1048 Prevalence Of Positional Obstructive Sleep Apnea (OSA) In Patients With OSA-COPD Overlap Syndrome

Introduction

Positional OSA (non-supine apnea-hypopnea index [AHI] < 5 events/hr) is present in 30% of patients with OSA. We demonstrated that in patients with OSA- COPD overlap syndrome the AHI inversely correlated with the degree of gas trapping, suggesting a stabilizing effect on the upper airway. We hypothesized that sleep position would be less important, resulting in a lower prevalence of positional OSA.

Methods

Patients underwent a polysomnogram that demonstrated OSA (AHI > 5 events/hr). To confirm COPD, patients had spirometry performed and a chest computed tomography for measurements of percent gas trapping.

Results

Sixteen patients [6 (38%) males, 55±7 years/old, FEV1 1.2±0.5 L, FEV1 % Predicted 45±19%, FVC 2.3±0.8 L, FVC % Predicted 69±20%, FEV1/FVC 51±12%, BMI 33±9 kg/m2)] were diagnosed with OSA (AHI 15±12 events/hour). Four patients (25%) had positional OSA (AHI 13±6 events/hr, non-supine AHI 1±1 event/hr) compared to 12 patients who were non-positional [AHI 16±13 events/hr (p=0.95)]. There was no difference in age [52±8 and 56±7 yrs (p=0.3)] or severity of obstruction in those with and without positional OSA [FEV1 1.4±4 L and 1.1±0.5 L, (p=0.3), FEV1 % predicted 50±17% and 44±20%, (p=0.7), FVC 2.9±0.8 L and 2.1±0.8 L (p=0.1), FVC % predicted 78±21% and 66±20%, (p=0.3), and FEV1/FVC 50±11% and 51±12%, (p=0.8), respectively]. However, patients with positional OSA were less heavy than those with non-positional OSA [BMI 23±3 and 37±8 kg/m2, respectively (p=0.005)]. Finally, there was no difference in the CT-Derived % Gas Trapping in those with and without positional OSA [48±37% and 36±25%, (p=0.6), respectively].

Conclusion

The prevalence of positional OSA in patients with OSA-COPD overlap is similar to OSA patients without COPD. Despite the presence of obstructive disease and gas trapping that may affect upper airway stability, other factors including body position and BMI remain important determinants for developing OSA in patients with COPD.

Support

R01-HL089856, R01-HL089897

0567 Hypoxic Burden and Apnea-Hypopnea Duration in Patients with Positional Obstructive Sleep Apnea

Introduction

Recently, the measurement of the hypoxic burden and apnea-hypopnea duration has been shown to correlate with mortality in patients with obstructive sleep apnea (OSA). We hypothesized that in patients with mild positional OSA (apnea-hypopnea index [AHI] < 5 events/hr in the non-supine position) the hypoxic burden would be increased and apnea-hypopnea duration shortened and similar to patients with non-positional OSA.

Methods

Fourteen patients with positional OSA and 24 patients non-positional OSA with similar severity of OSA based on the respiratory event index (REI) were included. All patients had a home sleep apnea test for suspected OSA. The hypoxic burden was calculated by the multiplication of REI and the mean area under the desaturation curves.

Results

Thirty-eight patients [12 (35%) males, 50±12 yrs, BMI 35±7 kg/m2, Epworth Sleepiness Scale (ESS) 11±8, REI 10±3 events/hr, apnea-hypopnea duration 19±4 sec, mean SaO2 94±2%, lowest SaO2 79±8%, % total sleep time (TST) SaO2 < 90% 11±16%, hypoxic burden 30±17 %min/hr] completed the study. Fourteen patients [9 (64%) males, 46±14 yrs, BMI 31±6 kg/m2, ESS 7±5, REI 9±3 events/hr, mean SaO2 94±2%, lowest SaO2 81±6%, %TST SaO2 < 90% 4±6%] had positional OSA (supine REI 16±7 events/hr, non-supine REI 3±1 events/hr) and 24 patients had non-positional OSA [3 (13%) males, 52±10 yrs, BMI 38±7 kg/m2, ESS 12±9, REI 10±3 events/hr, mean SaO2 94±2%, lowest SaO2 77±9%, %TST SaO2 < 90% 14±19%]. The hypoxic burden was elevated in both the positional and non-positional OSA patients with no difference between the groups (26±19 %min/hr and 32±15 %min/hr, respectively, p=0.13). The apnea-hypopnea duration was similar in positional and non-positional OSA patients (20±3 sec and 18±4 sec, respectively, p=0.08 sec).

Conclusion

In patients with mild positional OSA the hypoxic burden, which has been associated with cardiovascular mortality, is elevated and similar to patients with non-positional OSA.

Support

None

0610 Prevalence of Positional Obstructive Sleep Apnea Based on 3% Vs 4% Oxygen Desaturation Using Home Sleep Apnea Testing

Introduction

Approximately 30% of patient with obstructive sleep apnea (OSA) have positional OSA [non-supine apnea-hypopnea index (AHI) < 5 events/hr]. However, the prevalence is based on variable definitions for hypopneas related to the degree of oxygen desaturation. In addition, use of a home sleep apnea test (HSAT) to identify positional OSA is limited. We hypothesized that in patients evaluated with an HSAT, using a definition for hypopneas based on 4% compared to 3% oxygen desaturation will significantly decrease the percentage diagnosed with positional OSA.

Methods

Fourteen patients with positional OSA based on a non-supine respiratory event index (REI) < 5 events/hr were included. The initial diagnosis was determined based on a hypopnea definition of ≥ 3% oxygen desaturation. The studies were reanalyzed using a hypopnea definition of ≥ 4% oxygen desaturation.

Results

Fourteen patients [9 (64%) males, 46±14 yrs, BMI 31±6 kg/m2, ESS 7±5, REI 9±3 events/hr, mean SaO2 94±2%, lowest SaO2 81±6%, %TST SaO2 < 90% 4±6%] were identified with positional OSA (supine REI 16±7 events/hr, non-supine REI 3±1 events/hr) using a hypopneas definition of ≥ 3% oxygen desaturation. When reanalyzed using a hypopnea ≥ 4% oxygen desaturation there was a significant decrease in the REI to 7±2 events/hr (p<0.001). Three patients (21%) no longer were considered to have OSA. These patients were younger (32±14 vs. 50±11yrs, p=0.03) and had less severe OSA (REI 6±1 vs. 9±3 events/hr (p=0.04), but there was no difference in BMI (32±11 vs. 31±5 kg/m2, p=0.9) or mean and lowest SaO2 (96±0.4 vs. 94±2%, p=0.13, and 82±8 vs. 81±6%, p=0.9, respectively).

Conclusion

In patients with mild positional OSA, using a hypopnea definition of at least 4% vs. 3% oxygen desaturation on a HSAT will have a significant effect on the overall REI and often exclude patients who would otherwise be treated for OSA.

Support

None.

Quality of life in survivors of prolonged mechanical ventilatory support

OBJECTIVE

To examine the long-term quality of life (QOL) in a group of patients after prolonged mechanical ventilatory support.

DESIGN

Prospective cohort study.

SETTING

Outpatient follow-up.

PATIENTS

Survivors of prolonged mechanical ventilatory support who were discharged from a ventilator rehabilitation unit (VRU).

INTERVENTIONS

Measurement of health-related QOL using the Sickness Impact Profile (SIP).

MEASUREMENTS AND MAIN RESULTS

Forty-six patients were contacted approximately 2 yrs after their discharge from the VRU and asked to complete the SIP. Twenty-five patients (age, 59 +/- 17 yrs; duration of mechanical ventilatory support, 45 +/- 36 days [mean +/- sd]) agreed to participate in this study and completed the SIP questionnaire 23 +/- 18 months after their discharge from the VRU. Patients' VRU stay was 29 +/- 21 days. Two patients were discharged with nocturnal ventilatory support, and the rest were completely weaned of mechanical ventilatory support before discharge. Fifteen patients (60%) were discharged to home, eight patients (32%) were discharged to a rehabilitation facility, and two patients (8%) were discharged to a skilled-care facility. Most patients had mild dysfunction, and the global SIP score was 12 +/- 10, the physical dimension score was 12 +/- 12, and the psychosocial dimension score was 9 +/- 11 (SIP scores range from 0 to 100, with higher scores indicating worse QOL). Subgroup analysis showed that postoperative patients had lower SIP scores compared with patients with chronic respiratory diseases (global SIP, 7 +/- 6 vs. 19 +/- 8; p <.05). Moreover, the patients in the postoperative group were older, but had similar SIP scores as patients who had acute lung injury (17 +/- 15). Global SIP scores correlated with age (r = -.40; p =.046), but not with duration of mechanical ventilatory support (r = -.23) or VRU admission Acute Physiology and Chronic Health Evaluation II scores (r = -.39; p =.06).

CONCLUSIONS

In survivors of prolonged mechanical ventilatory support, using specific selection criteria shows that there is minimal impairment in the QOL at long-term follow-up. Although some patients continue to have moderate to severe limitations, it is the cause of respiratory failure and the underlying disease, rather than duration of ventilatory support, that have a significant impact on QOL.

Acute respiratory failure after lung volume reduction surgery

In this study, we characterized patients who developed respiratory failure postoperatively after lung volume reduction surgery (LVRS). We retrospectively reviewed the records of 72 patients who underwent LVRS from February 1995 to February 1998, examining perioperative variables and complications. Twenty-one patients (29%) developed postoperative respiratory failure, five due to hypoxemia, nine due to hypercapnia, and seven secondary to hemodynamic instability. The hospital mortality was 33% among patients who developed respiratory failure. No preoperative clinical or physiologic variable (including percent ideal body weight, serum albumin, prednisone use, lung function, maximal O(2) uptake on exercise testing, 6-min walk distance, and hemodynamic parameters) was predictive of postoperative respiratory failure. Patients who developed respiratory failure were older (63 +/- 7 versus 57 +/- 8 yr, p = 0.01), had longer anesthesia time (188 +/- 96 versus 127 +/- 56 min, p = 0.001), had a higher incidence of coronary artery disease (40% versus 10%, p = 0.001) and performance of concomitant surgical procedures during the LVRS operation (40% versus 2%, p < 0.001) compared with those without respiratory failure. All patients who underwent simultaneous surgery, which were mostly for cardiac disease, developed respiratory failure. Risk factor analysis confirmed that older patients and those undergoing cardiac surgery combined with LVRS are at increased risk for postoperative respiratory failure.

Myocardial ischemia and weaning failure in patients with coronary artery disease: an update

OBJECTIVE

To determine the frequency and effects of weaning-related myocardial ischemia on weaning outcomes in patients with coronary artery disease.

DESIGN

Prospective cohort study.

SETTING

Medical and cardiac intensive care units of a 300-bed teaching community hospital.

MEASUREMENTS AND MAIN RESULTS

Three-lead ST segments, heart rate-systolic blood pressure products, and respiratory rate/tidal volume ratios were obtained for patients with coronary artery disease just before and during their initial trials of weaning from mechanical ventilation. ST segments were interpreted by a blinded cardiologist. Eighty-three patients with a mean age of 72.4 +/- 1.1 years (mean +/- SEM), a mean Acute Physiology and Chronic Health Evaluation II score of 16.4 +/- 0.8, and a mean duration of mechanical ventilation of 4.6 +/- 0.9 days were studied. Eight patients showed electrocardiographic evidence of ischemia during weaning, and seven of these patients failed to be liberated on their first day of weaning. The presence of ischemia significantly increased the risk of weaning failure (risk ratio, 2.1; 95% confidence interval, 1.4-3.1). The rate-pressure product for the group as a whole increased significantly during weaning, from 11.9 +/- 0.4 to 13.5 +/- 0.5 mm Hg x beats/min x 10(3) (p < .01). The increase in rate-pressure product tended to be greater in patients who became ischemic (12.8 +/- 0.9 to 17.3 +/- 2.0 mm Hg x beats/min x 10(3)) than in patients who were not ischemic during weaning (11.8 +/- 0.4 to 13.0 +/- 0.5 mm Hg x beats/min x 10(3); p = .05). The rate/volume ratio did not change significantly during weaning, but the rate/volume ratios after both 1 min (65.6 +/- 4.6 vs. 98.0 +/- 9.4 breaths/min/L; p < .05) and 30 mins (68.6 +/- 4.3 vs. 91.1 +/- 8.9 breaths/min/L; p < .05) of unassisted breathing were lower in successful than in unsuccessful patients.

CONCLUSION

Electrocardiographic evidence of myocardial ischemia occurs frequently and is associated with significantly increased risk of first-day weaning failure in patients with coronary artery disease.

Effects of a medical intensivist on patient care in a community teaching hospital

OBJECTIVE

To determine the effect of adding a trained intensivist on patient care and educational outcomes in a community teaching hospital.

MATERIAL AND METHODS

We retrospectively reviewed outcomes for patients admitted to the medical intensive-care unit (MICU) of a 270-bed community teaching hospital between July 1992 and June 1994. Mortality rates and durations of stay were determined for the year before (BD, 1992 through 1993) and the first year after (AD, 1993 through 1994) introduction of a full-time director of critical care. Performance of resident trainees on a standardized critical-care examination was measured for the same periods.

RESULTS

Overall, 459 patients in the BD period were compared with 471 patients in the AD period. The mix of cases and severity of illness (acute physiology and chronic health evaluation or APACHE II scores) on admission were similar for the BD and AD periods. MICU mortality decreased from 20.9% during the BD to 14.9% during the AD period (P = 0.02), and in-hospital mortality decreased from 34.0% to 24.6% (P = 0.002). Disease-specific mortalities were lower during the AD period for most categories of illness. Detailed analysis of a subgroup of patients (those with pneumonia) demonstrated no differences in distribution of patients by gender, race, or acuity of illness (APACHE II scores). The mortality rate due to pneumonia decreased from 46% during the BD period to 31% during the AD period. This decrease was consistent across categories of APACHE II scores. From BD to AD periods, mean durations of total hospital stay decreased from 22.6 +/- 1.4 days to 17.7 +/- 1.0 days, and mean MICU stay decreased from 5.0 +/- 0.3 days to 3.9 +/- 0.3 days (P < 0.05). Critical-care in-service examination scores for 22 residents increased from 53.8 +/- 1.7% to 67.5 +/- 2.2% (P < 0.01), and AD scores were significantly higher than BD scores for residents at similar levels of training.

CONCLUSION

Addition of a medical intensivist was temporally associated with improved clinical and educational outcomes in our community teaching hospital.

The unassisted respiratory rate/tidal volume ratio accurately predicts weaning outcome in postoperative patients

OBJECTIVE

To evaluate the accuracies of the respiratory rate/tidal volume ratio (rate/volume ratio), minute volume, and negative inspired force in predicting weaning outcome in postoperative mechanically ventilated patients.

DESIGN

A prospective, observational study.

SETTING

Surgical intensive care unit of a 270-bed community teaching hospital.

PATIENTS

One hundred eighty-three postoperative, mechanically ventilated patients.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

The spontaneous minute volume, unassisted respiratory rate/tidal volume ratio, and negative inspired force were measured just before weaning. The rate/volume ratio was remeasured after 30 to 60 mins of weaning. Weaning was conducted by the patients' primary physicians. Weaning success was defined as unassisted breathing for >24 hrs. Predictive characteristics were computed using threshold values of 100 breaths/min/L, 10 L/min, and -20 cm H2O for the rate/volume ratios, minute volume, and negative inspired force, respectively. Receiver operating characteristic curves were also constructed to assess each parameter. Sensitivities for the initial rate/volume ratio, rate/volume ratio after 30 mins, minute volume, and negative inspired force were 0.97, 0.96, 0.76, and 0.96, respectively. Specificities were 0.33, 0.31, 0.40, and 0.07, respectively. Areas (+/- SD) for receiver operating characteristic curves were 0.76 +/- 0.08, 0.75 +/- 0.06, 0.54 +/- 0.08, and 0.62 +/- 0.07, respectively. The rate/volume ratio after 30 mins correlated with the initial rate/volume ratio; the rate/volume ratio after 30 mins did not add significant, additional predictive information.

CONCLUSIONS

The rate/volume ratio measured at the beginning and after 30 mins of weaning is more highly predictive of weaning outcome than the negative inspired force and minute volume. The principal weakness of the rate/volume ratio is false-positive results.

The unassisted respiratory rate-tidal volume ratio accurately predicts weaning outcome

PURPOSE

To assess the accuracies of four commonly used parameters in predicting weaning outcome and whether breathing pattern changes during weaning.

PATIENTS AND METHODS

We prospectively examined the predictive accuracies of four weaning parameters in mechanically ventilated patients in the medical and cardiac intensive care units of a 270-bed community teaching hospital. The spontaneous respiratory rate:tidal volume ratio (RVRi), negative inspiratory force (NIF), and spontaneous minute volume (VE) at the onset of weaning, and the RVR at 30 to 60 minutes of weaning (RVR30) were measured. Weaning decisions were made by patients' primary physicians independent of this study. Threshold values for computations of predictive values were as follows: RVR 100 < or = breaths per minute/L, NIF < or = -20 cm H2O, VE < or = 10 Lpm. Receiver operator curves were generated for each parameter.

RESULTS

One hundred medical/cardiac intensive care unit patients were studied. Their mean age was 64.6 +/- 15.8 years, mean APACHE II score of 15.8 +/- 6.7 and mean duration of mechanical ventilation before the study of 4.9 +/- 8.1 days. RVRi sensitivity was 89%, specificity was 41%, positive predictive value was 72%, negative predictive value was 68%, and accuracy was 71%. The RVR30 sensitivity was 98%, specificity was 59%, positive predictive value was 83%, negative predictive value was 94%, and accuracy was 85%. Accuracies for the NIF and VE were 66% and 62%, respectively. The area under the receiver operator curve of the RVR30 (0.92 +/- 0.03) was higher than the RVRi (0.74 +/- 0.05), NIF (0.68 +/- 0.06) and VE (0.54 +/- 0.06) (p < 0.05).

CONCLUSIONS

The RVR is more accurate than other commonly utilized clinical tools in predicting the outcome of weaning from mechanical ventilation. The RVR measured at 30 minutes is superior to the RVR in the first minute of weaning. The predictive accuracy and unique simplicity of the RVR justify its use in the care of mechanically ventilated patients.

Cardiac ischemia during weaning from mechanical ventilation

In this prospective study, we measured the ST segments, heart rate-systolic BP product (RPP), respiratory rate to tidal volume ratio (RVR), and pulse oximetry saturations of patients in our medical/cardiac ICUs before and during weaning from mechanical ventilation. Ninety-three patients were enrolled with a mean age of 66.5 +/- 15.0 years (mean +/- SD), mean acute physiology and chronic health evaluation (APACHE) II score of 16.0 +/- 6.9, and mean duration of mechanical ventilation of 5.2 +/- 8.6 days. Forty-nine patients had coronary artery disease (CAD). Six of 93 patients (6.4%) experienced ECG evidence of ischemia during weaning. Five of these six had a precedent history of CAD and four failed initial weaning attempts (22% of patients with CAD who failed weaning). The RPP, for the group as a whole, increased significantly during weaning from 12.0 +/- 3.1 to 13.4 +/- 4.0 mm Hg.bpm.10(3) (p<0.01). The rate to volume ratio did not change significantly during weaning, except in the subgroup of patients who failed to wean, in whom it increased from 98.4 +/- 45.2 to 124.9 +/- 54.9 bpm/L (p<0.05). Oxygenation also decreased significantly from 0.98 +/- 0.02 to 0.96 +/- 0.03 and was significantly associated with weaning failure (risk ratio [RR]=3.9; 95% confidence interval [CI]=1.7 to 9.0). Thirty-seven patients failed the initial weaning attempt. Cardiac ischemia (RR= 1.8; 95% CI=1.0 to 3.4) and an increased RVR (RR=1.7; 95% CI=0.9 to 3.4) tended to increase the risk of weaning failure. Cardiac ischemia, although infrequent (6%) in the general population of weaning medical/cardiac ICU patients, should be considered in patients with CAD who fail to wean.

Effect of cooling on oxygen consumption in febrile critically ill patients

Hyperthermic critically ill patients are commonly cooled to reduce their oxygen consumption (VO2). However, no previous studies in febrile humans have measured VO2 during cooling. We cooled 12 febrile, critically ill, mechanically ventilated patients while measuring VO2 and CO2 production (VCO2) by analysis of inspired and expired gases. All patients were mechanically ventilated for hypoxemic, hypercapneic, or shock-related respiratory failure and had a mean APACHE II score of 22.4 +/- 7.7. As temperature was reduced from 39.4 +/- 0.8 to 37.0 +/- 0.5 degrees C, VO2 decreased from 359.0 +/- 65.0 to 295.1 +/- 57.3 ml/min (p < 0.01) and VCO2 decreased from 303.6 +/- 43.6 to 243.5 +/- 37.3 ml/min (p < 0.01). The respiratory quotient (RQ) did not change significantly, and calculated energy expenditure decreased from 2,481 +/- 426 to 1,990 +/- 33 kcal/day (p < 0.01). In 7 patients with right heart catheters, cardiac output decreased from 8.4 +/- 3.2 to 6.5 +/- 1.8 L/min (p < 0.01) as the oxygen extraction fraction also tended to decrease from a mean of 28.2 +/- 6.8 to 23.4 +/- 4.7% (p = 0.12) during cooling. Accordingly, cooling the febrile patient unloads the cardiorespiratory system and, in situations of limited oxygen delivery or hypoxemic respiratory failure, may thus facilitate resuscitation and minimize the potential for hypoxic tissue injury.

Treatment of bronchospasm by metered-dose inhaler albuterol in mechanically ventilated patients

beta 2-agonist bronchodilators delivered by metered-dose inhalers (MDI) are commonly used in the treatment of bronchospasm in both intubated and nonintubated patients. Substantial data support the effectiveness of MDI delivery systems in nonintubated patients. However, few studies have examined the effectiveness of MDIs in intubated, mechanically ventilated patients. MDIs are often used in conjunction with a spacing device that may enhance delivery of drug to the airways, but few in vivo data have demonstrated efficacy of this delivery method in ventilated patients. We studied ten critically ill patients who had a peak (Ppeak) to pause (Ppause) gradient of more than 15 cm H2O during sedated, quiet breathing on assist control ventilation. We administered 5, 10, and 15 puffs (90 micrograms per puff) of MDI albuterol through a specific spacer (Aerovent) at 30-min intervals, while measuring resistive pressure (defined as Ppeak-Ppause) before and after treatments. Resistive airway pressure after 5 puffs decreased in nine of ten patients, from 25.1 +/- 7.2 to 20.8 +/- 5.6 cm H2O (p < 0.12). The addition of 10 more puffs further reduced resistive pressure in nine of nine patients from 20.8 +/- 5.6 to 19.0 +/- 4.4 (p < 0.01). Fifteen more puffs (30 cumulative puffs) did not result in further improvement (p > 0.5). A toxic reaction occurred in one patient (systolic blood pressure decreased 20 mm Hg) after 5 puffs of albuterol. We conclude that MDI administered through this specific spacer is effective in mechanically ventilated patients in doses up to 15 puffs, and that therapy should be titrated to effectiveness and toxicity.

Prolonged weakness after the withdrawal of atracurium

Prolonged weakness after the use of neuromuscular blocking agents is an increasingly recognized iatrogenic complication of critical care. Vecuronium and pancuronium have been associated with this syndrome, but few cases of atracurium-associated prolonged paresis have heretofore been reported. Many practitioners have thus chosen to use atracurium instead of longer-acting neuromuscular blockers when muscle relaxation is required in the supportive management of critically ill patients. We here report a case of prolonged paresis after the administration of atracurium.