Indicators for Tracheostomy in Patients with Traumatic ... Clubs/2007-03/Tracheostomy in... · Indicators for Tracheostomy in Patients with Traumatic Brain Injury MYSTAN A. GURKtN,

Indicators for Tracheostomy in Patients withTraumatic Brain InjuryMYSTAN A. GURKtN, M,D., MANESH PARIKSHAK, M.D., KURT A, KRALOVICH, M,D., H. MATHILDA HORST, M.D.,VtKAS AGARWAL, B,S., NICOLE PAYNE, B.S.

From the Division of Trauma ami Critical Care, Henry Ford Hospital, Detroit, Michigan

Our objective was to develop criteria to identify patients with traumatic brain injury (TBI) whorequire a tracheostomy (TR), From January 1994 to May 2000 all TBI patients requiring intubationon presentation and who survived >7 days were identified from our trauma registry. Demograph-ics, Glasgow Coma Score (GCS), Injury Severity Score (ISS), and ventilator days, ICU days,hospital days, need for TR, and development of pneumonia were statistically analyzed. Of 246patients with TBI 211 without TR and 35 with TR were identified (mean time to TR 13,3 ± 7.0days). Logistic regression analysis identified presenting GCS <8, ISS >25, and ventilator days >7as significant predictors for TR. Applying these three predictors to our population Identified 48patients (21 with TR, 18 without TR, and nine who died on the ventilator without TR) with asensitivity of 60 per cent, a specificity of 87 per cent, a positive predictive value of 44 per cent, anda negative predictive value of 93 per cent. Patients with TR had lower presenting GCS and higherventilator, ICU, and hospital days (P < 0.05). Pneumonia rates were similar. Time to neurologicrecovery (GCS >9) was longer for the TR patients as compared with the patients without TR. Weconclude that patients with TBI presenting with a GCS <8, an ISS >25, and ventilator days >7 aremore likely to require TR, Performing TR late did not reduce pneumonia rates or ventilator, ICU,or hospital days. By identifying the at-risk population early TR could be performed in an attemptto decrease morbidity and length of stay.

C ONTROVHRSY EXISTS ovHR the indications and tim-ing of tracheostomy (TR) in patients with trau-

matic brain injury (TBI). To avoid secondary braininjury from hypoxia and to provide a secure airwaymost patients with Glasgow Coma Score (GCS) ^ 8are intubated atid ventilated in the trauma bay duringtheir evaluation and resuscitation. In addition to theseverity of the btain injury associated injuries, opera-tions, and subsequent pulmonary complications deter-mine the length of time these patient.s with TBI remainintubated and ventilated. Some of these patients re-quire only a secure airway whereas other patients needprolonged ventilatory support. Early and late TR areboth recommended in the literature.' '^ The time rangerecommended lor conversion to TR is from 3 days to21 days-"̂ - ^ Broad general guidelines from a consensusconference on artificial airways recommends the useof translaryngeal tubes for airways up to 10 days and

Presented at the 44th Annual Meetinjz. Midwest Surgical Asso-ciation, Lake Cieneva. Wisconsin, Aujzust 12-13, 2001.

Address correspondence and repriiu rctiuests to H. MathildaHorsi. M.D.. Division of Trauma and Critical Care. Clara FordPavilion, 4lh Moor. Henry Ford Hospital, 2799 West Grand Bou-levard. Detroit, MI 48202.

if the anticipated need for artificial airway will begreater than 21 days TR is preferred.'' Advocates ofearly TR (within one and 7 days) claim the advantageof lower incidence of pneumonia and shorter durationof mechanical ventilation, intensive care unit (ICU)stay, and hospital stay.'"' TR provides a .secure airwayand facilitates airway suetioning, mouth care, and pa-tient mobility. Furthermore TR decreases the damageto the vocal cords seen with endotracheal tubes.̂ -" -̂̂

Advoeates of late TR claim the avoidance of unnec-es.sary operations in patients with uncertain progno-sis.^ Given time many patients will be weaned andextnbated. whereas other patients may die early fromthe TBI or other associated injuries. Complications ofTR including pneumothotax, bleeding, and trachealstenosis are also cited as reasons to delay conversionto TR.̂

The development of percutaneous TR techniques,which converts a procedure usually done in the oper-ating room to a bedside piocedute. has renewed inves-tigation into the timing of TR conversion. Because thedemand for percutaneotis TR has increased in our in-stitution we wanted to examine our current practice forairway conversion in our patients with TBI with theintent to develop criteria that would identify those pa-tients who will most likely require TR.

324

No. 4 TRACHEOSTOMY IN TRAUMATIC BRAIN INJURY Gurkin et al. 325

Methods

The trauma registry of an urban Level I hospital wasretrospectively queried for all neurologically injuredpatients from January 1994 through May 2000. Of thispopulation only patients with TBI requiring intubationand ventiiatioti upoti presentation and surviving tnorethan 7 days were selected for our study. Patients werealso excluded if they sustained isolated spine, face, orsoft-tissue neck injuries, weic incon'eclly classified ashaving neurologic injuries, were transferred to our in-stitution with a TR in place, or wore unable to becompletely evaluated becatise of unavailable tnedicalrecords. For each patient, we recorded demogiaphicitiforrnation; mechanism of injury: presenting GCS:Injury Severity Score (ISS): ventilator, ICU. and hos-pital days; discharge disposition; need for TR; day oftracheostomy: development of pneumonia; and num-ber and type of operations. Beyond the presentingGCS a daily modified GCS (m-GCS) was calculatedfor each patient using the regression model determinedby Meredith et al.'̂ which calculates a modified totalGCS from motor and eye scores for intubated (non-verbal) patients:

m-GCS - Motor + Eye + [-0.3756+ (Motor X 0.5713) + (Eye x 0.4233)j

This score was tracked daily until extubation, TR, oi-14 days of mechanical ventilation. Student's t test, chisquare test, Wilcoxon two-sample test. Fisher's exacttest, and logistic regression were used to cotnpate thepatients who underwent TR with those who did not. AP value less than 0.05 was considered sigtiificant.

Results

From January 1994 through May 2000 96! neuro-logically injured patients were identified from thetrauma database. Of these 961 patients 246 (26%) metthe study criteria. Excluded from the study were 285patients (30%) who did not require ventilatory sup-port; 129 patients (13^/r) who died early; and 301 pa-tients (31%) who did not have TBI, who presented toour hospital with a TR in place, or whose medicallecords were unavailable lor review (Eig. 1). The 246patients had tnultiple injuries with a tnean ISS of 22.1± 9.9 and a presenting GCS of 8.7 ± 4.3. There were199 males and 47 females with an average age of 34.4± 20.9 years. Of the 246 patients 126 (51.2%) requiredoperations. There were 84 craniotomies. 21 face andneck operations, one thotacic operation, five abdomi-nal operations, and 33 extremity opet-ations. TR wasperformed on 35 of the 246 patients (14%). Twenty-two (62.8%) of the patients with TR tequired opera-tions (13 craniototTiies. two face surgeries, three ab-dominal surgeries, and eight extremity surgeries). Thetiming ol' conversion to TR varied and ranged fromDay one to Day 30. The mean time to TR was 13.3 ±7.0 days and the mean time from TR to removal fromventilatory support was 6.9 ± 8.4 days.

Data from the 35 patients who required TR werecotnpared with data from the 211 patients without TR(Table 1). No study patient was removed from venti-latory support on Day I. By Day 7 149 of 211 patientswithout TR (71%) weie extubated. Between Day 1and Day 7, seven patients with TR (20% of TRs)were converted to TR. The remaining 28 patients were

961 patients withneurologic injury from

trauma database

285 patients did notrequire ventiiatory

support

t29 patients died < 7davs

301 patients requiringventilatory support and

surviving > 7 daysexcluded:

246 patients requiringventilalory support and

surviving > 7 daysincluded in study

82 patients withoutneurologic injury

88 patienls withisolated spine trauma

41 patients withisolated face trauma

33 patients withisolated soft tissue

injury to neck

9 patients transferredto our institution with atracheostomy in place

48 patients whosemedical records wereunavailable for review

l-ici. 1. Distribution ot" 961 pa-tienls evaluated cil" study populalicm.

326 THE AMERICAN SURGEON April 2002 Vol. 68

TABLE L Demographic, Presenting, and Outcome Data for Study Populaiion^

With TR, n = 35 Without TR, n = 211

AgePresenting GCSISSMale:FemaleVentilator daysICU daysHospital daysPneumonia rale

37.4 ± 18.56.5 ± 4.0

30.5 ± 8.626 (74%):9 (26%)

20.] ±9.823.5 ± 9.337.2+ 17.432 (91.4%)

33.9 ±21.39.7 ± 4.3

20.8 ± 9.5173(82%):38(I8%)

5.4 ±5.78.5 ±7.1

14.9 ±9.666(31.3%)

P Value

0.370.0010.0030.37

<O.OOI

<().()() I

* Data are presented as means ± standard deviation or as number (%).

converted to TR after Day 7. The patients with TRwere more severely injured {P ^ 0.003) and had alower GCS on presentation {P ^ O.OOI). Patients withTR also had significantly longer mechanical ventila-tion times, more ICU days, and longer hospital stays{P < 0.001). There was a higher incidence of pneumo-nia in the patients with TR [P < 0.001) with 30 ofthe 35 patients (867f) diagnosed with pneumoniabefore TR.

Logistic regression analysis was used to identifyfactors that would predict the progression to TR. Thisanalysis identified presenting GCS <8, ISS ^25 . andventilator days greater than 7 as predietors of TR(Table 2). Applying these three criteria to the 246patients identified 48 patients including 21 of the 35patients with TR (60%) and 27 patients without TR(Table 3). Of the 27 patients without TR 18 patients(67%) were successfully weaned from mechanicalventilation and extubated whereas nine patients (33%)died on the ventilator without TR. For the 21 patientswith TR the mean time to TR was 13.7 ± 6.4 days witha mean time from TR to removal of ventilatory supportof 7.8 ± 10.2 days. A sensitivity of 60 per cent, aspecificity of 87 per cent, a positive predictive value of44 per cent, and a negative predictive value of 93 percent were calculated for these three predictors of TR.The demographic, presenting, and outcome data forthe 48 patients identified by logistic regression eriteriaare presented in Table 3. The data of the nine patientswho died on the ventilator without TR are presentedfor comparison but are not included in the statisticalanalysis. When patients wilh TR were compared withpatients successfully weaned without TR ISS andpneumonia rates were not different. However, venti-

TABLE 2. Logistic Re}>ression Analysi.s Results jbrPredictors of TR

Odds 95% ConfidenceClinieal Predictor Ratio Interval P Value

Presenting GCS <8 34 1.5-7.9 0.004ISS >25' ' 7.2 2.9-18.1 <0.00lVentilator days >7 21.45 7.3-63.5 <O.OOI

lator days {P < 0.001). ICU days {P = 0.02). andlength of hospital stay {P = 0.03) were significantlylonger and presenting GCS was significantly loweriP = 0.006) in patients with TR.

To fuither examine these 48 patients identified bylogistic regression criteria the daily m-GCS scores foreach group were compared (Table 4). The patientswithout TR on average reached neurologic recovery(GCS >9) by Day 7 as compared with the patientswith TR who had neurologic recovery at day 14. Therewas a significant difference in the m-GCS on Days 10to 14 for the.se two groups. The patients who died onthe ventilator without TR showed a steady decline intheir m-GCS over Days one to 14.

Calculating the average change in m-GCS from Dayone to the day of extubation. TR, or death identifiedthree distinct patterns in these 48 patients. For thepatients undergoing TR the average m-GCS on the dayof TR was 9.0 ± 2.5 with an average change in m-GCSfrom Day one to day of TR of +2.2 ± 4.0. In contrastthe patients without TR had a significantly higher av-erage m-GCS on the day of extubation of 12.8 ± 2.2with a significantly higher average change in m-GCSfrom Day one to extubation of-1-4.7 ± 2.3. The patientswho died on the ventilator without TR had an averagem-GCS on the day of death of 4.1 ± 3.2 with anaverage change in in-GCS from Day one to death of-3.3 ± 3.7 (Fig. 2).

Discussion

The incidence of conversion to TR was 14 per centin this study of patients with TBI and is similar to theincidence of 14 to 18 per cent reported by Gordon etal.'" in a series of 1132 head injury patients. Mostsurgeons would convert to TR when there is a need forprolonged mechanical ventilation (greater than 21days), a high cervical spinal cord injury, or extensivefacial trauma. In fact five of our seven early TRs (<7days) were performed for concurrent facial or cervicalspine injury.

The concept of early TR in brain-injuted traumapatients remains controversial. Two retrospective

No. 4 TRACHEOSTOMY IN TRAUMATIC BRAIN INJURY Gurkin et at. 327

TABLL 3. Dcin»i;rapliir. Presenting, and Outcome Data for 4H Patient.s with Presenting GCS <A' and ISS ^25 andRequiring Ventilator Support Greater than 7 Days

AgePresenting GCSISSMale:FemaIeVentilator davsICU daysHospital daysPneumonia rate

With TRn = 21

33.1 ± 16.04.4 ± 1.4

34.3 ± 7.216 (76%):5 (24%)

21.4 ±9.924.6 ± 9.540.2 ± 19.020 (95.2%)

Without TRn = 18

33.1 ±23.15.8 ±1.5

31.4 ±8.214 (78%):4 (22%)

12.9 ±2.518.4 ±7.629.1 ±7.814 (77.8%)

No TR/No EX*n = 9

42.1-^23.04.7-t- 1.9

28.2 -H 4.47 (78%):2 (22%)

15.9-1- 10.415.9-t- 10.415.9 ± 10.45 (55.6%)

P Valuet

0.75

0.080.99

<0.0010.020.030.16

Data are presented as mean ± standard deviations or as number (%).* Patients who died on the ventilaior wiihoiit TR.t P values are comparing the paiicnt.s with TR with those without TR.

TABLE 4. m-GCS Day One to 14 for Population withPresenting GCS ^H and ISS ^25 and Rcqiiiring VentilatorSupport Greater than 7 Days

WithTRn = 21

Without TRn = 18

No TR/Notl = 9

PValuev

Day 1Day 2Day 3Day 4Day 5Day 6Day 7Day 8Day 9Day 10Day 11Day 12Day 13Day 14

6.84 ±3.407.36 ±3.157.07 ±2.447.27 ±2.947.51 ±2.717.83 ± 2.747.16±3.IO7.80 ±2.868.15±3.1()8.47 ± 2.338.73 ±3.228.48 ±3.018.74 ±3.02

10.02 ±2.30

8.07 ±2.338.09 ±2.498.59 ±1.318.37 ±2.308.89 ±3.288.88 ±3.189.24 ±3 . II9.65 ± 3.66

10.43 ±4.0111.74 ±3.2711.96 ±2.6812.17±3.2313.37 ±1.0712.68 ±1.93

7.44 ±3.448.12 ±3.327.93 ±3.957.58 ±4.095.38 ±2.764.86 ±3.264.36 ±2.414.I9±2.4S3.14± 1.573.93 ± 2.094.58 ± 2.363.80 ±2.363.80 ±2.364.19 ±2.72

0.2040.430

0.2260.1630.2910.0600.1090.0880.{X>4

0.010

0.025

Data are presented as mean ± standard deviations.* Patient.s who died on the ventilator without TR.t P values are comparing the patients with TR with those

without TR.

studies and one prospective analysis suggest thai pneu-monia rate, mechanical ventilator support. ICU days,and hospital days arc less wilh early TR. However, theTR must be petformed by Day 7 for the patient toexperience these beticfits. Lesnik et al.^ performedearly TR (<4 days) in nuillitrauma patients with sig-nificant reduction in ventilator days and pneumoniarate.s. All of the patients undergoing early TR wereliberated from the ventilator by 7 days. D'Amelio etal.' found reductions in ventilator. ICU. and hospitaldays for head-injured patients undergoing TR ^ 7days. Siinilar results wore achieved in a prospectivestudy of trauma patients by Rodriguez et al.^ In ourseries with the time to TR conversion at 13.3 ± 7.0days these potential benefits were not realized. Ourpatients did not have reduced pneumonia rates or ven-tilator. ICU. or hospital days. Our findings ate similarto those of a series of patients with inlVaientorial le-sions reported by Qureshi et al." who found that TR

• W ithout TR

U Wilh TR

^•'No TR/No Ex

Mean m-GCS Change inni-G(S

FIG. 2. Mean ni-GCS on day of exiiibiilion. TR, or death andchange in m-GCS from Day one to day of cxtubation. TR. or dcalli.

at Day 10 resulted in more ventilator. ICU. and hos-pital days.

To realize the benefits of early TR one must identifythose patients who will most likely progress to TR.The GCS serves as a guide to the severity of braininjury, correlates to outcotne. and has been used as anindicator of the need for TR.'"- '̂ - '̂ Patients with aGCvS from 13 to 15 have a minor brain injury, goodneurologic recovery, and a low mortality rate.'-^ Intu-bation with mechanical ventilation is not tequired fortheii- btain injury but may be used because of associ-ated injuries. Patients with a GCS ^8 have severebrain injury and a mortality rate of 24 per cent. Thesepatients require intubation and ventilation lor theirbrain injury. However, even in this patient populationthere is a 55 per cent gt)od neurologic recovery tate. a14.5 per cent severe disability rate, and a 6.5 per centvegetative rate.'" Koh et al.'"^ found that a GCS <8 atDay 7 in head-injury patients was a good indicator ofprogtession to TR. There was a reintubation rate of 22per cent in patients with GCS <8, which was thoughtto be secondary to airway laxity and tracheal secre-tions. In sutgical intensive care patients Johnson etal.'-^ coticluded that a GCS ^ 9 at Day 2 was a pre-dictor of mechanical ventilation greater than 14 dayswith a positive predictive value of 91 per cent and anegative predictive value of 96 per cent. These results

328 THE AMERICAN SURGEON April 2002 Vol. 68

correspond with our findings that the patients pro-gressing to extnbation showed neurologic recoveryearlier (Day 7 versus Day 14) than those requiring TR.

Our time to conversion to TR (13.3 days) is veryconservative and is clearly considered to be in thecategory of late TR. Our bias has been to give patientsa chance for extubation and if extnbation is unsuccess-ful a TR is performed. Based on the literature and ourdata it appears that the patients with GCS ^ 8 and ISS>:25 may benefit from early TR. Examination of thepatients presented in Table 4 shows that if we per-formed TR at 48 hours on patients with GCS ^ 8 somepatients who had later deterioration of brain functionwould have received TR. However, if the decision toconvert to TR was delayed to Day .̂ i, 6. or 7 a cleardelineation of patients with deteriorating brain func-tion versu.s improving neurologic status would havebeen made.

Conclusions

Patients with TBI presenting with a GCS <8. anISS ^25, and need for ventilator suppott for longerthan 7 days are more likely to require a TR. Lack ofneurologic recovery by Day 7 can also aid in the de-cision to perform a TR. Performing TR late (Day 13)does not reduce ventilator, ICU. or hospital days nordoes it reduce pnetimonia rates. Early identification ofthe at-risk population could lead to earlier TR in anattempt to decrease morbidity and length of stay.

REFKRENCKS

1. D'Anielin I.F. Hammond JS. Spain DA. Stityak .IP. Trache-ostomy and percLitaneoLis cndo.scopic ya.strostomy in the manage-ment of the head-injured trauma patleiii. Am Stirg l994;60:l80-5.

2. Rodriguez JL. Steinberg SM. Luelietii FA, et al. Early tra-cheoslomy for primary airway management in the surgical criticalcare setting. Surgery l99();!O8:f)5.')-9.

3. Lcsnik 1. Rappaport W. Fulgiiiiti J. Wit/ke D. The role ofearly tracheostomy in blunt, multiple organ trauma. Am Surg1992:58:346-9.

4. Heffner JE. Timing of tracheostomy in mechanically venti-lated patients. Am Rev Respir Dis 1993:147:76^-71.

5. Scheinhom DJ, Stearn-Hassenpfkig M. Provision of long-term mechanical ventilation. Crit Care Clin 1998; 14:819-32.

6. Plummer AL. Giacey DR. Consensus conferenee on artificialairways in patients receiving meehanical ventilation. Chest 1989;96:178-80.

7. Marsh HM. Gilespie DJ. Baumgartner AE. Timing of tra-cheostomy in the critically ill patient. Chest l989;96:l90-3.

8. Gibbons KJ. Tracheostomy: Timing is everything. Crit CareMed 2000;28:i663^.

9. Meredith W. Rtitledge R. Fakhry SM, et al. The eonundrumof the Glasgow Coma Scale in iuttihated patients: A linear regres-sion prediction of the Glasgow verbal score for the Glasgow eyeand motor scores. J Trauma 1998;44:839-45.

10. Gordon E, von Hoist H. Rudehill A. Outcome of head in-jury in 2298 patients treated in a single clinic during a 21-ycarperiod. J Neurosurg Anesthesiol 1995;7:235-47.

11. Qureshi AI, Suarez Jl, Parekh PD, Bhardwaj A. Predictionand timing of tracheostomy in patients with inlVaientoria! lesionsrequiring mechanical ventilatory support. Crit Care Med 2O0U:28;1383-7.

12. Pal J. Fleis/er D. The valtieol'the Glasgow Coma SealeandInjury Severity Score: Predicting outcome in multiple trauma pa-tients with head injury. J Trauma ]989;29;746-8.

13. Johnson SB. Kearney PA, Barker DE. Early criteria predic-tive of prolonged mcchLinical ventilation. J Trauniii I992;33:95-100.

14. Koh WY. Lew TW. Chin NM. Wong MF. Traehcostomy ina neuro-intensive care setting: Indications and timing. AnaesthIntens Care I997;2.'>:365-8.

DISCUSSION

DR. ROXIE M. ALBRFXHT (Albtiquerqtte. NM): Thisis a very timely paper in this day of limited hospital beds,ICU beds, and nurses to take care of our ICU patients. Theauthors have really searched for predictors and indicators toidentify patients that would benefit thetn fix)ni early trache-ostomy and decrease these ICU stays and hospital stays. Dr.Gurkin and her colleagues have found thai GCS of less than8 and an ICS of greater than 25 as well as ventilator days ofover 7 arc significant predictors of tracheotomy. I wouldreally like lo see if we could decrease those diiys evenfurther before the patients get their first episode of pneu-monia. As you have shown a number of those patients de-velop piieutnonia and that seemed to delay their trache-otoiny until their first ptietitnoniLi episode was over.

I do have a couple questions. Regarding the eight patientswhd underwent early tracheotomy before Day 7 did they

have a reason other than the truumatie brain injury thatrequired the tracheostomy? Did these patients have less hos-pital stay days and less ICU days and did they have lessincidence of pnetimonia? I know it"s a small number, but 1think it's important to know that.

What Type of sedation do you usually use in your ICU anddid that have any inOuenee on the recovery of their (jlasgowComa Scale over that period of time?

Did you look at any components of the ISS. such as theAbbreviated Injury Score (AIS) of the head or the AIS ofthe cerebrtim as a potential determinant for the need for thetracheotomy?

DR. J. WAYNK MEREDITH (Witiston-Suiem. NC):Are you making a specific recommendation as to whoshould get a tracheostomy in the first 5 days? Everybodywith Glasgow Coma Scale of less than 7? Everybody withan ISS greater than 25? And if so what do you do with thosepatients who are going to be extubatable and you can't get

No. 4 TRACHEOSTOMY IN TRAUMATIC BRAIN INJURY Gurkin et al. 329

them into a nursing home because now they have a trachand that requires a skilled facility and they are not availableand all of that?

DR. SCHOLTEN (Moderator): Did you identify thephenomenon of what is termed by the neurosurgeons as the"wake-up tnich?" That is. after tracheostomy is done withina day or two the patient comes off the ventilator with animproved mental status?_ DR. MYSTAN A. GURKIN (Closing Discussion}: Asfor Dr. Albrecht's questions we had seven patients that un-derwent tracheostomy before or on day seven. Five of thosepatients had concomitant high cervical spine injuries and/orfacial fractures which provided a more concrete indicationfor early airway control.

Those patients that underwent early tracheostomy didhave lower ventilator. ICU, and hospital days as well aslower pneumonia rates, but due to the small numbers theseresults would not reach any statistical significance.

As to sedation in our neurosurgical ICU, the neurosur-geons are very strict. We do not give any sedation exceptcodeine and an occasional small dose of Versed for anyprocedures. As to the iSS breakdown we did not look at the

individual components of the ISS but will do this in follow-up studies. We did evaluate which patients underwent op-erations. Fifty per cent of the total population required op-erations. There were 144 operations performed in 126patients.

As to the questions about specific recommendations, asnoted in our presentation the three criteria are very goodnegative predictors and they reliably eliminate patients thatwill eventually progress to extubation. Of the 211 patientsthat did not require tracheostomy. 70 per cent were extu-bated by Day 7. Those patients that did not require trache-ostomy tended to rapidly progress toward extubation.

So yes, if you did use these as absolute recommendations,there would be 18—well. 27—patients that included thenine that died later on that would have a tracheostomy thateventually reqtiired extubation. However, the number ofvent days for those patients was also on the longer side.

As to the wake-up tracheostomy our average number ofdays from tracheostomy to liberation from the ventilatorwas 7 days. I don't have the range on me here, but for themost part, our patients that needed tracheostomy did need itat that time.