Ultrasonographic Diagnostic Criterion for Severe Diaphragmatic Dysfunction After Cardiac Surgery

Original Research ULTRASONOGRAPHY

Ultrasonographic Diagnostic Criterion for Severe Diaphragmatic Dysfunction After Cardiac Surgery* Nicolas Lerolle, MD; Emmanuel Gueroi, MD; Saoussen Dimassi, MD; Rachid Zegdi, MD, PhD; Christophe Faisy, MD, PhD; Jean-Yves Fagan, MD, PhD; and Jean-Luc Diehl, MD

Background: Severe diaphragmatic dysfunction can prolong mechanical ventilation after cardiac surgery. An u1trasonographic criterion for diagnosing severe diaphragmatic dysfunction defined by a reference technique such as transdiaphragmatic pressure measurements has never been determined. Methods: Twenty-eight patients requiring mechanical ventilation > 7 days postoperatively were studied. Esophageal and gastric pressures were measured to calculate transdiaphragmatic pressure during maximal inspiratory effort and the Gilbert index, which evaluates the diaphragm contribution to respiratory pressure swings during quiet ventilation. Ultrasonography allowed measuring right and left hemidiaphragmatic excursions during maximal inspiratory effort. Best E is the greatest positive value from either hemidiaphragm. Twenty cardiac surgery patients with uncomplicated postoperative course were also evaluated with ultrasonography preoperatively and postoperatively. Measurements were performed in semirecumbent position. Results: Transdiaphragmatic pressure during maximal inspiratory effort was below normal value in 27 of the 28 patients receiving prolonged mechanical ventilation (median, 39 em H 20; interquartile range [IQR] 28 cm H 20). Eight patients had Gilbert indexes S 0 indicating severe diaphragmatic dysfunction. Best E was lower in patients with Gilbert index S 0 than > 0 (30 mm; IQR, 10 mm; vs 19 mm; IQR, 7 mm, respectively; p 0.001). Best E < 25 Dun had a positive likelihood ratio of 6.7 (95% confidence interval [CI], 2.4 to 19) and a negative likelihood ratio of 0 (95% CI, 0 to 1.1) for having a Gilbert index S O. None of the patients with uncomplicated course had Best E < 25 mm either preoperatively or postoperatively. Conclusions: U1trasonographic-based determination ofhemidiaphragm excursions in patients requiring prolonged mechanical ventilation after cardiac surgery may help identify those with and without severe diaphragmatic dysfunction as defined by the Gilbert index. (CHEST 2009; 135:401-407)

=

Key words: cardiac surgery; diagnostic technique; diaphragmatic paralysis; postoperative complication; transdiaphragmatic pressure; ultrasonography Abbreviations: Best E == greatest positive value from either hemidiaphragm on maximal inspiratory effort; CI == confidence interval; E-Ieft == left hemidiaphragm excursion on maximal inspiratory effort; E-right == right hemidiaphragm excursion on maximal inspiratory effort; IQR == interquartile range; ROC == receiver operating characteristic

postoperative cardiac patients are generally able to resume spontaneous breathing as soon as they recover from anesthesia. However, 3 to 6% of these

patients require prolonged mechanical ventilation.l-s Some degree of diaphragmatic dysfunction related to phrenic nerve injury frequently occurs following car-

'From Service de Reanimation Medlcale (Drs. Lerolle, Cuerot, Dimassi, Faisy, Fagon, and Diehl), and Departement de Chirurgie Cardiovasculaire (Dr. Zegdi), Hopital Europeen Georges Pompidou, Assistance Publiquc-Hopitaux de Paris, Universite Paris Descartes, Faculte de Medecine, Paris, France. The authors have no conflicts of interest to disclose. Manuscript received June 19, 2008; revision accepted August 4, 2008.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal, org/mtsc/repnnts.shtml). Correspondence to: Nicolas Lerolle, MD, Service de Reanimation Medscale, Hopital Europeen Georges Pompidou, 20, roe Leblanc, 75908 Paris Cedex 15, France; e-mail: [email protected] DOl: 10.1378/chest.08-1531

www.chestjournal.org

CHEST/135/2/ FEBRUARY, 2009

401

diac surgery. Prevalence of electromyographic abnormalities as high as 26% has been reported, but fortunately diaphragmatic dysfunction remains modest in most patients and does not translate into significant clinical impairment.3-fi On the opposite, the more severe forms of diaphragmatic dysfunction can be responsible for delayed weaning leading to prolonged stay in the leU.i Severe diaphragmatic dysfunction generally comes into question in cardiac postoperative patients requiring prolonged ventilation after cure or exclusion of usual causes of weaning failure such as pneumonia or volume overload. Diagnosing such severe dysfunction is crucial because it has been shown to carry a high risk of subsequent complications such as pneumonia, sudden respiratory arrest, and prolonged mechanical ventilation." Transdiaphragmatic pressure measurements allow functional evaluation of the diaphragm and has proved to be relevant in such patients to diagnose severe diaphragmatic dysfunction. i This technique, however, is semiinvasive and does not discriminate which side, or both, of the diaphragm is dysfunctional. Ultrasonography enables totally noninvasive visualization of the motion of each hemidiaphragm." However, the significance of the morphologic data obtained by ultrasonography is not known: no correlation has been made between ultrasonographic observations and functional data obtained by transdiaphragmatic measurements. The difficulty to extrapolate morphologic observation to functional evaluation is illustrated by the fact that 6% of asymptomatic subjects have a dyskinetic hernidiaphragm." This observation relates to the fact that one efficient hemidiaphragm is sufficient to generate adequate transdiaphragmatic pressure during quiet breathing. H.g Therefore, evaluating the motion of the better hemidiaphragm by ultrasonography may be more relevant to functional evaluation than detecting an immobile or dyskinetic hemidiaphragm. The aim of this study was to determine a quantitative ultrasonographic criterion of diaphragm motion for the diagnosis of severe diaphragmatic dysfunction as defined by transdiaphragmatic pressure measurements used as a reference method. Thus, we studied cardiac surgery patients with postoperative prolonged mechanical ventilation with both techniques. Additionally, the ultrasonographic criterion was assessed in cardiac surgery patients with uncomplicated postoperative course, both before and after surgery. MATERIALS AND METHODS This study was conducted prospectively in the medical ICU of Ht>pital Europeen Georges Pompidou, Paris, France. a tertiary

402

university hospital. From December 2004 to March 2006. 211 consecutive patients requiring prolonged uu-chanical ventilation > 7 days postoperatively after cardiac surgery underwent diaphragmatic ultrasonography when they met the criteria for a first weaning trial: (I) Glasgow coma scale score ~ 14: (2) hemodynamic stability without vasopressor use; (3) inspired fraction of oxygen < 50%. and positive end-expiratory pressure level 5: 5 em II zO ; and (4) no suspected or ongoing sepsis or uncompensated cardiac failure. As part of standard care in our ICU. such postcardiac surgery patients systematically undergo transdiaphragmatic pressure measurements,' and ultrasonography is performed on the same day. Two independent operators measured transdiaphragmatic pressure (J-L.D) and performed ultrasonography (N.Ll, each blinded to the results of the other. All measurements were obtained after 10 min breathing on Tvpiece. Patients were allowed to rest on mechanical ventilation for 2 h hetween each measurement. The order of measurements was randomly determined. Date of measurements could not he standardized in reference with surgery because criteria fulfillment depended on iudividual patient course and was screened by the primary physician independently of this study. Because treating physicians were not aware of the ultrasonography results. they obviouslv did not influence care. T;J further assess the predictive value of ultrasonography to diagnose or exclude severe diaphragmatic dysfunction, 20 consecutive cardiac surgery patients underwent diaphragmatic ultrasonography on the day before surgery, and on day 2 or day 3 postoperatively. All of these patients had uncomplicated immediate postoperative courses (successfully extubated in the 12 h following surgery) and were considered to represent "standard" cardiac surgery patients regarding mechanical ventilation weaning. Every patient was informed of the study, and informed consent was obtained. The study protocol was approved by the Ethics Committee of the French Society for Intensive Care Medicine.

Ultrasonography Doppler ultrasonography was performed with a 7.5-MHz transducer (Envisor: Philips Medical Systems: Bothell, \VA). All measurements were obtained with subjects in semi recumbent position (head 45° from bed). Longitudinal images parallel to body axis were obtained on left and right midaxiIIary lines to visualize the longest course of the corresponding hemidiaphragm (Fig 1). The excursion of the leading edge of the diaphragmatic echo was recorded where the diaphragm had the largest excursion. In the majority of postoperative cardiac surgery patients requiring prolonged mechanical ventilation. some amount of bilateral pleural effusiou and/or passive pulmonary atelectasis allowed easy visualization of the entire hemidiaphragm. In absence of pleural fluid or atelectasis, air-Hlled lung blocks ultrasound transmission, hut satisfactory observation of the hemidiaphragrns was possible through the liver or spleen. When the entire hemidiaphragm was not ohserved on the same ultrasonographic slice, we paid special attention to search several ultrasonographic windows for the greatest diaphragmatic excursion. Diaphragmatic excursion was measured from the end of normal expiration to end of maximal inspiratory effort (right hemidiaphragm excursion on maximal inspiratory eflort [E-rightJ; left hemidiaphragm excursion on maxima! inspiratory effort [E-leftJ; Fig I). Seven measurements were obtained, and the three best were retained and averaged for each side. Each value for excursion of the leading edge of the diaphragmatic echo value was defined as being positive when physiologic hemidiaphragmatic descent was ohserved during inspiration and negative when paradoxical hemidiaphragmatic ascent was observed during inspiration. The greatest positive value from either herrudiaphragm Original Research

FIGURE 1. Top left, A: The ultrasonography (US) probe was positioned on left or right midaxillary line to observe longitudinal images parallel to body axis. Top right, B: Schematic representation of the ultrasonographically observed structures. In the postoperative period, excellent ultrasonographic visualization of the entire length of the diaphragm is frequently permitted by presence of some amlJl~nt of pleural effusion and/or atelectasis. Bottom left, C, and bottom right, D: Ultrasonography 01 a hemidiaphragmatic assessment (right sided). Diaphragmatic excursion was measured from the end of normal expiration (bottom left, C) to end of maximal inspiratory effort (bottom right, D). At end-expiratory time, a mark is placed on the leading edge of the diaphragmatic echo. At end-inspiratory time, a second mark is placed, corresponding to the displacement of the first mark during inspiration. Distance between the two marks gives the hemidiaphragmatic excursion.

on maximal inspiratory effort (Best E) was the greatest positive value from either E-right or E-Ieft. In a prestudy, reproducibility of uItrasonographic measurements was assessed on 19 ICU patients. Two intensivists, each blinded to the results of the other, performed ultrasonography during two different spontaneous breathing trial sessions on the same day. Intraobserver and interobserver reproducibility (intraclass correlation coefficients and 9.5% confidence intervals [CIs]) were 0.99 (9,5% CI, 0.97 to 0.99) for E-Ieft and 0.88 (9.5% CI, 0.73 to 0.94) for E-right; and 0.99 (9.5% CI, 0.98 to 1.00) and 0.90 (9.5% CI, 0.79 to 0.9.5) for E-left; and 0.99 (9.5% CI, 0.98 to 1.00) and 0.92 (9.5% CI, 0.81 to 0.96), for Best E, respectfvely.!"

Transdiaphragmatic Pressure Measurements in Patients With Prolonged Mechanical Ventilation Transdiaphragmatic pressure measurements were performed as described elsewhere.'-u.1I In brief, esophageal and gastric www.chestjournal.org

pressures were measured with a double-balloon catheter (Marquat; Boissy Saint-Leger, France) connected to two differential pressure transducers (MP4.5; Validyne; Northridge, CAl. All measurements were obtained with subjects in semirecumbent position (head 4.5° from bed). Signals were sampled and digitized at 128 Hz, and data were entered into a computer, allowing the calculation of the following: (I) transdiaphragmatic pressure during a maximal static inspiratory effort, against an airway occluded at end-expiration (2.5-s occlusion); three attempts were done, and the highest value was retained; (2) the ratio of transdiaphragmatic pressure change during normal inspiration to the difference between peak inspiratory gastric pressure and end-expiratory gastric pressure during normal inspiration, also known as the Gilbert index, which evaluates the contribution of the diaphragm to respiratory pressure swings during quiet tidal breathing.!" The Gilbert index is normally> 0.30, with a value ::5 0 indicating a very severe diaphragmatic dysfunction.' 12 CHEST / 135 /2/ FEBRUARY, 2009

403

Clinical Data In patients requiring prolonged mechanical ventilation; age; Simplified Acute Physiology Score-If at ICU admission!"; type of cardiac surgery; time interval between surgery and measurements; outcome in ICU; total duration of mechanical ventilation; number of extubation failures (reintubation during the 48 h following extubation); number of ventilator-associated pneumonia episodes; need for tracheotomy; and duration of ICU stay were recorded. In cardiac surgery patients with an uncomplicated postoperative outcome, age, sex, and type of cardiac surgery were recorded.

Statistical Analysis Data are presented as median (interquartile range [IQR]) or number. Statistical analysis was performed using statistical software (Statview ,5.0; SAS Institute; Cary, NC; and Stata Software; StataCorp LP; College Station, TX). The Kruskal-Wallis test, Mann-Whitney U test, or Fisher exact test, as appropriate, were applied for between-group comparisons. The relationship between continuous variables was evaluated using Spearman correlation coefficient. We constructed receiver operating characteristic (ROC) curves to evaluate the diagnostic value of Best E for predicting severe diaphragmatic dysfunction (defined by a Gilbert index oS 0). The area under the ROC curve and its 95% CI were estimated. We chose the cutoff point giving the highest accuracy for the diagnosis of severe diaphragmatic dysfunction; p < 0.05 was considered significant.

RESULTS

Patients With Prolonged Mechanical Ventilation Twenty-eight patients receiving prolonged postoperative mechanical ventilation after cardiac surgery were studied. Median age was 70 years (IQR, 16 years), and 18 patients were male (mean Simplified Acute Physiology Score-II at hospital admission, 45 [IQR, 21]). Type of surgery was coronary artery bypass grafting in 11 patients, valve replacement in 9 patients, coronary artery bypass grafting plus valve replacement in 4 patients, and other procedures in 4 patients; 5 patients underwent reoperation. None had topical cardiac cooling with ice-cold solution in the pericardium. Normothermic cardioplegia had been applied in all patients. Prolonged ventilation

after surgery was attributed to pneumonia or other infections and/or cardiac failure. A median of 11 days (IQR, 16 days) elapsed between surgery and the day of diaphragmatic evaluation. Diaphragmatic testing was performed at distance (minimum 5 days) of administration of any paralyzing agent or sedative. All patients were awake and able to move their legs and arms on demand with strength evaluated ~ 4/5 on the day of diaphragmatic assessment, but no formal testing for ICU neuromyopathy was performed. Transdiaphragmatic pressure measurements displayed an abnormal maximal static inspiratory effort for all but one patient (39 em H 20; IQR, 28, em H 20; normal> 80 em H 20)./; The median Gilbert index was 0.16 (IQR, 0.44). This index was normal (> 0.30), excluding Significant diaphragmatic dysfunction, for nine patients; low positive « 0.30 but > 0), indicating moderate diaphragmatic dysfunction, for 11 patients; and very low (:5 0), indicating severe diaphragmatic dysfunction, for 8 patients. Of note, patients with such severe diaphragmatic dysfunction had Significantly worse outcomes than the other patients (Table 1), no difference for redo surgery was noted (n = 1 vs n = 4, respectively; p> 0.9). Ultrasonographic examinations of both hemidiaphragms were possible in all patients. Nine patients had paradoxical inspiratory motion (dyskinesia) or immobility of a hemidiaphragm (akinesia). Akinesia! dyskinesia was left sided in six patients and right sided in three patients. No patient had bilateral akinesia/dyskinesia, Akinesia/dyskinesia was encountered both in patients with Gilbert index :5 0 and > 0 (n = 5 and n = 4, respectively). Nineteen patients had physiologic inspiratory descent of both hemidiaphragm. Ultrasonographic parameters are presented in Table 2. Best E correlated Significantly with Gilbert index (p = 0.64, P = 0.001; Fig 2). Indeed, Best E was Significantly different in patients with Gilbert index :5 0 and > 0 (Table 2, Fig 3). ROC curve for Best E as a criterion for having a

Table l--Outcome of Patients Receiving Prolonged Mechanical Ventilation* Variables Duration of mechanical ventilation before the day of diaphragmatic evaluation, d Duration of mechanical ventilation after the day of diaphragmatic evaluation. d Extubation failures after the day of diaphragmatic: evaluation. No. of episodes Need for tracheotomy, No. of patients] Ventilator-acquired pneumonia, No. of patientst Length of ICU stay, dt ICU mortality, No.

Gilbert Index> O. n = 20

10 (16)

Gilbert Index n=8

oS

O.

p Value

13 (18)

0.83

1.5 (4)

28 (40)

0.5 (I) I 8

1.5 (2)

0.0004 0.02 0.004 0.01 0.01 >0.99

5 8

29 (27)

52 (83)

3

I

*Data are presented as median (IQR) unless otherwise indicated. tIndependently of the date of diaphragmatic evaluation. 404

Original Research

Table 2-Ultrasonographic Obseroations in Cardiac Surgery Patients With Uncomplicated Postoperative Course Evaluated Before and After Cardiac Surgery and in Patients Receiving Prolonged Mechanical Ventilation* Uncomplicated Postoperative Course, n = 20 I

Variables

Prolonged Mechanical Ventilation, n I

I

Gilbert Index < 0, n

= 20

=

28

Gilbert Index ::s 0, n

Before Surgery

After Surgery

0

3t

4

5§

47 (15) 50 (7) 51 (12)

27 (16H 32 (28H 37 (ll)t

19 (18)' 30 (10)# 30 (10)**

o (22)§ 15 (lOll 19 (7)11

Akinetic/dyskinetic hemidiaphragm, No. of patients E-Ieft, mm E-right, mm Best E, mm

I

=8

*Data are presented as median (IQR) unless otherwise indicated. Kruskal-Wallis test showed significant difference between the four groups for E-Ieft, E-right, and Best E (p < 0.001 for all parameters). tp = 0.07 for comparisons before vs after-surgery variables in patients with uncomplicated course. [p < 0.0001 for comparisons before vs after-surgery variables in patients with uncomplicated course. §p = 0.07 for comparisons of patients with Gilbert index> 0 and ::s o. lip = 0.001 for comparisons of patients with Gilbert index> 0 and ::s o. '1P = 0.03 for comparisons of after-surgery variables in patients with uncomplicated course vs patients with Gilbert index> O. #p = 0.5 for comparisons of after-surgery variables in patients with uncomplicated course vs patients with Gilbert index> O. **p = 0.007 for comparisons of after-surgery variables in patients with uncomplicated course vs patients with Gilbert index> O.

Gilbert index :s 0 had an area under curve of 0.93 (95% CI, 0.77 to 0.99). The cutoff point giving the highest accuracy for predicting Gilbert index :s 0 was Best E < 25 mm. This Best E value had a sensitivity of 100% (95% CI, 63 to 100%), a specificity of 85% (95% CI, 62 to 97%), a positive likelihood ratio of 6.7 (95% CI, 2.4 to 19), and a negative likelihood ratio of 0 (95% CI, 0 to 1.1) for having a Gilbert index :s O.

Patients With Uncomplicated Postoperative Course Twenty cardiac surgery patients were evaluated preoperatively with ultrasonography. All of them were weaned from the ventilator in the 12 h following surgery and were then examined on days 2 or 3 after surgery. Median age was 64 years (IQR, 20

DISCUSSION

55 50

rho

45 E E ui Ui QI m

•

=0.64, P =0.001 •

•

40

35 30 25 20 15 10 5

years), and 10 patients were male. Surgery was coronary artery bypass grafting alone in 10 patients, plus valve replacement or repair in 3 patients, valve replacement or repair alone in 5 patients, and other procedures in 2 patients. One patient had redo surgery, and cardioplegia was normothermic in all patients. Best E < 25 mm was not encountered in any patient either before or after surgery (Fig 3). A significant decrease in all of the ultrasonographic excursion parameters was observed after cardiac surgery (Table 2). Three patients had akinesia/dyskinesia of a hemidiaphragm postoperatively. E-Ieft and Best E were significantly higher than in patients requiring prolonged ventilation and with Gilbert index> O.

In this study, we observed that ultrasonographically determined Best E < 25 mm was associated with severe diaphragmatic dysfunction as defined by a Gilbert index :s 0 measured during transdiaphragmatic pressure measurements. This ultrasonographic threshold had an excellent negative likelihood ratio, which was confirmed by studying patients with uncomplicated postoperative course, none of them having Best E < 25 mm, either before or after

•

I· • i

1·

• -.8

-.6

-.4

-.2

0

surgery.

.2

.4

.6

.8

Gilbert Index

2. Craphic representation of the relationship between Best E and Gilbert index in the 28 postoperative patients receiving prolonged mechanical ventilation.

FIGURE

www.chestjournal.org

The 28 patients with prolonged mechanical ventilation reflect not the usual course of patients after cardiac surgery, but a subgroup of patients with an expected high frequency of severe diaphragmatic dysfunction. Indeed, in eight of these patients, transdiaphragmatic pressure measurements diagnosed seCHEST /135/2/ FEBRUARY, 2009

405

•

55

•

50 45

• •• •••• • --..•••

40

E E

-

u.i 1/1 Q)

lQ

35 30 25 20 15 10 5

••

•• • ••

I.

...

•• ••

--• •

.

•• ••• •

•• •

• Gilbert Index ~o

Gilbert Index

>0

Patients requiring prolonged mechanical ventilation

Patients with uncomplicated course: postoperative assessment

FIGURE .3. Graphic representation showing Best E in patients with prolonged mechanical ventilation. n = 28, and postoperative Best E in patients with uncomplicated course, n = 20. Horizontal hars indicate median values.

vere diaphragmatic dysfunction. It is important to underline that assessment of diaphragmatic function in patients with prolonged ventilation is not intended to predict success or failure of a first T-piece trial after cardiac surgery, but to find an explanation for prolonged mechanical ventilation after cardiac surgery and to invite attention to patients at high risk of further respiratory complications.' Indeed, we observed a very good association between clinical outcome parameters and the results of our reference method, transdiaphragmatic pressure measurements. Although Gilbert index was not measured in the 20 patients with uncomplicated course after surgery, absence of any weaning difficulty indicates that absence of severe diaphragmatic dysfunction can be assumed. Several studies have previously described the use of ultrasonographically observed hemidiaphragm motion to diagnose diaphragmatic dysfunction after cardiac surgery.4.14-I6 In these studies, ultrasonography was not compared to any evaluation of diaphragmatic function or with clinical outcome or, if com406

pared to such parameters, no severe diaphragmatic dysfunction was observed, limiting any conclusion on the utility of ultrasonography in the critical care setting. Gottesman and Mcf.ool!" studied ultrasonographically visualized thickening of the diaphragm, which correlates with its contraction. Nevertheless, ultrasonographic assessment of diaphragm thickening requires perfect visualization of the different diaphragmatic layers throughout the respiratory cycle, which is almost impossible in ICU patients undergoing routine evaluation. Considering the excellent negative likelihood ratio of Best E < 2.5 mrn, ultrasonography may be considered to exclude severe diaphragmatic dysfunction following cardiac surgery in daily practice with the advantages of being fully noninvasive and widely available in ICU. The good agreement between observers confirms the relevance of ultrasonography for daily practice. On the other hand, the dependency of Best E on maximal voluntary inspiratory effort in these patients limits the interpretation of a Best E value < 25 mm even if the positive likelihood ratio of ultrasonography was substantial. In a twostep strategy, Gilbert index determination, being independent of the patient effort, could therefore be used in patients with Best E < 25 mm for confirming severe diaphragmatic dysfunction with certainty. Positive predictive value of ultrasonography would merit to be refined on a larger validation cohort. Unfortunately, conducting such study is impeded by the infrequency of severe diaphragmatic dysfunction after cardiac surgery. Our results are restricted to patients after cardiac surgery with supposed surgeryrelated phrenic nerve injury, and cannot be extrapolated to other form of diaphragmatic dysfunction although ICU neuromyopathy may also have participated in the pathogenesis of diaphragmatic failure in our patients. In conclusion, ultrasonography-based determination of hemidiaphragmatic excursion in patients requiring prolonged mechanical ventilation after cardiac surgery is a useful technique for ruling out severe diaphragmatic dysfunction. REFERENCES Engoren M, Buderer NF, Zacharias A. Long-term survival and health status after prolonged mechanical ventilation after cardiac surgery. Crit Care Med 2000; 2R:2742-2749 2 LoCicero J III, McCann B, Massad M, et al. Prolonged ventilatory support after open-heart surgery. Crit Care Med 1992; 20:990-992 :3 Ahd AG, Braun NM, Baskin MI, et al. Diaphragmatic dysfunction after open heart surgery: treatment with a rocking bed. Ann Intern Med 19R9; 1I 1:R81-886 4 DeVita MA, Rohinson LR, Rehder J. et al. Incidence and natural history of phrenic neuropathy occurring during open heart surgery. Chest 1993; 103:R.50-8.56 .5 Markand ON, Moorthy 55, Mahomed Y, et al. Postoperative Original Research

6 7 8 9 10 II 12

phrenic nerve palsy in patients with open-heart surgery. Ann Thorac Surg 1985; 39:68-73 Wilcox P, Baile EM, Hards J, et al. Phrenic nerve function and its relationship to atelectasis after coronary artery bypass surgery. Chest 1988; 93:693-698 Diehl J-L, Lofaso F, Deleuze P, et al. Clinically relevant diaphragmatic dysfunction after cardiac operations. J Thorac Cardiovasc Surg 1994; 107:487-498 Tobin M, Laghi F. Monitoring of respiratory muscle function. In: Tobin M, ed. Principles and practice of intensive care monitoring. New York, NY: MacGraw-Hill, 1998; 497-544 Scillia P, Cappello M, De Troyer A. Determinants of diaphragm motion in unilateral diaphragmatic paralysis. J Appl Physio12004; 96:96-100 Shrout PE, Fleiss JL. Intraclass correlation: uses in assessing rater reliability. Psychol Bull 1979; 86:420-428 Baydur A, Behrakis PK, Zin WA, et al. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982; 126:788-791 Gilbert R, Auchincloss JH Jr, Peppi D. Relationship of rib

www.chestjournal.org

13

14 15 16

17

cage and abdomen motion to diaphragm function during quiet breathing. Chest 1981; 80:607-612 Le Gall J-R, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European! North American multicenter study. JAMA 1993; 270:29572963 Fedullo AJ, Lerner RM, Gibson J, et al. Sonographic measurement of diaphragmatic motion after coronary artery bypass surgery. Chest 1992; 102:1683-1686 Manabe T, Ohtsuka M, Usuda Y, et al. Ultrasonography and lung mechanics can diagnose diaphragmatic paralysis quickly. Asian Cardiovasc Thorac Ann 2003; 11:289-292 Merino-Ramirez MA, Juan G, Ramon M, et al. Electrophysiologic evaluation of phrenic nerve and diaphragm function after coronary bypass surgery: prospective study of diabetes and other risk factors. J Thorac Cardiovasc Surg 2006; 132:530-536 Gottesman E, McCool FD. Ultrasound evaluation of the paralyzed diaphragm. Am J Respir Crit Care Med 1997; 155:1570-1574

CHEST /135/2/ FEBRUARY, 2009

407