constraints. Consequently, translating non-disabled thresholds to SCI leads to inconsistent and potentially inappropriate training stimuli across individuals. Domain-based prescription anchored to individual thresholds (e.g., LT1; ideally CP/CS for heavy–severe transition) is more physiologically defensible. Although RPE values at domain transitions were similar between groups (LT1≈11; LT2≈15), inter-individual variability suggests RPE should be individualized and, where feasible, calibrated against physiological thresholds. These insights inform both research design—where controlling exercise dose across participants is crucial—and clinical/community practice, emphasizing the need for individualized, domain-based intensity prescription rather than fixed percentages.

Prescribing aerobic exercise as "moderate to vigorous" or using fixed %VO2peak/%HRpeak does not produce uniform intensity domain distribution in adults with SCI and should not be used. Instead, intensity should be individualized relative to physiologically determined domain transitions (e.g., LT1 and, preferably, CP/CS) to standardize stimulus across individuals. However, implementing individualized testing poses challenges for population-level guidelines. Future work should validate reliable methods to identify domain transitions in SCI, characterize VO2 kinetics by domain and mode, and refine user-friendly strategies (potentially incorporating calibrated RPE) for translating individualized prescriptions into practice.

- Heavy–severe transition was inferred using LT2 (LT1 + 1.5 mmol·L−1) rather than directly measured critical power/speed (CP/CS); limited evidence and multiple LT2 determination methods constrain conclusions about the severe domain threshold. - Data derived from competitive athletes; findings may not generalize to sedentary or low-active SCI populations, where variability may be greater. - Retrospective analysis; protocol heterogeneity across modes and discontinuous WCP submax testing (to allow blood sampling) could introduce variability. - Lack of direct assessment of VO2 kinetics across domains and absence of CP/CS testing in SCI participants restricts domain validation. - Inter-individual variability in RPE suggests caution if using fixed RPE values without individual calibration.