Adaptation to reduced lung function in children and young people with spinal deformity

Abstract

Background

Severe scoliosis can affect respiratory function in growing patients and produce cardiopulmonary complications, leading to significant morbidity. The development of spinal deformity may impact on young patients’ level of function and reported quality of life (QOL). The aim of this study was to investigate the relationship between lung function, exercise capacity and quality of life in young patients with spinal deformity.

Methods

This is a retrospective analysis of 104 patients (31% male, 69% female with mean age 14.9yrs). 77% of patients had an adolescent idiopathic scoliosis, with the remainder having other scoliosis diagnoses or Scheuermann’s kyphosis. Principal outcomes included Spirometry [FEV₁, FVC], Whole Body Plethysmography, Cardiopulmonary Exercise Testing [CPET] and patient outcome questionnaires (with SRS-22). CPET measures included maximal exercise capacity [VO_2peak] as well as VO₂ at ventilatory threshold [VT] expressed as %predicted VO_2max-a measure of physical conditioning, and minute ventilation [V_E] from which breathing reserve [BR] could be calculated.

Results

Mean (±SD) main thoracic scoliosis was 59.9⁰ (±15.2⁰), and mean kyphosis in those with Scheuermann’s condition was 95.3⁰ (±11.5⁰). No correlation was elicited between FEV₁ or FVC (%predicted) and VO_2peak (%predicted) in this patient cohort. Greater thoracic curves were associated with lower FEV₁ (%predicted), r = −0.343, p = 0.001, FVC (%predicted), r = −0.307, p = 0.003 and BR (%) at the end of exercise (r = −0.-0.459, p < 0.001). The patient cohort had a mean (sd) VO₂peak of 98(17) %predicted, with greater VO₂peak levels recorded in female subjects, those of younger age and those with higher scoliosis angles. Those with better lung function [FEV₁ (%predicted)] had better BR (%) at the end of exercise (r = 0.483, p < 0.001). SRS-22 scores correlated significantly with VO_2peak (%predicted) (total SRS-22 versus VO_2peak (%predicted), r = 0.336, p = 0.002).

Conclusion

Larger thoracic scoliotic curves are associated with poorer lung function but better exercise capacity, likely related to higher levels of physical conditioning. Higher QOL scores were recorded in patients who had greater VO_2peak levels, suggesting that exercise capacity may be a protective factor for emotional well-being in patients with spinal deformity.

1. Introduction

Spinal deformity in childhood and adolescence is common and may present primarily as scoliosis of varying aetiology or increased thoracic kyphosis. Scoliosis, although characterised by a lateral curvature of the spine, produces a 3-dimensional deformity with axial rotation and sagittal imbalance. Any spinal deformity, particularly affecting the thoracic spine, may have an impact on pulmonary function. If left untreated, curves can deteriorate over time as skeletal growth accelerates. In some patients this can lead to restrictive or obstructive lung disease and rarely death as a consequence of cor pulmonale.¹

Weinstein et al.² identified a direct negative correlation between respiratory function and severity of the thoracic curve in adults with untreated scoliosis, whilst several studies report reduced pulmonary function in subjects with adolescent idiopathic scoliosis (AIS) when compared with healthy controls.^{3, 4, 5, 6, 7}

Pulmonary function tests offer an accurate measurement of lung function at rest. The integration of respiratory, cardiovascular and metabolic parameters during exercise can be measured by Cardiopulmonary Exercise Testing (CPET). Maximal exercise capacity is assessed and factors influencing exercise limitation may be suggested. In healthy individuals, ventilation and gas exchange are sufficient to provide oxygenation to the point at which exercise is limited by cardiovascular function. Muscle de-conditioning may also limit exercise capacity and can be implied by early onset of anaerobic metabolism during a maximal exercise test.⁸ An individual who is de-conditioned has fewer muscle capillaries (allowing less oxygen transfer), a lower mitochondrial density (resulting in reduced oxygen transfer) in muscle and decreased oxidative enzyme concentrations (less ability for energy transformation). Thus, de-conditioning results in less efficient oxidative energy generation and an earlier switch to anaerobic metabolism. Studies have suggested that exercise capacity is limited in individuals with AIS, with patients achieving a lower oxygen uptake and minute ventilation at maximal exercise.^{9, 10, 11} These patients often have a normal breathing reserve at the end of exercise, indicating that despite the known pulmonary function impairment, exercise is not limited by ventilatory function.^10,11

It is recognised that the development of spinal deformity can affect patients’ quality of life, in particular relating to pain, mental health, self-image and function. The disease-specific quality of life scoring system SRS-22, developed specifically to assess patient perception on treatment outcomes in AIS, provides a reliable and valid measure of these quality of life indicators.¹² This is in routine clinical use both pre- and post-operatively in our centre.

In our study, we aimed to determine the relationship between type of spinal deformity, lung function, exercise capacity and quality of life in a cohort of young patients with spinal deformity.

2. Materials and methods

We analysed retrospectively a single surgeon’s series of children and young people who underwent spinal deformity correction at the Scottish National Spine Deformity Centre between February 2015 and February 2019.

Patients with spinal deformity were referred for lung function and exercise capacity assessment as part of their routine pre-surgical work-up and patients with AIS (in whom the tool is validated) were asked to complete SRS-22 questionnaires pre-operatively.

2.1. Anthropometric data

Body mass and standing height were determined using an electronic scale (Weymed, 500 series, H Faraday, London, England, UK) and a stadiometer (Harpenden Stadiometer, Holtain Limited, Crymych, Wales, UK) respectively, in triplicate or until 2 identical values. Height from arm-span was also measured and greatest height (standing height or measurement from arm-span) was used for lung function calculations to correct for any inaccuracies due to the presence of the spinal deformity.

2.2. Lung function

Spirometry and whole body plethysmography (Jaeger Masterscreen PFTPro) were performed prior to CPET. Spirometry measures of forced expiratory volume in 1 s (FEV₁) and forced vital capacity (FVC) were expressed as percentage predicted values for age and height as defined by Global Lung Function Initiative 2012 reference equations.¹³

2.3. CPET

All our patients performed cardiopulmonary exercise testing on an electromagnetically braked cycle ergometer (Ergoline Viasprint 200, Ergoline, Blitz, Germany) using an incremental exercise protocol. Work rate was increased by 10w, 15w or 20w each minute, whilst maintaining a cadence at 65 rpm. They breathed into a full face mask connected to a flow sensor and sampling lines using a calibrated metabolic cart (CareFusion UK, Basingstoke, England). Masks of known dead space were used (Carefusion 7400 and 8900), the choice dependent upon patient size and facial characteristics.

Breath-by-breath measurements of oxygen uptake (VO₂), carbon dioxide production (VCO₂), minute ventilation (V_E), and respiratory exchange ratio (RER) were made. Heart rate (HR) was also monitored continuously by a 12-lead electrocardiogram, blood pressure was measured at rest and every 3 min during exercise and transcutaneous oxygen saturation (SpO₂%) was measured by a pulse oximeter placed on the right ear or on the index or middle finger.

Peak aerobic capacity was calculated as the VO_2peak over the last 30sec of the test and was expressed corrected for body weight (mL.kg⁻¹.min⁻¹). Additional variables of interest from the CPET included measurement of VO₂ at ventilatory threshold, expressed as %predicted VO_2max. Criteria for accepting maximal test were either seeing true VO_2max - a plateau in VO₂ despite increasing workload, or were based on heart rate attainment, maximal exercise ventilation and RER data.

The ventilatory threshold (VT) is the point at which ventilation increases at a faster rate than oxygen uptake (VO₂) and reflects the point at which anaerobic metabolism begins to predominate with exponentially increasing carbon dioxide production and accumulation of fatigue-related metabolites including lactate. These effects on musculoskeletal and respiratory mechanisms serve to limit exercise capacity after VT.¹⁴

Breathing reserve was defined as BR=(Maximal voluntary ventilation [(MVV) - V_Epeak)/MVV] x 100, where MVV was estimated using equation 35 x FEV₁.¹⁵ Breathing reserve is the percentage of an individual’s expected maximal minute ventilation that remains at maximal exercise e.g. ‘How much gas is left in the tank.‘¹⁶

2.4. Data analysis

Data were collected in an anonymised Microsoft Access 2016 database and analyses were performed using SPSSv25 (SPSS Inc, Chicago, Illinois). Descriptive statistics were displayed as mean (±SD). Correlations were analysed using Pearson’s correlation co-efficient and comparisons between groups performed using Student’s t-test or analysis of variance (ANOVA). Statistical significance was assumed where p < 0.05.

2.5. Ethics

The study was a retrospective review of existing clinical testing and standardised practice. The study was discussed with the regional ethics committee and considered not to require formal ethics approval as was deemed to be service evaluation.

3. Results

3.1. Demographics

A total of 104 patients were included having undergone pre-operative spirometry and CPET. 72 patients (69%) were female and the mean age at CPET was 14.9 years (range: 10.6–18.6). The distribution of diagnoses within the group is depicted in Table 1.

Table 1.

Diagnoses among our study cohort (AIS: adolescent idiopathic scoliosis, JIS: juvenile idiopathic scoliosis, IIS: infantile idiopathic scoliosis).

Type of spinal deformity	Female n = 72		Male n = 32		Total n = 104
	n	%	n	%	n	%
AIS	57	79	23	72	80	77
JIS	5	7	0	0	5	5
IIS	1	1	0	0	1	1
Congenital	2	3	1	3	3	3
Syndromic	4	6	2	6	6	6
Scheuermann’s kyphosis	3	4	6	19	9	9

Patients with AIS mainly had either double thoracic and lumbar (35%) or single thoracic (28%) curves. 8 patients (10%) had only thoracolumbar/lumbar scoliosis. The mean main thoracic scoliosis angle was 59.9° (SD ± 15.2°) and the mean kyphosis angle in those patients with Scheuermann’s condition was 95.3° (SD ± 11.5°).

3.2. Respiratory statistics

Table 2 presents the baseline demographics of our study group.

Table 2.

Baseline respiratory data. Comparisons were made using the Student’s t-test.

	Female n = 72 Mean (SD)	Male n = 32 Mean (SD)	Overall group n = 104 Mean (SD)	p-value
Age	14.4 (1.9)	15.9 (1.6)	14.9 (1.9)	<0.001
FEV1 (%predicted)	77.8 (12.8)	76.3 (17.9)	77.4 (14.5)	0.636
FVC (%predicted)	79.7 (12.6)	79.0 (11.1)	79.4 (12.1)	0.781
VO2peak (%predicted)	102.2 (13.4)	88.7 (20.0)	98.1 (16.9)	<0.001
VO2 at VT (%predicted VO2max)	54.6 (8.2)	47.4 (9.7)	52.4 (9.3)	<0.001
Breathing Reserve (%)	28.7 (17.9)	23.4 (16.7)	27.1 (17.6)	0.161
Length of hospital stay (days)	6.5 (3.0)	5.9 (1.4)	6.3 (2.6)	0.317

The patient cohort had a mean (sd) FEV₁ of 77 (15) %predicted, and FVC of 79 (12) %predicted. The mean (sd) %predicted VO_2peak [98 (17)], %BR [27 (18)] and %predicted VO₂ at VT [52 (9)] were within normal range (>80%, >15% and >50% respectively). Female subjects had significantly better exercise capacity (p < 0.001) and later onset ventilatory threshold than males (p < 0.001).

3.3. Relationship between lung function and exercise capacity

No significant correlation between lung function (%predicted FEV₁ or FVC) and overall exercise capacity (%predicted VO_2peak) was found. However, lung function (%predicted FEV₁) was positively correlated with breathing reserve (%) (r = 0.483, p < 0.001). Similar correlations were seen when analysed with %predicted FVC.

3.4. Relationship between scoliosis size and lung function or exercise capacity

The relationships between the severity of main thoracic or most severe scoliosis and lung function or exercise variables were explored. More severe main thoracic curves were significantly associated with decreased %predicted FEV₁ (r = −0.343, p = 0.001) and %predicted FVC (r = −0.307, p = 0.003), as well as decreased %breathing reserve (r = −0.469, p < 0.001). No significant correlation was found with overall exercise capacity or VO₂ at VT (%predicted VO_2max). Correlations between the size of the most severe scoliosis and the lung function or exercise tolerance revealed similar results, however larger curves at any site were also positively associated with a later onset of ventilatory threshold (r = 0.249, p = 0.015) and with higher peak VO₂ (r = 0.203, p = 0.048).

3.5. Comparison between underlying diagnoses

We divided the patients into three subgroups (AIS, early onset scoliosis and Scheuermann’s kyphosis) in order to analyse the differences in demographics, spinal curvatures, lung function and exercise capacity depending on diagnosis (Table 3).

Table 3.

Comparison of demographics and curve data between those patients with AIS, early onset scoliosis and Scheuermann’s kyphosis. *ANOVA, **Student’s t-test. ˆn = 77, ˆˆn = 13. ^✝equal variances not assumed.

	AIS n = 80 Mean (SD)	Early Onset Scoliosis n = 15 Mean (SD)	Scheuermann’s kyphosis n = 9 Mean (SD)	p-value
Age (years)	15.0 (1.7)	13.9 (2.7)	15.8 (1.4)	0.051*
Main thoracic scoliosis (Cobb angle)	58.3 (14.4)ˆ	69.5 (16.5)ˆˆ	–	0.014**
Most severe scoliosis (Cobb angle)	61.2 (10.9)	72.0 (16.8)	–	0.029**✝

Patients with early onset scoliosis (juvenile idiopathic scoliosis, infantile idiopathic scoliosis, congenital or syndromic scoliosis) had significantly larger scoliotic curves compared to AIS (72° versus 61.2°, p = 0.029).

Comparison of lung function and exercise variables revealed significantly better lung function (%predicted FEV₁ and FVC) in patients with Scheuermann’s kyphosis compared with the 2 scoliosis groups (p < 0.001, ANOVA). The patients with Scheuermann’s kyphosis also had significantly better breathing reserve (%) compared with those with early onset scoliosis (p = 0.019, ANOVA). No other statistical difference was demonstrated (Fig. 1).

Fig. 1.

Comparison of lung function and exercise tolerance variables between diagnoses.

3.6. Relationship between self-rated QOL and scoliosis severity, lung function or exercise capacity

We analysed the data for any correlation between self-rated QOL and curve severity, lung function or exercise capacity in patients with an AIS. 76/80 patients with AIS had completed the pre-operative SRS-22 questionnaire. We found no significant correlation between QOL and lung function or scoliosis severity.

A significant positive correlation was identified between QOL (function: r = 0.273, p = 0.017; self-image: r = 0.296, p = 0.009; mental health: r = 0.344, p = 0.02; total score: r = 0.336, p = 0.002) and maximal exercise capacity - VO_2peak (%predicted).

3.7. Analysis of exercise limitation

Comparison of scoliosis size, lung function and exercise variables was made between patients with normal exercise capacity (VO_2peak >80%predicted) and those with exercise limitation (Table 4).

Table 4.

Comparison of degree of scoliosis, lung function and exercise variables in patients with normal exercise capacity and those with exercise limitation. Comparisons were made using the Student’s t-test. *n = 80, **n = 10, ***n = 83, ****n = 12, ˆequal variances not assumed.

	VO2peak >80% predicted, n = 89 Mean (SD)	VO2peak ≤ 80% predicted, n = 15 Mean (SD)	p-value
Age (years)	14.6 (1.8)	16.6 (1.5)	<0.001
Main thoracic scoliosis (Cobb angle)	60.1 (15.8)*	58.6 (8.2)**	0.768
Most severe scoliosis (Cobb angle)	63.6 (13.0)***	58.0 (7.3)****	0.037ˆ
FEV1 (% predicted)	77.7 (13.0)	75.2 (21.6)	0.528
FVC (%predicted)	79.3 (12.2)	80.4 (11.9)	0.756
VO2 at VT (% predicted VO2max)	54.9 (7.2)	37.7 (6.5)	<0.001
Breathing reserve (%)	25.5 (18.0)	13.4 (11.8)	0.025

Patients with normal exercise capacity were significantly younger and had later ventilatory threshold suggestive of better muscle conditioning. These patients also had significantly larger scoliotic curves. Breathing reserve was significantly higher in those patients with exercise limitation. There was no statistical difference demonstrated in lung function between the 2 groups.

4. Discussion

It is recognised that spinal deformities may affect lung function and volumes, however the effect of such deformities on exercise capacity is less well characterised. Exercise capacity along with lung function may be useful in identifying patients at higher risk of respiratory complications from surgery. Furthermore, CPET is a dynamic test that assesses lung health in combination with cardiovascular, musculoskeletal and metabolic systems to ascertain the ability to transfer and utilise oxygen during exercise.

Correlations demonstrated poorer lung function in those patients with larger scoliotic curves; however, such patients with a more severe deformity were suggested to have higher levels of physical conditioning, and a higher exercise capacity than those with smaller curves. Previous studies have confirmed the association between scoliosis size and lung function^{3, 4, 5, 6}; however, the inverse association between lung function and exercise capacity is a novel finding. One previous study identified a positive correlation between these variables.¹⁷ We suggest that this finding represents improvements in muscle conditioning in those patients with larger curves and poorer lung function as a result of having to work harder in order to perform activities of daily living – a training effect. Training is the antithesis of de-conditioning and results in improved cardiac stroke volume, as well as increases in muscle capillary numbers, mitochondrial density and oxidative enzyme concentrations. Thus, the exercising individual is better able to transfer oxygen at muscle level and more effectively utilises oxygen to transform into energy in the exercising muscle. This prolongation of oxidative energy generation results in a later VT (switch to predominantly anaerobic energy generation) with an ability to exercise for longer and achieve a greater VO_2peak.

Patients with higher exercise capacity (as measured by VO_2peak) also report higher quality of life. The underlying cause for this relationship is unclear. This may reflect better exercise levels as a result of less pain and increased functional ability or alternatively, it could represent an improved sense of wellbeing as a result of being less limited in activity e.g. ‘exercise makes one feel better’.

We acknowledge that our results may be influenced by the inclusion of different diagnoses within the same analysis. Our comparison of demographics, lung function and exercise variables between different diagnoses demonstrates that children with earlier onset scoliosis tend to have larger curves, poorer lung function and proportionally better exercise capacity and that patients with Scheuermann’s kyphosis have significantly better lung function but tend to have worse exercise capacity. It is acknowledged that the sample sizes are small for the less common diagnoses.

Patients with exercise limitation had significantly higher breathing reserve (more gas left in the tank) than those with normal exercise capacity e.g. such patients were not ventilatory-limited. It is acknowledged that this finding is also seen where patients do not work to maximal effort during the CPET, however the additional association with earlier onset anaerobic metabolism (lower VO₂ at VT) suggests deconditioning rather than poor effort as the cause for these results. Furthermore, strict criteria were used during testing and reporting to ensure submaximal tests were not accepted for analysis.

We suggest, therefore, that poorer muscle conditioning due to a less active lifestyle rather than restrictive lung disease may be the principal limiting factor for exercise capacity in patients with spinal deformity. However, having utilised a retrospective study design, we acknowledge that we are unable to collect data to identify patients’ activity levels at the time of testing, and are therefore unable to account for this potential confounding factor in our analysis. Future studies utilising physical activity (PA) questionnaires, or measuring PA by accelerometry could be helpful in this regard.

5. Conclusion

In a pre-operative group of patients with spinal deformity, we have demonstrated novel associations between scoliosis size, lung function and exercise capacity. These data demonstrate that patients with a more severe scoliosis had reduced lung function and that this in turn was associated with reduced breathing reserve on exercise. Interestingly, these findings were also associated with improved levels of physical conditioning as suggested by a later onset VO₂ at VT (%predicted VO_2max) and greater peak exercise capacity (VO_2peak) suggesting that adaptation (or a training effect) may occur in those patients who exercise with reduced lung function and a greater spinal deformity. In addition, we identified a positive association between VO_2peak and self-reported quality of life which may support the hypothesis that exercise promotes general well-being. We suggest that physical adaptation may occur in response to poor lung function in patients with spinal deformity and further hypothesise that the resultant improvement in exercise capacity may be protective to patients’ quality of life.

Funding

None to declare.

Ethics

The study was a retrospective review of existing clinical testing and standardised practice. The study was discussed with the regional ethics committee and considered not to require formal ethics approval as was deemed to represent service evaluation.

Declaration of competing interest

None to declare.