Abstract
Background & objective: Acute Respiratory Distress Syndrome (ARDS) is one of the leading causes of ICU admission and mortality. This study examined the clinical utility of the SpO2/FiO2 (SF ratio) as a substitution of PaO2/FiO2 (PF ratio) in diagnosis and classification of ARDS.
Design: It was a prospective observational study.
Setting: ICU of academic tertiary care government run hospital of Bangladesh.
Methods: All consecutive patients with ARDS on mechanical ventilator fulfilling the inclusion & exclusion criteria were included in the study. After diagnosis of ARDS as per Berlin definition (which uses PF ratio), PaO2 from arterial blood gas (ABG), SpO2 from pulse oximetry, and FiO2 from ventilator setting were documented at zero hour, at 24 hours and at 48 hours. All study patients were placed on PEEP ≥ 5 cm of H2O.
SF ratio was calculated in all study subjects corresponding to PF ratio. The relationship between SF and PF were described by linear regression equation after being plotted in scatter plot diagrams to see their correlation. To find out cut off value of SF ratio against PF ratio, best fit method was used. As we did not have any mild ARDS patient, the study subjects were either moderate or severe ARDS based on Berlin classification.
Results: In this study a total of 50 mechanically ventilated adult ARDS patients were enrolled following eligibility criteria. In this study, cut off value of SF ratio was found to be 123 and 122 at zero hour and at 24 hours respectively against PF ratio of 110 (according to best fit linear relationship). A validity test showed PF 110 (not PF 100) had the best specificity among PF 95, PF 100 and PF 110 at zero hour and 24 hours. So, we accepted SF ≤ 123 (with PEEP ≥ 5 cm H₂O) as indicative of Severe ARDS. At zero hour and 24 hours as there was no SF value against PF 200. So, scatter plot correlation (best fit) was not applicable. However, using scatter plot diagram and linear regression equation the cut off value of SF ratio against PF 200 at 48 hours were calculated and the best fit value for SF was 214.91. With value above PF 200 corresponding SF value was calculated to be 235± 00 and 234.98 ± 10.73 at 24 hours and 48 hours respectively using mean ± SD. So, we accepted SF 235 as the cutoff value for PF 200 and we defined SF > 123 to ≤ 235 (with PEEP ≥ 5 cm H₂O) as indicative of Moderate ARDS.
Conclusion: In spite of having limitations in our study we conclude that like PF ratio as is used in Berlin definition, SF ratio can also be reasonably used to diagnose and classify mechanically ventilated ARDS (severe and moderate) particularly in a resource limited setting where ABG facility is scarce but pulse oximetry facility is abundant