Retrospective analysis of patients with cardiopulmonary symptoms in the setting of Long COVID syndrome: investigating risk factors

Methods

A retrospective cohort study examined patients seen in the post–COVID-19 pulmonary clinic or post–COVID-19 virtual recovery center between April 2021 and September 2021. This study was reviewed by the ChristianaCare Institutional Review Board (IRB registration number: IRB00000480; FWA00004301) and deemed exempt (protocol number CCC-40191). Patients who were diagnosed with COVID-19 and developed persistent symptoms were subsequently referred to the post–COVID-19 pulmonary clinic or the post–COVID-19 virtual recovery center. Inclusion criteria for being evaluated in the clinic were COVID-19 diagnosis at least 12 weeks prior (8 weeks if patient was newly on home oxygen), recent chest x-ray or computed tomography (CT) scan of the chest, and available pulmonary function testing results. Exclusion criteria include symptoms attributable to diagnosis other than COVID-19, or failure to follow up in the clinic. We assessed patients’ COVID-19 severity by oxygen requirements while hospitalized, duration of oxygen support, and hospital length of stay. Treatment modalities were assessed and included steroids, intravenous immunoglobulin (IVIG), monoclonal antibodies, Interleukin-6 (IL-6) inhibitors, remdesivir, convalescent plasma, azithromycin, and nirmatrelvir-ritonavir (sold under the brand name Paxlovid). Symptoms that prompted referral included dyspnea, cough, fatigue, exercise intolerance, palpitations, and chest pain. Patients were screened for pre-existing pulmonary, cardiovascular, and diabetes symptoms. Smoking history was also considered. Analysis was performed on these data utilizing Cox regression analysis.

Results

Three hundred ninety (390) patients were referred to either the post–COVID-19 clinic or post–COVID-19 virtual recovery center. Among those patients, 144 patients were excluded. As a result, 246 patients were included in the retrospective analysis and identified with Long COVID symptoms greater than 12 weeks post–COVID-19 infection. Descriptive statistics of the cohort summarized in Table 1 including details regarding demographics, pre-existing comorbidities, smoking history, and symptoms prompting referral to post–COVID-19 follow-up. Table 2 summarizes data regarding rates of hospitalization, oxygenation modalities utilized while inpatient, and treatments received.

Cox regression analysis was performed showing an association in patients who had obstructive sleep apnea (OSA) and required oxygen, noninvasive ventilation, or intubation (odds ratio [OR] 3.6, 95 % confidence interval [CI], 1.70–7.65, p=0.0012). There was also an association with chronic obstructive pulmonary disease (COPD) and oxygen requirements in patients who had developed Long COVID symptoms (OR 12.19; 95 % CI, 2.38–62.33; p=0.0015). These data are summarized in Figure 1.

Figure 1:

Cox regression analysis showed that in patients who had an in-hospital oxygen requirement along with chronic OSA (OR 3.6; 95 % CI, 1.70–7.65; p=0.0012) or COPD (OR 12.19; 95 % CI, 2.38–62.33; p=0.0015), there was an association with the development of long COVID symptoms.

Discussion

Some of the earlier literature investigating the characterization of Long COVID syndrome, risk factors, and demographics was published from Sudre et al., [17] utilizing the COVID Symptom Study application. They were able to analyze and report on prospective observational data from a cohort of 4182 people who had tested positive for COVID and who participated in long-term symptom reporting [17]. Long COVID demographic predictors included increasing age, body mass index (BMI), and female sex [17]. They also found that experiencing more than five symptoms during the first week of illness was associated with Long COVID [17]. Unfortunately, their study was limited somewhat by selection bias, being limited to those who interacted with the symptom reporting application. Although age was the strongest predictor, their study population was relatively young, and they did not report specifics on the comorbidities that contributed to prediction in older adults (>70) [17]. Now, there are enough studies to invite large meta-analyses. Luo et al. [18] meta-analyzed 211 studies, 97 of which investigated Long COVID risk factors. They identified several commonly reported risk factors, but they did not pool ORs, and they did not distinguish between different pre-existing comorbidities [18]. Hou et al. [19] included 429 studies published from 2021 to 2024 in their meta-analysis. However, when analyzing risk factors, the number of studies available was much smaller overall, only allowing them to pool ORs from 5-22 different studies [19]. They were able to find a significant OR of 1.50 (95 % CI, 1.24–1.81) for pre-existing cardiovascular disease, and a nonsignificant OR of 1.33 (95 % CI, 0.94–1.88) for COPD from five studies each [19]. Interestingly, not much has been reported on pre-existing respiratory comorbidities as a risk factor, despite it having been shown to be a significant risk factor for long-term effects in the past with patients hospitalized for community-acquired pneumonia and other coronavirus infections [14], 15]. Recent studies have described a correlation between pre-existing lung disease such as asthma, COPD, and OSA with increased risk of developing Long COVID [20], [21], [22].

In our retrospective study, we found no specific association between cardiopulmonary symptoms of Long COVID and underlying cardiovascular conditions (CAD, hypertension [HTN], atrial fibrillation), diabetes mellitus, obesity, hospital oxygen requirement, or length of stay. Underlying COPD or OSA with a requirement of oxygen delivered by any modality was correlated in our study with the development of long-term cardiopulmonary symptoms of Long COVID. With COVID being known to exacerbate certain chronic conditions, there is a potential pathophysiologic basis for this association. Specifically, pre-existing parenchymal damage and predisposition to exacerbation of immune- and inflammatory-related lung pathology (like in the case of COPD) or chronic intermittent hypoxia and sympathetic activation (as in the case of OSA) could perhaps place patients at higher risk for developing clinically relevant long-term sequelae and therefore a diagnosis of Long COVID [21].

Long COVID continues to prove to be a challenging diagnosis owing to the diversity of presentation and symptoms, the difficulty distinguishing it from re-inoculation or new viral infection, and the fact that it presents itself in patients with varying severity of initial infection (even those with mild symptoms who were never hospitalized, or even those who were asymptomatic). Even at the time of our investigation, we have rapidly changing definitions and diagnostic criteria and varying results regarding pre-existing pulmonary disease as a risk factor [8], [19], [20], [21], [22]. Therefore, it is a condition that could benefit from further inquiry moving forward.

This study was limited by its retrospective nature and is subject to bias due to confounding factors. For example, those patients with pre-existing COPD or OSA may have had previous established relationships with pulmonologists who were more likely to suspect or diagnose Long COVID and refer them to the Long COVID clinic. We are also not able to tell if these patients had baseline symptoms prior to a COVID infection; despite being baseline, that may have contributed to a clinical diagnosis of Long COVID. In 2021, when these patients were being referred to the COVID follow-up clinic, we did not have access to specific criteria or indices for the diagnosis of Long COVID that exist today [8], 23]. Patients were referred to the clinic for dyspnea, cough, fatigue, exercise intolerance, palpitations, and chest pain. Although these symptoms align with our current understanding of Long COVID symptomatology, they are not inclusive of other Long COVID manifestations that we now know exist. Symptoms were identified subjectively by clinicians. Breathlessness or dyspnea in Long COVID could be related to anything from hypoxia, to breathing pattern disorder, to inducible laryngeal obstruction [24], 25]. Chest pain and palpitations are also symptoms related to a wide variety of syndromes. We relied on physician judgment to diagnose Long COVID based on the symptoms, and we excluded patients from the COVID clinic if their symptoms were attributed to another diagnosis.

The exclusion of alternative diagnoses continues to be a critical aspect of the diagnosis Long COVID, but it is an area of ambiguity [26]. This aspect is acknowledged but not directly addressed in current definitions and diagnostic indices [8], 23]. Our study population may have differed based on current definitions. Further investigation with larger, multi-center, forward-facing studies with benefits of current consensus and clinical definitions, blinding, and long periods of follow-up are potential next steps for this line of inquiry. An overall understanding of this disease would also benefit from further investigation into the pathophysiology of COVID-19 infection as it pertains to the long-term sequelae on lung parenchyma and how this affects patients with pre-existing lung disease, as well as pathophysiologic processes to help prevent the development of Long COVID. For example, future studies may include vitamin D screening because some existing literature suggests that vitamin D deficiency has been associated with severity and mortality of Long COVID given that 1,25(OH)2 D3 serves as a role in migration of dendritic cells and cytokine and chemokine production [27].

Conclusions

This analysis found that patients in this cohort who required oxygen support during their initial hospitalization with COVID-19 and with a previous history of OSA or COPD developed cardiopulmonary Long COVID symptoms at higher rates. This suggests that pre-existing pulmonary conditions and the severity of initial COVID-19 illness could be influential in the development of Long COVID. These findings may assist clinicians in identifying patients who are at higher risk of developing Long COVID syndrome (perhaps those with severe disease requiring oxygen and pre-existing lung disease such as COPD, and OSA), which may inform current management of patient expectations and future investigation into potential strategies for prevention.