Abstract
Respiratory illnesses are common indications for mechanical ventilation in children. The adequacy of ventilatory support for oxygenation is measured using arterial blood gas analysis and calculation of oxygenation index (OI). Due to invasive nature of arterial blood sampling needed to calculate OI, several researchers have replaced blood gas-derived partial pressure of oxygen values with oxygen saturation (SpO 2 ) obtained from pulse oximetry. This noninvasive index called oxygen saturation index (OSI) is found to be useful in neonates. Studies in pediatric population are lacking. In this prospective study on mechanically ventilated children, both OI and OSI were determined and compared against alveolar–arterial oxygen difference (AaDO 2 ). A total of 29 children were studied. Both OSI and OI had good correlation of 0.787 and 0.792 with AaDO 2 , respectively. OSI of 7.3 and 9.4 had good sensitivity and specificity for AaDO 2 cutoffs of 344 and 498, which represents moderate and severe respiratory illness, respectively. The correlation coefficients of both OSI and OI are similar against AaDO 2 . OSI can be used instead of OI for constant monitoring of children on mechanical ventilation. Arterial blood gas analysis and calculation of OI can be reserved for situations where SpO 2 measurement is unreliable.
Keywords: oxygenation, oxygenation index, arterial–alveolar oxygen difference, mechanical ventilation, children, critical care
Introduction
Respiratory illnesses are common indications for mechanical ventilation leading to significant morbidity and mortality in children younger than 5 years old. 1 Quantification of the severity and institution of appropriate therapeutic measures are necessary for successful outcomes in these children.
Clinical progress of critically ill children on mechanical ventilation depends on complex interaction of dynamic respiratory pathology with the therapeutic interventions. Hence quantification of the severity of the illness needs to include both the quantum of the therapeutic support and its consequent measurable clinical parameters.
To assess the severity of the respiratory illness in children, one must obtain values of arterial blood gas analysis and ventilatory parameters to make an objective conclusion on both oxygenation and ventilation separately. An overall picture of the magnitude of the pulmonary disease and adequacy of ventilatory support is often measured using various invasive metrics such as alveolar–arterial difference in oxygen (AaDO 2 ), partial pressure of oxygen (PaO 2 ): fraction of inspired oxygen (FiO 2 ) (PF ratio), and oxygenation index (OI). Noninvasive metrics include oxygen saturation SpO 2 :FiO 2 (SF ratio), and oxygen saturation index (OSI). Calculation of invasive metrics need blood gas analysis on an intermittent basis. AaDO 2 presents an overall picture of both disease and its response to treatment, and OI focuses on oxygenation component of disease response to treatment. Noninvasive OSI has a similar inference but uses SpO 2 obtained by pulse oximeter instead of PaO 2 obtained from arterial blood gas analysis.
The AaDO 2 2 is a comprehensive measure of the dysfunction of respiratory membrane resulting in a ventilation–perfusion mismatch. This can be derived by using the values obtained by blood gas analysis and the PaO 2 in the alveoli which is determined by the FiO 2 on ventilatory settings. Acceptable alveolar–arterial gradient is determined with the formula (age + 10)/4. 3 Both ventilatory parameters and consequent blood gas values are factored in calculation of AaDO 2 .
OI is one of the validated measures of severity of respiratory illness particularly in pediatric—adult respiratory distress syndrome (ARDS). Unlike, AaDO 2 , OI focuses on the oxygenation portion of mechanical ventilation. OI is derived using the pressure and oxygen parameters on the ventilator and the PaO 2 in arterial blood gas. 4
Arterial blood gas measurement is an invasive procedure requiring skilled personnel. Despite appropriate skill, there is a definitive failure rate of 8 to 9% and complications up to 20% of attempts associated with arterial punctures and indwelling arterial catheters. 5 6
Given the disastrous nature of the illness and the need for constant assessment of progress, it is essential to have a noninvasive, comprehensive bedside tool to measure the severity of illness. This will help the clinician to make appropriate management decisions.
Pulse oximetry is a noninvasive assessment tool that measures oxygen saturation (SpO 2 ), and is used for all children with respiratory illness needing mechanical ventilation for continuous monitoring. SpO 2 value alone without the information on the quantum of respiratory support is insufficient. Pulmonary disease quantification is done with inclusion of at least one other treatment parameter such as FiO 2 . 7 Several indicators based on pulse oximetry, such as saturation to FiO 2 (SF ratio) ratio and OSI have been sparingly used in last decade. 8 They are yet to become popular and universal probably because the relation between SpO 2 and PaO 2 is sigmoid and not linear throughout the range. SF ratio is limited in utility by not including pressure parameter on ventilator setting which can significantly influence the gas exchange and hence measured parameters. On the other hand, OSI needs mean airway pressure (MAP) for calculation.
OSI is like OI where the PaO 2 is replaced by SpO 2 . By virtue of this replacement, the index has become a useful noninvasive tool which has been validated and used in neonatology practice. 9 10 11 OSI has been sparingly used in the pediatric population beyond neonatal age and hence literature is scant. Knowledge about the strength of correlation of OSI with AaDO 2 and its comparison with the correlation between OI and AaDO 2 would help pediatric intensivists to utilize this tool in day-to-day practice.
Objectives
Primary Objective
The primary objective is to compare the correlation coefficients between OSI and OI against AaDO 2 in ventilated children.
Secondary Objective
The secondary objective is to obtain the cutoff value of OSI to represent AaDO 2 values for moderate and severe illness.
Materials and Methods
Study design: This was a prospective observational analytical study conducted in a tertiary level pediatric intensive care unit between October 2019 and November 2020.
Subjects: Children aged between 3 months and 16 years requiring intubation and mechanical ventilation for a primary respiratory cause were included. Children with associated congenital heart disease diagnosed on echocardiography (ECHO) were excluded.
Data collection: After an initial history and clinical examination, a chest X-ray was performed to ascertain the respiratory pathology. ECHO was done at admission to rule out congenital heart disease and estimate right ventricular systolic pressure. SpO 2 was measured using a properly fitting probe (Philips Sure sign VM4 Philips India Limited, Bengaluru, India). SpO 2 was recorded when the pleth waves were good and consistent. In case of SpO 2 values which were above 98%, appropriate FiO 2 adjustments on the ventilator were made and the values were recorded only when the SpO 2 was consistently between 70 and 98%. However, if the patient was not needing oxygen, SpO 2 was recorded as it is. Concurrent ventilatory parameters were noted. A blood gas sample was obtained by sterile technique arterial puncture from the radial artery. No suctioning, invasive procedures, or ventilatory settings were done for at least 10 minutes prior to arterial puncture. Blood gas analysis was done within 10 minutes of sampling (ABL800 FLEX blood gas analyzer, Radiometer Medical ApS, Denmark). All the subjects were ventilated in pressure-regulated volume control mode (targeted tidal volume of 5 mL/kg) which is the standard of care in our unit. The rest of the management was decided by the intensivist as per the unit protocol.
The OI is calculated by the equation OI = ([FiO 2 × MAP]/PaO 2 ) × 100.
The OSI is calculated by the equation OSI = ([FiO 2 × MAP]/SpO 2 ) × 100.
AaDO 2 = (FiO 2 %/100) × (P atm − 47 mm Hg) − (partial pressure of carbon dioxide [PaCO 2 ]/0.8) − PaO 2 .
Statistics
Sample size: We assumed a minimum correlation of 60%, an α error of 5%, and a power of 90% to calculate the sample size. We needed to study 25 subjects.
Statistical analysis: Quantitative data are summarized as mean (standard deviation) or median (interquartile range [IQR]). Qualitative data are summarized as proportions. Spearman's correlation was used to obtain correlation coefficients due to heteroscedasticity. The equation for OSI against OI and AaDO 2 was also obtained. OSI cutoff value for AaDO 2 values determined to represent moderate and severe disease was derived using receiver operating characteristic (ROC) curve.
Results
Twenty-nine children were included in the study. Thirteen (45%) were females. The median (IQR) age of our study subjects was 24 (8.3–72) months. Eight (27.6%) were infants and four (13.7%) were adolescents.
Table 1 depicts the various indications for mechanical ventilation of the study subjects. Thirteen (45%) of our subjects had septicemia along with their primary diagnosis.
Table 1. Various indications for mechanical ventilation.
| Indication for ventilation | n | % |
|---|---|---|
| Bronchopneumonia | 12 | 41.5 |
| Lobar consolidation | 6 | 20.8 |
| Aspiration pneumonia | 4 | 13.8 |
| Pneumothorax | 2 | 6.9 |
| Adult respiratory distress syndrome | 1 | 3.4 |
| Lung abscess | 1 | 3.4 |
| Pulmonary hemorrhage | 1 | 3.4 |
| Severe asthma | 1 | 3.4 |
| Bronchiolitis | 1 | 3.4 |
| Total | 29 | 100 |
The correlation coefficient (rho) for OSI versus AaDO 2 was 0.787 with a 95% confidence interval of 0.583 to 0.897. The correlation coefficient (rho) for OI versus AaDO 2 was 0.792 with a 95% confidence interval of 0.593 to 0.900. The correlation coefficient (rho) of P:F ratio versus AaDO 2 was −0.794 with a 95% confidence interval of −0.901 to −0.595.
There is no significant difference between the correlation coefficients of OSI versus AaDO 2 and OI versus AaDO 2 ( p = 0.9124). OSI showed a moderate linear correlation with OI, rho = 0.601 with 95% confidence interval of 0.291 to 0.797.
Using the cutoff established for pediatric ARDS as per the 2015 PALICC criteria, AaDO 2 cutoffs were obtained for OI values greater than 8 and 16 ( Fig. 1 ). An OI > 8 represents moderate disease and OI > 16 represents severe disease. AaDO 2 value of 344 had a sensitivity of 92.3 and specificity of 93.7% for an OI of 8. AaDO 2 value of 498 had a sensitivity of 87.5% and specificity of 82% for an OI of 16. OSI cutoff values for the above-derived AaDO 2 values were computed. OSI value of 7.3 had a good sensitivity of and specificity for the AaDO 2 value corresponding to moderate disease and a value of 9.4 had a similar diagnostic ability for severe disease ( Table 2 ).
Fig. 1.
ROC curve for AaDO 2 values representing OI of 8 and 16. AaDO 2 , alveolar–arterial oxygen difference; OI, oxygenation index; ROC, receiver operating characteristic.
Table 2. OSI cutoff values for moderate and severe respiratory disease.
| AaDO 2 —344 (corresponding to OI of 8) | AaDO 2 —498 (corresponding to OI of 16) | |
|---|---|---|
| OSI value | 7.3 | 9.4 |
| Sensitivity | 83.8% | 90.8% |
| Specificity | 89% | 80% |
Abbreviations: AaDO 2 , alveolar–arterial oxygen difference; OI, oxygenation index; OSI, oxygen saturation index.
OSI values of 7.3 and 10.6 were useful in predicting OI values of 8 and 16. Fig. 2 shows the ROC curves comparisons for various OSI values against OI and AaDO 2 .
Fig. 2.
Comparison of predictive ability of OSI values for OI and AaDO 2 representing moderate and severe disease. AaDO 2 , alveolar–arterial oxygen difference; OI, oxygenation index; OSI, oxygen saturation index.
The median (IQR) duration of ventilation was 3 (2–4.5) days. The minimum duration was 1 day, and the maximum was 22 days. The median (IQR) duration of hospital stay of our study subjects was 8 (3–11) days. Eight (27.6%) of our study subjects died. Eleven (37.9%) subjects improved and were discharged home. Ten (34.5%) of our subjects opted for transfer to a government facility due to financial constraints. The outcome of these transferred patients is not known.
Discussion
In this prospective study, we demonstrated that the correlation coefficient of the noninvasive OSI with AaDO 2 was not only excellent, but comparable to OI, which is obtained after an invasive procedure. OSI values of 7.3 and 9.4 represented moderate and severe respiratory illness, respectively.
Blood gas analysis is often done in children ventilated for respiratory illness. The procedure is not only painful and leads to vascular complications but can also result in the introduction of infection resulting in sepsis. 12 Minimizing this invasive procedure with a reliable and equally representative noninvasive measure is invaluable in pediatric critical care.
The relation between PaO 2 in the blood and SpO 2 of hemoglobin follows a sigmoid pattern. 13 There is a linear relation between the SpO 2 and PaO 2 between 20 and 80 mm Hg which is the clinically useful range. This linearity can be harvested to substitute the PaO 2 with SpO 2 .
Pulse oximetry relies on normal pulsatile flow for its signal. Hence impaired perfusion or vasoconstriction associated with hypothermia, vasopressor treatment for hypotension, tourniquet effect from blood pressure cuff, etc. can lead to falsely low readings. 14 15 This should be borne in mind while employing OSI or other SpO 2 -based measurements for monitoring critically ill children.
Thomas et al used SpO 2 in lieu of PaO 2 in the OI formula and came up with OSI. 16 Their initial studies in children up to 18 years of age demonstrated the usefulness of OSI in classifying acute lung disease in children. They reported an OSI of 7.8 to represent moderate pediatric ARDS. The OSI values in our study are closer to this study where OSI of 7.3 qualified for moderate illness. The minor difference is since the former has compared OSI against PF ratio and we have compared against AaDO 2 .
OSI has been consistently shown to be useful even in neonates where the oxygen dissociation curve is significantly influenced by higher quantity of fetal hemoglobin. 9 10 11 17 This supports the use of OSI in children where such an issue is less significant.
Unlike in neonates, there are only a few studies on OSI in children. 18 19 These studies were not exclusively designed to evaluate OSI as their primary focus. Ours is probably the first study in the pediatric population which has evaluated OSI as a primary tool of noninvasive monitoring of sick children. We have compared both OSI and OI against AaDO 2 which is a comprehensive measure of severity of illness unlike other studies which have compared against PF ratio which is primarily designed to determine oxygenation status.
Limitations of our study are limited sample size, not using mixed model equations for repeated measures. Other limitation is not studying the effect of parameters such as pH, PaCO 2 , temperature on OSI and OI. However, these variables were well controlled and the practical significance of the influence on these parameters on OSI and OI is minimal.
Conclusion
OSI strongly correlates with AaDO 2 in the pediatric population. The correlation coefficients of OSI and OI with AaDO 2 are not significantly different implying OSI can be used in lieu of OI for monitoring critically ill children. OSI of 7.3 and 9.4 correlated with good sensitivity and specificity for AaDO 2 cut-offs of 344 and 498, which represent moderate and severe respiratory illness, respectively.
Funding Statement
Funding None.
Conflict of Interest None declared.
Ethical Approval
The institutional ethical committee approved this study. Informed written consent was taken by the parents for the participation of their ward in this study.
Authors' Contributions
S.M.D. generated the research question and did the study design. He was involved in statistical analysis and fine tuning of manuscript. S.C. collected the data and did the literature search and initial analysis. J.K.K. helped in study design and preparation of manuscript.
References
- 1.Prabhakaran P, Sasser W, Borasino S. Pediatric mechanical ventilation. Minerva Pediatr. 2011;63(05):411–424. [PubMed] [Google Scholar]
- 2.Hantzidiamantis P J, Amaro E. Treasure Island (FL): StatPearls Publishing; 2020. Physiology, alveolar to arterial oxygen gradient. [PubMed] [Google Scholar]
- 3.Sharma S, Hashmi M F, Burns B. Treasure Island (FL): StatPearls Publishing; 2020. Alveolar gas equation. [PubMed] [Google Scholar]
- 4.Pediatric Acute Lung Injury Consensus Conference Group Khemani R G, Smith L S, Zimmerman J J, Erickson S.Pediatric acute respiratory distress syndrome: definition, incidence, and epidemiology: proceedings from the Pediatric Acute Lung Injury Consensus Conference Pediatr Crit Care Med 201516(5, Suppl 1):S23–S40. [DOI] [PubMed] [Google Scholar]
- 5.Yee K, Shetty A L, Lai K. ABG needle study: a randomised control study comparing 23G versus 25G needle success and pain scores. Emerg Med J. 2015;32(05):343–347. doi: 10.1136/emermed-2014-203600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Soderstrom C A, Wasserman D H, Dunham C M, Caplan E S, Cowley R A. Superiority of the femoral artery of monitoring. A prospective study. Am J Surg. 1982;144(03):309–312. doi: 10.1016/0002-9610(82)90007-1. [DOI] [PubMed] [Google Scholar]
- 7.National Institutes of Health, National Heart, Lung, and Blood Institute ARDS Network . Rice T W, Wheeler A P, Bernard G R, Hayden D L, Schoenfeld D A, Ware L B. Comparison of the SpO2/FIO2 ratio and the PaO2/FIO2 ratio in patients with acute lung injury or ARDS. Chest. 2007;132(02):410–417. doi: 10.1378/chest.07-0617. [DOI] [PubMed] [Google Scholar]
- 8.Pediatric Acute Lung Injury and Sepsis Network Investigators (PALISI) . Khemani R G, Thomas N J, Venkatachalam V et al. Comparison of SpO2 to PaO2 based markers of lung disease severity for children with acute lung injury. Crit Care Med. 2012;40(04):1309–1316. doi: 10.1097/CCM.0b013e31823bc61b. [DOI] [PubMed] [Google Scholar]
- 9.Doreswamy S M, Chakkarapani A A, Murthy P. Oxygen saturation index, a noninvasive tool for monitoring hypoxemic respiratory failure in new-borns. Indian Pediatr. 2016;53(05):432–433. [PubMed] [Google Scholar]
- 10.Rawat M, Chandrasekharan P K, Williams A et al. Oxygen saturation index and severity of hypoxic respiratory failure. Neonatology. 2015;107(03):161–166. doi: 10.1159/000369774. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Khalesi N, Choobdar F A, Khorasani M, Sarvi F, Aski B H, Khodadost M. Accuracy of oxygen saturation index in determining the severity of respiratory failure among preterm infants with respiratory distress syndrome. J Matern Fetal Neonatal Med. 2019;34(01):1–6. doi: 10.1080/14767058.2019.1666363. [DOI] [PubMed] [Google Scholar]
- 12.Raffin T A. Diagnostic decision. Ann Intern Med. 2019;105(03):390. doi: 10.7326/0003-4819-105-3-390. [DOI] [PubMed] [Google Scholar]
- 13.Hall J E, Guyton A C. 13th ed. Philadelphia: Elsevier Publications; 2016. Transport of oxygen and carbon dioxide in blood and tissue fluids; pp. 527–536. [Google Scholar]
- 14.Talke P, Stapelfeldt C. Effect of peripheral vasoconstriction on pulse oximetry. J Clin Monit Comput. 2006;20(05):305–309. doi: 10.1007/s10877-006-9022-3. [DOI] [PubMed] [Google Scholar]
- 15.Wille J, Braams R, van Haren W H, van der Werken C. Pulse oximeter-induced digital injury: frequency rate and possible causative factors. Crit Care Med. 2000;28(10):3555–3557. doi: 10.1097/00003246-200010000-00036. [DOI] [PubMed] [Google Scholar]
- 16.Thomas N J, Shaffer M L, Willson D F, Shih M C, Curley M AQ. Defining acute lung disease in children with the oxygenation saturation index. Pediatr Crit Care Med. 2010;11(01):12–17. doi: 10.1097/PCC.0b013e3181b0653d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Muniraman H K, Song A Y, Ramanathan R et al. Evaluation of oxygen saturation index compared with oxygenation index in neonates with hypoxemic respiratory failure. JAMA Netw Open. 2019;2(03):e191179. doi: 10.1001/jamanetworkopen.2019.1179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pediatric Acute Lung Injury and Sepsis Investigators . Willson D F, Thomas N J, Markovitz B P et al. Effect of exogenous surfactant (calfactant) in pediatric acute lung injury: a randomized controlled trial. JAMA. 2005;293(04):470–476. doi: 10.1001/jama.293.4.470. [DOI] [PubMed] [Google Scholar]
- 19.Curley M A, Hibberd P L, Fineman L D et al. Effect of prone positioning on clinical outcomes in children with acute lung injury: a randomized controlled trial. JAMA. 2005;294(02):229–237. doi: 10.1001/jama.294.2.229. [DOI] [PMC free article] [PubMed] [Google Scholar]