Kyphosis

Study objectives: To investigate the relationship of thoracic kyphosis following tuberculosis to the development of ventilatory failure and to assess the efficacy of nocturnal noninvasive ventilatory support. Design: Retrospective consecutive case series with crossover from a phase without noninvasive ventilatory support to a phase with this treatment. Setting: The Respiratory Support and Sleep Centre, Papworth, Hospital, Cambridge, England. Patients: Seven patients with thoracic kyphosis following tuberculous osteomyelitis which had been contracted by the age of 4 years were studied. Their mean age was 53 (SD 7.1) years and the mean angle of kyphosis was 113.60. All patients were in ventilatory failure. Interventions: The patients were treated with nocturnal noninvasive ventilation with either an individually constructed cuirass shell and a negative pressure pump or nasal intermittent positive pressure ventilation using a volume preset ventilator. Measurement and results: Each patient underwent an initial clinical assessment along with radiologic studies of the spine, pulmonary function tests, daytime arterial blood gas tensions and overnight recordings of arterial saturation, and transcutaneous carbon dioxide tension. They were reassessed in detail at a mean of 5 years after starting ventilatory support. Symptoms, vital capacity, daytime carbon dioxide tension, and overnight oximetry had all improved following treatment. Temporary withdrawal of ventilatory support led to severe sleep fragmentation in four patients and the appearance of central apneas and hypopneas in the other three. Six of the 7 patients were alive at a mean of 5.7 years after starting nocturnal ventilation. Conclusion: These results show that ventilatory failure may develop, after an interval of many years, in patients with a severe thoracic kyphosis due to tuberculosis in childhood. Noninvasive nocturnal ventilatory support can control the symptoms of ventilatory failure, improve the physiologic abnormalities, and is associated with prolonged survival. (CHEST 1996; 110:1105-1110)

Key words: kyphosis; mechanical ventilators; respiratory insufficiency

Abbreviations: DI=desaturation index (number of dips in [SaO.sub.2] >4% per hour); Pemax=maximal expiratory mouth pressure; Pimax=maximal inspiratory mouth pressure; [PtcCO.sub.2]=transcutaneous partial pressure of [CO.sub.2]; REM=rapid eye movement; [SaO.sub.2]=arterial oxyhemoglobin saturation, TLC=total lung capacity; [TSaO.sub.2] <90%=percentage of time overnight when [SaO.sub.2] <90%.

Over a period of 10 years, 7 patients have been referred to the Respiratory Support and Sleep Centre, Papworth Hospital, Cambridge, England, with ventilatory failure associated with kyphosis and no other respiratory disorder. No previous report has analyzed this association or the outcomes of treatment. We therefore examined the clinical and physiologic factors which might have caused ventilatory failure in this group of subjects and present the results of treatment with long-term nocturnal-assisted ventilation in the home.

Materials and Method

Subjects

The records of all patients referred to the Respiratory Support and Sleep Centre who had kyphosis and subsequently received assisted ventilation were examined. Patients with an associated scoliosis, a neuromuscular disease, previous pulmonary tuberculosis, or another pulmonary disorder, such as chronic bronchitis, were excluded. Seven patients fulfilling the entry criteria for the study were referred to the center between May 1983 and July 1993. All patients gave their informed consent for inclusion in the study.

Study Design

Assessment at Initial Referral: Information was obtained regarding age of onset of the spinal deformity, duration, and nature of symptoms related to ventilatory failure, and tobacco consumption. A complete physical examination and an ECG were performed. Plain radiographs of the thoracic spine were obtained to determine the level of the kyphosis, and the angulation was estimated using Cobb's method.[1] The hemoglobin concentration and the hematocrit level were measured. [PaO.sub.2] and [PaCO.sub.2] were measured with the patient at rest breathing room air. The [FEV.sub.1] and FVC were measured using a dry spirometer (Vitalograph;). Functional residual capacity was measured by helium dilution (P. K. Morgan, Gillingham, United Kingdom) and total lung capacity (TLC) and residual volume were calculated. All volumes were corrected to body temperature, pressure, and saturation and expressed as a percentage of the predicted value calculated on the arm span rather than height to account for the spinal deformity.[2] Maximum mouth pressures were recorded using a calibrated variable reluctance pressure transducer (Si-Plan Electronics Research; Stratford-upon-Avon, England) and expressed as a percentage of predicted values calculated from regression equations according to a previously described method.[3]

The arterial oxyhemoglobin saturation ([SaO.sub.2]) was recorded continuously overnight using either an ear oximeter (Hewlett Packard; San Diego) or a pulse oximeter (Biox 3700; Ohmeda, Herts, UK) and the transcutaneous partial pressure of [CO.sub.2] ([PtcCO.sub.2]) was recorded using a heated polarographic electrode (TCM3; Radiometrics; Copenhagen). A two-channel recorder, with the paper speed set at 1 mm/min, was used to produce a hard copy. From the paper tracings, the mean and minimum [SaO.sub.2] levels, the percentage of time overnight when [SaO.sub.2] was less than 90% ([TSaO.sub.2] <90%), the desaturation index (DI [number of dips in [SaO.sub.2] greater than 4% per hour]), and the mean and minimum [PtcCO.sub.2] were calculated and recorded. The indications for the initiation of assisted ventilation and details of the technique of ventilatory support were recorded. Any changes in the method of assisted ventilation during subsequent admissions and the reasons for such changes were recorded.

Reassessment Admission: Between July 1993 and May 1994, the 7 patients were reassessed in detail. All of the investigations performed at initial referral were repeated. In order to observe abnormalities in nocturnal ventilation and to reassess the need for treatment, a controlled withdrawal of assisted ventilation was performed. The patients were monitored on the first night using the ventilator system which they used at home. On the second and third nights, assisted ventilation was withdrawn. On all three nights, [SaO.sub.2] and [PtcCO.sub.2] were recorded continuously. On the 3rd day, arterial blood gas tensions were measured and on the 4th night polysomnography was performed (Biomedical Monitoring Systems amplifiers and Sleepmaster computer). Chest and abdominal movements were recorded using inductance plethysmography bands (Respi-trace) and nasal-oral airflow was measured with a thermistor. [SaO.sub.2] was measured using a pulse oximeter. The EEG was staged using standard criteria[4,5] the record was scored for sleep efficiency, and the apnea-hypopnoea index was calculated.

Statistical Analysis

Correlation between the angle of the kyphosis and pulmonary function tests at presentation, the daytime blood gas tensions, hemoglobin levels, and overnight measurements of [SaO.sub.2] and [PtcCO.sub.2] were examined using Kendall's ranked correlation coefficient. Differences were sought between the results of the laboratory investigations at first referral and those on the first day of the reassessment admission using the Wilcoxon's matched-pairs signed-rank test. The analysis was performed using the SPSS package (SPSS, Chicago). Probability results of less than 0.05 were accepted as significant.

Results

Initial Referral

Clinical Details at Initial Referral: The 7 patients (4 men) had a mean age at referral of 53 years (SD 7.09). All had had tuberculous osteomyelitis of the spine before the age of 4 years (mean age, 2.2 years; SD 1.1). This had affected the upper thoracic spine in two patients and the midthoracic spine in five. The mean angle of kyphosis at presentation was 113.6 [degrees] (SD, 14; range, 95 [degrees] to 135 [degrees]). The patients had been treated with bedrest and plaster braces. None had received corrective surgery, and only one had received antituberculous chemotherapy.

All patients were ex-smokers (mean number of pack-years was 14; SD 12.6), but none complained of a productive cough or had wheezing on examination. All 7 patients described increasing dyspnea on exertion over a period of at least 4 years. At referral four had dyspnea walking on the flat surface and three had dyspnea at rest. Two patients had morning headaches and all seven described disrupted sleep and daytime somnolence. Six patients were being treated with diuretics, but three of these had persisting peripheral edema.

Physiologic Measures at Initial Referral: All 7 patients had p pulmonale, right axis deviation, and evidence of right ventricular hypertrophy on an ECG. The results of the arterial blood gas estimations are shown in Table 1. The daytime resting [PaCO.sub.2] was greater than 50 mm Hg in 6 patients (mean 53.7 mm Hg; SD 7.30). Results of lung function tests at initial referral are shown in Table 2 and demonstrate a restrictive pattern. In 6 patients TLC and residual volume were both reduced to less than 50% of predicted. Maximal inspiratory mouth pressure (Pimax) was less than 50% of predicted in 2 patients and greater than 85% of predicted in only 2. In 5 patients, maximal expiratory mouth pressure (Pemax) was recorded only as greater than 100 em [H.sub.2]O and so the percentages of predicted values could not be calculated. However, using a figure of 100 cm [H.sub.2]O for these 5 patients, Pemax was at least 85% of predicted for 6 patients and 36% of predicted for the remaining patient. The complete results of overnight monitoring at initial referral are given in Table 2. Of particular note is the low level of the mean overnight [SaO.sub.2] (mean, 81%, SD, 4.2) and [TSaO.sub.2] <90% (mean, SD, 12.3).

[TABULAR DATA OMITTED]

There was no correlation between the angle of kyphosis on the one hand and the hemoglobin concentration, daytime arterial blood gas tensions, the overnight oximetry measures, or maximal mouth pressures on the other hand. There was no overall correlation between the Cobb angles and the measures of lung volume. However, one patient, who did not have any other unusual measurements, did have larger lung volumes despite having the most pronounced kyphosis. If the data of de patient are excluded, there is a strong inverse correlation between the angle of kyphosis and [FEV.sub.1], TLC, and in particular FVC (r=0.88; p=0.02) with each expressed as a percentage of the predicted value (Fig 1).

Progress: Noninvasive nocturnal-assisted ventilation was commenced in all patients. In six patients, the indication was stable ventilatory failure with an elevated daytime [PaCO.sub.2]. The remaining patient had relatively normal daytime blood gas tensions (PaO.sub.2], 70 mm Hg; [PaCO.sub.2], 41.4 mm Hg) but had a mean overnight [SaO.sub.2] value of 79% and a [PtcCO.sub.2] level of 53 mm Hg associated with peripheral edema that did not resolve with treatment with diuretics. Cuirass ventilation using customized shells, and the Newmarket negative pressure pump (Si-Plan Electronics Research, Stratford-upon-Avon, England) was begun for the three patients referred to the center before 1988. The other 4 patients, referred after 1988 began receiving nasal intermittent positive pressure ventilation using standard masks (Respironics, Murrysville, Pa) and the Monnal D positive pressure ventilator in assist-control mode (Taema; Paris, France). None of the patients required additional oxygen therapy at this stage. The dose of diuretics was reduced in all patients taking them, and the diuretics were discontinued completely in two patients.

All patients have been followed up at regular intervals. One patient who had been receiving cuirass ventilation deteriorated after an interval of 29 months with a history suggestive of obstructive sleep apnea and peripheral edema. Overnight monitoring showed a DI of 28, a mean [SaO.sub.2] of 83%, and a [TSaO.sub.2] <90% of 88%, although daytime arterial blood gas values were normal. He was converted to nasal intermittent positive pressure ventilation using a Monnal D ventilator with an improvement in his symptoms and resolution of the peripheral edema. Subsequent overnight monitoring revealed a DI of 4 and a mean [SaO.sub.2] of 90%. One other patient who had been receiving cuirass ventilation deteriorated after a period of 34 months with recurrent hypoxemia and peripheral edema (overnight mean [SaO.sub.2], 81%; [TSaO.sub.2] <90%, 100%) but no evidence of obstructive sleep apnea. The overnight [PtcCO.sub.2] showed that he was well ventilated, and daytime [PaCO.sub.2] while the patient was breathing air was only 43.6 mm Hg. Oxygen therapy was introduced to correct hypoxemia. He died of bronchopneumonia a little more than 9 years after his initial referral. Postmortem examination confirmed right ventricular enlargement and dilated main pulmonary arteries, which was consistent with pulmonary hypertension.

Reassessment Admission

The mean period between the initial referral and the withdrawal studies was 5.08 years (SD 2.50). With the exception of one, all the patients thought that their symptoms were well controlled using nocturnal nasal ventilation with the Monnal D or cuirass ventilation. Six patients reported breathlessness on inclines only and had no morning headaches or daytime somnolence. The remaining patient, who died 3 months after this admission, was breathless on minimal exertion and complained of insomnia. None of the patients had peripheral edema. Improvements were demonstrated in the daytime arterial blood gas values with a statistically significant fall in the [PaCO.sub.2]. The dynamic lung volumes improved in all patients and the differences were statistically significant <p<0.05) although the changes were small. There were no statistically significant changes in maximal mouth pressures. in four patients, however, the Pimax was lower at reassessment, and one patient was too breathless to repeat the measure. Overnight monitoring showed significant improvements in the mean and minimum [SaO.sub.2] levels and in the [TSaO.sub.2] <90%. The complete results are given in Tables 1 and 2. The ECG had not altered in any of the patients.

The arterial blood gas levels on the 1st and 3rd days of the reassessment admission were not significantly different. However, there were differences in the overnight oximetry traces on the first and fourth nights, with and without assisted ventilation, respectively. Using Wilcoxon's matched-pairs sum-ranked test, the mean overnight [SaO.sub.2] was significantly lower (p=0.04;mean difference, 6.7%) as was the minimum [SaO.sub.2] (p=0.03, mean difference, 17.3%). The other major finding on the fourth night, when polysomnography was performed, was that 3 of the 7 patients slept very poorly without ventilatory support (sleep efficiency <70%) and one did not sleep at all. As a consequence, the studies were terminated early in these four patients and ventilatory support was restarted during the night. In the 3 patients who spent the whole night without assisted ventilation, the mean sleep efficiency was 83% and mean apnea-hypopnea index was 14.7/h. None of the seven patients had obstructive apneas, but central apneas and hypopneas were demonstrated which were most frequent in rapid eye movement (REM) sleep.

Discussion

We have reported 7 patients with ventilatory failure associated with thoracic kyphosis. This has only been previously documented in a few case reports,[7-10] but it may be that the association is more common than is apparent from the literature. Many series report patients with kyphoscoliosis and ventilatory failure.[11,12] A true kyphoscoliosis is very rare, and while the term usually is used as a synonym for scoliosis, it is possible that some patients with a genuine kyphosis have been included with the scoliotic subjects in previous series.

This distinction is important because both the etiology of kyphosis and scoliosis and the abnormalities of the respiratory mechanics observed in them are very different. All seven of our patients had had tuberculous osteomyelitis in the thoracic spine. This is recognized to cause a sharp kyphosis or gibbus while scoliosis is usually absent or trivial. The normal thoracic kyphosis has a range of 20 [degrees] to 40 [degrees],[13 while in our patients the mean value was 113.6 [degrees]. In all of the seven subjects, the kyphosis involved the upper or midthoracic vertebrae. Tuberculosis usually infects the vertebrae of the thoracolumbar junction, but it is only when it damages the thoracic vertebrae that important respiratory consequences would be anticipated.

The effects of thoracic scoliosis on respiratory mechanics have been extensively investigated.[14] There is a large decrease in the compliance of both the lungs and the chest wall in adults with a Cobb angle of greater than 100 [degrees]. The respiratory muscles on either side of the asymmetrical chest are affected differently, and the overall influence on maximal mouth pressures appears to depend on the severity of the angulation. In mild scoliosis, with an angle of less than 30%, Pemax has been shown to be reduced more than Pimax,[15] while in more severe scoliosis with an angle greater than 60 [degrees], Pimax has been found to be considerably reduced while Pemax was preserved.[16] With kyphosis, the chest wall may be greatly deformed but remains symmetrical and the changes in orientation of the respiratory muscles are different from those seen with scoliosis. In our patients at referral, Pimax was less than 85% of predicted in all but 2 patients, while Pemax was greater than 85% of predicted for all but 1 patient. It seems likely from our findings that despite the small lung volumes, diaphragm function is impaired, possibly because it is at a mechanical disadvantage due to its abnormal orientation.

In all of our patients, the kyphosis was apparently by the age of 4 years. This has been a feature of the previously reported cases, and an increased risk of ventilatory failure has been demonstrated in patients with scoliosis of early onset.[17] The number of alveoli normally increases until around the age of 8 years but it only does so in response to the mechanical forces generated by the chest wall. The failure of the thoracic cavity to develop, the presence of abnormal respiratory mechanics due to the kyphosis itself, and the ankylosis of the costovertebral joints may have prevented the usual number of alveoli form being formed.[18] This would contribute to the restrictive deficit seen in our patients.

Ventilatory failure only developed 43 to 61 years after the kyphosis was noticed, suggesting that it may have been precipitated by changes associated with aging, such as reduction in respiratory drive or chest wall compliance. Loss of respiratory muscle strength probably was a contributory factor because the Pimax was reduced in all but two patients. In other patient groups treated with assisted ventilation, maximal mouth pressures have improved, and it has been argued that this may reflect the relief of respiratory muscle fatigue.[19] Because there was no overall improvement in mouth pressures after effective treatment in these patients with kyphosis, we would argue that respiratory muscle fatigue was not an important factor in the development of ventilatory failure.

The patients have been followed up for a mean of 5.7 years (SD, 2.52), and 6 of the 7 patients are still alive. One patient, age 62, died a little more than 9 years after the initiation of assisted ventilation. Right heart failure (cor pulmonale) has been reported previously in patients with kyphosis,[20] and in one series all four patients died.[21] Peripheral edema was present in three of our patients at the time of first referral. This resolved in each case once assisted ventilation had been introduced. The electrocardiographic changes of right ventricular hypertrophy did not however improve and the patient who died had postmortem evidence of pulmonary hypertension, which was probably a factor in his death.

At reassessment, the mean [SaO.sub.2] at night improved compared with that at presentation; the values increased from 81 to 92% (p=0.0001). The daytime [PaCO.sub.2] fell significantly between presentation and the reassessment admission (mean interval, 5.08 years). Similar improvements in daytime blood gas tension have been demonstrated previously with assisted ventilation in other chest wall disorders, such as scoliosis,[11] and after thoracosplasty[19] but not in kyphosis. Interestingly, on the fourth night of the reassessment admission in the present study, only three of the seven patients were able to tolerate the whole night off their ventilators. In all patients, there was a deterioration in the mean overnight [SaO.sub.2] after this short period without assisted ventilation. In a previous study of the withdrawal of ventilatory assistance for a period of 15 days, similar changes in nocturnal gas exchange values were found.[22] Our results suggest that even shorter periods of withdrawal may be deleterious.

None of our patients demonstrated obstructive sleep apnea when self-ventilating at night. The records of the three subjects who managed to complete the fourth night of the reassessment admission without assisted ventilation showed central apneas and hypopneas which were most frequent in REM sleep. There have been no previous polysomnographic studies in kyphotic subjects to compare with the findings of the present study. The present observations suggest that either the reduction in biochemical ventilatory drive in REM sleep compare with no-REM sleep and wakefulness or the loss of accessory muscle activity in REM sleep was responsible for these apneas and hypopneas.

The findings suggest that ventilatory failure with kyphosis is associated with an onset of deformity in early childhood and may be more common than has been previously recognized. The cause of the development of ventilatory failure after an interval of over 40 years remains uncertain, but it seems that respiratory muscle weakness is a factor and that abnormalities in nocturnal ventilation preceded the deterioration in daytime arterial blood gas values. We have shown that treatment with nocturnal assisted ventilation is well tolerated and can effectively control symptoms and improve physiologic abnormalities in the majority of patients.

References

[1] Cobb JR. Outline for the study of scoliosis. Am Acad Orthop Surg 1948; 5:261-75. [2] Linderhom H, Lindgren U. Prediction of spirometric values in patients with scoliosis. Acta Orthop Scan 1978;49:469-78 [3] Wilson SH, Cooke NT, Edwards RHT, et al. Predicted normal values for maximal respiratory pressures in caucasian adults and children. Thorax 1984; 39:535-38. [4] Ryan CF, Lowe AA, Li D, et al. Magnetic resonance imaging of the upper airway in obstructive sleep apnea before and after chronic nasal continuous positive airway pressure therapy. Am Rev Respir Dis 1991; 144:939-44 [5] Rechtschaffen, A, Kales A. A manual of standardised terminology, techniques and scoring system for sleep stages of human subjects. Bethesda, Md: US Government Printing Office; 1968. National Institutes of Health publication 204 [6] Brown L, Kinnear WJM, Seargent KA, et al. Artificial ventilation by external negative pressure: a method for making cuirass shells. Physiotherapy 1985; 71:181-83 [7] Brille D, Hatzfeld C, Lejeune F. Le traitement de l'hypoventilation alveolaire dans la defaillance cardio-respiratoire des gibbeux. J Fr Med Chir Thor 1963; 17:181-90 [8] Gimenez M, Pham QT, Vittoz-Polu E. Aspects particuliers de la reeducation des gibbeux, insuffisants respiratoires. Ann Med Physique 1969; 12:9-25 [9] Prowse CM, Gaensler EA. L'insuffisance respiratoire aigue des gibbeux. Anesthesiology 1965; 26:381-92 [10] Turino GM, Goldring RM, Fishman P. Cor pulmonale in musculoskeletal abnormalities of the thorax. Bull NY Acad Sci 1965; 41;959-80 [11] Ellis ER, Grunstein RR, Chan S, et al. Noninvasive ventilatory support during sleep improves respiratory failure in kyphoscoliosis. Chest 1988; 94:811-15 [12] Guilleminault C, Kurland G, Winkle R, et al. Severe kyphoscoliosis, breathing, and sleep. Chest 1981; 79:626-30 [13] Moe JH, Winter RB, Bradford DS, et al. Scoliosis and other spinal deformities. Philadelphia: WB Saunders, 1978; 10 [14] Baydur A, Milic-Emili J. Respiratory mechanics in kyphoscoliosis. Monaldi Arch Chest Dis 1993; 48:69-79 [15] Smyth RJ, Chapman KR, Wright TA, et al. Pulmonary function in adolescents with mild idiopathic scoliosis. Thorax 1984; 39:901-04 [16] Cooper DM, Velasquez Rojas J, Mellins RB, et al. Respiratory mechanics in adolescents with idiopathic scoliosis. Am Rev Respir Dis 1984; 130:16-22 [17] Branthwaite MA. Cardiorespiratory consequences of unfused idiopathic scoliosis. Br J Dis Chest 1986; 80:360-69 [18] Berend N, Marlin GE. Arrest of alveolar multiplication in kyphoscoliosis. Pathology 1979; 11:485-91 [19] Jackson M, Smith IE, King MA, et al. Long term non-invasive domiciliary assisted ventilation for respiratory failure following thoracoplasty. Thorax 1994; 49:915-19 [20] Hanley T, Platts MM, Clifton M, et al. Heart failure of the hunchback. Q J Med 1958;27:155-71 [21] Coombs CF. Fatal cardiac failure occurring in persons with angular deformity of the spine. Br J Surg 1930; 18:326-28 [22] Jimenez JFM, de Cos Escuin JS, Vicente CD, et al. Nasal intermittent positive pressure ventilation: analysis of its withdrawal. Chest 1995; 107:382-88

COPYRIGHT 1996 American College of Chest Physicians
COPYRIGHT 2004 Gale Group