Progression of Organ Failure in Patients Approaching Brain Stem Death


  • Support was provided solely through institutional funds

* Corresponding author: Mark T. Keegan,


We performed a retrospective cohort study to document the progression of organ dysfunction in 182 critically ill adult patients who subsequently met criteria for brain stem death (BSD). Patients were admitted to intensive care units (ICUs) of Mayo Medical Center, Rochester, MN, between January 1996 and December 2006. Daily sequential organ failure assessment (SOFA) scores were used to assess the degree of organ dysfunction. Serial SOFA scores were analyzed using analysis of variance (ANOVA). Mean (standard deviation, SD) SOFA score on the first ICU day was 8.9 (3.2). SOFA scores did not significantly change over the course of ICU stay. 67.6% of patients donated one or more organs after BSD was declared. The median time from ICU admission to declaration of BSD was 18.8 h (interquartile range 10.3–45.0), and in those who donated organs, the time from declaration of BSD to organ retrieval was 11.8 h (9.5–17.6). The fact that mean SOFA scores did not change significantly over time, even after BSD occurred, has implications for the timing of retrieval of organs for transplantation.


Organ transplantation is an effective treatment for end-stage organ failure. Increased public awareness, improved efficiency of the donation process, greater expectations for transplantation, expansion of the donor pool and the development of donor management protocols have led to unprecedented organ procurement and transplantation (1–8). In 2006, more than 28 000 patients received organ transplants from more than 14 000 donors in the United States (9). Although donation after cardiac death has particularly increased since the beginning of the Organ Donation Breakthrough Collaborative in 2003 (8), most organ donors have brain stem death (BSD). The demand for organs continues to far outweigh the supply, and more than 17 patients die each day while waiting for a transplant (10). Further understanding and characterization of the progression of critically ill patients toward BSD is necessary.

Optimization of physiological parameters is an integral part of the management of all ICU patients, but is especially important in potential organ donors. Of further importance is the optimal timing of organ procurement. There is a perception that once BSD occurs, extracranial organ dysfunction rapidly follows. In the past, retrieval of organs for transplantation occurred as soon as possible after declaration of BSD. Recent evidence suggests, however, that a compromise between optimization of organs and avoidance of deterioration is needed (11,12).

The aim of our study was to track the progression of organ dysfunction in this patient population, before and after declaration of BSD. We anticipated that our results would provide further information to the transplant community regarding the timing of organ retrieval.


After institutional review board approval, we performed a retrospective cohort study. Adult patients admitted to the intensive care units (ICUs) of Mayo Medical Center, Rochester, MN, a tertiary referral center, between January 1996 and December 2006, who subsequently met BSD criteria were identified. Patients in whom authorization for review of medical records for research was not given and those patients under the age of 16 were excluded. Institutional policy dictates that the local organ procurement organization is contacted by the nursing or medical staff when a patient meets or may meet the criteria for BSD or is otherwise deemed to be a potential organ donation candidate.

Data were abstracted from the medical record. Demographic details were collected, as were all variables required for the calculation of sequential organ failure assessment (SOFA). The time of ICU admission, duration between admission and declaration of BSD, duration between admission and organ donation if it occurred, and duration between declaration of BSD and donation were ascertained. All data were abstracted by a single researcher (FTL).

Originally developed for use in patients with sepsis, the SOFA has been used as a marker of organ dysfunction in critical care (including neurocritical care) and has been of benefit in prognostication (13,14). The SOFA score is derived from six physiological variables and is calculated for each 24 h period (15,16). The elements of the SOFA score are detailed in Table 1. SOFA scores were calculated as described by Vincent and colleagues for each 24 h period during which a patient was present in the ICU (15,16). A SOFA score was also calculated for the period between the declaration of BSD and organ retrieval, if the patient became an organ donor. The worst physiological value for each 24 h period was recorded. Values obtained during apnea testing were excluded for calculation of SOFA.

Table 1.  Calculation of the sequential organ failure assessment (SOFA) score
SOFA score1234
  1. Adrenergic agents administered for at least 1 hour and doses in mcg per kg per minute.

  2. NE = norepinephrine; MAP = mean arterial pressure (mm Hg).

with respiratory support
 HypotensionMAP <70Dopamine ≤5 or dobutamine any doseDopamine >5 or epinephrine ≤0.1 or NE ≤0.1Dopamine >15 or epinephrine >0.1 or NE >0.1
Central nervous system
 Glasgow coma score13–1410–126–9<6
 Creatinine (mg/dL) or urine output1.2–1.92.0–3.43.5–4.9 or <500 mL/day>5.0 or <200 mL/day
 Bilirubin (mg/dL)1.2–1.92.0–5.96.0–11.9>12.0
 Platelets (×103/mm3)<150<100<50<20

Data regarding the organs retrieved from each patient and medium-term follow-up of graft function were obtained, where available, from the local organ procurement agency.

Continuous data were summarized as mean (standard deviation, SD) or median (interquartile range, IQR). Normally and nonnormally distributed continuous variables were compared using t-tests and Wilcoxon rank sum tests, respectively. Chi-square analysis was used to compare categorical variables. Analysis of variance was used to compare serial SOFA scores. Data were analyzed using JMP statistical software version 6.0 (SAS Institute, Cary, NC).


After exclusion of 6 patients in whom authorization for use of medical records for research was not given, 182 patients who met brain death criteria during the study were included. The mean (SD) age was 46.5 (±18.7) years. The majority of patients (92.9%) had suffered an isolated intracranial catastrophe. The remainder had multiple trauma. Fifty-four patients (29.7%) underwent a surgical procedure (35 neurosurgical procedures, 13 procedures for trauma and 6 other surgical procedures). The prevalence of preexisting extracranial comorbidities was low (10 patients with cancer and one each with cirrhosis and congestive heart failure). Twenty-seven patients (14.8%) required cardiopulmonary resuscitation prior to ICU admission. Sixty-nine percent of the patients had a Glasgow coma score (GCS) (best motor response, best verbal response and eye opening) of 3 on admission to the ICU. Although continuous infusions of pressors or inotropes were being administered to only 18% of the patients on admission to the ICU, by the end of the first ICU day 74% of patients required pressors or inotropes. For subsequent ICU days, between 67% and 94% of patients required pressor/inotrope administration.

The mean SOFA score on the first ICU day was 8.9 (SD 3.2). The mean SOFA scores for each ICU day, along with the number of patients for whom SOFA scores were calculated on each day, are shown in Figure 1. SOFA scores did not significantly change over the course of the ICU stay (p = 0.41). Mean scores for the individual components of the SOFA score for the first 10 ICU days are presented in Table 2.

Figure 1.

Mean SOFA scores for each ICU day for 182 patients who developed brain stem death.

Table 2.  Mean scores for individual components of the SOFA score for the first 10 ICU days
DayNumber of patientsRespiratoryCVSNeurologicRenalLiverCoagulation
  1. CVS = cardiovascular.

 3 671.
 4 351.

One hundred twenty-three patients (67.6%) donated one or more organs after BSD was declared. The reasons for nondonation included refusal by family (or previous statement by the patient) in 45 patients, autopsy required or medical examiner involvement in 5 patients, malignancy in 2 patients, lack of suitable recipients for organs in 2 patients, infection in 1 patient, pregnancy in 1 patient and unclear reasons in 3 patients. The median time from ICU admission to declaration of BSD was 18.8 h (IQR 10.3–45.0) and from declaration of BSD to organ retrieval was 11.8 h (9.5–17.6). There were no statistically significant differences between donors and nondonors with regard to demographics, SOFA scores, or physiologic variables (except for a nonclinically significant higher admission GCS in the donors) (Table 3).

Table 3.  Comparison of organ donors versus nonorgan donors in 182 patients who developed brain stem death
CharacteristicDonor group (n = 123)Non-donor group (n = 59)p-Value
  1. NA = not applicable; GCS = Glasgow coma score; CPR = cardiopulmonary resuscitation; BSD = brain stem death; IQR = interquartile range; SOFA = sequential organ failure assessment.

Age (years), mean (SD)45.6 (17.8)48.3 (20.5)0.36
Surgery, number (%)37 (30.1%)17 (28.8%)0.86
Admission GCS, median (IQR)3.0 (3.4)3.0 (3.3)0.01
GCS on admission to ICU
 419 4 
 5 2 2 
 6 5 2 
 7 1 1 
 ≥819 1 
Preadmission CPR, number (%)18 (14.6%)9 (15.3%)0.78
Pressors before ICU, number (%)21 (17.1%)13 (22.0%)0.42
Pressors at end of first ICU day number (%)86 (69.9%)48 (81.4%)0.10
Day 1 SOFA, mean (SD)8.6 (3.2)9.5 (3.2)0.07
ICU admission to BSD (hours), median, IQR 18.7 (10.7, 44.7)35.1 (8.9, 45.8) 0.50 
BSD to donation (hours), median, IQR11.8 (9.5, 17.3)NANA

Of the 123 patients in whom organs were retrieved for donation, information regarding the specific organs retrieved was available for 112. Organs retrieved included 184 kidneys, 105 livers, 44 hearts, 35 lungs, 37 pancreases and 3 intestines. Information regarding graft function was only available for organs retrieved between July 2004 and 2006 (30 patients) and data were incomplete even for that limited period. The small number of patients and organs for whom data were available made it impossible to discern any trends regarding survival of transplanted organs.


In our study, the mean SOFA score on admission to the ICU for patients who subsequently progressed to BSD was 8.9. The SOFA scores did not change significantly over the course of the ICU stay. Furthermore, the SOFA scores did not differ significantly between the period immediately prior to BSD compared with the period after BSD and before donation. The average time between declaration of BSD and organ retrieval was approximately 12 h.

Intracranial pathology may lead to nonneurologic organ dysfunction (13,14,17). Mass effect may lead to brainstem herniation and a complex physiologic response (18–20). The mixed sympathetic storm and increased vagal outflow, thermoregulatory and hormonal impairment, and spinal cord ischemia can profoundly affect all organs (18,19,21–23). The management of patients that proceed to BSD is complex and the organ donation process is challenging (18,21–24). It has been reported that up to 17–25% of potential donors can be lost due to complete circulatory collapse (25). However, in our study, the severity of organ dysfunction in the patients did not change significantly before—or even after—the declaration of BSD.

The potential implications of this finding are interesting. Much attention is focused on the maximal utilization of available donor organs. Early, aggressive and appropriate management of donors is associated with increased organ procurement and transplantation. For example, left ventricular regional wall motion abnormalities, common after BSD and a reason for nondonation, have been shown to improve on repeat echocardiogram and with intravenous β-blockade (26). What seems less clear is the ideal time for organ procurement after BSD. Kunzendorf et al. found that kidney grafts retrieved from donors with longer duration of brain death (>470 min) exhibited a higher incidence of primary graft function and better graft survival compared with grafts retrieved from donors with shorter duration of brain death (<470 min), and concluded that ‘[retrieving] a potential organ transplant from a donor as early as possible is probably not the best procedure in all situations’ (27). Longer periods between declaration of BSD and organ retrieval was also not associated with primary liver allograft dysfunction (28).

Delay of organ retrieval to allow optimization of organ function may be particularly important for maximization of lung procurement, which has remained at a low rate of 17% (10). Recent animal studies have shown that early hemodynamic damage during BSD is associated with increased reperfusion injury and that delaying retrieval may allow the lung to more fully recover (29). This period of recovery may also include a shift from more cerebroprotective ventilatory strategies to a more lung-protective strategy (30,31). Lung procurement has been delayed to allow time for bronchial hygiene and recruitment maneuvers in marginal donors, which may lead to improved oxygenation (as measured by PaO2 to FiO2 ratio) and subsequent suitability for transplantation (2,32).

Perhaps, on the basis of the results of our study, there is more time to maximize donor potential than has been traditionally considered. Optimization of organs prior to retrieval with potential increase in the yield of transplantable organs is the focus of initiatives by organ procurement organizations. Our study lends support to the increasingly common practice of waiting to optimize donor organs rather than rushing to the operating room for organ retrieval soon after BSD is declared.

Our study is limited by the fact that it is a single-center study. Management of patients approaching and following BSD will influence the progression of SOFA scores, and may vary between institutions. For example, the use of vasopressin is relatively common in this patient population, but is not accounted for in the SOFA score. Additionally, although the SOFA score did not change significantly following BSD and prior to organ procurement, the time was relatively short (median 11.8 h), and further investigation is warranted.

The decrease in the number of patients available for analysis as the days go by limits the power of the analysis. While our total of 182 patients reflect a large number of patients with BSD, it is relatively small when placed in the context of the number of patients who develop BSD every year.

Although the SOFA score has been validated in multiple studies to represent the degree of organ dysfunction present in a critically ill patient, it has not been validated as a mechanism to predict outcomes of transplanted tissue. Furthermore, we have not demonstrated a correlation between SOFA score and the performance of a transplanted organ. Nonetheless, we believe that demonstration of stable SOFA scores is evidence against the perception that patients with BSD need to proceed urgently to the operating room to avoid rapid decompensation and loss of all organs. Stability of SOFA over time may allow for aggressive interventions to improve organ function.

We recommend a prospective study of organ yield rates after implementation of protocols that serve to optimize potential donor organs. Ideally, a prospective randomized trial of early versus delayed (after optimization) organ retrieval would be performed, though the practicalities and ethics of such a study are debatable.


SOFA scores do not change significantly over the course of the ICU stay in patients who develop BSD. This may have implications for the time available to optimize organ donors prior to organ retrieval.


The authors appreciate the assistance of Ms. Jennifer Hull in the acquisition of the data relating to donated organs.