Background Occult hypoperfusion (OH) is defined as hypoperfusion in the presence of normal vital signs. It is associated with increased length of stay (LOS) and increased mortality.
Objectives To compare four methods of detecting OH in adult major trauma patients at a level 1 trauma centre—base excess (BE), non-invasive cardiac index (CI), shock index (SI) and rate over pressure evaluation (ROPE).
Method Patients meeting the Victorian Trauma Registry entrance criteria who presented with normal vital signs were enrolled. CI was obtained half hourly using an USCOM monitor. BE, SI and ROPE were obtained clinically.
Results Sixty-four patients were enrolled. Mean injury severity score (ISS) was 19 (SD 11) and mean hospital LOS was 10 days (SD 8). Two patients (3%) died in hospital. Ten patients (16%) had OH detected by CI, seven (11%) by BE, four (6%) by SI and two (3%) by ROPE. There was a significant association between hospital LOS and BE (p<0.005). Agreement between BE and CI in detecting OH was poor to fair, κ=0.25.
Conclusion OH is associated with increased hospital LOS and occurs in up to 16% of patients. BE performed best as a detection method.
- cardiac index
Statistics from Altmetric.com
Some trauma patients, despite having normal vital signs, have evidence of occult hypoperfusion (OH) resulting in insufficient tissue perfusion and oxygenation. The presence, degree and duration of OH has been repeatedly related to adverse outcomes.1 2 Detecting OH is difficult. A patient's central haemodynamics cannot be determined by clinical examination3 and even the definition of abnormal vital signs in the trauma patient is being questioned.4
OH can be detected in a variety of ways in trauma patients. Base excess (BE) has been demonstrated to be a better predictor of outcome than blood pressure in two large database reviews.4 5 Prospectively, it has been shown to differ between survivors and non-survivors in blunt trauma6–10 and those with major vascular injuries.11 12 It has also been shown to maintain this difference in the presence of ethanol intoxication.13 BE has been linked with neutrophil expression14 and chemiluminesence15 as well as multisystem organ failure,16 17 acute respiratory distress syndrome,18 transfusion requirements9 and fluid requirements19 in trauma patients. While a level of −6 mEq/l appears to be the best predictor of outcome9 10 18, larger retrospective studies have demonstrated increased mortality for levels less than −2 mEq/l.4 5
In patients monitored with the pulmonary artery catheter (PAC) cardiac index (CI) values have been shown to differ significantly between survivors and non-survivors of trauma. This has been demonstrated in series of patients with multiple trauma,20 blunt trauma,21 in elderly trauma patients22 23 and in burns patients.24 Shoemaker et al25 demonstrated a strong relationship between decreased CI and non-survival in a prospective study of nearly 1000 trauma patients monitored using non-invasively (transthoracic electric bio-impedance; TEB). Two hundred and sixty-seven of these patients had simultaneous monitoring with both the PAC and TEB and the relationship between decreased CI and non-survival was independent of the monitoring method. The USCOM (Coefficient Systems P/L, Coffs Harbour, Australia) is a non-invasive, continuous wave Doppler CI monitor. The degree of agreement with the PAC has been reported several times,26 27 including in trauma patients,28 and its ease of use has been demonstrated.29
Shock index (SI) is defined as the ratio of heart rate/systolic blood pressure (normal values 0.5–0.7) and was first described in 1967 by Allgower et al30 They also demonstrated that SI increases in proportion with gastrointestinal haemorrhage.30 Oestern et al20 then demonstrated that there was a significant difference in SI and CI between survivors and non-survivors in multiple trauma. Two large retrospective studies of trauma patients have demonstrated that SI outperforms vital signs in the prediction of mortality, transfusion requirements and injury severity.31 32 A SI value greater than 0.8 performed best in predicting trauma mortality31 and the value of 0.9 was found to predict outcome best in a mixed group of emergency department (ED) patients, including trauma patients.33
The rate over pressure evaluation (ROPE) is calculated by dividing the heart rate by the pulse pressure (heart rate/systolic blood pressure−diastolic blood pressure).34 The ROPE evaluation has only been described in one paper of 184 trauma patients with a level of 3.0 or greater being predictive for developing decompensated shock.34 As far as we are able to discern, no previous paper has been published comparing these four methods for detecting OH.
We aimed to determine the prevalence of OH among trauma patients at presentation and its impact on outcome and also to compare the four methods of detecting OH.
This observational pilot study was undertaken in an adult, level 1, trauma centre between January and July 2007. The study was approved by the Human Research and Ethics Committee, Royal Melbourne Hospital and the Standing Committee for Ethics and Research in Humans, Monash University.
Major trauma patients (using the Victorian State Trauma Registry definition—table 1) presenting with normal vital signs were enrolled.
OH was defined as a heart rate of 110 bpm or less, systolic blood pressure of 100 mm Hg or greater, temperature greater than 35°C and either CI less than 2.6 l/minute per square metre, BE of −3 mEq/l or less, SI of 0.9 or greater or ROPE of 3.0 or greater.
We selected a heart rate of 110 bpm as we were concerned about the effects of inadequate prehospital analgesia. The systolic blood pressure of 100 mm Hg was selected due to recent studies demonstrating increased mortality at levels well above 90 mm Hg.4 5 The definition of hypoperfusion used the lower limit of normal for CI,35 the level of BE at which mortality has been demonstrated to increase,4 5 the level of SI best shown to reflect outcome32 33 and the value for ROPE at which decompensated shock is likely to develop.34
We chose to use BE rather than lactate as a marker of hypoperfusion as it is routine for Ambulance Victoria to use Ringer's lactate solution as their principal crystalloid, and the use of lactate-containing solutions has been shown to confound the analysis of lactate levels.36
Patients had their vital signs and CI measured as soon as practicable after presentation and at half hourly intervals for the initial 4 h or until discharge from the ED, whichever occurred first. Blood gases were taken at a mean time of 75 minutes post-arrival (SD 63 minutes).
Sixty-four subjects were enrolled (table 2). Two were excluded from analysis as a result of an inability to obtain CI readings.
Ten subjects (16.1%) had OH detected by CI, seven (11.2%) by BE, four by SI (6.4%) and two by ROPE (3.2%). Table 3 demonstrates outcomes stratified by the method of detecting OH.
One of the possible limitations of this study was the high heart rate (110 bpm) used to define the upper limit of normal. However, analysis using a heart rate of 100 bpm or less to define normality demonstrates similar findings to the initial analysis. This is presented in table 4.
The Wilcoxon rank sum test showed a significant association between length of stay (LOS) and OH detected by BE (Z=−2.818, p<0.005), but not by CI (Z=−0.375, p = 0.71). The association between LOS and OH detected by BE was also demonstrated by univariate analysis by Cox regression (Z=−2.46, p=0.014) and by multivariate analysis (Z=−2.63, p=0.008). OH detected by CI, an injury severity score (ISS) of 15 or greater, age, gender and time in the ED were not associated with LOS.
Agreement between BE and CI in detecting OH was measured using the κ statistic: κ=0.25, indicating poor to fair agreement. Agreement between the other methods was not tested due to the small frequencies. The frequency table of agreement between BE and CI is presented in table 5.
Our population was older, had less penetrating trauma, a lower prevalence of OH and a lower mortality than previous studies.1 37 The mean ISS in our study was higher or comparable with the populations in those studies, indicating that differences in OH and mortality were not caused by injury severity. Given our small sample size the significance of these findings is uncertain.
The reported positive association between OH and morbidity is supported by this study. OH detected by BE had a strong relationship with LOS and was present in both patients who died. CI detected OH in only one of the patients who died. Neither SI nor ROPE was positive in sufficient subjects to allow a meaningful analysis.
Our study suggests that biochemical testing is the best way to identify OH. This would confirm previous reports,1 and we would recommend this method. However, given the small number of participants presenting with OH involved in this study, we also recommend that all methods should be examined further in larger clinical studies.
OH was detected in up to 16% of major trauma patients presenting to our institution. There was fair to poor agreement between CI and BE in detecting OH. OH detected by BE is associated with LOS in major trauma patients presenting to a level 1 trauma centre.
Funding This study was supported by a project grant from the Transport Accident Commission (TAC). The TAC had no involvement in the study design, collection, analysis and interpretation of data, in the writing of the report or the decision to submit for publication.
Competing interests None.
Ethics approval This study was conducted with the approval of the Human Research and Ethics Committee, Royal Melbourne Hospital and the Standing Committee on Ethics in Reseach on Humans, Monash University.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.