A short course of prednisone reduced early relapse rates after the treatment of acute asthma in the emergency room, an effect limited to the period of steroid administration.

i. Ventilatory response to CO 2 was measured regularly by a rebreathing technique in nineteen patients with severe asthma from the day of presentation to the time of clinical recovery.2. Ventilatory response to CO 2 increased during recovery in sixteen patients and the increased ventilatory response correlated well with increase of FEVr-Among these sixteen patients only one showed elevation of arterial CO 2 tension at the time of presentation.3. Ventilatory response to CO 2 failed to increase during recovery in three patients despite increasesin FEVr-All three patients showed elevation of arterial CO 2 tension at the time. of presentation.4. In five patients (including three of the four with initial hypercapnia) ventilatory response to CO 2 after recovery remained below the previously reported lower limit for normal subjects. The limits of normality were explored by examining ventilatory response to CO 2 in seventeen outstanding athletic performers. Values for ventilatory response to CO 2 both above and below the previously defined 'normal range' were found. The normal ventilatory response to CO 2 covers a 14-fold range from 0'57 to 8'171 min"" mmfIg"! Pco 2 • It is well known that ventilatory response to inhaled CO 2 is impaired in patients with chronic obstructive airways disease. Less is known of the patterns of ventilatory response to CO 2 when the airways obstruction is both acute and reversible. Read (1967) developed a 4 min rebreathing test suitable for serial studies of ventilatory response to CO 2 in sick patients. In the present study this rebreathing method has been used to study ventilatory response to CO 2 in patients recovering from attacks of severe asthma. Each patient was studied day by day during the course of clinical recovery so that, as airways obstruction lessened, each patient served as his own control.The range of ventilatory response to CO 2 in normal subjects was also further explored.

S U M M A R Y1. Changes in ventilation during progressive exercise were measured in eleven normal subjects. Ventilatory response to carbon dioxide at rest was measured in the same subjects using a rebreathing method.2. The range of ventilatory response to exercise was 16.6-32.0 litres min-' (litres CO, min-')-' (mean 22.7; SD 5.35); response to 0, uptake was 17.0-43-9 litresmin-'(litres 0, min-')-' (mean29.02; SD 9.07). Ventilatory response to CO, (Sco,) ranged from 0.81 to 3.22 litre min-' mmHg-' (mean 1-87; SD 0.62).3. There was a highly significant (P

Ear oximetry is widely used to detect arterial O2 desaturation during exercise in patients with cardiopulmonary disease. Although oximeters have been evaluated for accuracy, response time, and the influence of skin pigmentation, tests of their reliability have not been reported during strenuous exercise. Accordingly, we compared arterial O2 saturation (Sao2) measurements obtained by Hewlett-Packard (HP, model 47201A) and Biox II oximeters with those determined directly from arterial blood in six healthy volunteers during progressive exercise while rebreathing hypoxic gas mixtures. The relationship between the HP oximeter value and blood Sao2 was described by the equation: HP = 0.93 (Sao2) + 5.37 and for the Biox II: Biox = 0.55 (Sao2) + 38.97. With these equations, at a blood Sao2 value of 90%, the underestimation by both oximeters was less than 2%. At a blood value of 70%, the HP oximeter overestimated blood Sao2 by 0.7%, whereas the Biox II showed an overestimation of 10.7%. Below blood Sao2 of 83%, the Biox II tended to overestimate blood Sao2 by an amount greater than the error of the instrument, whereas the HP estimations were within the error of the instrument over all levels of blood Sao2 studied. We conclude that the HP oximeter provides valid estimates of Sao2 during exercise but that the Biox II oximeter, although reflecting qualitative changes in oxygenation that occur during exercise, does not provide accurate records of the degree of desaturation.

Each antibiotic has a characteristic penetration into respiratory secretions, influenced by serum concentration, protein binding, transport systems, and the presence of infection. Whereas previous studies have used only bronchial secretions and blood, in the present study, blood, bronchial secretions, and bronchoalveolar lavage (BAL) fluid were analyzed for tobramycin levels. In 10 studies in 9 patients, serum levels were significantly related to BAL fluid levels (r = 0.8, p less than 0.01) when both were expressed as a function of creatinine (mean BAL level +/- SD = 144 +/- 124 micrograms/mg creatinine; serum level, 293 +/- 216 micrograms/mg creatinine). The level of drug penetration in BAL fluid, expressed by the slope of the relationship between blood and fluid, was 0.5. The penetration of tobramycin into bronchial secretions ranged from 0 to 1.4 micrograms/ml, the ratio of secretions to serum being 0.2 (r = 0.68; p less than 0.05). It is concluded that the disposition of tobramycin in bronchial secretions and BAL fluid differ. Thus, sampling both fluids offers a more suitable method to study antibiotic pharmacokinetics in bronchi and alveoli.