Results
The mean score for tonsillar hypertrophy on the Brodsky scale was
2.9 ± 0.7. Ten children had an SHS above the 2.75 threshold (mean
2.72 ± 1.22).
PSG analysis showed a median OAHI of 3.4 [0.9; 9.2] per hour. Of the
16 children, 10 were diagnosed with OSAHS (OAHI ≥ 1.5). Four patients
had mild OSAHS, two had moderate OSAHS and four had severe OSAHS
(Table 2). The median ODI for the study population was 2.8 [1.8;
8.3] per hour.
A full analysis of the second sleep cycle (mean time of 72 ± 13 minutes)
was performed in all children. All the videos were of good quality and
centered on the child’s head and chest throughout the recording.
Analysis of all the videos yielded median values of 4.5 [0.5; 9.5]
apneas, 6 [4; 10] whole body movements and 1 [0; 3] awakening.
In one of the five children with observed intercostal retractions, the
movement was continuous. Seven children were occasional mouth-breathers.
The mean RS30 was 4.70 ± 3.04 and the mean RS10 was 3.72 ± 2.96
(Table 3).
Children without OSAHS had a median RS30 of 3.00 [1.00; 4.50].
Patients with OSAHS had a median RS30 of 3.01 [1.23; 5.88] if mild,
7.62 [6.09; 9.15] if moderate and 7.25 [4.75; 9.26] if severe.
Children without OSAHS had a median RS10 of 1.71 [1.00; 3.00].
Patients with OSAHS had a median RS10 of 1.17 [0.08; 4.29] if mild,
6.37 [4.36; 8.38] if moderate and 7.17 [4.25; 7.92] if severe.
Associations between OSAHS severity and RS30 or RS10 values were not
significant (p=0.09 for both).
The correlation coefficient between the RS30 and the OAHI was 0.71
(p = 0.002) and −0.63 (p = 0.008) between the RS30 and the BMI. The
correlation coefficient between the RS10 and the OAHI was 0.59
(p = 0.02) and −0.58 (p = 0.02) between the RS10 and the BMI. No link
between the RS30 or RS10 and patient age was identified.
The RS30 and RS10 area under the ROC curve for an OAHI ≥ 5 was 0.88
(95%CI: 0.70 to 1.00).
According to Youden’s index, a RS30 threshold of 6.09 was predictive of
an OAHI ≥ 1.5 with a sensitivity of 60% (95%CI: 26 to 88%) and a
specificity of 100% (95%CI: 54 to 100%). The positive and negative
predictive values were equal to 100% and 60% respectively. A RS10
threshold of 6.40 was predictive of an OAHI ≥ 1.5 with a sensitivity of
50% (95%CI: 19 to 81%) and a specificity of 100% (95%CI: 54 to
100%). The positive and negative predictive values were equal to 100%
and 55% respectively.
When considering an OAHI ≥ 5, a RS30 threshold of 6.09 had a sensitivity
of 83% (95%CI: 36 to 99%) and a specificity of 90% (95%CI: 56 to
99%). The positive and negative predictive values were equal to 83%
and 90% respectively. A RS10 threshold of 6.50 had a sensitivity of
67% (95%CI: 22 to 96%) and a specificity of 100% (95%CI: 69 to
100%). The positive and negative predictive values were equal to 100%
and 83% respectively.
The correlation coefficient between the RS30 and the ODI was 0.70
(p = 0.003), 0.57 (p = 0.02) between the RS10 and the ODI, and 0.29
(p = 0.28) between the ODI and the OAHI. Greater OAHI values were not
significantly associated with greater ODI, SHS, number of apneas, number
of movements, number of waking episodes, RS30 and RS10 values (Figure
1).
Fifty percent (5/10) of patients with an SHS ≥ 2.75 had an OAHI ≥ 5 per
hour. The correlation coefficient
between the OAHI and the SHS was 0.43 (p = 0.09), 0.62 (p = 0.01)
between RS30 and the SHS, and 0.65 (p = 0.007) between RS10 and the SHS.