CORRESPONDENCE criteria using PaO2/FIO2 in patients with different severity of respiratory provide a graph with the respiratory rate on the x-axis and the failure will further help researchers in the future. n SpO2/FIO2 on the y-axis, plot the failures and the successes in different symbols, and mark the isopleth with a slope of 4.88. Such a graph allows the reader to see the positions of Author disclosures are available with the text of this letter at www.atsjournals.org. successes and failures in relation to the cut point of 4.88 and the role that tachypnea and hypoxemia played in those Habib M. R. Karim, M.D., D.N.B., I.D.C.C.M. positions. I also suggest to the authors testing the index with the All India Institute of Medical Sciences respiratory rate squared. The range of the respiratory rate is Raipur, India narrow. This transformation increases the range of the Antonio M. Esquinas, M.D., Ph.D.* denominator and might create clearer separation between the Hospital Morales Meseguer successes and the failures. n Murcia, Spain ORCID IDs: 0000-0002-6632-0491 (H.M.R.K.); 0000-0003-0571-2050 (A.M.E.). Author disclosures are available with the text of this letter at www.atsjournals.org. *Corresponding author (e-mail: antmesquinas@gmail.com). Aiman Tulaimat, M.D.* Cook County Health and Hospital Systems References Chicago, Illinois 1. Kang BJ, Koh Y, Lim CM, Huh JW, Baek S, Han M, et al. Failure of high- ORCID ID: 0000-0001-8085-9402 (A.T.). flow nasal cannula therapy may delay intubation and increase mortality. Intensive Care Med 2015;41:623–632. *Corresponding author (e-mail: atulaimat@cookcountyhhs.org). 2. Helviz Y, Einav S. A systematic review of the high-flow nasal cannula for adult patients. Crit Care 2018;22:71. 3. Roca O, Caralt B, Messika J, Samper M, Sztrymf B, Hernández G, et al. References An index combining respiratory rate and oxygenation to predict outcome of nasal high flow therapy. Am J Respir Crit Care Med [online 1. Roca O, Messika J, Caralt B, Garcı́a-de-Acilu M, Sztrymf B, Ricard J-D, ahead of print] 21 Dec 2018; DOI: 10.1164/rccm.201803-0589OC. et al. Predicting success of high flow nasal cannula in pneumonia 4. Brown SM, Grissom CK, Moss M, Rice TW, Schoenfeld D, Hou PC, patients with hypoxemic respiratory failure: the utility of the ROX et al.; NIH/NHLBI PETAL Network Collaborators. Nonlinear imputation index. J Crit Care 2016;35:200–205. of Pao2/Fio2 from Spo2/Fio2 among patients with acute respiratory 2. Roca O, Caralt B, Messika J, Samper M, Sztrymf B, Hernández G, et al. distress syndrome. Chest 2016;150:307–313. An index combining respiratory rate and oxygenation to predict 5. Madan A. Correlation between the levels of SpO2 and PaO2. Lung India outcome of nasal high flow therapy. Am J Respir Crit Care Med [online 2017;34:307–308. ahead of print] 21 Dec 2018; DOI: 10.1164/rccm.201803-0589OC. 3. Yang KL, Tobin MJ. A prospective study of indexes predicting the Copyright © 2019 by the American Thoracic Society outcome of trials of weaning from mechanical ventilation. N Engl J Med 1991;324:1445–1450. Copyright © 2019 by the American Thoracic Society Predicting the Outcome of Nasal High-Flow Therapy: A Proposed Representation of the Data and a Supplemental Analysis Reply to Tatkov, to Karim and Esquinas, and To the Editor: to Tulaimat I read with interest the paper by Roca and colleagues that From the Authors: followed up on their initial publication in 2016 (1, 2). One aspect of the report that makes interpreting the results We read with great interest the letter by Stanislav Tatkov and thank difficult is that the authors did not provide graphs with the him for his interest in our work. Dr. Tatkov’s thoughts are individual data points for the respiratory rate, the oxygen interesting, and the figure he provides is insightful. saturation as measured by pulse oximetry (SpO2)/FIO2, and The figure shows that two distinct combinations of the ROX to compare successes and failures, and only respiratory rate and FIO2 (which may reflect two different provided comparisons of summary data in the tables. In clinical situations in terms of disease severity)—a a manner similar to the graph used by Yang and Tobin respiratory rate of 20 with FIO2 of 0.8, and a respiratory rate when they validated the frequency-to-VT ratio to predict of 40 with FIO2 of 0.5—provide the same ROX index. The extubation success (3), I suggest that Roca and colleagues figure further shows that the ROX index is unlikely to drop This article is open access and distributed under the terms of the Creative This article is open access and distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives License 4.0 Commons Attribution Non-Commercial No Derivatives License 4.0 (http://creativecommons.org/licenses/by-nc-nd/4.0/). For commercial usage (http://creativecommons.org/licenses/by-nc-nd/4.0/). For commercial usage and reprints, please contact Diane Gern (dgern@thoracic.org). and reprints, please contact Diane Gern (dgern@thoracic.org). Originally Published in Press as DOI: 10.1164/rccm.201903-0702LE on Originally Published in Press as DOI: 10.1164/rccm.201903-0571LE on April 17, 2019 March 21, 2019 Correspondence 117 CORRESPONDENCE below 4.88 with FIO2 up to 0.5. Although the construction of 400 Success the curves is unquestionable, the combination Dr. Tatkov Failure highlights is more than unlikely. It is indeed quite exceptional to 300 SpO2/FIO2 encounter a patient with a respiratory rate of 40 breaths/min 12 hours after admission to an ICU for acute respiratory failure 200 (which is the time at which the ROX yielded the best value). For example, in the FLORALI (Clinical Effect of the Association of 100 Noninvasive Ventilation and High Flow Nasal Oxygen Therapy in Resuscitation of Patients with Acute Lung Injury) study, the 0 mean respiratory rate of patients under high-flow nasal cannula 10 20 30 40 (HFNC) treatment at 12 hours was only 27 breaths/min. In Respiratory rate (bpm) addition, only 6.5% of the patients had a respiratory rate >40 Figure 1. Scatterplot of respiratory rate and SpO2/FIO2 with the isopleth (minimum, 40; maximum, 44). Similarly, in the ROX validation corresponding to a ROX index value of 4.88. SpO2 = oxygen saturation as cohort, only one patient had a respiratory rate of 40 breaths/min measured by pulse oximetry. at 12 hours of HFNC, and the mean respiratory rate at this time point was 24 breaths/min. The second component of the might expect that the diagnostic accuracy of the ROX index combination is the FIO2. Once again, it is rare to encounter a using Pa O 2 /F I O2 in predicting HFNC outcomes would be patient with FIO2 below 0.5 within the first 12 hours of better than the one observed with Sp O2 /F I O 2 . However, two treatment. This observation is supported by solid data. For important issues should be taken into account. First, in the example, in the validation cohort, only 5.2%, 9.2%, and 21.3% of validation cohort, less than half of the patients had a Pa O 2 the patients after HFNC onset needed a FIO2 , 0.5 after 2 value available, which suggests that patients undergoing hours, 6 hours, and 12 hours of treatment, respectively (1). HFNC treatment are mainly monitored noninvasively using Even lower percentages were observed in the FLORALI cohort: Sp O 2 . Second, no significant differences were observed when 3.9%, 7.1%, and 15.8% at 1 hours, 6 hours, and 12 hours, the diagnostic accuracy of the ROX index constructed with respectively (2). Pa O 2 /F I O2 was compared with the same index constructed The second point raised by Dr. Tatkov is the fact that the with SpO2 /F I O 2 at any time point (P = 0.652, P = 0.122, and ROX index would get under 4.88 when the FIO2 was .0.8 P = 0.407 at 2 hours, 6 hours, and 12 hours of HFNC treatment, for respiratory rates of >20 breaths/min. It is our opinion respectively). Thus, according to these data, it seems that that the initial settings should favor high levels of FIO2 in a more practical and feasible approach for measuring the patients with more severe disease. However, most of the patients ROX index would be to use the Sp O2 , as it will be available who needed FIO2 . 0.8 presented respiratory rates of >20 in all patients treated with HFNC at the bedside and it breaths/min (83.1%, 79.6%, and 85.7% after 2 hours, 6 hours, does not predict worse than the modified ROX index using and 12 hours of treatment in the validation cohort). It should the Pa O 2 . also be noted that the ROX index measured after 12 hours Finally, we thank Dr. Tulaimat for his interest in our of treatment predicted better HFNC outcomes than the manuscript. As he suggested, we provided the scatterplot FIO2 value (area under the receiver operating characteristic curve with the respiratory rate on the x-axis and the SpO2/FIO2 [AUROC], 0.752 [0.664–0.840] vs. 0.672 [0.574–0.770]; P = 0.008). on the y-axis, with the successes and failures in different colors Therefore, even though mathematically the figure is totally valid, and the isopleth with a slope of 4.88 marked in black (Figure 1). from a practical standpoint, most patients may benefit from ROX As the range of the respiratory rate is narrow, we have also estimation rather than direct use of the set FIO2. tested the diagnostic accuracy of the ROX index with the On the other hand, we also provided other cutoff points for respiratory rate squared. However, no differences in the AUROC predicting HFNC failure: 2.85, 3.47, and 3.85 at 2 hours, 6 values between the ROX index and the squared ROX index hours, and 12 hours after initiation of HFNC therapy. We were observed (AUROC, 0.752 [0.664–0.840] vs. 0.753 suggest that the ROX index should be monitored over time. In [0.665–0.840]; P = 0.968). Therefore, as the use of a squared fact, if the value of the ROX index is between 3.85 and 4.88 after respiratory rate for the calculation of the ROX index does not 12 hours of treatment, the ROX index could be reassessed in 1 or provide any additional benefit in terms of prediction of HFNC 2 hours. Then, if the score has increased, the patient may have a success or failure, the use of the normal ROX index may be greater likelihood of success. In contrast, if it has decreased, the recommended. n patient may be more likely to fail. Finally, if the score is unchanged, another reassessment could be performed after one more hour. We also thank Drs. Karim and Esquinas for their interest Author disclosures are available with the text of this letter at in our paper. We fully agree that oxygen saturation as measured www.atsjournals.org. by pulse oximetry (SpO2)/FIO2 does not always correlate well Oriol Roca, M.D., Ph.D.* with PaO2/FIO2. In fact, when this correlation was assessed in Vall d’Hebron University Hospital patients with acute respiratory distress syndrome, measurements Barcelona, Spain with SpO2 values . 97% were excluded from the analysis and because the oxyhemoglobin dissociation curve is flat above these Instituto de Salud Carlos III levels (3). Therefore, as the authors pointed out, theoretically, one Madrid, Spain 118 American Journal of Respiratory and Critical Care Medicine Volume 200 Number 1 | July 1 2019 CORRESPONDENCE Berta Caralt, M.D. 3. Rice TW, Wheeler AP, Bernard GR, Hayden DL, Schoenfeld DA, Ware Vall d’Hebron University Hospital LB; National Institutes of Health, National Heart, Lung, and Blood Barcelona, Spain Institute ARDS Network. Comparison of the SpO2/FIO2 ratio and the and PaO2/FIO2 ratio in patients with acute lung injury or ARDS. Chest Universitat Autònoma de Barcelona 2007;132:410–417. Barcelona, Spain Copyright © 2019 by the American Thoracic Society Jean-Damien Ricard, M.D., Ph.D. Hôpital Louis Mourier Colombes, France UMR 1137 Paris, France Erratum: Healing Pulmonary Rehabilitation in the and United States: A Call to Action for ATS Members Sorbonne Paris Cité Paris, France There is an error in the editorial by Garvey and colleagues (1), ORCID IDs: 0000-0003-3841-4551 (O.R.); 0000-0002-1224-4761 (B.C.); published in the April 15 issue of the Journal. The third sentence 0000-0003-1828-2299 (J.-D.R.). of the third paragraph lists a group of advocacy organizations; *Corresponding author (e-mail: oroca@vhebron.net). the American Thoracic Society was incorrectly omitted from this list. n References Reference 1. Roca O, Caralt B, Messika J, Samper M, Sztrymf B, Hernández G, et al. An index combining respiratory rate and oxygenation to predict 1. Garvey C, Novitch RS, Porte P, Casaburi R. Healing pulmonary outcome of nasal high flow therapy. Am J Respir Crit Care Med [online rehabilitation in the United States: a call to action for ATS members. ahead of print] 21 Dec 2018; DOI: 10.1164/rccm.201803-0589OC. Am J Respir Crit Care Med 2019;199:944–946. 2. Frat JP, Thille AW, Mercat A, Girault C, Ragot S, Perbet S, et al.; FLORALI Study Group; REVA Network. High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure. N Engl J Med 2015;372:2185–2196. Copyright © 2019 by the American Thoracic Society This article is open access and distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives License 4.0 (http://creativecommons.org/licenses/by-nc-nd/4.0/). For commercial usage and reprints, please contact Diane Gern (dgern@thoracic.org). Correspondence 119