COVID Convalescence-A Boon or Bane in Cardiac Surgery?: A "Second Hit" Hypothesis.

The coronavirus disease 2019 (COVID-19) pandemic continues to affect healthcare services. As elective cardiac surgical services resume, clinicians will encounter COVID-19-recovered patients for cardiac surgery. The hyperimmune pathophysiology of COVID-19 and exposure to the inflammation of cardiac surgery, cardiopulmonary bypass, mechanical ventilation, blood transfusion, and perioperative infections could lead to exacerbated responses, exemplified by systemic inflammatory response syndrome and cascade to multiorgan dysfunction syndromes. The authors present a patient with coronary artery disease undergoing off-pump coronary artery bypass surgery after the institutional protocol of two COVID-19 reverse-transcriptase polymerase chain reaction tests reported negative. Intraoperatively, unexplained hypoxemia was observed, which warranted cardiopulmonary bypass support to complete the grafting. After multiple attempts of failed weaning, intra-aortic balloon pump and high inotropes helped to wean. The patient had a stormy postoperative course, with low oxygenation, bleeding, low-cardiac-output syndrome, rhabdomyolysis of lower limb muscles, requiring multiple blood and blood product transfusion, and renal replacement therapy. Despite the corrective measures, severe hyperkalemia and cardiac arrest ensued. IgG antibodies to the severe acute respiratory distress syndrome coronavirus-2 virus were tested considering the unexplained hypoxemia. A “convalescent COVID-19” patient with “first hit” at primary infection, encountering a “second hit” of surgery and perioperative insults, might have a hyperimmune response. This “second hit” hypothesis should be considered when COVID-19 convalescent (COVID-19 symptomatic or asymptomatic) patients undergo cardiac surgery and present with unusual complications.

Case Report

A 62-YEAR-OLD male, with coronary artery disease, was scheduled for elective coronary artery bypass grafting (CABG). He was a long-term diabetic and hypertensive on regular medications. The preoperative evaluation of other organ systems was unremarkable. Investigations of renal function, liver function, chest X-ray, and coagulation parameters were within normal limits. However, the blood count revealed lymphopenia that had no clinical correlation. Echocardiography demonstrated normal functionality but coronary angiography showed significant triple-vessel disease. As per institutional protocol, screening was done twice before surgery for severe acute respiratory distress syndrome coronavirus-2 (SARS-CoV-2) by reverse-transcriptase polymerase chain reaction (RT-PCR) test. These tests were done seven days apart, with a negative report for coronavirus disease 2019 (COVID-19) in both instances. A high-resolution computer tomography scan of the chest was performed that revealed nothing overt. He was retained in hospital-based quarantine isolation for the waiting period of the RT-PCR testing and subsequently transferred to the cardiac unit on the eve of scheduled off-pump CABG surgery.

On transfer to the operating room, his baseline vitals recordings were unremarkable, with an oxygen saturation of 97% on room air. Under local anesthesia, an invasive radial arterial blood pressure monitoring line was inserted and arterial blood gas (ABG) analysis at this stage was normal. Induction of anesthesia was accomplished with a standardized narcotic, benzodiazepine, and muscle relaxant-based technique. Subsequently, he was intubated with an appropriately sized endotracheal tube (ETT). After confirmation of the ETT position, the patient was connected to the anesthesia circuit. Controlled ventilation with volume-controlled mode was chosen, with a tidal volume of 8 mL/kg. This generated a peak airway pressure of less than 25 cmH2O. Adequacy of ventilation parameters was verified by ABG. Heparin was instituted in a dose of 400 IU/kg to facilitate grafting and monitored with activated clotting time. Left internal mammary artery-to-left anterior descending artery anastomosis was completed uneventfully. However, during verticalization for the obtuse marginal artery grafting, the attending anesthesiologist observed significant desaturation. Common causes of desaturation were evaluated that included checking of the ventilator and circuits, ETT, pulse oximetry probe position, and end-tidal carbon dioxide monitoring. Both pleural cavities were opened by the surgeon, revealing adequate lung expansion. The fraction of inspired oxygen (FIO2) was increased to 1, ETT suctioning was done to clear out secretions, and positive end-expiratory pressure was increased to 8 cmH2O. A repeat ABG showed marginal improvement of the arterial partial pressure of oxygen (PaO2) (Table 1 ). Transesophageal echocardiography was performed to rule out intracardiac shunts, patent foramen ovale causing right-to-left shunt, severe tricuspid regurgitation, new-onset mitral regurgitation, right ventricular (RV) and left ventricular (LV) dysfunction, and acute pulmonary embolism. Grafting to the obtuse marginal artery was performed on a beating heart. Considering the low PaO2:FIO2 ratio, a decision was made to establish CPB to graft the posterior descending artery. An attempt to separate from CPB resulted in hypoxia, worsening hemodynamics, and ventricular fibrillation. CPB was re-established and the heart was defibrillated. The grafts when examined appeared to be functioning well, and in the absence of any mechanical or metabolic cause of instability, a decision was made to put in additional vein grafts to the distal left anterior descending artery and OM vessels. A second attempt to wean off CPB was unsuccessful due to low cardiac output and low PaO2. An intra-aortic balloon pump (IABP) was inserted in the right femoral artery to augment cardiac function. After a further 45-minute rest on CPB, the patient was successfully weaned. The sternum was left open with a retractor in situ, due to borderline hemodynamics and high vasoactive support (epinephrine 0.1 µg/kg/min, norepinephrine 0.1 µg/kg/min, levosimendan 0.1 µg/kg/min). A pulmonary artery catheter was floated that revealed normal pulmonary artery pressure, pulmonary capillary wedge pressure, and cardiac index (CI). The patient was transferred to the intensive care unit (ICU), with a PaO2:FIO2 ratio of 170 mmHg and a CI of 2.3 L/min/m2. Hematuria was noticed.

Table 1

Serial Arterial Blood Gas During the Perioperative Period

Timeline	PaO2:FIO2 (mmHg)	PaCO2 (mmHg)	PEEP (cmH2O)
Baseline on spontaneous ventilation (T0)	220	35	-
Post induction mechanical ventilation (T1)	220	35	5
During OM grafting off pump (T2)	83	60	5
After maneuvers to improve oxygenation (T3)	90	42	8
At chest closure attempt (T4)	70	52	8
Final successful weaning from CPB (T5)	208	58	12
POD 1 in ICU (T6)	128	41	10
POD 2 in ICU just before RRT initiation (T7)	127	39	10

Abbreviations: CPB, cardiopulmonary bypass; FIO2, fraction of inspired oxygen concentration; OM, obtuse marginal artery; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure; POD, postoperative day; RRT, renal replacement therapy; T0-T7, timeline.

In the ICU, the patient was placed on controlled of ventilation, with an FIO2 of 100% and a positive end-expiratory pressure of 15 cmH2O. His vital parameters stabilized over a few hours, with a modest dose of vasoactive support (epinephrine 0.05 µg/kg/min and levosimendan 0.05 µg/kg/min) and an IABP. Inhaled nitric oxide was commenced in a dose of 40 PPM to counter borderline PaO2, but did not significantly improve oxygenation. The serial ABG during the perioperative period is depicted in Table 1. The patient was atrioventricular sequentially paced at 90 beats/min, with an underlying sinus rhythm of 45 beats/min. Multiple transfusions of packed red cells and blood products were needed in view of generalized bleeding. A cell saver was attached to the mediastinal drains to salvage lost blood. All transfusions were directed with thromboelastographic (TEG) and coagulation screen monitoring. Interestingly, the patient persistently had a prolonged reaction time and a low maximum amplitude on the TEG despite targeted transfusions. The activated clotting time readings were persistently high, between 600 and 700 seconds. He subsequently developed a low-cardiac-output state secondary to high drain output a few hours later, necessitating re-exploration that revealed generalized bleeding. His vasoactive support levels had significantly increased (epinephrine 0.12 µg/kg/min, norepinephrine 0.12 µg/kg/min, vasopressin 0.1 µg/kg/min, and dobutamine 3 µg/kg/min). The CI varied between 1.5 and 1.8 L/min/m2, and the systemic vascular resistance ranged between 650 and 750 dynes/s/cm−5, indicating myocardial pump failure and a vasodilatory state. The following morning the chest was washed to relieve a tamponade. Serum IgG and D-dimer levels were investigated suspecting COVID-19 convalescence state, as the perioperative hypoxia was unexplained. The IgG was positive, indicating prior infection with COVID-19. His D-dimer level and CRP level were reported high at 1,028 ng/mL and 141 mg/L, respectively. Heparin-induced thrombocytopenia screen was negative. Due to the coagulopathic state of the patient, no anticoagulant or antiplatelet therapy was instituted. The patient developed oliguria, with gradually worsening lactic acidosis, along with hyperkalemia. Hence, renal replacement therapy (RRT) was commenced. On the third postoperative day, the patient manifested bilateral lower limb ischemic changes and muscle rigidity. High creatine kinase (34,460 U/L) and lactate dehydrogenase levels (1998 U/L) confirmed the diagnosis of rhabdomyolysis. An arterial Doppler study revealed poor flow in both the lower limbs. The hyperkalemic state and metabolic acidosis continued to worsen on RRT and culminated in a cardiac arrest. Extracorporeal membrane oxygenation was considered but not offered, given the patient's advanced critical state with no reversibility quotient.

Discussion

The authors report this case to highlight the perioperative complications encountered in managing a convalescent COVID-19 patient during elective cardiac surgery. Asymptomatic COVID-19 disease continues to be a predominant presentation in patients infected with SARS-CoV-2. Elective cardiac surgical services have resumed, with perioperative protocols aimed at minimizing risk to the healthcare professional and providing safety to patients from acquiring COVID-19 during their hospital stay. Varied protocols exist depending on endemic factors and the prevalence of the disease. Screening with RT-PCR for SARS-CoV-2 is an integral part of the preoperative workup. Testing for antibodies as a screening modality has been reviewed without any conclusion. The patient in context was cleared for active COVID-19 disease by two negative RT-PCR tests as per institutional protocol. But the unexplained perioperative hypoxemia, excessive bleeding, and raised D-dimer values triggered IgG antibody testing for SARS-CoV-2. The presence of these antibodies confirmed the diagnosis of convalescence of COVID-19.

In the current global COVID-19 pandemic, patients presenting for elective cardiac surgery can lie in the symptomatic/asymptomatic spectrum. Recently recovered COVID-19 patients have cardiac myocardial inflammation, myocardial edema, fibrosis, and RV dysfunction. A German study, encompassing 100 recently recovered COVID-19 patients with the confirmed disease (tested by RT-PCR), was followed up with cardiac magnetic resonance imaging (CMRI). The majority of these patients received home-based therapy based on disease severity. Seventy-eight percent of this cohort had reductions of LV ejection fraction, LV volumes, late gadolinium, and pericardial enhancement. At the time of the CMRI, 70% of these patients had tested positive for high-sensitive troponin T, whereas only 15% had high-sensitive troponin T during their disease period. Endomyocardial biopsy of severely affected patients has revealed active lymphocytic inflammation. This observation suggested a similar severity of cardiac dysfunction in hospitalized and home-based therapy. RV dilatation and dysfunction were predominant echocardiographic findings in comparison to LV pathology, indicative of RV involvement secondary to pulmonary lesions. Biventricular cardiac involvement and secondary changes to pulmonary pathophysiology result in heart failure, which is a complex endpoint in COVID-19 patients. Whether these findings are of significance in convalescent patients and lead to increased perioperative adverse events is speculative.

Microvascular COVID-19 lung vessels obstructive thrombo-inflammatory syndrome (Micro CLOTS) is a hypothesis proposed to explain the complement cascade mediated by massive alveolar epithelial and vascular endothelial damage microvascular thrombosis. Alveolar capillary microthrombi were nine times more prevalent in comparison to influenza H1N1 infections. Alveolar microthrombi were observed in lung biopsies of asymptomatic carriers of COVID-19 as an incidental finding. COVID-19 is a prothrombotic state, with reports of venous, arterial, and catheter-related thrombosis. Acute pulmonary embolism continues to be the most common thrombotic manifestation of COVID-19. Lower limb ischemia and rhabdomyolysis are potential complications of IABP-initiated ischemia complicated by a prothrombotic state. Cross-links between inflammation and prothrombotic state are evident by laboratory markers of raised D-dimers, fibrinogen, factor VIII, von Willebrand factor, and decreased antithrombin levels.

The exacerbation of inflammatory cascade in recovered COVID-19 patients undergoing surgery in convalescence can be explained by a “first and second hit” theory. The contained inflammation in the primary disease (first hit) can be exacerbated by stimuli of mechanical ventilation, infection, stasis, thrombosis, bleeding, ischemia, hypoxia, blood transfusions, and transfusion-related acute lung injury (second hit), and cause a second uncontrolled inflammatory cascade, culminating in multiple organ dysfunction syndromes. In such scenarios, avoiding the second hit by reducing inflammation with lung-protective strategies, thromboprophylaxis, early prone ventilation in the postoperative period, antifibrinolytic therapy and peripheral vascular screening are evolving concepts.

In the current clinical setting, with the presence of IgG antibodies for SARS-CoV-2 and 2 preoperative negative RT-PCR, the authors postulated a convalescent phase of recent past COVID-19 asymptomatic disease in their patient. RT-PCR continues to be the gold standard for diagnosing COVID-19 to date. False-negative results vary from 2% to 29% across the globe due to factors related to viral load, sampling techniques, and timing of testing in the disease course. Pretest probability is the clinical likelihood of a person having COVID-19 depending on local COVID-19 prevalence, SARS-CoV-2 exposure history, and symptoms. Pretest probability helps validate the result of RT-PCR or guide further testing. Repeat testing is known to overcome the limitations in RT-PCR sensitivity, but it should be at the discretion of the treating team. Institutional protocol, with testing RT-PCR twice 7 days apart to rule out false-negative results, was followed. With seroprevalence of 0.22% to 47% across the population, the use of antibodies for testing as a routine preoperative screening protocol is far from being cost-effective.11, 12, 13 But, using antibodies to test inconclusive RT-PCR and RT-PCR not correlating with the clinical scenario is recommended, as was performed in the presented patient in the postoperative period. ELISA for antibodies specific to SARS-CoV-2 were performed and resulted positive for IgG in the presented patient. There are seven types of human coronavirus infections including SARS-CoV-2, SARS-CoV, and MERS-CoV. Cross-reactivity of antibodies across the SARS virus family does exist, but the small number of patients infected per year and the younger age population affected more with the low pathogenicity SARS virus make it clinically less likely for this cross-reactivity to affect the decision-making in the COVID-19 pandemic.

The first episode of desaturation and hypoxemia possibly was an exacerbated inflammation due to mechanical ventilation (second hit) potentiated by altered hemodynamics during the verticalization of the heart, mitral regurgitation, worsening diastolic function, increased pulmonary capillary wedge pressure, and subsequent pulmonary venous congestion. The pulmonary Micro CLOTS pathophysiology evolved extensively during the case, with the lowest recorded PaO2:FIO2 ratio of 70 mmHg. Cardiac dysfunction secondary to residual COVID-19 is speculative unless preoperative CMRI defines myocardial inflammation. Limb ischemia and rhabdomyolysis are extensions of the prothrombotic state complicated by IABP insertion and high-dose vasopressor. Thrombotic microangiopathy was observed to be disproportionately high in COVID-19 patients, even in those not hospitalized. Thrombotic microangiopathy is characterized by low platelet count, elevated D-dimer, and LDH levels. These laboratory values are difficult to interpret in the settings of cardiac surgery and CPB, but a preoperative assay might alert the clinician. The role of perioperative dexamethasone in cardiac surgery is not outlined in this setting. The benefits of plasma exchange on CPB and plasma adsorbent filtration on RRT are yet to be ascertained in the perioperative cardiac surgical patient.

To conclude, convalescent COVID-19 patients increasingly will present for elective cardiac surgical procedures in the near future. They pose a higher risk in the perioperative phase. Negating the “second hit” by avoiding CPB, ventilation-associated lung injury, postoperative infection states, and transfusion-related acute lung injury is not always feasible. These cases could very easily “snowball” into high resource utilization with a low success rate. Routine testing of IgG antibodies preoperatively, prophylactic steroid therapy, and restratifying surgical techniques are debatable options. Extended evaluation of the heart (CMRI), pulmonary function testing, and coagulation testing (D-dimer and TEG) can be of help in COVID-19-recovered patients for perioperative risk stratification. By default, in the current COVID-19 pandemic, the patient manifesting a severe perioperative inflammatory response should trigger a suspicion of COVID-19 even though they were deemed negative preoperatively by RT-PCR. It is prudent to outline these risks while seeking consent from patients for surgery, as the pathophysiologic manifestations of the COVID-19 disease process during the convalescence phase continue to evolve, with no cardiac risk stratification scoring system in place to encompass this scenario.