ABSTRACT
When deciding whether to initiate elective mechanical ventilatory support for patients with chronic respiratory failure, the decision can be facilitated by the clinician having objective information about the patient’s nocturnal ventilation and gas exchange. Some of this information may be deduced from the patient’s daytime arterial blood gases (ABG), especially if hypoventilation is present, as the carbon dioxide tension (PCO2) will be elevated. However, in many patients, the daytime gases underestimate the severity of nocturnal hypoventilation (1). The recognition of respiratory changes during sleep is important as it may result in an approach that could reduce morbidity and mortality. The relationship between altered pulmonary function and disordered sleep is not a precise one. In patients with amyotrophic lateral sclerosis, neither forced vital capacity nor negative inspiratory pressure could predict sleep-disordered breathing (2). In patients with myasthenia gravis, pulmonary function was documented as an associated risk factor for abnormal breathing during sleep (3). In our experience (4), the vital capacity of 30 patients with kyphoscoliosis and respiratory failure was 27% 9% predicted, their maximum inspiratory pressure was 25% 14% predicted, and their maximum expiratory pressure was 73% 50% predicted. As the underlying conditions progress, impaired mechanics are accompanied by altered resting awake arterial gas tensions, with PCO2 levels of 56 to 61 mmHg and PaO2 levels of 37 to 70 mmHg in representative studies (5-7). Not surprisingly, the relationship between altered arterial blood gas tensions and pulmonary function is not precise. A study, designed to determine at what level of mechanical impairment respiratory failure was likely to occur, measured 53 patients with proximal myopathies. The authors reported that hypercapnia occurred when respiratory muscle strength fell below 30% predicted in uncomplicated myopathies and when the vital capacity fell below 55% predicted in those with or without parenchymal involvement (8).
