Capillary leak syndrome is defined as the combination of severe hypoalbuminemia with diffuse pitting edema, exudative serous cavity effusions, non-cardiogenic pulmonary edema, and hypotension which, in the most serious cases, can progress to resistant hypovolemia and shock [7,50,51]. Its incidence depends on the underlying cause; however, as these symptoms are frequently encountered in the course of many diseases, the syndrome is often underdiagnosed. Regardless of the etiology, however, the exaggerated endothelial permeability to proteins lies in the center of its pathophysiology [52], which, in turn, seems to be the result of an increase in pro-inflammatory cytokines in most cases. Cadherin is a fundamental component of adherent junctions, the main type of junctions through which endothelial cells bind to neighboring cells. Mild inflammatory stimuli cause cadherin internalization, weakening these junctions; as inflammation proceeds, adherent junctions are further disrupted, leading to gaps between endothelial cells and a massive increase in protein permeability [53]. In a study by Atkinson et al., it was shown that, in the acute phase of the syndrome, substances of up to 900 kDa in molecular weight were extravasated into the interstitial space; under physiological conditions, the endothelial barrier is permeable to less than 5% of albumin, with a molecular weight of only 66.5 kDa [54].

Many diseases have been associated with capillary leak syndrome, including certain drugs, hemorrhagic fevers, ovarian hyperstimulation syndrome, and others [7]. However, sepsis leading to hemophagocytic lymphohistiocytosis (HLH) is the most prominent risk factor, and it seems that this is the case in severe COVID-19 disease. HLH is a condition that affects multiple systems and is characterized by an aberrant, uncontrollable immune response, which results in increased levels of pro-inflammatory cytokines such as IL-6, IL-2, (MCP-1), and TNF-α [55,56,57]. In COVID-19 infection, the virus binds to the ACE2 receptor in type I and type II alveolar epithelial cells in the lungs and then enters the cells through serine proteases such as transmembrane Serine Protease 2 (TMPRSS2). This enhances cell apoptosis and the production of inflammatory mediators; at the same time, increased apoptosis of lymphocytes leads to lymphocytopenia [58]. In addition, the apoptosis of alveolar epithelial cells leads to activation of the NOD-like receptor family pyrin domain-containing 3 (NLRP3) inflammasome [59]. The combined effect of these actions is a fulminant hypercytokinemia known as “cytokine storm”, a state which resembles HLH. In people with diabetes, the already pre-existing increased pro-inflammatory cytokine profile and oxidative stress could further aggravate the already aberrant immune response to COVID-19, while the immunosuppression per se could facilitate viral entry into endothelial cells [34]. Innate and adaptive immunity are compromised in diabetes and obesity, with an imbalance of CD4+ T helper cells toward Th17 and Th22 pro-inflammatory subsets and a persistent NET formation after infection, which prolongs and enhances the inflammatory cascade [60]. Finally, NLRP3 expression is increased in insulin-resistant states through the activation of the NF-kB pathway, which is combined with the release of C3a and C5a anaphylatoxins during SARS-CoV-2 infection [59].

Idiopathic systemic capillary leak syndrome (ISCLS—Clarkson’s disease) is another type of syndrome which seems to have an association with some cases of COVID-19 infection. First described in 1960 [61], it is an extremely rare disease, with about 260 cases reported worldwide until recently [62]. It is characterized by recurrent flares, with the trigger behind them not always being clear; however, inflammation has a robust association, and viral infections, such as influenza and respiratory syncytial virus, have been recognized in about half of the cases in recent cohorts [63]. With general, flu-like preceding symptoms being a frequent but inconsistent finding, patients with ISCLS rapidly develop the SCLS diagnostic triad, composed of the “3 H’s”: hypotension, hemoconcentration (hematocrit > 49–50% in men, 43–45% in women), and hypoalbuminemia (<3.0 g/dL), with shock and generalized edema due to huge plasma extravasation [62]. After a mean period of 1–3 days, the capillary endothelial barrier is spontaneously restored, with rapid fluid remobilization into the intravascular space, an effect that is often fatal due to subsequent pulmonary edema. Other causes of death include multi-organ failure due to hypovolemic shock and thromboembolic events due to hyperviscosity and hypercoagulability. While about 80% of the patients have a history of monoclonal gammopathy of undetermined significance (MGUS) [64], the pathogenetic role of the monoclonal protein is not clearly defined. In fact, vascular endothelial hyper-permeability is the main pathogenetic mechanism of the disease, although persistent structural abnormalities have not been found in biopsy specimens, indicating a reversible defect in endothelial integrity [62]. Elevated levels of pro-inflammatory cytokines such as CXCL-10 and IL-6 have been found in the acute serum of patients with ISCLS, and incubation of microvascular endothelial cells with acute serum reduced cadherin expression and thus led to a loosening of adherent junctions and endothelial gaps [65,66]. This effect implied the presence of a soluble factor that is responsible for the endothelial disruption, with possible candidates being VEGF and angiopoietin-2 (Ang-2), whose levels are elevated during disease flares [65,67]. These factors contribute to endothelial permeability in states such as sepsis [68,69,70]; at the same time, their levels are increased in diabetes and obesity and contribute to complications such as nephropathy and retinopathy, where endothelial hyperpermeability and degradation of the endothelial glycocalyx is a key factor [71,72]. Although an association between metabolic diseases and ISCLS has not been recognized, it is possible that the chronic endothelial dysfunction in these states could be a favorable background in a hyperinflammatory state like COVID-19.

Many studies have shown an increased incidence of hypoalbuminemia in severe COVID-19 patients; in fact, low albumin levels have a strong, positive association with poor disease outcomes [73]. Since the first months of the pandemic, a strong correlation between low albumin levels and adverse clinical outcomes was shown [73], and in a more recent meta-analysis by Soetedjo et al. [8], hypoalbuminemia was a risk factor for poor prognosis (OR: 6.97). While the mechanisms linking hypoalbuminemia with an adverse prognosis of patients with COVID-19 have not been directly addressed, several mechanisms have been proposed. In patients with sepsis, albumin modulates inflammatory status and oxidative damage and may also regulate effective plasma volume, circulation, and systematic perfusion [74,75,76]. Low albumin levels are also associated with lower oxygenation, especially in patients with acute respiratory distress syndrome [77].

However, the mechanisms underlying hypoalbuminemia in patients with COVID-19 are not fully elucidated. It has been demonstrated that impaired liver synthesis was not the trigger, as hypoalbuminemia occurred within the first 3 days of patients’ admission, far shorter than the half-life of serum albumin, while albumin concentrations were inversely correlated with inflammation markers such as CRP, suggesting that systemic inflammation and capillary leak syndrome was the main pathogenetic mechanism [78]. Taking into account the previously described correlations between metabolic disease, chronic endothelial dysfunction, and COVID-19, it would be expected that hypoalbuminemia could have a positive association with the presence of cardiometabolic co-morbidities in these patients. Indeed, in the study by Wu et al., median serum albumin concentrations were significantly low in patients requiring intermediate or intensive care hospitalization (20 g/L and 28 g/L, respectively), with the median BMI in both groups being 27.8 kg/m² and 26.6 kg/m², respectively; even more importantly, the presence of diabetes was associated with a higher rate of ICU admission [79]. No other significant correlations were found for other co-morbidities, implying that the endothelial dysfunction of insulin-resistant states was crucial in hyperpermeability and disease severity. In another cohort, hypertension was found in 38.7% of the patients with hypoalbuminemia compared to 17.1% in patients with normal albumin levels (p < 0.001), and the results were similar for the presence of diabetes (22.6% vs. 5.7%, p < 0.001) [78]. As in the previous study, no association was established with impaired liver function tests, pointing towards capillary leak as the main trigger. Serum albumin levels < 30 g/L were found in almost half of COVID-19 hospitalized patients in a report by Bassoli et al. [80]. Hypertension was present in 43.8% in the group with albumin < 30 g/L, compared to 24.5% in those patients with higher levels (p: 0.003); on the contrary, no such association was noted for diabetes, although it must be noted that its incidence was considerably lower in the total cohort. A statistically significant association between hypoalbuminemia and BMI, but not for diabetes, was found by Hundt et al., again independent of abnormal liver function [81]. Chen et al. in 482 hospitalized patients where hypoalbuminemia had a ratio of 53.7%, showed a statistically significant higher prevalence of diabetes in the hypoalbuminemia group, while hypertension was also more frequent, albeit marginally insignificant [82]. Even more interestingly, Viana et al. showed that hypoalbuminemia on admission (<34 g/L) was more frequent in COVID-19 non-survivors than survivors, and the same applied to the incidence of diabetes, hypertension, and dyslipidemia, all hallmarks of cardiometabolic disease [83]. In addition, serum albumin concentration showed linear correlation with highest hs-CRP (r = −0.306, p < 0.001), highest procalcitonin (r = −0.286, p < 0.001), and lowest lymphocyte count (r = 0.264, p < 0.001), indicating a clear link with the hyperinflammatory response in COVID-19, especially in the setting of chronic inflammation in metabolic disease. The fact that hypoalbuminemia was associated with mortality irrespectively of hs-CRP levels further indicates that albumin is not just a biomarker used in risk stratification but that capillary leak and the subsequent hypoalbuminemia and thrombo-inflammation have a direct pathogenic role in severe COVID-19 cases. Finally, in a report by Hirashima et al., albumin was significantly lower in critical and severe COVID-19 patients compared to milder cases, and hypertension and diabetes were again more frequent in the first group (p: 0.014 and <0.001, respectively) in the absence of any liver disorders [84]. Contrary to these results, the strong correlation of capillary leak syndrome in COVID-19 patients to the presence of cardiometabolic disorders has not been indicated in some other reports, where hypoalbuminemia was consistently associated with increased disease severity [8,85,86]; however, this possibly only implies that COVID-19 infection exerts its deleterious effects with several other immune and inflammatory mechanisms, with capillary leak being crucial, but not the only one.

Apart from incident hypoalbuminemia in the course of severe disease, another interesting link between cardiometabolic co-morbidities and capillary leak in COVID-19 is indicated through several SCLS cases. Pineton de Chambrun et al. were the first to report a fatal relapse of SCLS during a SARS-CoV-2 infection in a patient with a history of MGUS [87], and Lacout et al. reported the first SCLS episode elicited by COVID-19 in a patient with no medical history [88]. Since then, a few other case reports have been published, mainly regarding patients with a known history of Clarkson’s disease [89,90,91,92,93,94,95]. However, in the report by Case et al., the patient’s medical history was important only for the presence of systemic hypertension [92]. Similarly, in the report by Cheung et al., a 59-year-old woman died due to acute SCLS, beginning 6 days after the onset of respiratory symptoms due to COVID-19 infection; her medical history again included only hypertension, with no known history of SCLS or previous episodes of peripheral edema [90]. The presence of hypertension was also noted in a 58-year-old man who also had an SCLS history [89], while, in the report by Beber et al., obesity and Sjogren’s disease were noted in the absence of any previous syndrome flares [93]. On the other hand, the incidence of SCLS episodes after COVID-19 vaccination is higher, especially with adenoviral vector vaccines [94]; however, although in the few case reports thoroughly described in the literature, a previous history of SCLS or MGUS is a frequent finding, medical history, and other details are not known for many other cases reported by national regulatory authorities. However, a case by Yatzusuka et al. must be noted where the patient had a medical history of hypertension, obesity, and psoriasis, with the first two conditions having a well-established role also in psoriatic exacerbations apart from their possible role in aggravating the COVID-19 endothelial dysfunction and the subsequent disease flare [95]. In general, the possible contribution of cardiometabolic disease to the capillary leak in COVID-19-mediated SCLS flares is not yet robust, albeit notable. A possible explanation for this might be the fact that COVID-19 flares and typical SCLS episodes have some remarkable differences; for example, SCLS is not generally characterized by detectable damage to endothelial cells, while, on the other hand, endothelial injury in COVID-19 might not always be the aftermath of capillary leak, but rather the result of thrombo-inflammation and hypercoagulability [79,85].