Download study of placing a second intravenous line in trauma

STUDY OF PLACING A SECOND INTRAVENOUS LINE IN TRAUMA
Mark A. Merlin, DO, Emily Kaplan, MD, Jeffrey Schlogl, MD, Heather Suss, MD,
Frank D. DosSantos, DO, Pamela Ohman-Strickland, PhD, Adam Shiroff, MD
is modeled after the Advanced Trauma Life Support
(ATLS) course. An additional IV line, in a trauma setting, is placed to protect against the potential loss of intravascular volume. However, only small percentages
of trauma patients present with hemodynamic instability and require aggressive fluid resuscitation for improved outcome.2
Studies have demonstrated that internal hemorrhage
may be worsened by aggressive resuscitation when a
mean arterial pressure (MAP) of 60–80 mmHg is chosen for the resuscitation endpoint.3–6 Aggressive fluid
management may result in a number of undesired
complications, which include lowering the patient’s
body temperature to a point of hypothermia, dilution
of essential coagulation factors, and the emergence of
visceral swelling with pulmonary edema. Capone et
al. studied fluid resuscitation in rats with uncontrolled
hemorrhage and demonstrated frequent coagulopathy
and rebleeding if resuscitated too aggressively.7 In addition, prehospital IV lines have been demonstrated
to increase on-scene times with subsequent worsened
outcomes.8–10
Aside from the detrimental effects of aggressive fluid
management, practical reasons for avoiding placement of a second IV line exist. Obtaining IV access in the prehospital arena is especially difficult
in a moving ambulance. This level of difficulty can
be detrimental to the patient, potentially resulting
in a vascular or nerve injury while attempting to
place a second line.11 Additional IV lines also increase the chance of accidental needlestick to the
emergency medical services (EMS) provider as well
as the possibility of introducing bacteria into the
patient.11
This study was conducted to evaluate the effect of
placement of one versus two IV lines on the endpoints
of heart rate, blood pressure, volume of fluid infused,
pulse oximetry, mental status (Glasgow Coma Scale
[GCS] score), rehospitalization rate, and 30-day mortality. Furthermore, we hypothesized that patients who
received additional IV fluids would have no difference
in these endpoints. Heart rate, systolic blood pressure,
and diastolic blood pressure were measured both at the
scene (prior to IV fluids) of the emergency and on arrival in the emergency department (ED) (after IV fluids). Injury Severity Score (ISS) was used to verify that
the patients in the two groups had experienced similar
injuries.
ABSTRACT
Objective. We evaluated the benefit of emergency medical
services providers’ placing a second intravenous (IV) line in
the prehospital trauma setting. Our hypothesis was that the
placement of a second IV catheter in trauma does not result in an improvement in heart rate, blood pressure, rehospitalizaton rate, or 30-day mortality. Methods. A retrospective chart review of 320 trauma patients in a one-year period
was conducted at our level I trauma center. All trauma patients who had vascular access obtained prehospitally were
included. Results. Patients with two IV lines received an average of 348.4 mL more fluid (95% confidence interval [CI]:
235.6, 461.1; p < 0.0001). No change in heart rate, pulse
oximetry, Glasgow Coma Scale score, systolic blood pressure, rehospitalization rate, or 30-day mortality was noted.
These effects persisted for patients who were initially tachycardic (heart rate 3.92 bpm; 95% CI –3.01, 10.82; p = 0.27)
or hypotensive (blood pressure 22.00 mmHg; 95% CI −4.17,
48.16; p = 0.10). Conclusions. Redundant prehospital IV
lines provided no noticeable benefit in physiologic support
for trauma patients. When controlling for confounding variables, no significant outcome difference was noted, even in
the hypotensive patients. The traditional approach for establishment of a secondary IV line in prehospital trauma
patients should not be followed in a dogmatic fashion.
Key words: emergency medical services; intravenous;
trauma; ambulance; prehospital
PREHOSPITAL EMERGENCY CARE 2011;15:1–6
INTRODUCTION
According to Prehospital Trauma Life Support (PHTLS) guidelines, it is standard of care to provide every trauma patient with two large-bore intravenous
(IV) lines in the prehospital setting.1 This curriculum
Received January 28, 2010, from the Department of Emergency
Medicine (MAM), Robert Wood Johnson Medical School, New
Brunswick, New Jersey; and the Departments of Medical Education
(EK, JS, HS, FDDosS), Public Health (PO-S), and Surgery (AS), University of Medicine and Dentistry of New Jersey–Robert Wood Johnson Medical School, New Brunswick, New Jersey. Revision received
October 16, 2010; accepted for publication October 18, 2010.
Address correspondence and reprint requests to: Mark A. Merlin,
DO, Robert Wood Johnson Medical School, Emergency Medicine, 1
Robert Wood Johnson Place, MEB 104, New Brunswick, NJ 08901.
e-mail: [email protected]
doi: 10.3109/10903127.2010.545612
1
2
PREHOSPITAL EMERGENCY CARE
APRIL/JUNE 2011
VOLUME 15 / NUMBER 2
FIGURE 1. State trauma triage criteria. ALS = advanced life support; AVPU = alert, verbal, painful, unresponsive; BP = blood pressure.
METHODS
This was a retrospective cohort study. It was approved by our university’s institutional review board,
which has a subcontract with our hospital. A chart
review was performed on 320 patients who were
evaluated between May 1, 2008, and April 30,
2009. The search terms “trauma” and “IV attempt”
were used in our electronic medical record (EMR)
(www.emscharts.com, West Mifflin, PA) to select our
initial patient population. The inclusion criteria were
all trauma patients who received IV access in the prehospital setting. Patients were identified as trauma vic-
tims by our prehospital providers if they met state
trauma guidelines (Fig. 1). This database was crossreferenced with the ED database, Sunrise Clinical Manager (SCM) (Eclipsys Corporations, Atlanta, GA).
Information for ISSs was obtained upon arrival of
patients to our trauma center. Rehospitalization rates
and 30-day mortality data were obtained by research
associates who called the individual household phone
numbers given upon arrival to our trauma center.
This study was conducted at an American College of Surgeons (ACS)-verified level I trauma center. Our trauma volume is approximately 1,500
patients per year, with 18% having experienced
Merlin et al.
3
TWO IVS IN TRAUMA
TABLE 1. Summaries of Raw Data
Variable
Age, years
Gender —female
Any fluids
Volume of fluids, mL
GCS score at 15
GCS score
Heart rate, bpm
Systolic blood pressure, mmHg
Diastolic blood pressure, mmHg
Pulse oximetry
Time (If
Applicable)
On-scene
Arrival
On-scene
Arrival
Change
On-scene
Arrival
Change
On-scene
Arrival
Change
On-scene
Arrival
Change
Overall (N = 236)
No. Missing
(% of Total)
3
2
5
5
8
22
3
22
23
2
3
4
7
7
12
25
16
30
53
(1%)
(1%)
(2%)
(2%)
(3%)
(9%)
(1%)
(9%)
(10%)
(1%)
(1%)
(2%)
(3%)
(3%)
(5%)
(11%)
(7%)
(13%)
(22%)
All Mean (±SD) or
n (% of Observed)
37.3
47
75
174.7
157
156
13.6
13.6
0.0
93.9
90.5
–3.6
131.8
133.3
0.7
81.1
79.3
–2.3
98.7%
( ± 19.3)
(20%)
(32%)
( ± 422.6)
(69%)
(73%)
( ± 3.1)
( ± 3.4)
( ± 1.6)
( ± 20.5)
( ± 17.9)
( ± 9.3)
(22.3)
( ± 23.5)
( ± 19.3)
( ± 15.3)
( ± 14.2)
( ± 17.5)
( ± 1.9%)
One IV Line Mean
(±SD) or n
(% of Observed)
37.6
38
44
76.7
116
114
13.8
13.7
–0.1
93.9
90.7
–3.5
132.5
132.6
0.3
82.0
78.9
–3.2
98.7%
( ± 20.0)
(23%)
(27%)
( ± 187.7)
(72%)
(76%)
( ± 3.0)
( ± 3.4)
( ± 1.5)
( ± 21.9)
( ± 18.9)
( ± 9.6)
(21.0)
( ± 23.1)
( ± 19.5)
( ± 14.1)
( ± 13.0)
( ± 14.0)
( ± 2.0%)
Two IV Lines Mean
(±SD) or n
(% of Observed)
36.5
9
31
425.1
41
42
13.2
13.3
0.2
93.8
90.1
–3.7
130.1
135.1
1.6
78.9
80.5
0.2
98.6%
( ± 17.5)
(14%)
(48%)
( ± 679.9)
(62%)
(67%)
( ± 3.4)
( ± 3.4)
( ± 1.9)
( ± 16.8)
( ± 16.0)
(8.8)
(25.3)
( ± 24.3)
( ± 18.9)
( ± 17.8)
( ± 17.0)
( ± 24.2)
( ± 1.6%)
p-Value for
One versus
Two
0.68
0.15
0.0020
0.0001
0.16
0.19
0.19
0.51
0.28
0.98
0.81
0.90
0.51
0.48
0.62
0.23
0.49
0.32
0.81
GCS = Glasgow Coma Scale; IV = intravenous; SD = standard deviation.
penetrating trauma. The county population of approximately 800,000 residents is made up of 68.4% white,
13.9% Asian, 13.6% Hispanic, and 9.1% African American residents. The county occupies 323 square miles
with a combination of urban cities and suburban communities. Our EMS system covers 85% of the county.
In the county, there are five acute care hospitals and
three advanced life support (ALS)-providing agencies.
All trauma patients in the region who meet state criteria are brought to our level I trauma center (Fig. 1).
The EMS system is two tiered, composed of a combination of paid and volunteer basic life support (BLS)
units and paid hospital-based ALS units that contain
two paramedics each. There are eight BLS units and
six ALS units that respond to approximately 30,000 dispatches per year, 6,500 of these being patients treated
by ALS in the system.
Our state paramedic protocols for trauma patients
recommend placement of two IV lines prehospitally.
However, paramedics insert the second IV line with
variability based on time to the hospital, subjective determination of the patient’s degree of illness, and difficulty obtaining the IV placement. In critical patients
when an IV line cannot be started, the paramedics are
allowed to use an intraosseous needle. The decision to
place the second IV line or to use the intraosseous needle is made by the paramedics at the scene.
An investigator trained in Microsoft Access (Microsoft Corp., Redmond, WA) and the Emergency Department Information Management (EDIM) database
collected all data. Patient care reports (PCRs) generated on a laptop computer during the course of patient interactions were transferred to a Microsoft Access database after completion.
Statistical analysis was carried out utilizing SAS 9.1
TS level 1M0, XP PRO platform (SAS Institute Inc.,
Cary, NC) and MINITAB 15 (MINITAB Inc., State College, PA). Summaries of all covariates and outcomes
are presented in Table 1, means and standard deviations for continuous variables and numbers and frequencies for categorical variables. p-Values are given
for differences between groups, based on observed
data. Table 2 contains the estimated unadjusted and
adjusted differences in means between listed outcomes
for patients with one versus two IV lines as well as the
regression model for volume of infused fluid. Sensitivity analyses were conducted when evaluating outcomes measured as changes from on-scene time to arrival at the trauma center, such that only subjects who
had low GCS scores (less than 15), low systolic or diastolic blood pressure (less than the median), or high
heart rate (greater than the median) were included in
each set of appropriate analyses. Only subjects who
had these low GCS scores, low systolic and diastolic
blood pressure, or high heart rate were included, resulting in parametric data. The cutoff values of blood
pressure and heart rate were made based on current
literature and the authors’ clinical experience.1,12 Systolic and diastolic blood pressures as well as heart
rates were normally distributed both on scene and on
arrival. In addition, changes in these measures were
normally distributed. The distribution of GCS scores
was highly right skewed, whereas the distribution of
pulse oximetry values was highly left skewed at both
time points; however, changes were unimodal and
symmetrical. The distributions of these changes were
light tailed; methods used for analysis were robust
and not expected to be influenced by these light tails,
4
PREHOSPITAL EMERGENCY CARE
APRIL/JUNE 2011
VOLUME 15 / NUMBER 2
TABLE 2. Estimated Difference in Adjusted and Unadjusted Means (Confidence Intervals and p-Values for Testing No
Difference) between Subjects with One and Two Intravenous Lines
Variable
Pulse oximetry, %
Volume of fluids, mL
Change in GCS score
All subjects
GCS score <15 on scene
Change in heart rate, bpm
All subjects
HR >100 bpm on scene
Change in systolic BP, mmHg
All subjects
Systolic BP <132 mmHg on scene
Systolic BP <100 mmHg on scene
Change in diastolic BP, mmHg
All subjects
Diastolic BP <80 mmHg on scene
Diastolic BP <60 mmHg on scene
Unadjusted
Adjusted for Age, Gender, and
Whether below Median on Scene
(When Appropriate)
–0.05 (–0.64, 0.54) 0.87
348.4 (235.6, 461.1) < 0.0001
–0.10 (–0.69, 0.49) 0.74
357.2 (224.1, 470.4) < 0.0001
Adjusted for Age, Gender, Whether
below Median on Scene (When
Appropriate), and Volume of Fluids
0.20 (–0.45, 0.84) 0.55
–
0.29 (–0.17, 0.74) 0.22
0.72 (–0.60, 2.04) 0.29
0.28 (–0.19, 0.74) 0.24
0.73 (–0.64, 2.11) 0.30
0.34 (–0.18, 0.86) 0.21
0.91 (–0.68, 2.51) 0.26
–0.21 (–2.89, 2.46) 0.88
2.65 (–3.55, 8.86) 0.64
0.40 (–2.13, 2.92) 0.76
2.88 (–3.27, 9.02) 0.36
0.87 (–1.94, 3.67) 0.54
3.92 (–3.01, 10.86) 0.27
3.42 (–2.10, 8.94) 0.22
4.26 (–3.24, 11.75) 0.27
4.30 (–13.71, 22.31) 0.64
4.34 (–1.09, 9.77) 0.12
5.41 (–2.08, 12.89) 0.16
2.50 (–18.59, 23.58) 0.82
2.93 (–2.83, 8.69) 0.32
2.55 (–5.86, 10.97) 0.55
22.00 (–4.17, 48.16) 0.10
5.32 (1.02, 9.62) 0.015
8.18 (2.16, 14.19) 0.0077
13.47 (–7.43, 34.37) 0.21
5.27 (1.10, 9.45) 0.013
8.58 (2.79, 14.37) 0.0037
6.86 (–20.41, 34.13) 0.62
5.41 (0.76, 10.07) 0.023
9.89 (3.38, 16.40) 0.0029
31.63 (2.20, 61.05) 0.035
BP = blood pressure; GCS = Glasgow Coma Scale; HR = heart rate.
particularly with the moderately large sample sizes
used.
Correlations between outcome measures were also
calculated. Based solely on observed data, differences
between patients with one versus two IV lines were
assessed using either Pearson chi-square tests or twogroup t-tests allowing for unequal variances, with a
Satterthwaite method for estimating the appropriate
degrees of freedom.13,14
Regression models estimated the unadjusted effects
of the number of IV lines as well as the effects adjusted
for age and gender; when appropriate (for GCS score,
heart rate, systolic blood pressure, and diastolic blood
pressure), whether the patient was above or below the
median and the previously listed covariates; and the
volume of fluids infused. Estimates, confidence intervals (CIs), and p-values are given for these unadjusted
and adjusted effects of number of IV lines on changes
in GCS score, heart rate, systolic blood pressure, and
diastolic blood pressure from on-scene to arrival time
points as well as pulse oximetry and volume of fluids. In addition, regression models estimated the effects of volume of infused fluids, unadjusted and adjusted for age and gender as well as, when appropriate
(for GCS, heart rate, systolic blood pressure, and diastolic blood pressure), whether the person was above or
below the median on a particular outcome on scene.
In these regression analyses, multiple imputation was
used to account for the small amount of missing
data.15,16
The above analyses were repeated for individuals.
The ISS means were compared between the groups
using differences of means and rates of rehospitalization and 30-day mortality using differences of
proportions.
RESULTS
From these analyses, we determined that the effect of
number of IV lines was even larger (8.18 mmHg; 95%
CI: 2.16, 14.19; p = 0.0077) when considering only patients who had low diastolic blood pressures when
their on-scene care began. Also, patients with elevated
heart rates (greater than 100 bpm) did not benefit from
the additional IV fluids (3.92 bpm; 95% CI; –3.01, 10.86;
p = 0.27), nor did the patients with low systolic blood
pressures (<100 mmHg).
Volumes of fluids differed significantly between the
numbers of IV lines. Those with two IV lines received
on average of 348.4 mL more fluids (95% CI: 235.6,
461.1; p < 0.0001). Change in diastolic blood pressure
showed a significant difference between patients with
one versus two IV lines, an average of 5.32-mmHg
greater increase (95% CI: 1.02, 9.62; p = 0.015) than
that for patients with only one IV line. The difference
persisted after adjusting for covariates, including the
volume of fluids infused. This effect persisted with
diastolic blood pressures below 60 mmHg. No other
outcomes significantly differed between patients with
one versus two IV lines.
No differences were found in ISS between the two
groups (15.609 in those with one IV line versus 18.806
in those with two IV lines; p = 0.1255). No differences
were found in rehospitalization rates (9.7% in those
with one IV line versus 9.2% in those with two IV lines
[46 of 76 patients contacted had one IV line versus 30 of
76 patients contacted had two IV lines]) or 30-day patient mortality (8.1% in those with one IV line versus
8.2% in those with two IV lines [same patient information obtained]). Follow-up was based on only 23% of
patients who were able to be contacted.
Merlin et al.
5
TWO IVS IN TRAUMA
DISCUSSION
Standards of care in the prehospital setting need to
be constantly reevaluated. We sought to evaluate the
utility of placing a second IV line in trauma patients
on the endpoints of blood pressure, heart rate, GCS
score, rehospitalization rate, and 30-day mortality. As
in all protocols, risks versus benefit must be weighed
to determine the utility of maintaining a practice. Since
most trauma patients are treated with a sense of urgency, the potential risk of needlestick is increased. In
addition, IV lines increase on-scene times while worsening outcomes.
Significant literature exists questioning the standard
practice of aggressive fluid management in the prehospital and hospital settings.17–21 This is the first study in
the literature to evaluate the placement of one versus
two IV lines in this arena. We demonstrated changes
in both the volume of fluid infused and the diastolic
blood pressure for those patients who received a second IV line. While a statistically significant difference
in diastolic blood pressure was found, the cause of this
difference is difficult to determine. It is possible that
another unaccounted-for variable caused the diastolic
blood pressure to rise, independent of the underlying
condition. It is also possible that patients’ blood pressures would have been lower if they had not received
any IV fluids.
The difference in heart rate, as a potentially sensitive marker of low volume states, was consistently
not significant between the patients who received one
versus two IV lines. Most importantly, for tachycardic
patients, the extra fluid did not yield improvement
in vital signs. This finding persisted in hypotensive
patients. Since this is the most sensitive endpoint of
volume status in our study, it provides the best evidence that additional fluid difference is not altering vital signs on heart rate.
LIMITATIONS AND FUTURE RESEARCH
There are limitations to our study. This was a singleinstitution study at one of three level I trauma centers in our state and, therefore, extrapolations can be
made only to similar settings. The data entered by the
prehospital care providers may have had some slight
inaccuracies. The information obtained from the Access databases was entered during emergency situations when time was limited and, therefore, could
potentially contain reporting errors. Some discrepancies were identified when matching patients between
databases, including errors in spelling and dates. The
data were thoroughly reviewed to identify these errors, but the possibility exists that not all errors were
detected. Additionally, this study was conducted retrospectively. Therefore, no cause-and-effect relationships could be established. Although associations were
found between patient treatment groups, there exists
the possibility that these may be due to confounding
variables.
Our endpoints represented data points that were
easily accessible and standard; however, vital signs
vary based on medical conditions, medications such
as atrioventricular (AV) nodal blockers, diseases of the
vascular system, and fitness levels. Blood pressure in
particular is not known to be a very sensitive marker of
volume status until later stages of shock. Additionally,
blood pressures vary between health care providers
taking the measurements, calibration of machines, and
general margins of error of the sphygmomanometer.
A significant number of patients were lost to
follow-up. Approximately 50% did not provide correct phone numbers or provided numbers that were
disconnected. Another 23% did not return our phone
calls after multiple messages were left. It is possible
that mortality and rehospitalization rate would yield
different results based on improved follow-up. Also, it
is possible that patients who responded to follow-up
were less severely injured than the rest of the population. We attempted to review all records in our
system to verify that the patients did not return to our
regional trauma center.
Future research should evaluate how the placement
of no IV line compares with the placement of one and
two IV lines. It was extremely rare for patient in our
study population to have no prehospital peripheral access, so it was not possible for us to entertain this hypothesis.
CONCLUSION
Redundant prehospital IV lines provided no difference
in physiologic outcomes for trauma patients. When
controlling for confounding variables, no significant
outcome difference of heart rate or volume infused
was noted, even in the hypotensive patients. The traditional approach for establishment of a secondary IV
line in prehospital trauma patients should not be followed in a dogmatic fashion.
References
1. American College of Surgeons. Prehospital Trauma Life Support Manual, 6th edition. St. Louis, MO: Mosby, 2006.
2. Cannon WB, Fraser F, Cowell EM. The preventative treatment
of wound shock. JAMA. 1918;47:618–21.
3. Capone AC, Safar P, Stezoski W, Tisherman S, Peitzman AB.
Improved outcome with fluid restriction in treatment of uncontrolled hemorrhagic shock. J Am Coll Surg. 1995;180:49–56.
4. Marshall HP Jr, Capone A, Courcoulas AP, et al. Effects of
hemodilution on long-term survival in an uncontrolled hemorrhagic shock model in rats. J Trauma. 1997;43:673–9.
5. Burris D, Rhee P, Kaufmann C, et al. Controlled resuscitation
for uncontrolled hemorrhagic shock. J Trauma. 1999;46:216–23.
6
6. Bickell WH, Wall MJ Jr, Pepe PE, et al. Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries. N Engl J Med. 1994;331:1105–9.
7. Capone A, Safar P, Stezoski SW, Peitzman A, Tisherman S. Uncontrolled hemorrhagic shock outcome model in rats. Resuscitation. 1995;29:143–52.
8. Seamon MJ, Fisher CA, Gaughan J, et al. Prehospital procedures
before emergency department thoracotomy: “scoop and run”
saves lives. J Trauma. 2007;63:113–20.
9. Birk HO, Henriksen LO. Prehospital interventions: on-scene
time and ambulance-technicians’ experience. Prehosp Disaster
Med. 2002;17:167–9.
10. O’Gorman M, Trabulsy P, Pilcher DB. Zero-time prehospital IV.
J Trauma. 1989;29:84–6.
11. Bhananker SM, Liau DW, Kooner PK, et al. Liability related to
peripheral venous and arterial catheterization: a closed claims
analysis. Anesth Analg. 2009;109:124–9.
12. Edwards M, Ley E, Mirocha J, Hadjibashi AA, Margulies DR, Salim A. Defining hypotension in moderate to
severely injured trauma patients. Am Surgeon. 2010;76:1035–
8.
13. Satterthwaite FE. An approximate distribution of estimates of
variance components. Biometric Bull. 1946;2:110–4.
PREHOSPITAL EMERGENCY CARE
APRIL/JUNE 2011
VOLUME 15 / NUMBER 2
14. Neter J, Wasserman W, Kutner MH. Applied Linear Statistical
Models. Burr Ridge, IL: Richard D. Irwin, Inc., 1990.
15. Little RJA, Rubin DB. Statistical Analysis with Missing Data.
Hoboken, NJ: J. Wiley & Sons, 1987.
16. Rubin DB. Inference and missing data. Biometrika.
1976;63:581–92.
17. Owens TM, Watson WC, Prough DS, Uchida T, Kramer GC.
Limiting initial resuscitation of uncontrolled hemorrhage reduces internal bleeding and subsequent volume requirements.
J Trauma. 1995;39:200–7.
18. Hein LG, Albrecht M, Dworschak M, Frey L, Bruckner UB.
Long-term observation following traumatic-hemorrhagic shock
in the dog: a comparison of crystalloidal vs. colloidal fluids.
Circ Shock. 1988;26:353–64.
19. Krausz MM, Hirsh M. Bolus versus continuous fluid resuscitation and splenectomy for treatment of uncontrolled
hemorrhagic shock after massive splenic injury. J Trauma.
2003;55:62–8.
20. Kreimeier U, Prueckner S, Peter K. Permissive hypotension.
Schweiz Med Wochenschr. 2000;130:1516–24.
21. Solomonov E, Hirsh M, Yahiya A, Krausz MM. The effect of
vigorous fluid resuscitation in uncontrolled hemorrhagic shock
after massive splenic injury. Crit Care Med. 2000;28:749–54.