Cdn-links.lww.com



Long-Term Survival in Adults Treated with Extracorporeal Membrane Oxygenation for Respiratory failure and SepsisCritical Care MedicineViktor von Bahr*Jan HultmanStaffan EksborgBj?rn FrencknerH?kan Kalzén* Department of Physiology and Pharmacology, Section for Anesthesiology and Intensive Care Medicine, Karolinska Institutet, Stockholm, Sweden.E-mail address: viktor.von.bahr@ki.seSupplemental Digital Content 3 – Methods and resultsThis supplemental content is written in an attempt to clarify some important aspects of our methods and results.On included and excluded patientsSince our main goal was to provide survival data on as many patients from our center as possible, with the longest possible follow-up time, we decided to include all respiratory failure and sepsis patients (n=280). According to the discussions below, in these patients the ELSO indication criteria were followed most of the time, while ECPR patients and cardiac patients were not of interest in this study. Non-Swedish citizens were excluded since they lacked personal identification number and hereby were impossible to follow.On patients with several ECMO runsFor patients who had more than one support run outcome was attributed to the first instance of mechanical support. This was chosen as a pragmatic approach. E.g. in a patient treated with VV ECMO for aspiration pneumonia with a P/F ratio of 65 mm Hg, and at a later instance with VA ECMO for pneumomediastinum with another P/F ratio, this patient was grouped based on the first instance of mechanical support, i.e. belonging to the group aspiration pneumonia, with P/F 65 mm Hg.On ECMO systems and the long time span of patients included in the study It is important to remember that ECMO treatment is a constantly evolving field, with new pump technology, heparin coated tubing systems and new ECMO and ventilator treatment strategies that have been developed during the follow-up time. The reported survival data should therefore, as with the early data provided in the ELSO registry, mainly work as a historic reference.When our program started a roller pump was used together with Avecor silicone membrane oxygenator (16). Today only centrifugal pumps are used (CentriMag, Levitronix, Zurich, Switzerland and in some instances CARDIOHELP system, Maquet Cardiopulmonary). Oxygenators have been changed to polymethylpentene membranes with MEDOS 7000 LT (Medos Medizintechnik, Stolberg, Germany) as standard. The ventilator settings and approach to sedation have however remained more or less unchanged since the beginning in 1995. This means that as a standard procedure, the ventilator settings, including fraction of inspired oxygen (FiO2), are reduced within 3 hours from ECMO start in order to prevent further ventilator-associated lung injury. This ‘gentle ventilation’ is maintained during lung recovery until ECMO discontinuation. Sedation is kept at a minimum and the patients are generally awake, triggering their own pressure supported breaths and being able to communicate when possible. After ECMO is discontinued, the patients are ventilated conventionally with low settings of pressure support and inspired fraction of oxygen until breathing spontaneously (16).Changes of equipment have been gradual and during periods there has been a significant overlap. A separation of the data into specific eras was for this reason considered impossible. On the ECMO Center Karolinska and diagnostic groupsThe ECMO Center Karolinska is a high-volume center, and has treated more than 1,000 patients with ECMO since 1987. Initially, only neonatal and pediatric patients were treated at the center. The first adult patient was treated in 1995, and since 2007 adult patients constitute the most numerous age group. *n=320, including the 280 patients treated for respiratory failure or sepsis described in this study, 28 ECPR patients, five cardiac patients, four patients with multiple ECMO runs (i.e. same patient counted several times) and three patients with <90 days of follow-up timeEarly experiences with 17 adult ARDS patients treated at our center were described in 2000 (16), and 13 adult patients treated for the H1N1 virus were described in 2010 (6).The wide spectrum of patients treated with ECMO makes grouping of patients difficult. The main problems are heterogeneity within groups, and that many patients – because of this - risk ending up in groups called “other”. The ELSO criteria for grouping of patients are routinely followed at many centers, with the main categories “Respiratory”, “Cardiac” and “ECPR” (i.e. cardiopulmonary resuscitation with ECMO) (4). The ECMO Center Karolinska is a dedicated ECMO ICU, and treats patients in all three main categories, but no developed ECPR program exist (some ECPR patients are however treated as an exception, see below). By tradition, another ICU at the Karolinska hospital (the thoracic ICU) treats most adult patients with subacute cardiac indications, primarily with perioperative cardiac failure. Therefore, the approach of our center is best described as an “everything but perioperative adult cardiac, if possible” approach, rather than an “only respiratory” one, and the vast majority (>90%) of patients treated are patients with respiratory failure or sepsis. When patients are treated for cardiac failure or ECPR at our center, ECMO is often used as a salvage therapy (since our center has the longest experience, exclusively treats ECMO patients and all members of staff are highly trained in handling the equipment, some referrals regarding cardiac and ECPR patients are received, often hyper acute as a last-resort-action). The ELSO respiratory subgroups are the following (15):Viral pneumonia Bacterial pneumoniaAspiration pneumoniaARDS ARF/non-ARDS Other?In our ICU database, discussions often arise about the overlap between ARDS, pneumonia and sepsis patients, where the same patient could suffer from all three simultaneously. Should sepsis patients have their own group, subdivided into “Septic shock” and “Sepsis without circulatory failure”? In that case, in what group does a septic pneumonia patient who fulfills the ARDS criteria belong? Also, a vast number of patients are grouped as “Other”. Patients with non-pulmonary infections have no natural group within the ELSO system, and patients with traumatic lung contusion who developed respiratory failure due to e.g. shock lung, pneumothorax, lung bleed and rib fractures, differ a lot from patients in the ELSO category “ARDS post-operative/trauma” that developed ARDS as part of a systemic inflammatory process. Therefore modified ELSO subgroups for respiratory failure were created, to avoid syndrome diagnoses such as sepsis or ARDS, and to, as far as possible, group patients by the etiology of respiratory failure (see also Table 1 of the main text):PneumoniaViral pneumonia Bacterial pneumonia Aspiration pneumonia Non-pulmonary infection (new)Severe inflammatory response (new)Traumatic chest/lung contusion (new)Other pulmonary etiology Patients who developed respiratory failure due to confirmed or highly suspected infectious organisms (i.e., sepsis) were classified as pneumonia (bacterial and viral pneumonia) or non-pulmonary infection. Aspiration pneumonia patients, who can suffer from both chemical and infectious pneumonitis, were included as a subgroup to pneumonias. Patients with a systemic inflammation causing respiratory failure were classified as severe non-infectious inflammatory response. The initial intention was to create a subdivision into “postoperative/trauma” and “other”, but these groups were merged since there was only one surviving patient in the latter subgroup. Traumatic contusions causing immediate respiratory failure were given their own group. Two ECMO physicians reviewed all unclear cases based on the patients’ charts, and met later for discussion, coming to an agreement on the appropriate diagnostic group. In future studies, with patients that differ from ours, more groups can be created. Nevertheless, heterogeneity within groups is hard to avoid, with different levels of illness, cannulation, background, age and complications. We think it is reasonable to further discuss the need for a diagnostic strategy that would improve the homogeneity of patients within each group, and thereby create more reliable and comparable data. On the criteria for starting ECMO, and the choice of cannulationThe ELSO criteria have generally been used since start of the adult program in 1995, and these have somewhat changed over time (16, 38). However, since ECMO is constantly evolving and offers a relatively new (at least in the late 90’s) and unique way of supporting both the circulatory and respiratory system, there are cases where ECMO is used as a salvage therapy, i.e. as a last resort in patients with a potentially reversible condition, but with many unknowns. Also, since patients are referred to our center from all over Sweden with the undisputable risk of errors or worsening during long distance transportation, sometimes VA ECMO is applied even when there is a slight risk of right heart failure during transport. This is another reason why ECPR and cardiac patients were not included, and why comparison of survival data between centers is cumbersome.On survival based on cannulationThe following data can be extracted from table 3:Survival estimates by Cannulation technique (from Table 3)nSurvives treatmentSurvives to dischargeSurvives 90d1Y survivala5Y survivala1Y conditionalb survival5Y conditionalb survivalVV-ECMO135103 (76%)99 (73%)84 (62%)60% NAR 7656%NAR 3996%90%VA-ECMO7046 (66%)45 (64%)38 (54%)51% NAR 3444%NAR 1097%84%Converted to VVd1310 (77%)10 (77%)9 (69%)69% NAR 841% (3YS)e100%60%Converted to VA379 (24%)9 (24%)8 (22%)19% NAR719%NAR 488%88%Groups with fewer than 5 patients are written in italic. Definition of abbreviations: Y=year; NAR=number at risk; VV=veno venous cannulation; VA=veno arterial cannulation. Results are presented as number (percentage).a Survival calculated by means of the Kaplan Meier method.b Survival for patients that were alive 90 days after treatment.d Includes VA-VV (i.e. patients converted from an initial VA cannulation to VV) and VV-VA-VV. e Follow-up time was 4.5 years, in median 2.3 years. NAR (3 years)=3.Given the discussion above, one must bear in mind that the presented cannulation groups do not fully represent the ideal separation into ”pure” circulatory support with VA ECMO vs VV respiratory support.Classifications are always to some extent simplifications, which affect the reliability when interpreting the data in each group.For patients converted to VA ECMO, the survival to discharge was very poor. This need for conversion in many cases probably represents a complicating circulatory failure and deterioration despite VV ECMO support, i.e. if the patient’s heart fails in addition to respiratory failure, the survival prognosis is dismal. In line with other groups presented in this study, survivors to 90 days after decannulation was good, except in the very small “converted to VV” group, where two late deaths strongly affected the 5 year conditional survival. The good short-term survival in this group probably represents the opposite situation of the VA group, i.e. in patients needing VA ECMO support, if stable enough to convert to VV support only, the prognosis for survival is good. These results support the intuitive impression described by many centers, but needs to be further evaluated in future studies. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download