PURPOSE: Obesity hypoventilation syndrome (OHS) is characterized by PaCO2 ≥45mm Hg and Body mass index ≥30Kg/m2, with no other etiology of hypoventilation. It is diagnosed by clinical suspicion, polysomnography and Arterial Blood gas analysis (ABG) which is a Gold standard to document daytime hypercapnia. Transcutaneous carbon dioxide measurement (PtcCO2) is an upcoming modality to measure arterial PaCO2 non-invasively. We compared the accuracy of PtcCO2 with PaCO2 from ABG and studied the use of PtcCO2 as a screening tool in suspected OHS patients.
METHODS: Prospectively we followed patients coming to our Sleep center for the evaluation of Obstructive sleep apnea from October 2007 till March 2008. From this patient population, we selected 19 patients with clinical suspicion of OHS and subjected them to ABG and PtcCO2 using TOSCA 500 system simultaneously prior to an overnight polysomnography. Fischer exact test was used for statistical analysis.
RESULTS: Using linear regression method, the estimated bias between PaCO2 and PtcCO2 was -1.1053 ± 4.6mmHg with a tight correlation (R2 = 0.89, p< 0.001). Out of 19 patients 5 had PaCO2 ≥45mmHG on ABG. PtcCO2 diagnosed 3 of these 5 patients (Figure1) with a sensitivity of 60%, specificity of 71.4 %, positive predictive value (PPV) of 42.9% and negative predictive value (NPV) of 83.3% (p<0.236). Based on the SD of ±4.6mmHg, if the cutoff value for PtcCO2 for diagnosing OHS is taken as ±40.4mmHg, then the sensitivity changes to 100%, the specificity drops to 35.7%, PPV is 35.7% and NPV goes up to 100% (p<0.172).(Figure2).
CONCLUSION: We conclude that PtcCO2 measurement cannot replace conventional ABG for diagnosing OHS due to wide variabilty in the readings. Although, the statistical significance was not obtained, if the cutoff for PtcCO2 is taken as < 40.4mmHg, it has a very good NPV. Therefore, it may be useful in excluding diagnosis of OHS and avoiding invasive ABG analysis.
CLINICAL IMPLICATIONS: Single value PtcCO2 measurement cannot replace conventional ABG for diagnosing OHS. More studies with larger patient population need to be done.
DISCLOSURE: Sivashankar Sivaraman, No Financial Disclosure Information; No Product/Research Disclosure Information