Hypoalbuminemia: an underestimated, vital characteristic of hospitalized COVID-19 positive patients?

The COVID-19 pandemic has led to the greatest worldwide health crisis in decades. The number of infected patients with severe SARS-CoV-2 (COVID-19) disease has overwhelmed the capacity of almost all health care systems around world. Hypoalbuminemia has now been reported in patients with severe disease seeking help in the emergency room because of COVID-19 infection. In the past, hypoalbuminemia was considered to be a negative prognostic marker, not only in patients with chronic liver disease, but also in patients with SARS and MERS infections. Albumin is the major serum protein synthesized by the liver. A low serum albumin level is an ominous clinical sign. Introduction of amino acids to a patient’s diet is of fundamental importance to hepatic albumin synthesis in different clinical situations. This highlights the importance of nutritional support during the early phases of COVID-19-infection. Furthermore, albumin synthesis in the hepatocyte is downregulated at a pretranslational level by the direct interaction of the major acute-phase cytokines which are released into the circulation during the cytokine “storm” induced by the viral effects on the lungs. Both mechanisms contribute to severe hypoalbuminemia which, combined with massive fluid losses due to the fever, is responsible for severe hypovolemia and shock commonly observed in patients with COVID-19 in critical care settings.

cornonavirus-1 and Middle East respiratory syndrome coronavirus caused primarily pulmonary diseases. HuCoV 229E, 0C43, NL63 and HKU1 are mainly responsible for the common cold, but can also cause lethal nonspecific pneumonias [1] . However, SARS-CoV-2 has a wide range of clinical presentations, with acute respiratory distress syndrome being the often fatal pulmonary complication [2][3][4] .
Most of the publications reporting clinical characteristics for patients with SARS-CoV-2-infection originate from China, many from the city of Wuhan. These publications are descriptive retrospective case series about patients hospitalized with the virus or who died in intensive care units (ICU) [5,6] . The symptoms reported mainly concern the reason for hospitalisation. The spectrum of all symptoms, and key timings from when patients first felt unwell is less well reported [7,8] . In fact, far less is known about the symptomatology at the time of first appearance of the disease in hospitalized patients and in infected persons who remained at home, and who may had even died there.
Parameters indicating liver damage include prothrombin time, serum transaminase and bilirubin levels, acute-phase response markers such as leukocyte count. C-reactive protein, procalcitonin, and several serum cytokine levels have been reported in patients with SARS-CoV-2, together with changes in serum albumin levels [2][3][4][5]9,10] . Previous experiences in patients with SARS or MERS suggested that hypoalbuminemia, lymphopenia, a serum CRP level greater than 4 mg/dL, plus elevated lactate dehydrogenase on hospital admission were predictive for pneumonia progressing to respiratory failure [11][12][13][14] . Low serum albumin levels have now been found to be an important predictor of progression to severe disease and increased mortality in hospitalised SARS-CoV-2 positive patients of older age [15,16] .

PATHOPHYSIOLOGICAL ASPECTS OF HYPOALBUMINEMIA AND CLINICAL RELEVANCE OF ALBUMIN INFUSION
Albumin is a single chain protein with a molecular weight of 66 kDa made of 585 amino acids which represents more than 50% of the serum proteins and represents an important component of interstitial fluid. The albumin fraction was first separated from the other components of the plasma in 1944 by Edwin Cohn [17] , who also appreciated its strong oncotic properties. This characteristic of albumin was also confirmed by Scatchard et al. [18] in 1944. Serum albumin levels are used as useful surrogates of liver function [19] . Soon after the fractionation studies, intravenous albumin administration was performed in patients with advanced liver disease. This was done in the United States during the 1940' s [21,22] and also in the United Kingdom at the beginning of the 1960' s by Wilkinson and Sherlock et al. [22] .
The beneficial effect of prolonged administration was first demonstrated in a clinical trial by the group of Paolo Gentilini in Florence [23] , and more recently by Caraceni et al. [24] in Bologna.
The positive diuretic effect of albumin infusion in three patients with liver cirrhosis was published by Patek et al. [25] . This finding was subsequently corroborated in a group of ten patients [26,27] , showing that albumin infusion in patients with liver cirrhosis and ascites (without spontaneous bacterial peritonitis) increased sodium excretion in the urine, and led to weight reduction and a reduction in diuretics required.
It was shown that repeated daily intravenous administration of albumin was able to avoid the requirement for transjugular stent placement into the portal tract trough the hepatic vein (TIPS) [28] . A similar experience, in a larger patient numbers, was published by Trotter et al. [29] .
The positive effects of albumin infusion in cirrhotic patients with low levels of serum albumin was shown by Bajaj et al. [30] who observed a normalisation in serum sodium concentration in patients with liver cirrhosis and hyponatriemia. Infusion of intravenous albumin solution in decompensated cirrhotic patients was also able to reduce encephalopathic episodes and associated mortality [31] .
The prognostic importance of serum albumin levels in patients with liver disease is demonstrated by the inclusion of this parameter in the Child-Turcotte-Pugh score, used to assess the prognosis of chronic liver disease, mainly cirrhosis. This score was introduced by surgeons in 1963 [32] . In addition, serum albumin level is a key nutritional parameter used to estimate the grade of malnutrition, and to predict survival in patients with liver cirrhosis. Malnutrition is an independent risk factor for transplantation, and improves the prognostic value of the Child-Turcotte-Pugh score, reported by Alberino et al. [33] .
While administration of albumin in patients with advanced liver disease and hypoalbuminemia is now a standard therapy, albumin administration in critically ill patients with or without liver disease in the ICU is controversial [34][35][36] .  [37] The liver is the sole source of serum albumin [37] [Figures 1 and 2] which represents more than 50% of all proteins synthesized in the liver. Under normal conditions albumin synthesis in the hepatocytes is regulated by the amount of proteins reaching the intestine after each meal, and the amount of amino acids transported into the liver through the portal system.
During fasting, reduced albumin synthesis is due to a reduced uptake of amino acids into the hepatocytes [38] , which may be in part compensated by using amino acids from muscle proteins.
During acute phase situations, characterised by tissue damage induced by different insults such as trauma, bacterial infection, or viral infections such as SARS-CoV-2, the defence mechanisms of the body concentrate on eliminating the aggressive agent at the site of tissue entry and/or the damaged tissue. The main systemic reactions during the COVID-19 illness are fever, weakness and loss of appetite. In addition vomiting, diarrhea and abdominal discomfort [39] , which may be accompanied by loss of taste [40] and loss of smell (anosmia) [41,42] , may be also be present. At the beginning of the illness a dry cough and sometimes dyspnoea may be present. The systemic defence reaction may last for a few days and the consequences may not be clinically noted if the person continues to stay home and recovers promptly. If the symptoms last for a week or longer, two major consequences have to be considered: (1) severe fluid losses leading to dehydration and ultimately hypovolaemic shock; (2) reduction in caloric intake which worsens symptoms of weakness, and accelerates a rapid loss in body weight [43] .
These changes may be aggravated by the simultaneous intake of antihypertensive medication, including diuretics, as might be encountered in older patients and/or those patients with multiple comorbidities [44] .
The systemic reaction, a major component of body defence strategy, is induced by different cytokines that originate the main site of injury, e.g., the lungs. The so called "major acute-phase mediators" are Interleukin-6, Interleukin-1, TNF-alpha, and IFN-gamma, which are all synthesized in different amounts, depending on the quality (organ and damaging agent) and the quantity of tissue damage.
The acute phase cytokines are responsible for the central regulation of body temperature [45] , reduction in appetite, and associated adynamia and mental confusion [46] .
The reduction of appetite (anorexia) on the one hand, and abdominal discomfort on the other, can also be attributed to the direct action of the cytokines on the intestinal neurons, with alterations in the mobility  [37] of the large and small bowel [47,48] . The liver, as the source of the majority of the serum proteins, is the main target of the acute phase cytokines. These cytokines induce pretranslational modification of gene expression through direct interaction with the hepatocytes [49] [ Figure 3]. There are positive and negative acute phase proteins [45] .
As most of the major acute-phase proteins have a low molecular weight, measurement of their serum level may not correspond to a real increase in hepatic synthesis. This is due to the rapid elimination via the urine. Hepcidin was first identified in the urine [59] .
Albumin is the main negative secretory acute phase protein [ Figure 4] [49] , whilst ferroportin-1 and hemojuvelin belong to the negative intracellular acute-phase protein group [52][53][54][55] . In a rat model, albumin mRNA in the liver was reduced by 50%, while total mRNA was increased by 50%, 2 days after infection with live Escherichia Coli [60] . During the 2 days rats ate only 5%-10% of the amount of food consumed prior to injection by the bacteria. This was followed by a further aggravation of the reduction of albumin synthesis [60] , further demonstrated in isolated liver perfusion studies [61] , and in humans under caloric restriction [62] . The amount of the acute-phase cytokines released into the circulation, and the concentration needed for the systemic appearance of the symptoms and of the metabolic changes, are different in different patients. They may be regulated differently by the drug administered, especially in the acute diseases. However, the response is mainly proportional to the extent of the tissue damage.  [49] In summary, two main mechanisms act in reducing albumin serum concentration in patients with severe COVID-19-infection: (1) reduction in albumin synthesis due to reduced food intake; (2) inhibition of specific mRNA-synthesis in the hepatocellular nuclei induced by the direct interaction of the cell with the acute-phase cytokines.
The acute-phase cytokines induce up-regulation of gene-expression of several positive hepatic acute-phase proteins, and in extrahepatic organs [63] [ Figure 5], but the changes in serum level are influenced by their synthesis in liver cells [45] . This mechanism is not only active in cases of tissue damage caused by bacterial, but also by viral infections [64] . The order of magnitude of variations in the serum level of the acute-phase proteins caused by viral infections is lower than that induced by bacterial infections.  (24 h) and the longest was SAA (21 days). While synthesis of albumin was inhibited by increasing doses of human recombinant IL-1, synthesis of factor B and of SAA were increased at the same time in the hepatocyte reproducing the process taking place in the liver in vivo during an acute phase situation. It is understandable that the serum concentrations of the acute-phase cytokines produced at extrahepatic sites has to be quite high to induce changes of protein synthesis in the liver until these can become measurable. This is also the case for those proteins whose constitutive gene-expression is almost undetectable, as is the case for SAA or CRP in humans. SAA: serum amyloid A. 1985;162:930-42. (reprinted with permission) [49] Physical examination results obtained in hospitalized patients are not reported in the different publications, but most of the patients who were transferred from the emergency room to the ICU will likely have presented with clear signs of exsiccosis, hypotension and eventually malnutrition as testified by the low serum albumin levels. This should be highlighted in the guidelines for the initial supportive management of patients with COVID-19. If not recognized and promptly treated, progression to the second stage of the disease, with deterioration in respiratory function, will likely occur.
Patients suffering from mild disease who presented with normal serum albumin levels, even those who have developed a deterioration, maintained normal serum levels and could be released from the hospital [15,16,65] .
Although albumin administration is not recommended in patients with low serum albumin levels being treated in the ICU [35,36] , previous positive experiences [66] with repeated administration of 200-400 mL of convalescent plasma showed positive effects in some critically ill COVID-19-patients [67][68][69][70] . The positive effect of convalescent plasma infusion could be attributed not only to the COVID-19-specific immunoglobulins, but also to the other components of the plasma e.g., albumin [71] . Coli LPS as a model to induce an acute-phase reaction. The filters containing the tissue-RNA were hybridised with radio-actively labelled cDNAs specific for factor B, for SAA and for actin as control. In all organs factor B-and SAA-gene-expression was up-regulated in a dosedependent manner. The different time of exposure of the x-ray film demonstrate the different abundance of gene-expression of factor B and SAA in the different organs. SAA: serum amyloid A; LPS: lipolysaccharide. J Immunol 1985;135:3645-7. (reprinted with permission) [63] DECLARATIONS