This study demonstrates that certain vasculopathies are associated with significant differences in short-term outcomes for COVID-19 hospitalizations. Specifically, Kawasaki disease (aOR 7.9115, 95% CI [1.4612–42.8362]; p = 0.0164), microscopic polyangiitis (aOR 3.5589, 95% CI [1.1356–11.1529]; p = 0.0295), thrombotic microangiopathy (aOR 3.9213, 95% CI [2.1959–7.0027]; p = 4.00E-06), other specified necrotizing vasculitis (aOR 20.9423, 95% CI [4.5566–96.2510]; p = 9.44E-05), unspecified necrotizing vasculopathy (aOR 16.6108, 95% CI [2.0106–137.2299]; p = 0.0091), and cryoglobulinemia (aOR 8.5863, 95% CI [2.2848–32.2669]; p = 0.0015) were all found to have significant positive associations with all-cause in-hospital mortality. Many of these vasculopathies similarly carried significant positive associations with regards to mechanical ventilation, specifically microscopic polyangiitis (aOR 7.4010, 95% CI [2.9088–18.8311]; p = 2.72E-05), thrombotic microangiopathy (aOR 3.5112, 95% CI [2.1457–5.7459]; p = 6.06E-07), other specified necrotizing vasculitis (aOR 34.9851, 95% CI [6.5912–185.6968]; p = 3.06E-05), and unspecified necrotizing vasculopathy (aOR 18.3357, 95% CI [2.2010–152.7467]; p = 0.0072). Furthermore, the relative difference in length of stay was 53% higher in COVID-19 hospitalizations with vasculopathies than those without (RD 1.53, 95% CI [1.35–1.74]; p = 6.18E-11); similarly, the relative difference in total hospital charges (2020 US dollars) was 97% higher (RD 1.97, 95% CI [1.58–2.46]; p = 2.18E-09); these reported secondary outcomes may serve as a proxy for morbidity in these populations. Such observed positive associations may be explained by the interplay of known disease mechanisms.
One proposed method of comorbidity between COVID-19 and vasculopathies is thrombogenesis. COVID-19 infection is known to result in the activation of both platelets and the coagulation cascade, increasing the risk of thromboembolic events [3]. While thrombosis is a feared complication of any vasculopathy, infection with COVID-19 for patients with vasculopathy may compound the likelihood of venous and arterial thrombotic events and result in worse outcomes. This is suggested by the observed positive association of thrombotic microangiopathy (aOR 3.9213, 95% CI [2.1959–7.0027]; p = 4.00E-06) with mortality in this study. Meanwhile, the formation of neutrophil extracellular traps (NETs), which release proinflammatory proteins, due to COVID-19 infection may induce or exacerbate autoimmune pathologies, such as vasculitides [14]. Prior studies have linked the accumulation of NETs to the pathogenesis of ANCA-associated vasculitides, and our study demonstrates a statistically significant positive correlation of deaths and respiratory ventilation in microscopic polyangiitis. Interestingly, no statistically significant correlation was observed for eosinophilic granulomatosis with polyangiitis and granulomatosis with polyangiitis.
COVID-19 infection has been shown to induce endothelial dysfunction. Studies have demonstrated viral tropism directed towards the vascular endothelium that disrupts anti-inflammatory mediators, increases monocyte activation, and releases reactive oxygen species [1]. As a result, this may further compromise endothelial integrity and precipitate or exacerbate endotheliitis in the form of vasculitis [16]. An important consideration is that a number of hospitalizations in this study likely contained immunocompromised individuals due to ongoing immunosuppressive therapy including systemic steroids, a treatment of choice for many vasculitides. Paradoxically, this standard of care may have unfortunately predisposed them to worse outcomes. Interestingly, while Kawasaki disease has been implicated in a number of COVID-19 case reports and has a known predisposition to affect the pediatric population, our data demonstrates a statistically significant positive correlation with mortality in adults with COVID-19 (aOR 7.9115, 95% CI [1.4612–42.8362]; p = 0.0164).
However, several vasculopathies appeared to demonstrate no statistically significant effect on the examined outcomes for COVID-19 hospitalizations. Notably, both large-vessel vasculitis, giant cell arteritis (aOR 1.3055, 95% CI [0.7109–2.3975]; p = 0.3901) and Takayasu arteritis (aOR 3.3389, 95% CI [0.3049–36.5621]; p = 0.3236) did not have statistically significant correlations with all-cause in-hospital mortality. There was also no statistically significant association between granulomatosis with polyangiitis and outcomes for COVID-19 hospitalizations, which is notable, considering the lower respiratory involvement that is characteristic of granulomatosis with polyangiitis and strongly implicated in COVID-19. Meanwhile, thromboangiitis obliterans (aOR 0.0002, 95% CI 0.0001–0.0005; p = 5.67E-55), Behçet's disease (aOR 0.0004, 95% CI 0.0002–0.0008; p = 2.16E-107), IgA vasculitis (aOR 0.0009, 95% CI 0.0003–0.0024; p = 1.62E-44), and livedoid vasculitis (aOR 0.0016, 95% CI 0.0002–0.0117; p = 1.93E-10), other vasculitis limited to the skin (aOR 0.0002, 95% CI [0.0001–0.0006]; p = 4.60E-73), unspecified vasculitis limited to the skin (aOR 0.0007, 95% CI [0.0002–0.0023]; p = 1.76E-32), and hypersensitivity angiitis (aOR 0.0002, 95% CI [0.0001–0.0005]; p = 7.34E-85) were associated with improved mortality outcomes. Though the results may imply that these vasculopathies confer some form of protective effect on outcomes for patients with COVID-19, it is more likely that these results may be a function of the limited sample size examined for these conditions, as many of these cohorts contained 0 hospitalizations in their respective mortality and mechanical ventilation outcome groups over the brief time period examined. As such, further investigation with subsequent data is required to draw more definitive conclusions.
In this study, several baseline characteristics were skewed towards those with vasculopathies. Specifically, those that were female (61.70% vs 47.39%, respectively), Asian/Pacific Islander, (4.26% vs 3.24%, respectively), and Native American (2.13% vs 0.95%, respectively) were more likely to suffer from vasculopathies, consistent with existing socioeconomic disparities of health [15].
The relative difference in length of stay (days) was 53% higher in COVID-19 hospitalizations with vasculopathies than those without (RD 1.53, 95% CI [1.35–1.74]; p = 6.18E-11); similarly, the relative difference in total hospital charges (2020 US dollars) was 97% higher (RD 1.97, 95% CI [1.58–2.46]; p = 2.18E-09). These positive secondary outcome associations were likely due to escalated treatment and medication costs as a result of disease burden.
The NIS is the largest publicly available national database that delineates the outcomes of hospitalizations with the diseases of interest to this study. However, there are several limitations to this study linked to the nature of this database. First, the NIS has only compiled and released one year’s worth of data related to the COVID-19 pandemic, limiting this study to only analyze short-term outcomes. As such, for those who may have contracted the disease towards the middle or end of the year, the outcome endpoints reflect less than a full year’s worth of time after infection. Furthermore, vaccines began to be administered towards the end of 2020, albeit in very limited quantities, which likely had a sizable impact on measured outcomes. Next, the NIS relies on ICD-10 coding for determining diagnoses. COVID-19 was represented with the ICD-10 code U07.1, but some healthcare providers may have used codes such as ICD-10 B97.29 (for other coronaviruses, used as an interim code before April 1, 2020), ICD-10 B34.2 (Coronavirus infection, unspecified), or facility-specific interim codes, which may have resulted in a non-fully exhaustive inclusion criteria. On a similar note, this study relies on accurate ICD-10 coding of the selected vasculopathies. In addition, several of the ICD-10 vasculopathies with significant results are either general or unspecified categories (e.g. other specified necrotizing vasculitis and unspecified necrotizing vasculopathy). These limiting factors may prevent a more refined understanding of which vasculopathies in particular are truly associated with worse morbidity and mortality in COVID-19 hospitalizations. Lastly, the NIS does not provide patient-specific information, which prevents our analysis from accounting for disease timing, disease severity, treatment regimen, or cause of death in this analysis.
Future studies that adjust for more granular patient information and outcomes may provide a more robust analysis to help guide treatment decisions that consider both diagnoses. Additionally, pursuing the emerging subtype of vasculitis related to COVID-19 infection, referred to as COVID-19-associated vasculitis, may yield additional insight into the mechanism behind its pathogenicity [18]. It may also be worthwhile to examine the impact of COVID-19 vaccination in patients with vasculopathies, given that several cases of vasculitides have been described in response to COVID-19 vaccination [2, 6, 7]. Finally, additional studies are required to corroborate the associations noted in our analyses, since a number of vasculopathies recorded zero observations of primary outcomes.