85 / 280
controversy in the suitability of TLA for more severely
affected infants
6,7
along with the interference of tongue
mobility and the possibility of tongue dehiscence,
8
this pro-
cedure is not used at our institution. Tracheostomy is effec-
tive at bypassing the obstruction but doesn’t address the
cause of the airway obstruction and requires extensive main-
tenance. Until recently, this has been the standard treatment
of moderate to severe PRS.
9
Mandibular distraction (MD)
consists of performing an osteotomy on the ramus of the
mandible and gradually pulling it forward, correcting the
micrognathia and tongue-based airway obstruction by pro-
viding more space for the tongue and oropharyngeal airway.
An important consideration with these surgical interven-
tions is the associated cost for patients’ families and the
health care system. In our institution, patients with PRS
receiving MD seemed to have shorter hospital stays and
fewer subsequent interventions than those receiving tra-
cheostomy. We therefore hypothesized that MD would be
associated with significantly lower costs than tracheostomy.
To test this we performed a retrospective cohort study com-
paring the costs associated with MD and tracheostomy for
infants with PRS, both isolated and syndromic.
Additionally, costs for neonates with PRS who underwent
tracheostomy and secondarily underwent MD within the
first year of life were compared.
Methods
Data Collection
A retrospective chart review was performed on neonates
with PRS whose treatment at Cincinnati Children’s Hospital
Medical Center (CCHMC) began between 2001 and 2009.
This study was approved by the Institutional Review Board
(IRB) at CCHMC (#2009-0162). A multidisciplinary airway
team including neonatologists, geneticists, otolaryngologists,
pulmonologists, speech therapists, and plastic surgeons eval-
uated patients identified with PRS in the neonatal ICU.
Workup typically included bedside nasopharyngoscopy,
cephalogram, pulse oximetry monitoring, feeding assess-
ment, and a sleep study. Mildly abnormal sleep studies
despite repositioning lead to discharge with supplemental
oxygen as appropriate and close follow-up. Moderately/
severely abnormal sleep studies are followed by further ima-
ging including microlaryngoscopy, bronchoscopy, and/or
CINE MRI to evaluate for multilevel obstruction. Patients
with moderate-severe sleep studies and additional risk fac-
tors (eg, multilevel obstruction, neurologic delay) or those
requiring early intervention (eg, ex utero intrapartum
[EXIT] to airway) are often referred for tracheostomy.
Others receive tracheostomy or MD based on team recom-
mendations. Seventy neonates (defined as infants less than 1
year old) were identified with PRS who underwent MD or
tracheostomy. These included a subset of patients for whom
our group recently published separate outcomes data.
10
Patients with incomplete billing records or incomplete
follow-up charges (
\
3 years) were excluded (n = 23). One
syndromic patient who received both tracheostomy and
subsequent MD was excluded due to lengthy cardiac ICU
stay unrelated to PRS.
The CCHMC billing department provided records of all
charges to patients over a 3-year period. These included
daily inpatient fees (for all admissions over 3 years includ-
ing patient-specific nursing care, mechanical ventilation,
enteral feeding, radiologic studies, medications, and labora-
tory tests), surgical fees (gastrostomy, microlaryngoscopy
and bronchoscopy [ML&B], tracheostomy, mandibular
osteotomies and distractor placement/adjustment/removal
including distractor hardware costs), anesthesia fees, inpati-
ent consultation fees, outpatient clinic fees, emergency
room visits, and radiologic and sleep studies. Operations,
imaging studies, and emergency room and clinic visits unre-
lated to the PRS diagnosis were excluded. Charges prior to
2009 were adjusted for inflation using an annual rate of 3%.
All patients discharged with a tracheostomy received
home tracheostomy care. The monthly cost for home tra-
cheostomy care was estimated based on a patient’s level of
respiratory support (CPAP vs ventilator), the number of
months with tracheostomy before decannulation, estimated
equipment rental and tracheostomy supply costs, and indivi-
dualized home nursing care recommendations. A common
recommendation provided 8 hours of home nursing care per
night for 8 weeks. A list of the home nursing care and tra-
cheostomy rental and supply rates used may be found in
Supplemental Table S1 at
www.otojournal.org.
Data Analysis
Data distributions for continuous data were assessed using
means with standard deviations and medians with ranges
(minimum and maximum) and interquartile ranges.
Categorical data were reported as frequencies and percen-
tages. Comparisons of median costs (adjusted for inflation)
across the 3 groups (mandibular distraction only, tracheost-
omy only [Trach], and tracheostomy with subsequent man-
dibular distraction [Trach
1
MD]) were made using the
Kruskal-Wallis test. Post hoc pairwise comparisons between
groups were conducted using a Wilcoxon rank sum test with
a Bonferonni adjustment. Total costs for year 1 were also
adjusted for the number of days in the ICU using a general
linear model (with least square means reported as the
adjusted means). The data did not follow a Gaussian distri-
bution, and therefore a log transformation was conducted on
total costs for year 1 in order to control for number of days
in the ICU, and the results were back transformed into
whole dollar amounts for the purpose of interpretation.
Adjusted mean total costs were reported with 95% confi-
dence intervals.
Results
Forty-seven patients with PRS were identified who were
treated with mandibular distraction (MD, n = 26), tracheost-
omy (Trach, n = 12), or tracheostomy with subsequent MD
(Trach
1
MD, n = 9) and who met inclusion criteria (
Table 1
).
The MD group had a higher percentage of patients with non-
syndromic PRS (82%) compared to the Trach (58%) and
Otolaryngology–Head and Neck Surgery 151(5)
63