Prehospital Intravenous Fluid Administration is Associated

With Higher Mortality in Trauma Patients:

A National Trauma DataBank Analysis

(Annals of Surgery 2011;00(00):1-7)

Helicopters and the civilian trauma system:

National Utilization Patterns Demonstrate Improved

Outcomes After Traumatic Injury

(Trauma 2010;69(5):1030-4)

Background/rationale• Helicopter transport of the injured

patient has been an integral component of trauma care in the United States since the 1970s, due in large part to the military aeromedical experience

• Helicopter transport in the civilian trauma population remains controversial with some asserting that it is over used and others questioning its benefits

• However, previous studies looking at this issue have been limited by small patient populations, relying on single centre or meta-analysis design

• The administration of intravenous (IV) fluids is a key component of the prehospital treatment of trauma patients intended to haemodynamic-ally stabilize trauma patients by replacing intravascular volume and maintaining vital organ perfusion

• An increasing body of evidence has demonstrated that IV fluid administration does not improve survival in trauma and may actually be of harm in certain subsets of trauma patients

• Given the previous data, we chose to examine the impact of prehospital IV catheter placement and fluid administration on survival following traumatic injury using the NTDB - the largest dataset of trauma patients ever created

Objectives

To compare patient outcomes after transport from the scene of injury by helicopter transport and ground transport

To determine the influence of prehospital IV fluid administration on patient outcome

Study design

Retrospective study using 1 year (2007) worth of data from the American College of Surgeons National Trauma Data Bank (Version 8)

Retrospective study using 5 years (2001–2005) worth of data from the American College of Surgeons National Trauma Data Bank (Version 6.2)

What is the National Trauma Data Bank (NTDB)?

• Established by the American College of Surgeons as a public service to be a repository of trauma related data voluntarily reported by participating trauma centres

• Follows on from the major trauma outcomes study (MTOS) 1982-1989• Contains data on approx 3 million cases from over 900 US trauma centres• It is not a population-based data set• It is a convenience sample and is thus subject to various forms of bias

– Selection bias• Has a disproportionate number of larger hospitals with younger and more severely

injured patients– Information bias

• Inconsistency in the way clinical variables are measured - ISS has been found to calculated differently by different registry programs

• 31 of the 173 level I trauma centres in NTDB 7.0 submitted no complications associated with trauma incidents at all

– Missing data• In most cases NTDB data are not missing at random and analyses are subject to bias if

missing data are ignored • The results may be misleading when excluding all observations with missing data

Eligibility

• Inclusion criteria:– admitted in 2007 (n=507,262)– complete transport data

(n=427,493)– transported directly to a

trauma centre from the scene of injury by helicopter or by ground ambulance (n=258,989)

• Exclusion criteria:– patients who were dead on

arrival (n=602)

Final n = 258,387

• Inclusion criteria:– admitted between 2001 and

2005 (n=1,466,887)– complete prehospital files

(n=776,734)

• Exclusion criteria:– Incomplete data? (n=465,663)

Final n = 311,071

Variables• Primary outcome - survival to

discharge– Secondary outcome - discharge to

home

• Helicopter transport vs ground ambulance transport

• Variables adjusted for– age, gender, mechanism, ISS,

hypotension, insurance status, GCS<9

plus– abnormal RR, mechanical

ventilation, emergent operation, ICU admission, trauma centre designation

• Primary outcome - in-hospital mortality

• Prehospital IV fluid vs no prehospital IV fluid

– Prehospital IV - defined as patients with the word “intravenous” or “IV” in the prehospital procedure file

• Variables adjusted for– age, gender, mechanism, ISS,

hypotension, insurance status, GCS<9

plus – race, endotracheal intubation,

military antishock trousers , spine immobilization, splinting, and chest decompression

Study size

• A total of 776,734 patients with complete prehospital procedure files were identified from the 1,466,887 total patients in the National Trauma Data Bank

• Multivariable logistic regression was used to examine the relationship between prehospital IV and mortality in the 311,071 patients with complete data

Statistical methods• Univariate and multivariate logistic

regression analysis• X2 tests were used to compare

categorical variables, and nonparametric Mann-Whitney U tests were used to compare continuous variables.

• Covariates were included in the model if associated with survival at a level of p < 0.2

• A p value < 0.05 was considered significant

• Analysis was conducted using SAS JMP version 7.0 (Cary, NC)

• Univariate and multivariate logistic regression analysis

• Logistic regression analysis was repeated in numerous subgroups

– after excluding DOA patients– blunt patients, penetrating

patients, GSW, normotensive patients, hypotensive patients, hypotensive blunt patients, hypotensive penetrating patients, hypotensive GSW patients, patients with GCS < 9, patients with severe head injury (GCS<9 and head AIS 3–5) and need for immediate surgery

– based on ISS category• Analysis was conducted using

Stata/Multi-Processor 10.0 (StataCorp, College Station, TX)

Helo study

IV study

IV study

Results• Unadjusted survival was lower in

patients transported by helicopter (92.5% vs. 95.6%; p < 0.01)

• Adjusted multivariable analysis demonstrated that patients transported by helicopter were more likely to survive (OR 1.22; 95% CI 1.18–1.27)

– HT was a predictor of discharge to home (OR 1.05; 95% CI 1.02–1.07) after adjustment for covariates

It’s easier to interpret a logistic regression model presented in terms of the less common binary event (e.g. death rather than survival) as this results in odds that are similar to the corresponding probabilities and makes the effects of covariates more apparent. For example if the probability of death is 5% or 0.05 then the odds of death is 0.05/0.95 = 0.053, whereas if the probability of survival is 95% or .95 then the odds of survival is 0.95/0.05 = 19.

• Unadjusted mortality was higher in patients receiving prehospital IV fluids (4.8% vs. 4.5%; P < 0.001)

• Adjusted multivariable analysis demonstrated that patients receiving IV fluids were more likely to die (OR 1.11; 95% CI 1.05–1.17)

– The association was identified in nearly all subsets of trauma patients but was especially marked in patients with penetrating mechanism (OR 1.25; 95% CI 1.08–1.45), hypotension (OR 1.44; 95% CI 1.29–1.59), severe head injury (OR 1.34; 95% CI 1.17–1.54), and patients undergoing immediate surgery (OR 1.35; 95% CI 1.22–1.50)

Results• Unadjusted survival was lower in

patients transported by helicopter (92.5% vs. 95.6%; p < 0.01)

• Adjusted multivariable analysis demonstrated that patients transported by helicopter were more likely to survive (OR 1.22; 95% CI 1.18–1.27)

– HT was a predictor of discharge to home (OR 1.05; 95% CI 1.02–1.07) after adjustment for covariates

It’s easier to interpret a logistic regression model presented in terms of the less common binary event (e.g. death rather than survival) as this results in odds that are similar to the corresponding probabilities and makes the effects of covariates more apparent. For example if the probability of death is 5% or 0.05 then the odds of death is 0.05/0.95 = 0.053, whereas if the probability of survival is 95% or .95 then the odds of survival is 0.95/0.05 = 19.

• Unadjusted mortality was higher in patients receiving prehospital IV fluids (4.8% vs. 4.5%; P < 0.001)

• Adjusted multivariable analysis demonstrated that patients receiving IV fluids were more likely to die (OR 1.11; 95% CI 1.05–1.17)

– The association was identified in nearly all subsets of trauma patients but was especially marked in patients with penetrating mechanism (OR 1.25; 95% CI 1.08–1.45), hypotension (OR 1.44; 95% CI 1.29–1.59), severe head injury (OR 1.34; 95% CI 1.17–1.54), and patients undergoing immediate surgery (OR 1.35; 95% CI 1.22–1.50)

Numbers needed

DiscussionLIMITATIONS• First are those inherent to a

retrospective design, [such as] the limited number of variables and detail of information accessible for analysis which limits the type of data that can be used in regression modelling to control for potential confounders

• Second are those outlined by the American College of Surgeons Committee on Trauma for use of the NTDB

ADVANTAGES• The main advantage of the NTDB

is the access afforded to a large number of patients in a national sample

LIMITATIONS• This retrospective study suffers

from some inherent limitations, largely due to potential residual confounders, which were not available

– prehospital transport times– urban versus rural care

• The database did not differentiate between IV catheter placement and IV fluids

• “Due to these limitations, we could not determine the potential causal pathway of the higher mortality”

ADVANTAGES• However, the overall advantages

of using the large NTDB probably outweigh the potential limitations

Conclusions• Trauma patients transported

by helicopter were more severely injured, had longer transport times, and required more hospital resources than those transported by ground

• Despite this, HT patients were more likely to survive and were more likely to be discharged home after treatment when compared with those transported by ground

• Despite concerns regarding helicopter utilization in the civilian setting, this study shows that HT has merit and impacts outcome

• The harm associated with prehospital IV fluid administration is significant for victims of trauma

• The routine use of prehospital IV fluid administration for all trauma patients should be discouraged