To determine the likelihood that clinicians know carbon dioxide levels before administering supplemental oxygen to patients with neuromuscular disorders, to quantitate the effect of oxygen therapy on carbon dioxide retention, and to explore hypercapnia contributing to the need to intubate and use of continuous noninvasive ventilatory support to avert it.
Basic proceduresA retrospective chart review for patients with neuromuscular disorders intubated or having intubation averted by using continuous noninvasive ventilatory support with carbon dioxide known pre- and during oxygen administration.
Main findingsFor only 2 of 316 patients who were intubated did clinicians know carbon dioxide levels prior to administering oxygen. For four cases, intubation was averted by continuous noninvasive ventilatory support and mechanical insufflation–exsufflation despite severe hypercapnia and acidosis. After initiating oxygen therapy, patients’ carbon dioxide partial pressures increased 52.1±42.0mmHg in over as little as 20min.
Principal conclusionsClinicians should attempt to use continuous noninvasive ventilatory support and mechanical insufflation–exsufflation rather than supplemental oxygen to normalize blood gases for neuromuscular ventilatory failure and should be prepared to intubate hypercapnic patients for whom oxygen is administered.
Although administering supplemental oxygen (O2) to hypercapnic patients with cardiopulmonary disease can decrease dyspnea, pulmonary hypertension, polycythemia, exercise intolerance, permit greater ventilator-free breathing particularly for chronic lung disease patients, and prolong life,1–5 it can also depress hypoxic ventilatory drive and exacerbate hypercapnia.2,3 With the advancing respiratory muscle weakness of patients with neuromuscular disease (NMD), however, it can render noninvasive ventilation ineffective,6 which can cause carbon dioxide (CO2) narcosis, coma, and acute on chronic respiratory failure.
In intensive care units (ICUs), since it is far more common to encounter patients with lung disease than NMDs, it is generally assumed that O2 administration is harmless for anyone with respiratory symptoms and that NMD patients must accommodate elevated baseline CO2 levels anyway so emergency services tend to administer O2 immediately to all such symptomatic patients without first determining CO2 levels.7 The only publication that has thus far quantitated O2-induced hypercapnia in NMD is that of Gay and Edmonds in 1995,8 and they could only identify the pre- and post-O2 delivery CO2 levels for eight patients including three with polymyositis, three with motor neuron disease, and one each with inflammatory motor neuropathy and chronic poliomyelitis. The patients received low-flow O2 and within 0.8–144h, three became obtunded and required intubation but subsequently recovered; and two died when intubation was declined. Neither noninvasive ventilatory support (NVS) nor mechanical insufflation–exsufflation (MIE) was used to either avoid intubation or facilitate successful extubation.
Bi-level positive airway pressure (PAP) is generally used in acute and chronic care at less than full ventilatory support settings. However, many patients with little or no vital capacity (VC) require continuous NVS (CNVS) at volume or pressure settings that fully support respiratory muscle function. Two studies reported 166 patients with respiratory pump failure who required intubation after having been administered O2 supplementation.9,10 Another 61 patients underwent tracheotomy after being intubated following O2 administration for acute on chronic respiratory failure.11 None of these 227 patients had been offered CNVS or MIE10 to either prevent intubation or facilitate extubation prior to transfer to a specialized unit for extubation or decanulation to CNVS and MIE. The purpose of this case series is to determine the likelihood that clinicians obtain CO2 levels before administering O2 to patients with NMD, to quantitate subsequent CO2 levels as a contributing factor in the need to intubate, and to describe the use of CNVS and MIE to reverse hypercapnia and avert invasive airway intubation.
MethodsConsecutive outpatients to a NMD clinic from 1996 to 2015 were screened for having received O2 therapy, undergoing intubation then possibly tracheotomy, then being extubated or decanulated to CNVS and MIE. In addition, we screened for others whose hypercapnia had been documented either by arterial blood gas sampling or by end-tidal CO2 (EtCO2) measurements before O2 therapy but who were spared intubation by using CNVS and MIE. The MIE was administered by both hospital staff via invasive airway tubes and by the patients’ relatives post-extubation/decanulation in the acute care setting up to every 20–30min during waking hours via oronasal interfaces at insufflation and exsufflation pressures of 50–60cm H2O until ambient air oxyhemoglobin saturation (O2 sat) baseline remained greater than 94%.
ResultsOf 2804 patients with NMD, 316 were identified who had undergone intubation, then 61 of 316 tracheotomy. All were subsequently extubated or decanulated to CNVS and MIE.9–11 Of these, only nine cases had had CO2 levels available both pre- and during O2 administration, and for only cases 1 and 4 was the CO2 known by the clinicians who administered the O2 and intubated the patients. For cases 2, 3, 7, and 9 whose hypercapnia worsened during O2 therapy, intubation was averted by using CNVS and MIE.
Case presentationsCase 1: A 27-year-old woman with nemaline rod myopathy had failed extubation following scoliosis surgery at age 10, but she was successfully extubated to CNVS and MIE upon transfer to our intensive care unit (ICU).9,10 From age 10 to 27, she used bi-level PAP at an inspiratory (I)PAP of 18cm H2O, expiratory (E)PAP of 4cm H2O, and rate 13breaths/min only during intercurrent upper respiratory tract infections (URI) despite having a VC as low as 340mL. In December 2014, she presented to emergency services with a URI and partial pressure of CO2 in arterial blood (PaCO2) of 47mmHg despite continuous use of bi-level PAP. Then, she was started on supplemental O2. Over a 2-h period, her PaCO2 increased to 67mmHg. She became obtunded, was intubated, and failed one extubation attempt to IPAP 8cm H2O, EPAP 4cm H2O, and supplemental O2. After nine total days of intubation, she was transferred to our ICU, where she was successfully extubated to CNVS and MIE. Her PaCO2 remained below 40mmHg, and she was discharged home 4 days later, where she eventually weaned to nocturnal nasal NVS.
Case 2: A 32-year-old wheelchair-dependent man with Becker muscular dystrophy began to use nocturnal nasal NVS in November 2013 to treat fatigue, sleepiness, difficulty concentrating, and hypercapnia. He had a VC of 550mL and an EtCO2 of 77mmHg. His symptoms cleared and diurnal EtCO2 decreased to 54mmHg, but since he refused to use NVS during daytime hours, he again developed mild confusion and presented to emergency services in February 2014. He was placed on 100% fraction of inspired O2 (FiO2) and became obtunded over the next 20min. An arterial blood gas (ABG) revealed a PaCO2 of 177mmHg, bicarbonate of 41meq/L, and pH of 6.98. The patient was manually resuscitated and immediate intubation was recommended. However, the father insisted that the author (Bach) be called first, and when it was discovered that the patient's portable ventilator was in the trunk of his car, the father retrieved it and administered nasal CNVS at a volume-preset of 1150mL and rate 12breaths/min. A few minutes later, the patient transitioned to mouthpiece CNVS and, 20min later, was discharged home with normal O2 sat, fully coherent, and in no distress. At his next outpatient visit in May 2014, now using CNVS, the patient's EtCO2 remained below 40mmHg and ambient air O2 sat was normal. When breathing unassisted for a few minutes, his EtCO2 increased to 46mmHg.
Case 3: A 15-year-old boy with spinal muscular atrophy type 1 (SMA), who was CNVS-dependent with a VC of 10mL and received all nutrition via a nasogastric tube since 8 months of age, was admitted to a pediatric ICU for respiratory distress despite pressure-controlled continuous mandatory ventilation (CMV) at 24cm H2O, rate 12breaths/min. At 5PM, his PaCO2 was 21mmHg and partial pressure of O2 in arterial blood (PaO2) of 42mmHg. He was placed on FiO2 of 32% along with CNVS. Thirteen hours later, his PaCO2 increased to 56mmHg with a PaO2 of 59mmHg. As his aspiration pneumonitis resolved, his FiO2 was reduced to 28%. Two hours later, his PaCO2 decreased to 49mmHg. He was discharged home with normal O2 sat and EtCO2 of 35mmHg. At a 2-week follow-up outpatient visit, his EtCO2 was 30mmHg, which had been his baseline.
Case 4: A 4-year-old boy with SMA2, baseline VC of 480mL, and EtCO2 of 46mmHg was admitted to a pediatric ICU for distress due to a URI. He was using volume-preset CMV at 750mL, rate 16breaths/min. His PaCO2 was 41mmHg. He was placed on FiO2 of 50% and became disoriented over the next 2h. A repeat ABG drawn on FiO2 of 40% showed a PaCO2 of 73mmHg and PaO2 of 67mmHg. He was then intubated and his PaCO2 normalized. Two days later, on FiO2 of 21%, he was extubated back to CNVS on CMV at 20cm H2O, rate 16breaths/min.
Case 5: A 60-year-old woman with facioscapulohumeral muscular dystrophy using nocturnal CMV (Vt 1000mL, rate 12breaths/min, FiO2 21%) had a VC of 640mL (sitting) and 480mL (supine) and EtCO2 of 39mmHg. On a routine visit to her pulmonologist, despite normal pulse oximetry, she was placed on 2L/min O2 and bronchodilators. Three weeks later, she was admitted for CO2 narcosis, intubated, and underwent tracheotomy. She used continuous tracheostomy mechanical ventilation (CTMV) for 4 months until she was decanulated to CNVS on her previous settings. Her EtCO2 remained 32–40mmHg using CNVS over the next six outpatient visits spanning 5 years despite her VC decreasing to 250mL.
Case 6: A 7-year-old boy with SMA1 and CNVS dependence since 4 months of age, who had an EtCO2 of 19–34mmHg over 11 outpatient visits and a VC of 15mL, was hospitalized for gastroenteritis. Although he had no respiratory symptoms, he was placed on FiO2 of 50%. Eleven hours later, on repeat ABG, his PaCO2 had increased from 43 on admission to 62mmHg with a PaO2 of 48mmHg, and he was intubated for 3 days. Once stabilized with normal O2 sat and CO2 on ambient air, he was extubated back to CNVS and discharged home 5 days later.
Case 7: A 32-year-old man with myotonic dystrophy using CNVS was admitted for aspiration pneumonia. On admission, with a PaCO2 of 40mmHg and PaO2 of 79mmHg, he was placed on FiO2 of 32%. Three days later, his PaCO2 was 58mmHg with PaO2 53mmHg. Then, O2 was discontinued and the PaCO2 normalized to 39mmHg. Three days later, he was discharged home using nocturnal-only nasal NVS.
Case 8: A 71-year-old woman with post-poliomyelitis ventilatory insufficiency, baseline VC of 500mL, and EtCO2 45mmHg was placed on 2L/min O2 to treat exertional dyspnea. Six weeks later, she was hospitalized for CO2 narcosis with a PaCO2 of 106mmHg. She was intubated, failed conventional weaning/extubation, and transferred for extubation to CNVS and MIE.9,10 Three days later, she was discharged home using sleep NVS on CMV at Vt 950mL, rate 12breaths/min. At a 3-day follow-up, her EtCO2 was 40mmHg with normal O2 sat.
Case 9: A 12-month-old CNVS dependent infant with SMA type 1 used nasal CNVS at 20cm pressure and her PaCO2 ranged from 37 to 43mmHg over a 5-day period. Ventilator weaning was attempted, but her PaCO2 increased to 59mmHg and O2 sat decreased at which point she was placed back on CNVS with low-flow O2. Three hours later, her PaCO2 was 168mmHg. Intubation was averted by discontinuing the O2, and use of CNVS quickly normalized her CO2 level to 36mmHg.
DiscussionConsistent with the report of Gay and Edmonds,8 who found a mean increase in CO2 of 28.2±23.3mmHg with O2 therapy for eight patients with NMD, the CO2 of our cases increased from 43.8 to 95.9mmHg or 52.1±42.0mmHg (range of 21–59 to 56–177) over a mean 17.4h (range of 20min to 6 weeks). Whereas Gay and Edmonds did not use NVS or MIE, our cases 2, 3, 7, and 9 avoided intubation because of it.
Case 1 became hypercapnic because bi-level PAP settings were inadequate to fully rest or support a patient with little VC. Normally compliant lungs require full tidal volumes or pressure settings of 18–20cm H2O. While O2 therapy exacerbates hypercapnia by suppressing hypoxic drive for patients breathing spontaneously, for nasal bi-level PAP or NVS users like cases 1, 3, 6, and 9, it resulted in increased ventilation leakage out of the mouth.6 Patients with little or no VC can maintain normal lung ventilation during sleep by using open systems of NVS provided that ventilatory drive is adequate to reflexively prevent excessive air leakage to avoid severe hypercapnia. This reflex activity occurs at the nadir of O2 desaturations and is suppressed by O2 administration.6 Other mechanisms for O2-exacerbated hypercapnia include worsening ventilation-perfusion matching secondary to attenuation of hypoxic pulmonary vasoconstriction, depression of diaphragm function,9 and decreased affinity of hemoglobin for CO2 (Haldane effect), which results in increased bicarbonate and dissolved CO2.12 Although acute increases in PaCO2 and H+ stimulate ventilation, as baseline PaCO2 and bicarbonate levels increase,13 further increases result in respiratory depression, obtundation, and agonal breathing as nasally insufflated air leaks out of the mouth. Likewise, nasal bi-level PAP is also less effective in the acute setting for patients receiving supplemental O2 for lung disease14–16 as it is for patients receiving heavy sedation or narcotics.6
A limitation of this study is that for cases 2 and 8, baseline EtCO2 was compared with PaCO2 during O2 delivery. Since people with NMD generally have normal lung tissue diffusion, EtCO2 differs minimally from PaCO2. Reports suggest that EtCO2 is 117 to 2–6mmHg less than PaCO2 (R-value of 0.94–0.98) unless cardiac output and lung perfusion pressures are decreased.18,19 Thus, EtCO2 can be conveniently used for home sleep monitoring, whereas PaCO2 necessitates indwelling arterial lines.
Administering O2 and pressure-preset bi-level PAP rather than volume-preset NVS precludes the use of active lung volume recruitment (air stacking),20 which along with failure to use MIE, prevented sufficient cough flows for cases 1 and 4 to avoid pneumonia and intubation. Cases 5 and 8 demonstrate that even chronic low-flow O2 therapy (2L/min) can result in CO2 narcosis. Although CTMV dependent, decanulation to CNVS permitted case 5 to wean back to nocturnal-only NVS. This effect of weaning from CTMV to part-time NVS is especially pronounced when hyperventilated, hypocapnic CTMV users are decanulated to NVS.21,22 Cases 1, 4, 5, 6, and 8 were all successfully extubated or decanulated to CNVS and MIE in ambient air despite having no ventilator-free breathing ability.
ConclusionFor only 2 of 316 intubated patients were CO2 levels known to the clinician administering O2 and intubating the patient. Physicians need to obtain CO2 analyses before administering O2 to NMD patients with respiratory distress and to attempt to resolve the problem by using CNVS and MIE. The O2 should only be administered if CNVS and MIE have failed to normalize ambient air O2 sat, in which case they should be prepared to intubate. Whereas hypercapnic lung disease patients have poor prognoses with a 5-year survival of 30% for patients whose forced expiratory volume-one second (FEV1) is less than 750mL,23 patients with NMD have been reported to have survived 20 to over 60 years using CNVS.20 Thus, O2 therapy should not be used as a substitute for NVS and MIE to maintain normal O2 sat (>94%). Avoiding O2 also permits oximetry to gauge hypercapnia, airway secretion encumbrance, and intrinsic lung disease as well as the effectiveness of NVS and MIE in treating them.24
Ethical disclosuresProtection of human and animal subjectsThe authors declare that no experiments were performed on humans or animals for this study.
Confidentiality of dataThe authors declare that they have followed the protocols of their work center on the publication of patient data.
Right to privacy and informed consentThe authors have obtained the written informed consent of the patients or subjects mentioned in the article. The corresponding author is in possession of this document.
Conflicts of interestThe authors have no conflicts of interest to report.
FundingThere was no external funding available for this work.
