Abstract
Objective: The aim of the study was to evaluate the outcome of extracorporeal membrane oxygenation (ECMO) support in Hantavirus cardiopulmonary syndrome (HCPS) patients with a predicted mortality of 100%, and the complications associated with this treatment modality and different cannulation techniques. Methods: A total of 51 patients with refractory HCPS were supported with ECMO between April 1994 and June 2010. They were divided into group A consisting of the 26 patients treated between 1994 and 2000 and group B consisting of 25 patients treated between 2003 and 2010. No patients were treated between September 2000 and December 2003. Patients in group A were intubated when they became hypoxic and placed on ECMO when they became hemodynamically unstable, whereas patients in group B had elective insertion of vascular sheaths and were almost concurrently intubated and placed on ECMO when they decompensated. Cannulation of the femoral vessels was performed percutaneously in 18 (35.3%) patients and with an open technique in 33 (64.7%) patients. Results: Complications from percutaneous cannulation occurred in 4/18 (22.27%) patients: retroperitoneal hematoma in 2/18 (11.1%) and lower-extremity ischemia in 2/18 (11.1%) patients. Complications from open femoral cannulation occurred in 12/33 (36.3%) patients: bleeding in 10/33 (30.3%) patients and ischemia in 2/33 (6.1%) patients. The overall survival was 66.6% (34 of 51 patients); 56% (14/26) for group A and 80% (20/25) for group B (p = 0.048). There was no difference in mortality regarding the method of cannulation. A trend toward increased mortality in patients with cannulation complications was recognized, but it was not statistically significant. Mortality was not associated with ECMO duration (average 121.7 h, range: 5–276 h). All survivors recovered and were discharged from the hospital after a mean hospital stay of 19.8 days (range: 10–39 days). Conclusions: Two-thirds of 51 HCPS patients with a predicted mortality of 100%, who were supported with ECMO, survived and recovered completely. Survival was significantly higher in the second half of the study. Complications associated with both types of femoral cannulation were associated with a trend toward decreased survival, which was not significant.
1 Introduction
Hantaviral infection was first recognized in the Four Corners region of the American southwest in 1993, during an outbreak of unexplained deaths in previously healthy young individuals, who developed acute onset of respiratory failure, pulmonary edema, and shock [1]. This clinical syndrome was initially called Hantavirus pulmonary syndrome (HPS). As most deaths are caused by a combination of respiratory and cardiac failure, we – as well as other investigators – prefer the term ‘Hantavirus cardiopulmonary syndrome (HCPS)’ [2]. The Sin Nombre virus (SNV) has been identified as the etiologic agent of HCPS, and is carried by the deer mouse Peromyscus maniculatus. The SNV is transmitted primarily by inhalation of particles contaminated with saliva, urine, or feces of infected rodents [3]. Currently, there are no antiviral drugs, vaccines or immunotherapeutic agents available for treatment of the disease, and therapy is primarily supportive. Despite greater awareness and recognition of HCPS, and improvement in critical care management, contemporary mortality rates from 43% to 76% have been reported [4,5]. In 1994, our institution [6] introduced the use of extracorporeal membrane oxygenation (ECMO) as a rescue therapy in patients, who had developed irreversible shock or cardiac arrest as a consequence of cardiopulmonary failure due to HCPS. Since our report in 2008 [7] that included 38 patients with HCPS treated with ECMO, we have added 13 new patients. We divided the total of 51 patients with HCPS, who were refractory to maximal medical management, into two cohorts that were managed differently over two time periods. The purpose of this study is to evaluate the impact on survival of two different treatment strategies.
2 Materials and methods
2.1 Patient selection
We reviewed retrospectively the medical records of 51 consecutive patients with severe HCPS treated with ECMO at the University of New Mexico Hospital (UNMH) from April 1994 to June 2010. They were divided into two groups of patients who were managed differently over two time periods. Group A consisted of 26 patients treated between April 1994 and September 2000, who were intubated when they became hypoxic and were not cannulated for ECMO until they became hemodynamically unstable. Group B consisted of 25 patients treated between December 2003 and July 2010, who had elective insertion of percutaneous vascular sheaths shortly after arrival in the hospital, and were almost concurrently intubated and cannulated for ECMO when they became unstable. These groups were separated by a period of 3 years during which no HCPS patients were seen in the entire Four Corners, this being most likely attributed to cyclical variations of the rodent reservoir. During this hiatus between September 2000 and December 2003 without HCPS patients, general public health education and awareness of regional emergency physicians were intensified. Approval by the Institutional Review Board was obtained, and informed consent was obtained from each patient or the closest relative.
Patients were referred from three of the States of the Four Corners region (New Mexico (34), Arizona (11), and Colorado (six)) presenting with a history of fever, headache, and myalgias, followed by rapid deterioration, characterized by respiratory distress, and diffuse pulmonary infiltrates on chest X-ray. Most patients had history of exposure to rodent droppings while cleaning a shed or a building, 1 or 2 weeks before experiencing symptoms. Early presumptive diagnosis of Hantavirus infection was strongly suspected by the presence of thrombocytopenia, hemoconcentration, myelocytosis with absence of toxic granulations, and at least 10% lymphocytes with immunoblastic features in the peripheral blood smear [8]. Serologic tests were positive for SNV antibodies in 100% of the patients. All patients were admitted to the intensive care unit. No lines were placed in the right side of the neck, so the right jugular vein could be available for cannulation. Only 20 (39.2%) patients were breathing spontaneously on admission, but requiring increasing oxygen concentrations. Thirty-one (60.8%) patients were intubated prior to transfer to our hospital. The remaining patients required intubation immediately prior to cannulation.
Most patients with HCPS deteriorated over a short period of time. The current inclusion criteria for starting patients on ECMO are: a clinical presentation consistent with HCPS, a cardiac index less than 2.0 l min−1 m−2 despite maximal support and at least one of the following criteria found to be associated with a 100% mortality rate [6]: serum lactate of >4.0 mmol l−1 (normal range 0.0–2.2), a partial pressure of oxygen in the blood: fraction of inspired oxygen (PaO2:FiO2) ratio of ≪60, or cardiopulmonary deterioration, such arrhythmia or cardiac arrest. Exclusion criteria included irreversible neurological damage, mechanical ventilation longer than 5 days, and multi-organ failure syndrome. Criteria for ECMO support are summarized in Table 1 . Only one patient, who was 75 years old, and was in multi-organ system failure and neurologically compromised prior to transfer, was not considered for ECMO support.
Criteria for starting ECMO. (In addition to clinical presentation consistent with HCPS and a cardiac index less than 2.0 l min−1 m−2.).
A total of 51 patients who met the above criteria were placed on ECMO, and are the basis of this report. They represent approximately 75% of all the patients treated for HCPS at our institution. Their ages ranged from 9 to 69 years (mean age 39.6 years). There were 28 (55%) female and 23 (45%) male patients. Before cannulation, 51 (100%) patients were on multiple inotropic agents because of hemodynamic deterioration, and 10 (19.6%) required external cardiac compressions.
2.2 Cannulation and decannulation technique
Whenever possible, the ECMO cannulas were placed percutaneously into the femoral vessels. Otherwise, open cannulation by direct cut down was performed. In group B, vascular access sheaths (Boston Scientific Catalog number 15-711B) were prophylactically inserted percutaneously into the common femoral artery and vein in patients who had not met ECMO criteria yet. This approach facilitated percutaneous cannulation during urgent situations. Before cannulation, patients received an initial dose of heparin of 200 units kg−1. Bio-Medicus femoral arterial cannulae, ranging in size from 15 to 21 Fr/18-cm long (Medtronic Inc., Minneapolis, MN, USA) were placed using a Seldinger technique. Similarly, femoral venous cannulae, ranging in size from 22 to 29 Fr/76-cm long (Medtronic Inc., Minneapolis, MN, USA) were placed and advanced to the right atrium. Ten patients with pulseless electrical activity and/or ventricular fibrillation were cannulated using the open technique during external cardiac compressions.
To prevent ischemia of the lower extremity, we placed a distal perfusion cannula (size 8 or 10-Fr/10-cm long Bio-Medicus pediatric arterial cannula (Medtronic Inc., Minneapolis, MN, USA) in the superficial femoral artery (SFA) in all patients cannulated with the open technique. For patients cannulated percutaneously we inserted a small distal arterial cannula – under direct vision – in either the SFA or the posterior tibial artery. In addition, a distal venous cannula, size 14–20 Fr was inserted distally into the femoral vein in five patients to relieve venous congestion.
During ECMO, five (13%) patients had poor venous drainage because the venous cannula could not be advanced into the inferior vena cava using the femoral approach. In these patients, a Thin-Flex single-stage venous cannula (Edwards Lifesciences LLC, Irvine, CA, USA), ranging in size from 30 to 36 Fr was advanced into the right atrium through the right internal jugular vein, which was exposed using an oblique neck incision, anterior to the sternocleidomastoid muscle.
Patients were connected to the ECMO machine (Cobe Cardiovascular, Arvada, CO, USA), with an Avecor membrane oxygenator (Medtronic Inc., Minneapolis, MN, USA) and a Bio-Therm heat exchanger (Medtronic Inc., Minneapolis, MN, USA). The initial ECMO pump settings included a flow of 4.0–5.0 l min−1 delivered by a roller pump system and a FiO2 of 1.0. Anticoagulation was maintained with heparin, and the activated clotting time (ACT) was monitored hourly, with values ranging from 180 to 220 s. After a few days on ECMO, the inotropic agents and FiO2 were weaned, as the hemodynamic status and pulmonary function improved gradually. Subsequently, the patients were weaned from ECMO support when the pump flow was ≪1 l min−1, and 1-h trials off ECMO were tolerated. Clotting of the femoral artery or vein was not an issue because the patients were kept fully anticoagulated and flows reestablished until the cannulas were removed.
The cannulas were removed using an open approach in all cases. After removing the arterial cannula, a balloon embolectomy catheter was passed distally to retrieve any possible clots. The common femoral artery was repaired primarily in 24 patients, using continuous 6/0 polypropylene suture, or using a patch of autologous saphenous vein in 13 patients. Likewise, the common femoral vein was repaired primarily using continuous 6/0 polypropylene suture, or with a patch of autologous saphenous vein in three patients. Normal distal pulses were present after decannulation in all patients who survived ECMO.
2.3 Statistical analysis
Chi-square or Fisher’s exact test was used to compare several categorical variables associated with increased mortality during ECMO support (Table 3), and the Mann–Whitney test was used for comparison of continuous variables (Tables 4 and 5). A value of p ≪ 0.05 was considered significant.
3 Results
Duration of ECMO averaged 121.7 h. After decannulation, all patients remained intubated for an average of 7.1 days (range: 3–12 days), and four patients required a tracheostomy. The reason for prolonged intubation after discontinuation of ECMO is attributed to the fact that the cardiac failure phase usually recovers within 3–7 days, whereas the respiratory failure phase may last 2 weeks.
The hospital mortality was 33.3% (17 of 51 patients), of whom 15 patients died during ECMO support, and two additional patients died of multi-organ system failure after ECMO. We observed that survival improved in the second phase of the study: 80% (20/25) in group B compared with 53.9% (14/26) in group A (p = 0.048) (Fig. 1 ). There was no significant difference in mortality for patients supported with ECMO >7 days (5 of 11 died, mortality = 45.4%) when compared with patients on ECMO ≪7 days (12 of 40 died, mortality = 30%) (p = 0.753). Six (60%) of the 10 patients, who had cardiac arrest prior to cannulation, died from hypoxic–ischemic encephalopathy. Other causes of death included: brain hemorrhage in one, multi-organ system failure in six, and sepsis and renal failure in four patients (Table 2 ). Our results indicate that patients in group A, as well as severe lactic acidosis, were associated with significantly increased mortality (p ≪ 0.05) (Tables 3 and 4 ). Eighteen (52.9%) of the 34 survivors had an uneventful course on ECMO and no complications.
Yearly distribution and mortality of Hantavirus patients supported with ECMO from 1994 to 2010.
Causes of death.
Categorical variables associated with increased mortality during ECMO support, compared using Chi-square or Fisher’s exact test.
Continuous variables associated with increased mortality during ECMO support, compared using Mann–Whitney test.
All patients, who were successfully weaned and decannulated, survived and recovered completely. They were discharged from the hospital after a mean hospital stay of 19.8 days (range: 10–39 days).
No difference in survival was observed when the percutaneous and open techniques of cannulation were compared. Six (33.3%) of the 18 patients cannulated percutaneously, and 11 (33.3%) of the 33 patients who had open cannulation died. Complications from percutaneous cannulation occurred in six of 18 (33.3%) patients. Two of 18 (11.1%) patients developed a large retroperitoneal hematoma, two (11.1%) developed bleeding, and two (11.1%) patients developed lower-extremity ischemia. Adequate leg perfusion was reestablished with thrombectomy–embolectomy, and after insertion of an additional cannula in the superficial femoral artery.
Complications from open femoral cannulation occurred in 12 of 33 (36.3%) patients: severe bleeding occurred in 10 of 33 (30.3%) patients and two of 33 patients (6%) developed ischemia of the lower extremity. Two patients required a leg amputation, despite insertion of a distal perfusion cannula after the initial cannulation. The development of cannulation complications had a negative impact on the ECMO course and the outcome of the patients. Seven (38.8%) of the 18 patients with cannulation complications did not survive. By contrast, only 10 (30.3%) of the 33 patients without cannulation complications did not survive. However, the difference in mortality between the patients who had serious cannulation complications, and those who did not was not statistically significant (p = 0.80). Other serious complications that we observed during ECMO support included mechanical failures in three patients who did not survive (tubing rupture in two and pump malfunction in one).
4 Discussion
In 1951, American troops stationed in Korea were stricken by what was then thought to be a new and frightening illness. ‘Korean hemorrhagic fever’, as it came to be known, was characterized by fever, severe hemorrhagic manifestations, and renal failure. With no treatment available, about 10% of the infected soldiers died. Despite intense efforts directed toward finding the cause of the disease, it was not until 1976 that the etiological agent, the Hantaan virus, was identified [9]. Since then, many other relatives of the Hantaan virus have been recognized in several other countries of Asia and Europe. All of these ‘Old World’ viruses cause a similar clinical picture now known as ‘Hemorrhagic fever with renal syndrome (HFRS)’.
Hantaviral infection was first recognized in the United States in 1993 during an outbreak of unexplained deaths in the Four Corners area. These were previously healthy young individuals, who developed acute respiratory failure, pulmonary edema, and heart failure. The clinical course was characterized initially by mild flu-like symptoms followed by rapid progression to respiratory distress and myocardial depression and, in many instances, death. Serological and genetic studies demonstrated that the agent was a new type of Hantavirus, the SNV. Even though the majority of the cases continue to occur in the southwest region, a proliferation of the organism and the disease it produces have been recognized in other areas of USA and Canada (Fig. 2 ) as well as in South America. The clinical syndrome seen in the ‘New World’ is quite different from the previously known HFRS. The disease was initially called HPS, but as most deaths are caused by severe pulmonary and myocardial dysfunction, we, as well as others, prefer the term ‘HCPS’ [10].
Distribution of HCPS patients reported until July 1, 2010.
More than 95% of the HCPS patients reported in North America were caused by SNV. The main reservoir of the SNV in the United States is the deer mouse P. maniculatus. However, in other areas of the country, other rodent species can carry other related Hantavirus that causes HCPS. Infected mice shed the virus in their saliva, urine, and feces. Humans acquire the virus primarily by aerosol inhalation of particles contaminated by infected rodents. Person to person transmission of SNV has not been reported in North America. The viral antigens have been shown to be distributed within the endothelium of capillaries of various tissues, particularly the lungs and the heart. The infection causes a dramatic increase in endothelium permeability without overt cell death. Multiple lines of evidence suggest that the endothelial damage is a consequence of the initial immune response to the viral infection [10]. Given enough time, many patients are able to develop neutralizing antibodies that can clear the virus and eventually recover [2].
The HCPS is characterized by four clinical phases: prodrome, pulmonary edema and shock, diuretic phase, and convalescence [5]. Clinical presentation during the prodromal phase is characterized by high fever, headache, and myalgias for an average of 3–6 days, followed by dyspnea, which is frequently a sign of impending respiratory failure. The onset of respiratory distress and, later, the development of shock are abrupt. Risk of death is highest during the initial 24 h of the pulmonary edema and shock phase of the illness, which also last from 3 to 6 days.
Despite early diagnosis of HCPS, and improvement in critical care management, including mechanical ventilation and hemodynamic support with vasoactive drugs, mortality rates range from 43% to 76% [4,5]. In 1998, Crowley and colleagues [6] identified several criteria for non-survival, which include: refractory shock, lactate concentration > 4.0 mmol l−1, severe hypoxia (PaO2:FiO2 ratio ≪ 60), and cardiac arrest. In our present study, only those patients with a typical clinical presentation consistent with HCPS and a cardiac index that rapidly dropped to ≪2.0 l min−1 m−2 despite maximum inotropic support, and at least one of the predictive criteria for non-survival mentioned above, were considered for ECMO support. Early diagnosis of SNV infection was strongly suspected by characteristic hematologic findings, including thrombocytopenia, hemoconcentration, leukocytosis, lack of toxic granulations in neutrophils, and the presence of at least 10% immunoblasts in the peripheral blood smear [8]. The presence of four out of these five findings after the development of pulmonary decompensation has a sensitivity for HCPS of 96% and a specificity of 99%. The diagnosis of HCPS was confirmed by the presence of SNV antibodies. However, a major limitation of the serologic test is that it usually takes 8–12 h, whereas a recombinant immunoblot assay allows the diagnosis within 4 h [11].
In contrast to other forms of septic shock, patients with HCPS have a decreased cardiac index and elevated systemic vascular resistance [12]. Although the usual chest radiographic findings include interstitial edema early in the disease course, and lack of peripheral distribution of air space disease [13], the hemodynamic parameters in HCPS are unique because the pulmonary artery wedge pressure is low, unlike other forms of cardiogenic shock. Failure to improve with intravascular volume repletion differentiates these patients from other forms of hypovolemic shock [5]. Treatment of HCPS during the pulmonary edema and shock phase poses a dilemma because severe pulmonary capillary leak worsens with volume administration. Inotropic support and mechanical ventilation are indicated to improve cardiac output and oxygenation. Despite maximal therapy, many patients deteriorate and rapidly develop lactic acidosis and cardiac arrest.
Venoarterial cannulation for ECMO was used to treat our patients with HCPS because it provides support for both cardiac and respiratory failure. Venovenous ECMO, an alternative extracorporeal support for patients with severe respiratory failure not requiring cardiac support, was not applicable to treat patients with HCPS. A roller pump was used for all patients supported with ECMO, mainly because it provides constant flow independent of distal resistance, and is associated with a reduced risk of hemolysis [14].
Vascular access was accomplished percutaneously or by direct cut-down of the common femoral artery and vein. Whenever possible in group B patients, as an initial step prior to cannulation before hemodynamic deterioration occurred, we percutaneously placed vascular sheaths in the common femoral artery and vein. This simple maneuver facilitated percutaneous insertion of the ECMO cannulas once the decision was made to initiate circulatory support. However, cannulation of the femoral vessels was achieved percutaneously in only 18 (35.3%) patients, and an open technique was required in the other 33 (64.7%) patients. No significant difference in the use of percutaneous cannulation between the groups was observed. Cannulation during cardiac massage (10 patients), and difficulty inserting the venous cannula percutaneously were the most common reasons requiring open cannulation. Poor venous drainage was usually corrected by repositioning the tip of the venous cannula in the right atrium. Additional cannulation of the right internal jugular vein was required in five (10%) patients because the venous cannula could not be advanced into the right atrium using the femoral approach. Insertion of an additional cannula into the distal femoral vein was required in five patients to relieve venous leg congestion.
Ischemic injury of the lower extremity is a potential complication associated with prolonged ECMO support. In our experience, leg ischemia occurred in four (7.8%) patients, of whom two were cannulated percutaneously. To prevent this complication, we place a distal perfusion cannula in the superficial femoral artery in all patients cannulated with the open technique, as recommended by several authors [15,16]. For patients cannulated percutaneously, we place a small distal arterial cannula in the SFA or a distal perfusion cannula in the posterior tibial artery inserted under direct vision. Huang and colleagues [17] use a pressure criterion to select the patients at risk of distal leg ischemia, and place a perfusion catheter if the mean distal pressure is below 50 mmHg.
Bleeding is the most common complication of ECMO, mainly because of systemic heparinization [18]. In addition, thrombocytopenia and platelet dysfunction are incremental risk factors for bleeding in patients with Hantavirus infection. Severe hemorrhage may lead to profound hemodynamic decompensation during ECMO. Management of bleeding includes volume-for-volume replacement of blood loss, platelet transfusion to >125 000 mm−3, and decreasing the ACT to 180 s. Epsilon amino-caproic acid, aprotinin, and tranexamic acid have recently been reported to minimize bleeding complications in ECMO patients [18]. Despite aggressive treatment of the underlying coagulopathy, surgical re-exploration of the cannulation site is usually necessary. According to the Registry Report 2010 of the Extracorporeal Life Support Organization (ELSO), cannulation site bleeding occurred in 16% and 20.4% of adult patients on ECMO for respiratory failure and cardiac support, respectively. Magovern [19] reported that bleeding was noted in most patients with post-cardiotomy cardiogenic shock supported on ECMO, and 48% required re-operation for femoral vessel repair. In our series, severe bleeding requiring re-exploration occurred in 10 (30%) of 30 patients following open femoral cannulation and in two (11.1%) of 18 patients who developed a large retroperitoneal hematoma following percutaneous cannulation.
The overall survival in our series was 66.6% (34 of 51 patients). Given the fulminant and irreversible progression to cardiac failure seen in our patients and by criteria from our previous experience with HCPS patients, none of these patients would have survived without circulatory support. A trend toward improved survival was observed in the second group of patients. As our numbers increased, this difference has achieved statistical significance (Table 3). By comparison, survival of adults supported on ECMO for post-cardiotomy cardiogenic shock was 36% in the Allegheny experience [19], and 34% for adult cardiac failure and 53% for adult respiratory failure reported in the 2010 ELSO Registry.
In conclusion, two-thirds of the patients with severe HCPS, who were supported with ECMO, survived and recovered completely. The complications associated with both types of femoral cannulation may be attributed to the fact that all patients were in shock or in full cardiac arrest, and the procedure had to be done expeditiously. Several factors have contributed to the improved overall survival of patients in the second phase of our experience. These include: a treatment strategy consisting of insertion of percutaneous vascular sheaths that facilitates later cannulation, and almost concomitant intubation and placement on ECMO when the patient decompensated, increased awareness of the syndrome from regional emergency physicians leading to earlier recognition and referral as suggested by fewer group B patients intubated on admission (Table 5 ), as well as refinement of diagnostic and treatment algorithms as suggested by fewer patients with lactate >10 mmol l−1 prior to cannulation (Table 5). When confronted with patients suspected of SNV infection, we recommend early transfer to a center that can offer ECMO before pulmonary or cardiac deterioration.
Comparison of groups A and B using Mann–Whitney test.
Presented at the 24th Annual Meeting of the European Association for Cardio-thoracic Surgery, Geneva, Switzerland, September 11–15, 2010.
Acknowledgments
The authors are very grateful to Robert Dragan, Patrick Kelly, John Clement, and Will Cabaniss (Perfusion Services at University of New Mexico Health Sciences Center); to Denise M. Coleman, Dawn M. Joseph, Tito R. Monge, and Kristine M. Pleacher (Division of Pediatric Critical Care, Department of Pediatrics, University of New Mexico Health Sciences Center); and to Michel A. Boivin, Helen K. Busby, Betty Chang, Michelle S. Harkins, Gary Iwamoto, and Akshay Sood (Division of Critical Care Medicine, Department of Internal Medicine, University of New Mexico Health Sciences Center).
