75 / 232
DISE scoring systems have not been proven to correlate
well with sleep apnea parameters in adults.
7
It is still
unknown if DISE accurately identifies sites of obstruction
in children with polysomnogram (PSG)–diagnosed OSA. In
2014, we introduced and validated a new scoring system
for DISE (Chan-Parikh [C-P] score) in children with OSA
to identify the location and severity of obstruction.
8
This
study builds on that work by presenting new data on
patients undergoing DISE to determine if standardized
scoring correlates with PSG parameters. We hypothesized
that the C-P DISE score would correlate with PSG findings—
specifically, apnea-hypopnea index (AHI) and oxygen nadir,
broad indicators of OSA severity. As a secondary objective,
we sought to determine if age at the time of DISE, presence of
a syndrome, or history of adenotonsillectomy affect this
relationship.
Methods
Institutional Review Board approval was obtained from the
2 participating institutions: Seattle Children’s Hospital
(SCH) and the University of California–San Francisco
(UCSF). At both institutions, all patients undergoing DISE
are scored prospectively via the C-P score, and the findings
are recorded in a database. These databases contain basic
patient demographic information (date of birth, date of pro-
cedure, comorbidities), limited PSG parameters (AHI,
lowest O
2
saturation) if the patient underwent preoperative
PSG, and the C-P score. Records of all patients in these
databases were screened between January 1, 2011, and
December 31, 2014, to obtain past medical and surgical his-
tory. All patients who underwent PSG prior to DISE were
included in the study; there was no restriction on how far in
advance of DISE the PSG was performed. The decision to
perform DISE was based on clinical evaluation by the
attending surgeon. DISEs were typically performed on chil-
dren with small or absent tonsils or with clinical suspicion
for multilevel airway obstruction. Children who did not
have preoperative PSG were excluded from the study; PSG
was not always obtained prior to DISE, based on clinical
judgment, cost, and family decision making.
Subjects underwent PSG at an accredited sleep laboratory
as part of clinically indicated care, with results interpreted
by board-certified pediatric sleep medicine physicians.
PSGs were scored in accordance with the American
Academy of Sleep Medicine parameters.
9
AHI and lowest
recorded oxygen saturation were noted from the preproce-
dural PSG.
At both tertiary care facilities, sleep endoscopy is carried
out in standardized fashion, with all reports being categor-
ized per the C-P scoring system, which has been published
and validated.
8
The anesthetic technique for all DISE uti-
lized sevoflurane and propofol per institutional protocols.
The C-P score is based on 5 anatomic locations, with each
site graded on a 4-point scale according to severity of
obstruction. Sleep endoscopy scores were noted at the time
of surgery by the surgeon responsible for each case.
After all cases had been identified, chart review was per-
formed to acquire the demographic characteristics of the
participants, including age at time of sleep endoscopy, sur-
gical history, and presence of concomitant syndromal or
genetic disorder (eg, Trisomy 21). Univariate analyses were
performed to obtain descriptive statistics. Means were calcu-
lated for continuous variables, such as PSG and sleep endo-
scopy scores, along with average age at time of DISE.
Proportions were calculated for binary variables, such as the
presence of a syndrome. Mean AHI and lowest oxygen
saturation values for PSGs performed at UCSF and SCH
were compared with Student’s
t
test to ensure consistency
between studies performed at the separate institutions.
Spearman’s correlation coefficients (2-tailed) were then
calculated to determine the degree of linear correlation
between C-P score and each PSG result: AHI and oxygen
nadir. To determine if there was significant variability in
DISE results among the 8 attending surgeons, Spearman’s
correlation coefficients were also calculated for AHI and
lowest oxygen saturation for patients who underwent DISE
by the senior author (S.R.P.; n = 58) and the other 7 sur-
geons (n = 69). Two separate multivariable linear regression
models were then created, controlling for syndrome diagno-
sis, history of adenotonsillectomy, and age at time of sleep
endoscopy. AHI and oxygen nadir were used as the depen-
dent variables in these models, while C-P score was the
independent variable. For all tests,
P
\
.05 was considered
statistically significant. Stata 13.1 (Stata Inc, College
Station, Texas) statistical software was used for all analyses.
Results
A total of 127 children met inclusion criteria for the study.
The demographic composition of this patient population is
outlined in
Table 1
. The mean AHI value for PSGs per-
formed at UCSF was 11.7
6
12.8 (
6
SD), while at SCH
the mean AHI value was 14.1
6
20.9. There was no statisti-
cally significant difference in the mean AHI values between
the centers (
P
= .6). Mean lowest oxygen saturations were
also not significantly different between patients at the 2
institutions: mean O
2
nadir at UCSF was 82%
6
9.5%,
compared with 86%
6
9.4% at SCH (
P
= .1). The scatter-
plots represented in
Figure 1
demonstrate AHI and oxygen
nadir as a function of C-P score. Correlation analysis
Table 1.
Demographic Characteristics of the Study Population.
a
Subjects
127
Age at time of DISE, y
6.55
6
5.34
Syndromal or genetic disorder
56 (44)
Previous adenotonsillectomy
21 (16.5)
AHI, events/h
13.6
6
19.6
O
2
nadir, % (O
2
saturation)
85.4
6
9.4
Chan-Parikh score
5.9
6
2.7
Abbreviations: AHI, apnea-hypopnea index; DISE, drug-induced sleep
endoscopy.
a
Values presented as n (%) or mean
6
SD.
Dahl et al
53