Post-COVID-19 syndrome is a multisystem disease developing in patients with prior SARS-CoV-2 infection, characterized by a wide range of persistent clinical symptoms, occurring in hospitalized as well as in patients with relatively mild acute-phase illness (1–3). According to recent epidemiologic estimates, such an emerging condition is thought to affect 65–144 million individuals worldwide (4–6).

Since a notable portion of subjects affected by post-COVID syndrome reports lingering and debilitating symptoms (7, 8), such as dyspnoea and exertional intolerance (9), often associated with impairment of daily life activities (10), this new chronic health condition is expected to result in a considerable societal impact, potentially leading to economically relevant consequences, in terms of days off from work and utilization of healthcare resources and management (2, 6, 11). Appropriate management strategies specifically addressed to target post-COVID patients with clinically significant exertional dyspnoea should be established by healthcare systems and policymakers (1, 12–14). However, the limited capacity of healthcare systems would represent a paramount critical issue, in light of the significant restrictions and resource redirection from the usual chronic to acute healthcare settings, during the pandemic peaks (6, 15).

Furthermore, such a scenario is expected to become particularly challenging in those socioeconomic and/or geographic areas already facing shortages of medical equipment and care facilities. Many of Southern Italy’s regions were subjected to considerable cuts and healthcare restrictions to chronic respiratory disease management in the last decade, with consequent vulnerability to saturation of healthcare facilities (16, 17). When subjected to this additional strain, after the already-devastating pandemic waves, such healthcare systems might be led close to the risk of collapse (15–17). Knowledge of predictors for severe post-COVID syndrome-related dyspnoea might help identify high-risk patients and potentially relieve healthcare service from overload (1, 18).

However, the risk factors underlying dyspnoea associated with post-COVID syndrome are yet to be elucidated (1, 2, 19). Remarkably, an apparent lack of concordance between the presence of subjective exertional intolerance and results of pulmonary functional or radiological investigations has been observed in several studies (1, 18, 20), in that up to 35–65% of patients complained of dyspneic symptomatology despite normal pulmonary function test and chest CT imaging profile (14, 21–23). This study aimed to characterize a large cohort of post-COVID patients in Apulia, Italy, in the setting of a multi-disciplinary dedicated post-COVID Unit, to estimate the frequency of new or persistent dyspnoea in the post-acute-COVID-19-episode phase, in both hospital- and domiciliary-management patients, and to investigate predictors of post-COVID dyspnoea and reasons for lack of return to baseline health status at follow-up.

In the present study, we report a snapshot of the clinical features of a large cohort of patients referring to an outpatient service, specifically established for follow-up of Post-COVID Syndrome manifestations. At 3-month follow-up, patients displayed a wide range of clinical features, spanning from mild symptoms with little clinical relevance, or complete recovery, to the detection of long-term symptoms of considerable clinical significance. Our results evidenced that Post-COVID Syndrome, at 3-month post-infection remission, actually involved at least one reported clinical manifestation in a conspicuous portion of patients, with respiratory symptoms playing a pivotal role, as exertional dyspnoea revealed to be the commonest reported feature. PCED is a debilitating condition requiring targeted management, including frequent follow-up consultations and other-than-respiratory specialist evaluation (4, 12, 28, 29). A proper ongoing outpatient service is then needed in the appropriate healthcare setting, in the framework of a holistic multi-disciplinary approach, including pulmonary functional investigations, respiratory rehabilitation facilities (1, 15), and invasive and non-invasive imaging examinations (30–35). Such dedicated service would mitigate the post-COVID disease trajectory, thus potentially preventing the worsening of daily life and/or professional impairment, as well as avoiding future hospitalizations (15, 34).

Due to the high variability of PCED, comprehensive management cannot be achieved according to a “one-size-fits-all care” model (36). Rather, optimal care organization requires a personalized fashion, thus addressing different dyspnoea manifestations, severity degrees, and eventual co-morbid conditions (4, 6, 15). In this framework, it could be reasonable to think that PCED mainly develops in people with older age, with pre-existing respiratory conditions, or with acute-phase hospitalization courses (37–39). Although self-reported PCED was more frequent in patients with pre-COVID-19 respiratory co-morbidities in this cohort, our results suggest that a notable portion (57/190, 30.0%) of PCED-reporting patients were relatively young (under 50 yrs), most of whom were never admitted but remained at home and had an unremarkable medical history for pre-COVID-19 respiratory diseases (82.6%). Post-COVID multi-professional service should then sustain growing healthcare needs posed by different patients’ subsets, ranging from individuals with pre-existing respiratory conditions (whose potentially already altered lung parenchyma likely contributed to worsening acute-phase lung inflammation and consequent persistence of the exertional dyspnoea in post-COVID phase (40, 41)), to a considerable amount of relatively young and healthy dyspnoea-affected patients, who are expected to ask for increasing demand for adequate medical attention. Given their long life expectancy, the importance of ensuring adequate ongoing long-term outpatient post-COVID assistance in the latter patients’ subgroup is crucial (42, 43). Policymakers need to be aware that adequate supplies are required to potentiate healthcare delivery for PCED patients. However, expansion of healthcare capacity may be hard to achieve in those areas with resource limitations (44). In the last decade, the Apulia Region, Southern Italy, was subjected to diminished healthcare provision, reorganization of bed allocation, limitation of medical/healthcare personnel units, and reduction of respiratory rehabilitation infrastructures, with pandemic waves imparting further unprecedented strain, with consequent impaired capacity to sustain healthcare overload (16, 17).

As underlined in the results, an acute-phase hospitalization course is not a significant predictor of PCED. Furthermore, the persistence of acute-phase dyspnoea cannot fully account for the presence of PCED, since a notable portion of Post-COVID dyspneic subjects (19%) reported new-onset exertional dyspnoea, despite unremarkable acute-phase respiratory history. Our findings provide evidence that PCED is a heterogeneous nosological entity, with likely multiple underlying aetiologies. In the present study, instrumental characterization of respiratory function, based on 6-mwt, disclosed a high prevalence of PCED, even higher than subjectively reported dyspnoea, potentially due to either a higher sensitivity of instrumental 6-mwt-based approach or patients’ under-reporting during the medical interview. Noteworthy, we did not detect any statistically significant correlation between Post-COVID instrumental 6-mwt-based PCED and signs of actual pulmonary dysfunction, in terms of desaturation signs (DeltaSatHbO₂ ≤ −2%), thus indicating that PCED should not be regarded as a condition mainly involving lung damage or impairment. Rather, we found a statistically significant correlation between instrumental 6-mwt-based PCED and the frequency of post-COVID fatigue. Several studies also suggest that PCED is a wide-range disease, potentially encompassing several potential underlying conditions (6, 19, 43). Accordingly, at least two different phenotypes/mechanisms underlying PCED arise from our results. In a portion of individuals, PCED may represent a sequela associated with marked functional pulmonary involvement/damage and manifesting in the post-COVID phase as a walking drop in peripheral oxygenation. Such respiratory impairment might be primarily explained as chronic lung damage or, alternatively, as dysregulated inflammatory cytokine response in chest respiratory muscles or persistent pulmonary microvascular thrombosis and altered alveolar diffusion (37, 38, 42). Another subset of patients in our study had no detectable peripheral desaturation, thus clearly suggesting a different pathogenic mechanism. Accordingly, a number of recent reports described PCED as a consequence of decreased peripheral oxygen delivery or muscular consumption/deconditioning, in the absence of oxygen supply limitations (1, 20, 21). Several underlying mechanisms have been evoked, such as systemic microclotting/thromboinflammation, reduced metabolic oxidative capacity, virus-induced alterations in muscle tissue, and inactivity-induced muscle loss (18, 36, 45–49). Previous studies showed that a substantial proportion of Post-COVID Syndrome-affected patients with exertional dyspnea display radiological evidence of pulmonary interstitial disease, with heterogeneous aetiologies, such as pulmonary fibrosis conditions induced by the initial COVID-19 episode, previously identified interstitial fibrosis showing SARS-CoV-2-associated worsening/deterioration, as well as previously undiagnosed interstitial fibrosis unraveled by SARS-CoV-2 infection (50, 51). Furthermore, persistent interstitial disease and pulmonary embolism-related sequelae are also part of the dyspnea-associated spectrum post-COVID-related disease (33, 52, 53). It can be surmised that a fraction of dyspneic patients and 6-mwt-associated desaturation events might have been secondary to signs of a residual thin scattered area of the lung parenchyma involved by interstitial disease, although we could not test such an assumption since the radiological thoracic examination was not routinely part of our clinical protocol and data of chronic lung affection was not collected systematically in our cohort. Indeed, despite some chest CT abnormalities (ground-glass opacities, reticulations, interstitial thickening, fibrosis, and bronchiectasis) may persist 3 months after SARS-CoV-2 infection, only a weak correlation has been reported between abnormalities observed on “resting” investigations, such as imaging, and post-COVID exertional dyspnea (51–54).

Previous studies investigating risk factors behind PCED showed conflicting results (19, 55). In the present study, we set up a multivariate regression analysis, which disclosed a quite different risk profile between acute-phase dyspnoea and exertional dyspnoea in the post-COVID phase. Post-COVID fatigue, pre-existing respiratory co-morbidities, history of non-asthmatic allergy, age, and acute-phase dyspnoea emerged as independent statistically significant contributors to PCED, whereas hospitalization did not seem to play a significant role and only represented a risk factor for acute-phase dyspnoea, thus confirming findings of univariate analysis. Similarly, a clinical history of pre-existing cardiovascular co-morbidities represents a determinant for acute-phase dyspnoea, not for PCED, since it is well known that systemic hyper-inflammatory dysregulation in COVID acute phase is mainly responsible for rapid clinical deterioration in heart-disease patients, whereas it can be surmised that such condition will not play a major role in Post-COVID phase, characterized by a likely resolution of pro-inflammatory imbalance. In order to better characterize the risk factor profile, we also set up a sex-specific analysis. Our results showed a typical sex-oriented clinical profile in COVID, with increased risk of severe disease and hospitalization risk displayed by males during the acute phase, and greater frequency of long-lasting complaints occurring in females during the Post-COVID phase, in striking accordance with the literature (56–59). Sex-restricted multivariate regression analysis reveals a partially different risk profile underlying PCED, with post-COVID fatigue representing a female-specific risk factor, whereas pre-existing respiratory co-morbidities, age, and history of non-asthmatic allergy arose as male-specific determinants. Hence, considering the two above-described different aetiologies for PCED, the pulmonary-driven mechanism seems to display a male-to-female predominance. On the opposite, fatigue-driven PCED seems to be the predominant scenario in female patients, in both univariate and multivariate analysis. In accordance, fatigue is a manifestation that is typically more frequent in female patients, both hospitalized and at home, as shown in our study and previous studies (56). Female-to-male fatigue-driven increased risk of PCED seems to involve a relatively low pulmonary impairment, but likely represents a hallmark of Post-COVID Syndrome. A potential explanation for this female increased predominance finding could be related to potential hormonal influence (57–60). Based on the results reported in the present study, which disclosed the sex-specific independent predictors underlying dyspneic symptoms, we advocate that an optimization strategy should be deployed for personalized Post-COVID Syndrome exertional dyspnoea follow-up, according to a sex-specific protocol. In primary care settings for male patients with post-COVID dyspnoea, careful attention is needed for the presence of previous respiratory conditions, which should then prompt a higher priority toward specialty-care pulmonology management. In second-level pulmonary care, an appropriate work-up should include more invasive radiological and functional thoracic surveillance for these patients’ subgroups, in light of the male-specific greater risk of worsening previous respiratory diseases. On the other hand, since Post-COVID fatigue emerged as a major contributor to PCED in female patients, particular concern on this symptom is warranted by primary care physicians, aimed to activate higher priority scores for rapid evaluation in specialty-care settings, which would then be mainly focused on thorough investigation of fatigue in a female-specific protocol. In this framework, a comprehensive assessment of fatigue by a multi-professional healthcare team should consider functional muscle-weakness investigation, neuropsychological management, and non-invasive cardiopulmonary exercise testing, aside from standard pulmonology care.

Finally, our results are based on a cohort of patients infected during the first pandemic peaks characterized by the dominance of Alpha or Delta SARS-CoV-2 variants, whereas most recent pandemic waves (with the Omicron variant and its sublineages becoming dominant) displayed attenuated acute illness and mortality but increased spreading rate (61, 62). Despite more recent SARS-CoV-2 variants seeming to be associated with a reduced odd risk of post-COVID sequelae (63) compared to earlier variants, the high absolute numbers of infected people are expected to impose a non-negligible clinical burden and a considerable concern on healthcare organizations (64). Future studies are needed to assess the degree of overlap between the risk profiles for post-COVID in the different SARS-CoV-2 variants. Interestingly, a very recent inquiry highlighted female gender as the main post-COVID-underlying risk determinant independent of the viral strain (65), which would support a strategy based on sex-specific work-up protocol as a promising approach, not only for post-COVID management of people infected by initial-SARS-CoV-2 variants, but also for individuals suffering from long-term sequelae of more recent, or currently emerging, SARS-CoV-2 variants.

Our study has some limitations. Clinical data from the acute infection phase were retrospectively collected through medical history, which could lead to recall bias. Although the post-COVID service was available for all Post-COVID-Syndrome patients, regardless of symptoms or acute-phase healthcare setting, we cannot exclude potential referral bias towards dyspnea-affected patients as expected in a pulmonology clinical setting. Instrumental data are only based on 6-mwt technique, which mirrors an overall functional capacity and may potentially be affected by several confounders and was performed in a cross-sectional manner, thus resulting in an inevitably lacking pre-COVID assessment. It was not possible to systematically collect radiologic examinations in the recruited patients, which prevented us from more characterization of different phenotypic PCED clusters. Hence, most results are based on subjective patients’ reported outcomes, rather than objective assessment, since the observational nature of the study prevented us from carrying out a deeper investigation on metabolic aerobic fitness with a mainly experimental technique such as cardiopulmonary exercise testing.

Post-COVID exertional dyspnoea was revealed to be the commonest reported feature, potentially associated with a relevant clinical burden. A proper management strategy needs to be established by healthcare system institutions and policymakers, to sustain requirements for healthcare delivery and mitigate evolution towards chronicity. However, such growing healthcare demand will probably overload the insufficient capacity of the public health service in Italy and the Apulia Region, previously weakened by resource limitations and pandemic bursts. Knowledge of sex-specific risk-determining factors might help optimize personalized care management and healthcare resources.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Puglia Observatory for Epidemiology. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants' legal guardians/next of kin in accordance with the national legislation and institutional requirements.

ER: Conceptualization, Investigation, Writing – original draft. ECu: Conceptualization, Data curation, Methodology, Writing – original draft. PP: Data curation, Methodology, Validation, Writing – review & editing. CC: Conceptualization, Data curation, Investigation, Writing – review & editing. VS: Conceptualization, Data curation, Writing – review & editing. CS: Conceptualization, Methodology, Writing – review & editing. CP: Data curation, Formal analysis, Methodology, Writing – review & editing. FB: Data curation, Methodology, Writing – review & editing. MC: Data curation, Methodology, Writing – review & editing. MT: Data curation, Investigation, Methodology, Writing – review & editing. ECa: Data curation, Investigation, Methodology, Writing – review & editing. SL: Data curation, Investigation, Methodology, Writing – review & editing. LM: Data curation, Investigation, Supervision, Writing – review & editing. ST: Conceptualization, Supervision, Writing – original draft. OR: Conceptualization, Methodology, Writing – review & editing. GL: Conceptualization, Data curation, Formal analysis, Project administration, Writing – review & editing, Writing – original draft.