Levels of end-tidal carbon dioxide are low despite normal respiratory rate in individuals with long COVID

To the Editor,

Debilitating symptoms including fatigue, post-exertional malaise, brain fog, weakness, headache and dizziness can persist for over six months in individuals with long COVID [1, 2] (also referred to post-acute coronavirus disease 2019 (COVID-19) syndrome, post-COVID condition or post-acute sequelae of COVID-19) with many still experiencing these symptoms for 12 months after acute COVID-19 infection [1, 2]. Long COVID can occur despite less severe initial infection or hospitalization, and in the absence of obvious organ damage [1, 2]. Low end-tidal carbon dioxide (ETCO₂) in an alert, spontaneously breathing individual can result from hyperventilation, and has also been documented in patients with postural tachycardia syndrome, with or without hyperventilation [3]. Of note, some of the symptoms attributed to long COVID are also observed during hyperventilation and subsequent hypocapnia [4]. While there is potential for hypocapnia to be observed post COVID-19 infection in the presence of more severe sequelae such as microthrombi, pulmonary embolism and lung fibrosis [5], data are presented from a cohort of patients with long COVID who were mostly absent of severe sequelae following infection.

ETCO₂ and respiratory rate (RR) were measured in patients attending an outpatient clinic using an EMMA capnograph (Masimo Corporation, Irvine, CA). Approval for the collection and publication of this data was provided by the relevant Institutional Review Board (STUDY-20-001537). All patients were previously reviewed by a physician and diagnosed with confirmed (via PCR and/or antibody test) or probable (according to World Health Organization and Center for Disease Control and Prevention guidelines [6, 7]) past COVID-19 infection. The criteria for diagnosis of long COVID was the presence of persistent symptoms for at least 12 weeks after initial infection [2]. The capnograph was calibrated and a healthy control was used to check device readings at the commencement of each clinic. At the time of assessment, patients were seated and at rest, and had not performed other assessments or procedures immediately prior. Patients were instructed to breathe orally through the capnograph adapter at their resting rate and depth for a minimum of 15 s, and until the ETCO₂ level was stable for a further period of at least 5 s. Data were analyzed using Stata (StataCorp LLC, TX), with descriptive statistics and Pearson correlation coefficient provided.

In total, 82 patients with long COVID completed the assessment (table 1). The most common comorbidities reported were anxiety (n = 19, 23%), asthma (n = 11, 13%), depression (n = 10, 12%) and inflammatory bowel disease (n = 7, 9%). Six (7%) patients were hospitalized during acute COVID-19 infection, one (1%) patient had a pulmonary embolism, and three (3%) patients had subsequently developed pulmonary fibrosis. The median and interquartile range (IQR) for ETCO₂ were 32 and 8 mmHg respectively (table 1). The median (IQR) RR was 13 (6) breaths per minute. ETCO₂ was below 35 mmHg in 61 (74%) patients. The level of ETCO₂ and RR were moderately negatively correlated (r = −0.4, p < 0.01). In summary, the ETCO₂ levels observed were lower than the expected range of 35–45 mmHg, despite on average, RR falling within the normal range. It is apparent that hypocapnia is present in the absence of hyperventilation in patients with long COVID following less severe acute infection, potentially contributing to or as a manifestation of the underlying pathophysiology.

The possible causes of hypocapnia in this cohort may include increases in baseline respiration rate compared to pre-COVID-19 infection, disruption of the autonomic nervous system that may also be the cause of many long COVID symptoms, or organ damage that may not be detectable on routine clinical imaging and assessment. The potential limitations of these data include measurements performed in a clinical setting and in the absence of a control group, and a lack of pre-COVID-19 infection data. Further, while the capnograph used is generally comparable with traditional standing monitors, it has previously been demonstrated to slightly under-read ETCO₂ levels [8]. The causes and role of lower ETCO₂ levels in long COVID warrants further investigation.

This research was supported by the RTW Foundation.

The data that support the findings of this study are available upon reasonable request from the authors.