Reply: Inhaled Epoprostenol Through Noninvasive Routes of Ventilator Support Systems.

Abstract

Thank you Dr Li for your interest in our article, “Inhaled Epoprostenol Through Noninvasive Routes of Ventilator Support Systems,” and for the thoughtful comments regarding the role of high-flow nasal cannula (HFNC) on alveoli recruitment and drug delivery and the concerns regarding using PaO 2 /FiO 2 (PF) as an efficacy outcome. Although we agree that the use of HFNC could potentially reduce the rate of escalation to noninvasive or invasive ventilation in patients with acute respiratory failure, some randomized control trials did not confirm this benefit. Additionally, the exact duration of HFNC therapy that would prevent escalation of respiratory support has neither been validated nor well established. Unfortunately, in our retrospective study the duration of HFNC therapy prior to initiation of inhaled epoprostenol (iEPO) was not reported. However, at this medical center, the use of iEPO is generally used as a salvage therapy in patients who have failed traditional therapy. It would be extremely uncommon to initiate HFNC and iEPO simultaneously. This makes the findings of this retrospective analysis more pragmatic because clinicians would not withhold therapy until there is a response from HFNC therapy. We also agree that the flow rate of HFNC might influence the rate of aerosol deposition. At our institution, HFNC is managed by respiratory therapists based on a prespecified protocol, which titrates flow rates based on oxygenation. The logical workflow for a respiratory therapist would entail observing an improvement in oxygenation prior to decreasing flow. Hence, it is likely that the change in flow rates was a result of the response to iEPO, although negative feedback is certainly possible, where subsequent increased flow rates lead to even less drug delivery. The SpO 2 /FiO 2 (SF) ratio can be used as a surrogate PF ratio, which would have assisted with the missing PaO 2 values reported. However, there are several limitations with using SF ratio in our specific patient population. Patients included in this series had a baseline mean arterial pressure of 64 ± 11 mm Hg, and 25% of patients were receiving vasopressors. Administration of vasoactive agents decreases the accuracy of SpO 2 readings. It has been hypothesized that an oximeter reads the pulsatile venous flow because of the opening of the arteriovenous channels in septic shock patients. As such, oximeter readings might not be reliable in septic shock patients who are receiving vasopressors. Additionally, a high SpO 2 threshold is needed to detect significant hypoxemia with good sensitivity. A study by Khemani et al demonstrated that SF ratio can be a reliable surrogate for PF ratio as long as SpO 2 is 80% to 97%. Also, a target SpO 2 between 92% to 97% in noninvasive patients should be established. Patients included in this case series did not fit these criteria (baseline PaO 2 in HFNC and NIPVV were 69 ± 18 mm Hg and 79 ± 23 mm Hg, respectively, and overall SpO 2 at baseline was 91% ± 6%). Furthermore, other factors may alter the SpO 2 -PaO 2 correlation such as PCO2, pH, and temperature. These variables should be taken into account when evaluating SF ratio. As such, reporting SF ratio and SpO 2 would not be an accurate or reliable representation of the patients included in this evaluation. Finally, we would like to reiterate that the intent of the case series was not to prove a causal relationship between initiation of iEPO and respiratory response, but rather to illustrate the feasibility of such an approach and describe usual patient responses.