Faculty Bibliography

STUDY OBJECTIVES: To develop and demonstrate the utility of measures of sleep continuity based on survival analysis techniques. DESIGN: Retrospective. SETTING: University sleep laboratory. PATIENTS: Anonymous nocturnal polysomnograms from 10 normal subjects, 10 subjects with mild sleep disordered breathing (SDB) (apnea-hypopnea index [AHI], 15-30/hr), and 10 subjects with moderate/severe SDB (AHI > 30/hr). INTERVENTIONS: N/A. MEASUREMENTS AND RESULTS: Hypnograms were analyzed to measure the lengths of episodes of contiguous sleep and processed using several common survival analysis techniques. Using separate survival curves for each group to describe the durations of continuous epochs of sleep (sleep run lengths), statistically significant differences were found between all groups (p < .001) as well as between the normal and mild SDB groups (p < .001), suggesting differences in the stability of sleep. Using survival regression techniques applied separately to each subject, statistically significant differences were found among all three groups (p < .001) and, more importantly, between the normal and mild SDB groups (p < .005). Similarly, estimation of sleep continuity based on the pooled sleep run data for each group also showed statistically significant differences (normal vs mild, p < .001; Normal vs moderate/severe, p < .001). In addition, the latter technique showed that changes in the "stability" of sleep could be demonstrated as runs progressed. CONCLUSION: Survival curve analysis of the lengths of runs of contiguous sleep provides a potentially useful method of quantifying sleep continuity. The results suggest that sleep becomes more stable as sleep progresses in normal subjects and those with mild SDB and less stable in subjects with moderate/severe SDB.

OBJECTIVE: Obese patients without clinically apparent heart disease may have a high output state and elevated total and central blood volumes. Central circulatory congestion should result in elevated pulmonary diffusing capacity (DLCO) and capillary blood volume (Vc) reflecting pulmonary capillary recruitment; however, the effect on membrane diffusion (Dm) is uncertain. We examined DLCO and its partition into Vc and Dm in 13 severely obese subjects (BMI = 51 +/- 14 kg/m2) without manifest cardiopulmonary disease before and after surgically induced weight loss. RESEARCH METHODS AND PROCEDURES: DLCO and its partition into Vc and Dm [referenced to alveolar volume (VA)] as described by Roughton and Forster, total body water by tritiated water, and fat distribution by waist-to-hip ratio were performed. RESULTS: Despite normal DLCO (mean 98 +/- 16% predicted), Vc/VA was increased (mean 118 +/- 30% predicted), and Dm/VA was reduced (mean 77 +/- 34% predicted). Nine of 13 subjects were restudied after weight loss (mean 52 +/- 43 kg); Vc/VA decreased to 89 +/- 18% predicted (p = 0.01), and Dm/VA increased to 139 +/- 30% predicted (p < 0.01). Increasing total body water was associated with both increasing Vc (r = 0.74, p = 0.01) and increasing waist-to-hip ratio (r = 0.65, p = 0.02), indicating that circulatory congestion increases with increasing central obesity. DISCUSSION: Severely obese subjects without manifest cardiopulmonary disease may have increased Vc indicating central circulatory congestion and reduced Dm suggesting associated alveolar capillary leak, despite normal DLCO. Reversibility with weight loss is in accord with reversibility of the hemodynamic abnormalities of obesity.

INTRODUCTION: We examined pulmonary diffusing capacity (D(LCO)) and its partition in pulmonary vascular diseases without evident parenchymal disease to assess the pattern and proportionality of change in membrane diffusion (D(m)) and capillary blood volume (V(c)). Disproportionate reduction in D(m) relative to V(c) (low D(m)/V(c)) in these diseases has been attributed to associated alveolar membrane/parenchymal disease, thus providing a potentially important diagnostic tool. METHODS: Diseases included: idiopathic pulmonary arterial hypertension (n=6), chronic thromboembolic disease (n=5), and intravenous drug use (n=14), providing a spectrum of pulmonary vascular diseases. V(c) and D(m) were determined as described by Roughton and Forster. RESULTS: All diseases showed a reduced V(c) (59+/-10, 69+/-14, 71+/-21 % predicted, respectively) and D(m) (76+/-22, 53+/-19, 63+/-16 % predicted, respectively) with no differences between groups (p>0.05). Disproportionate reduction of D(m) (D(m)/V(c) % predicted <1) was seen in all diseases (range 0.36-1.89). A mathematical analysis is presented to illustrate that changes in vascular geometry may additionally influence the proportionality of changes in D(m) and V(c). The mathematical analysis suggests that when reduction in patency of some vessels co-exits with compensatory dilatation of the remaining vasculature, a disproportionate reduction in D(m) relative to V(c) may result. CONCLUSIONS: The balance between vascular curtailment and compensatory dilatation may contribute to the variability of the D(m)/V(c) relationship seen in pulmonary vascular disease. Disproportionate reduction in D(m) relative to V(c) may result from this imbalance and need not imply subclinical alveolar membrane and/or parenchymal disease.

Acute hypercapnia may develop during periodic breathing from an imbalance between abnormal ventilatory patterns during apnea and/or hypopnea and compensatory ventilatory response in the interevent periods. However, transition of this acute hypercapnia into chronic sustained hypercapnia during wakefulness remains unexplained. We hypothesized that respiratory-renal interactions would play a critical role in this transition. Because this transition cannot be readily addressed clinically, we modified a previously published model of whole-body CO2 kinetics by adding respiratory control and renal bicarbonate kinetics. We enforced a pattern of 8 h of periodic breathing (sleep) and 16 h of regular ventilation (wakefulness) repeated for 20 days. Interventions included varying the initial awake respiratory CO2 response and varying the rate of renal bicarbonate excretion within the physiological range. The results showed that acute hypercapnia during periodic breathing could transition into chronic sustained hypercapnia during wakefulness. Although acute hypercapnia could be attributed to periodic breathing alone, transition from acute to chronic hypercapnia required either slowing of renal bicarbonate kinetics, reduction of ventilatory CO2 responsiveness, or both. Thus the model showed that the interaction between the time constant for bicarbonate excretion and respiratory control results in both failure of bicarbonate concentration to fully normalize before the next period of sleep and persistence of hypercapnia through blunting of ventilatory drive. These respiratory-renal interactions create a cumulative effect over subsequent periods of sleep that eventually results in a self-perpetuating state of chronic hypercapnia.