Clinical Investigations
Central or Peripheral Catheters for Initial
Venous Access of ICU Patients: A Randomized
Controlled Trial
Jean-Damien Ricard, MD, PhD1,2; Laurence Salomon, MD, PhD3; Alexandre Boyer, MD4;
Guillaume Thiery, MD5; Agnes Meybeck, MD1; Carine Roy, MSc6; Blandine Pasquet, MSc6;
Eric Le Mière, MD1; Didier Dreyfuss, MD1,2
Objectives: The vast majority of ICU patients require some form
of venous access. There are no evidenced-based guidelines con-
cerning the use of either central or peripheral venous catheters,
despite very different complications. It remains unknown which
1
AP–HP, Hôpital Louis Mourier, Service de Réanimation Médico-chirurgicale,
Colombes, France.
2
Univ Paris Diderot, Sorbonne Paris Cité, UMRS-722, F-75018, Paris, France.
3
Département de Santé Publique, Assistance Publique—Hôpitaux de
Paris, Hôpital Louis Mourier, Colombes, France.
4
Service de Réanimation Médicale, Hôpital Pellegrin—Tripode, Bordeaux,
France.
5
Assistance Publique—Hôpitaux de Paris, Hôpital Saint-Louis, Service de
Réanimation Médicale, Paris, France.
6
Département de Biostatistique et Recherche clinique, Assistance Pub-
lique—Hôpitaux de Paris, Hôpital Bichat, Paris, France.
Registered at ClinicalTrials.gov: NCT00122707 (http://clinicaltrials.gov/
ct2/show/NCT00122707?term=ricard&rank=4).
Drs. Ricard, Salomon, and Dreyfuss designed the study. Drs. Boyer, Thiery,
Meybeck, and Le Miere collected data. Drs. Salomon, Roy, and Pasquet
analyzed the data. Drs. Ricard, Salomon, and Dreyfuss interpreted the
data. Drs. Ricard, Salomon, and Dreyfuss drafted the article. All authors
made substantial scientific input to the article. Drs. Ricard, Salomon, and
Dreyfuss finalized the article that was approved by all authors. Dr. Ricard
had full access to all the data in the study and takes responsibility for the
integrity of the data and the accuracy of the data analysis.
Supplemental digital content is available for this article. Direct URL cita-
tions appear in the printed text and are provided in the HTML and PDF
versions of this article on the journal’s website (http://journals.lww.com/
ccmjournal).
Supported by a grant devoted to research (Programme Hospitalier de
Recherche Clinique) from the French Ministry of Health (AOM02017). The
sponsor had no involvement in the study design, data collection, analysis
and interpretation, writing of the report, and in the decision to submit the
article for publication.
Dr. Ricard is a board member at Covidien, received payment for lectures
for Covidien and travel reimbursements from Fisher & Paykel. Dr. Dreyfuss
has received grant support from Pfizer. The remaining authors have dis-
closed that they do not have any potential conflicts of interest.
For information regarding this article, E-mail: jean-damien.ricard@lmr.aphp.fr
Copyright © 2013 by the Society of Critical Care Medicine and Lippincott
Williams & Wilkins
DOI: 10.1097/CCM.0b013e31828a42c5
2108 www.ccmjournal.org September 2013 • Volume 41 • Number 9
to insert in ICU patients. We investigated the rate of catheter-
related insertion or maintenance complications in two strategies:
one favoring the central venous catheters and the other peripheral
venous catheters.
Design: Multicenter, controlled, parallel-group, open-label ran-
domized trial.
Setting: Three French ICUs.
Patients: Adult ICU patients with equal central or peripheral
venous access requirement.
Intervention: Patients were randomized to receive central
venous catheters or peripheral venous catheters as initial venous
access.
Measurements and Results: The primary endpoint was the rate of
major catheter-related complications within 28 days. Secondary
endpoints were the rate of minor catheter-related complications
and a composite score-assessing staff utilization and time
spent to manage catheter insertions. Analysis was intention to
treat. We randomly assigned 135 patients to receive a central
venous catheter and 128 patients to receive a peripheral
venous catheter. Major catheter-related complications were
greater in the peripheral venous catheter than in the central
venous catheter group (133 vs 87, respectively, p = 0.02)
although none of those was life threatening. Minor catheter-
related complications were 201 with central venous catheters
and 248 with peripheral venous catheters (p = 0.06). 46%
(60/128) patients were managed throughout their ICU stay with
peripheral venous catheters only. There were significantly more
peripheral venous catheter-related complications per patient
in patients managed solely with peripheral venous catheter
than in patients that received peripheral venous catheter and
at least one central venous catheter: 1.92 (121/63) versus
1.13 (226/200), p < 0.005. There was no difference in central
venous catheter-related complications per patient between
patients initially randomized to peripheral venous catheters
but subsequently crossed-over to central venous catheter and
patients randomized to the central venous catheter group.
Kaplan–Meier estimates of survival probability did not differ
between the two groups.
 Clinical Investigations

Conclusion: In ICU patients with equal central or peripheral venous
access requirement, central venous catheters should preferably
be inserted: a strategy associated with less major complications.
(Crit Care Med 2013; 41:2108–2115)
Key Words: catheters; complications; nosocomial infection; quality
of care; venous access
MATERIAL AND METHODS
Objectives and Study Design
The main objective of this study was to compare the occurrence
of major complications related to venous access in a number
of predefined situations (in which there might be a clinical
preference for CVC insertion but no absolute indications for
it), with the following strategies: 1) systematic insertion of a
central venous access or 2) initial use of a peripheral venous
access (until predefined central venous access requirement cri-

C
ritical care medicine is one of the fields where evi- teria are met based on patient’s condition or treatment). The
dence-based guidelines have taken a major importance other objectives were to compare for both venous accesses, the
(1). This is the result of the completion of many ran- total occurrence of complications (major and minor), mortal-
domized controlled trials (RCT), which addressed diverse con- ity, and the time spent by medical or paramedical teams on
ditions such as the prevention of gastroduodenal bleeding (2) venous-access insertion and care.
or ventilator-associated pneumonia (3, 4) and the treatment This prospective, multicenter, randomized controlled
of septic shock (5) or acute respiratory distress syndrome (6– study was carried out in three French university hospital ICUs
9). Surprisingly, whereas most if not all ICU patients have— between March 2004 and January 2006.
at some time during their ICU stay—some form of venous
access, there has never been an adequate evaluation, including
Patients
RCT, of the respective advantages and complications of both
Consecutive adult patients hospitalized in ICU (whose dis-
forms of venous access. Furthermore, little is known on the
charge was not planned in the next 48 hours), who did not
distribution between central and peripheral venous catheters
already have a CVC but required a venous access that could
(CVC and PVC) and the reasons that govern the choice for one
either be a CVC or a PVC were eligible for enrolment. The
or the other. In 1995, the EPIC study showed that of the 10,038
CVC insertion site (jugular, subclavian, femoral) was left
patients surveyed, 78.3% had some form of IV catheteriza-
at the clinician in charge’s discretion. PVCs were true short
tion, but a wide range of types and uses of catheterization were
peripheral catheters (neither peripherally inserted central
included (10). Although many agree on some undisputable
catheters nor midline catheters). Inclusion criteria (complete
indications for central venous catheterization (11)—which
list in supplemental materials [SM] 1A, Supplemental Digi-
would render the conduct of an RCT disputable—clear con-
tal Content 1, http://links.lww.com/CCM/A648) fell under
sensus on the respective indications of CVC and PVC is lack-
two categories: 1) need for specific drugs known to be veino-
ing. This translates into considerable disparity in the frequency
toxic (epinephrine: dose less than or equal to 2 mg/hr; nor-
of CVC uses among ICUs, from 43% to 80% in one study (12)
epinephrine: dose less than or equal to 2 mg/hr; dopamine or
and indications for CVC (with over 20% considered nonjus-
dobutamine: dose not exceeding 10 mg/kg/min; amiodarone:
tified) (13). It has been generally accepted (although never
less than three ampoules [150 mg in 3 mL] per day, for an
proven) that the majority of ICU patients initially require a
expected period shorter than 3 d; vancomycin: discontinuous
central venous line. This attitude may not be devoid of risks.
infusion of a dose <1g/d; amphotericin B: for an expected
Recognized complications of CVC insertion or maintenance
period less than 3 d); 2) difficulties in peripheral venous cath-
may occur in more than 15% of patients with a CVC (14),
eter insertion or maintenance: two failed attempts to insert a
some of which may lead to death (11). Comparatively, far less
PVC; need to replace a PVC twice a day, 2 days in a row; need
is known about PVCs in the ICU. Fewer if any major complica-
to replace a PVC once a day, 3 days in a row. Noninclusion
tions seem to be reported with PVC insertion. Most complica-
criteria (detailed in SM 1B, Supplemental Digital Content 1,
tions relate to PVC maintenance of which phlebitis is the most
http://links.lww.com/CCM/A648) were as follows: age under
often reported with prevalence varying among studies between
18, pregnant or breastfeeding women, refusal to participate in
20% and 53% (15, 16). Contrary to CVC, PVC-related blood-
the study, contraindication to PVC, contraindication to CVC,
stream infections are rare, 40 times less frequent than with CVC
and need for any of the drugs listed above outside the bound-
(0.1% vs 4.4%) (17), and deep-vein thrombosis exceptionally
aries of doses or duration.
reported. Because of the differences in rates and severity of
complications associated with these venous accesses, compari-
son between both devices is difficult. If some situations clearly Randomization and Masking
require the use of one or the other device, many other settings We randomly allocated eligible participants in a one-to-one
may be taken care of with either one. Reasons that govern the ratio to receive a CVC or a PVC using a computer-generated
choice of one rather than the other in these instances may vary allocation sequence that included stratification by center.
from one physician to another. We therefore designed the first A statistician (L.S.) provided sealed envelopes containing
prospective, randomized, multicenter controlled study to com- allocated group (CVC or PVC) to each center. All envelopes
pare a strategy that favored PVC to one favoring CVC for the were collected to ensure respect of order and allocation. For
management of ICU patients. obvious reasons, the study was not blinded.

Critical Care Medicine www.ccmjournal.org 2109

Ricard et al
Strategy
All insertion, maintenance, and removal catheter procedures
were reviewed during design of the study so as to harmonize
patient preparation before catheter insertion and catheter care
in all participating ICUs. Patients were monitored until ICU
discharge or for 28 days.
Skin preparation was performed according to Centers for
Disease Control and Prevention recommendations with povi-
done-iodine (18). Central catheters were inserted using maxi-
mal sterile-barrier precautions including use of large sterile
drapes, surgical antiseptic hand wash, and use of sterile gown,
gloves, mask, and cap. CVC used were standard polyurethane
7F, 16 (6”) or 20 cm (8”), multilumen (2 or 3), noncoated,
nonimpregnated catheters.
CVCs were removed whenever they were no longer required
as recommended (18–20) and could be replaced with PVC if
a venous access was still necessary. PVCs used were 18 or 20
gauge, polyurethane catheters. PVCs were changed at least
every 72 hours, according to the Centers for Disease Control
and Prevention recommendations for the prevention of cath-
eter-related infections (18, 20). When their medical condi-
tion required it, or whenever PVC access was compromised,
patients in the PVC group could have a CVC inserted. In order
to standardize these crossovers, these situations were listed a
priori, and the doses of drugs requiring central access were also
defined (see below). Therefore, whatever the group of random-
ization and the initial strategy of catheter insertion, patients
could eventually have both types of venous access.
Crossover criteria
A priori defined crossover criteria were 1) increase in veinotoxic
drug infusion rate (doses and drugs, detailed in SM 1C, Supple-
mental Digital Content 1, http://links.lww.com/CCM/A648) or
2) impossibility or great difficulties in inserting or maintaining
a PVC (detailed in Supplemental Material SM 1C, Supplemen-
tal Digital Content 1, http://links.lww.com/CCM/A648).
Data Collection
The case report form included patient’s baseline characteristics
recorded at the date of patient inclusion. During patient’s fol-
low-up (i.e., during patient’s venous catheterization censored
to 28 d), we recorded all catheter complications (mechanical,
infectious, thrombotic, detailed in Supplemental Material
SM 2, Supplemental Digital Content 1, http://links.lww.com/
CCM/A648). Mechanical complications included insertion-
related CVC complications (necessity to change site insertion,
insertion success only after two changes of site insertion [two
failed attempts], failure to insert CVC after at least trials at two
different sites, arterial puncture, vessel injury [requiring surgi-
cal repair], pneumothorax, hemothorax, local hematoma of at
least 3 cm², mediastinal hematoma, gas embolism, embolism
of the wire); insertion PVC-related complications (impossi-
bility to place a PVC, more than five attempts to insert PVC).
Maintenance-related complications were also recorded during
CVC maintenance (gas embolism, hemorrhage at the inser-
tion site) and also during PVC maintenance (maintenance of
2110 www.ccmjournal.org September 2013 • Volume 41 • Number 9
catheter or catheter insertion impossible, subcutaneous dif-
fusion ≥ 5 × 5 cm or necrosis or blister ≥ 3 × 3 cm). Infectious
complications included the following: localized erythema
(>2 cm from insertion site); unexplained bacteremia (defined
as positive blood culture in the absence of obvious infec-
tious focus); catheter-related bacteremia (defined as presence
of a positive catheter tip culture (>103 colony-forming unit
[CFU]/mL), associated with secondary bacteremia due to the
same bacteria as retrieved from the catheter in the absence
of another focus of infection with the same bacteria); cath-
eter infection (defined as presence of a positive culture of the
tip of the catheter (>103 CFU/mL) in the presence of general
or local signs of infection, with at least partial regression of
symptoms on removal of the catheter). Thrombotic compli-
cations included complete vein (jugular, femoral, or subcla-
vian) thrombosis on ultrasound examination or deep-vein
thrombosis for a PVC. At the end of follow-up, the number of
catheters inserted, both peripheral and central, were recorded.
In the end, follow-up data were completed with patient vital
status and date at discharge from the ICU or at day 28. Trained
nurses and physicians implicated in the study were screened
for complications that were monitored in the study case report
form, as well as in the patient’s charts and medical records.
This monitoring was double-checked by each center’s princi-
pal investigator and site-dedicated clinical research assistant.
Outcomes
The primary outcome measure was the prevalence of major
catheter-related adverse events, measured as the number of
complications observed per patient. The proportion of patients
with at least one complication was also reported. Because
there is no validated classification of adverse events in ICU,
we performed two analyses to rate the complications related
to venous access. One simply compared for each strategy the
numbers of complications defined in the protocol (see above).
The major caveat of this analysis is that it does not take into
account the outcome of the complication. We therefore also
used a previously validated classification, the Common
Terminology Classification for Adverse Events (CTCAE-V3)
(21), to compare complications in the two strategies (SM 3,
Supplemental Digital Content 1, http://links.lww.com/CCM/
A648). We sought consensus for the application of the CTCAE
to the intensive care setting through a Delphi survey among
senior ICU physicians. Secondary outcomes were the prevalence
of minor complications, amount of medical and paramedical
time used, and mortality. For the amount of medical and
paramedical time used, a grading score was constructed using
the same philosophy as the CTCAE and validated through the
Delphi survey (SM 4, Supplemental Digital Content 1, http://
links.lww.com/CCM/A648, which details the score and the
grade).
Statistical Analysis
A computer program package (SAS 9.1, Cary, NC) was used for
statistical testing and management of the database. Continu-
ous variables are presented as mean ± sd and discontinuous
 Clinical Investigations

variable as median and 25–75 percentiles. Comparison of the Institutional Review Board (Comité de Protection des Per-
baseline characteristics among the patients in the two groups sonnes dans la Recherche Biomédicale) of Saint Germain en
was made with the Wilcoxon-Mann-Whitney test for quantita- Laye (project 03039) and by the ethics committee of the French
tive features and with Fisher exact test for qualitative variables. Society of Intensive Care (details in SM 5, Supplemental Digi-
An intent-to-treat analysis was performed, and the unit of tal Content 1, http://links.lww.com/CCM/A648).
analysis was the patient.
The number of complications per patient was compared RESULTS
between both groups using nonparametric Mann-Whitney Over the 2-year study period, 1,751 patients were screened for
tests. The comparisons of the proportions of complications potential inclusion, 1,485 had at least one noninclusion crite-
between the two groups were performed through chi-square ria, and 266 were randomized (Fig. 1). Three patients refused to
tests or Fisher exact tests when appropriate. All p values were two
give their pursuit consent after recovering consciousness, and
tailed, and the 0.05 level was considered statistically significant.
thus, 263 were included in the final analysis, 135 in the CVC
Survival data were analyzed using the Kaplan–Meier group and 128 in the PVC group. All the patients were analyzed
method with log-rank test for comparison of survival curves according to their initial allocation group (intention-to-treat).
and a hazard ratio estimate (95% CI) was calculated. The most frequent reasons for inclusion in the study were
need for vasopressor infusion (70%) and two failed attempts
Ethical Considerations of PVC insertion (13%). Median delay between ICU admission
The study protocol, which conformed to the ethical guide- and inclusion was 0 day (75 percentile: 2 d).
lines of the Declaration of Helsinki, was approved by the Baseline characteristics were similar in the two groups
(Table 1). Patients were severe
ICU patients with an average
Assessed for eligibility (n=1751) SAPS2 score of 56, providing
a 59.8% predicted hospital
mortality (22). The majority
was under invasive mechani-
Not included cal ventilation (86%). Ninety-
(no inclusion criteria, n=1485) four percent of the patients
Enrollment already had a PVC at the time
of inclusion.

Randomized (n=266) Crossover

According to protocol, 68
patients from the PVC group
had at least one crossover cri-
teria and 67 actually received
a CVC (one patient was trans-
ferred for urgent surgery
Allocation Allocated to PVC (n=129) Allocated to CVC (n=137)
Received PVC (n=110) Received CVC (n=135) before CVC insertion). Rea-
Did not receive PVC (n=19) Did not receive CVC (n=2) sons for crossover were mainly
(failure to insert PVC) increase in vasopressor dose
and impossibility to insert or
maintain a PVC (Table 2).

Number of Catheters
Received
Follow-up No lost to follow-up No lost to follow-up The mean number of CVC per
patient was 1.2 ± 0.55 in the
Withdrew consent = 1 Withdrew consent = 2
CVC group and 0.66  ± 0.67
in the PVC group. The mean
number of PVC per patient in
Intention-to-treat Analysed (n=128) Analysed (n=135) the CVC group was 1.48 ± 2.05
analysis (including 67 patients crossed and 2.86  ± 2.45 in the PVC
over to receive of CVC) group. Number of study days
in both groups was identi-
Figure 1. Patient flowchart indicating numbers of patients screened for eligibility, enrolled, allocated to each
group, lost to follow-up, and finally analyzed in the study. PVC = peripheral venous catheter, CVC = central cal: 12.17 ± 9.27 for CVC and
venous catheter. 12.06 ± 8.96 for PVC.

Critical Care Medicine www.ccmjournal.org 2111

Ricard et al

Table 1.Principal Baseline Characteristics received a CVC (n = 63; 60 patients randomized to the PVC,
of Patients Randomized to Receive Either who were not crossed over, 1 patient crossed over but did not
have time to receive a CVC, and two patients randomized to the
a Central or a Peripheral Catheter
CVC, who did not receive a CVC) than in patients who received
Centrala Peripherala at least one CVC: 1.92 (121/63) vs. 1.13 (226/200), p < 0.005).
(n = 135) (n = 128) There was no difference in complications per patient
Age: mean ± sd 63.42 ± 15.39 64.78 ± 15.97 related to CVC insertion or maintenance among patients ini-
tially randomized in the PVC group but subsequently crossed
Sex ratio (male/female) 1.53 1.98 over to CVC and patients randomized to the CVC group: 0.91
SAPS 2 55.9 ± 21.4 56.2 ± 21.4 (61/67) versus 1.07 (142/133), respectively, p = 0.3. There was
ODIN on admission 2.18 ± 1.10 2.21 ± 1.07
no statistical difference in mean time to first complication
among patients randomized to the CVC group and patients
Mechanical ventilation, n (%) 109 (83.2) 109 (89.3) randomized to the PVC but that crossed over to receive a CVC
ODIN on inclusion 2.42 ± 1.12 2.59 ± 1.00 (SM 6, Supplemental Digital Content 1, http://links.lww.com/
CCM/A648).
Medical/surgical patients (%) 95/5 91/9
SPAS2 = Simplified Acute Physiology Score 2 (19); ODIN = organ
dysfunction and/or infection score (30). PVC- and CVC-Attributable Complications
a
Initial group assignment (central or peripheral venous catheter). The numbers of major attributable mechanical and infectious
complications were greater in the PVC than that in the CVC
Primary Outcome group (SM 7, Supplemental Digital Content 1, http://links.
Major complications were more frequent in patients randomized lww.com/CCM/A648).
to the PVC group than to the CVC group (1.04 [133/128] vs
0.64 [87/135] complication per patient, respectively, p < 0.02; Complication Rates According to the Common
Table 3), although none of those was life threatening. There was Terminology Classification for Adverse Events
a trend toward a greater proportion of patients with at least one The number of complications per patient was higher in
complication in the PVC compared with the CVC group (47.7% patients allocated to PVC than in those receiving a CVC (1.54
[61/128] vs 36.3% [49/135], p = 0.06). Three pneumothoraces [198/128] vs 0.89 [120/135], respectively; p < 0.0001). Distri-
were observed in each group (in the PVC group, these occurred bution by grade is detailed in Table 5. The difference was due
in patients crossed over to CVC). Infectious and thrombotic to grade 1 or 2 complications (SM 8, Supplemental Digital
complications were equally distributed. Content 1, http://links.lww.com/CCM/A648 for further details
on the grading) (21).
Secondary Outcome
Minor complications were not different in patients ran- Medical and Paramedical Resource Required for
domized to the PVC than in those to the CVC group (1.94 Catheter Insertion
[248/128] vs 1.49 [201/135] complication per patient, respec- Medical and paramedical staff involvement was greater in PVC
tively, p = 0.08). There was no difference in the proportion of group compared with CVC group (SM 9, Supplemental Digital
patients with at least one minor complication (67.2% [86/128] Content 1, http://links.lww.com/CCM/A648).
vs 61.5% [83/135], p = 0.33) (Table 4).
Mortality
Early Versus Late Insertion of Catheters The Kaplan–Meier estimates of day-28 survival probability
There were significantly more complications per patient related did not differ between the two groups (HR, 1.30 [95% CI,
to PVC insertion and maintenance in patients who never 0.84–2·01]; p = 0.23). Six patients (three in each group) were

Table 2. Reasons for and Number of Occurrences of Crossover of Patients from

Peripheral to Central Catheter Group
Reason for Crossover Number (n = 67) Delay of Occurrence (d)

Need for > 2 mg/hr epinephrine 9

0 (0–1)
Need for > 2 mg/hr norepinephrine 22
Impossibility to maintain peripheral access 9
1 (0–3)
More than 5 peripheral venous catheter insertion failures 23
Continuous vancomycin infusion or > 2 g/d 1
Unknown 3
Median (quartiles).

2112 www.ccmjournal.org September 2013 • Volume 41 • Number 9

 Clinical Investigations

Table 3. List and Number of Occurrences of Major Complications in Each Allocation Group
Centrala Peripherala p

Patients without any complication, no. 86 67 0.06

At least one complication, no. patients
 Mechanical 42 51 0.14
 Infectious 18 23 0.30
 Thrombotic 1 5 0.09
Major complications, no. of complications 87 133 0.02
 Mechanical 63 92 0.06
  Pneumothorax 3 3
  Arterial puncture 7 4
  Hematoma 1 1
   Central venous catheter insertion site changes 34 9
   Peripheral venous catheter insertion difficulties 16 56
  Subcutaneous diffusion 2 19
 Infectious 23 36 0.25
   Erythema (>2 cm from insertion site) 8 20
  Phlebitis 1 1
  Unexplained bacteremia 9 6
  Catheter-related bacteremia 1 0
  Catheter infection 4 9
 Thrombotic 1 5 0.09
Initial group assignment (central or peripheral venous catheter).
a

discharged alive very early after ICU admission, and were lost
to follow-up regarding this outcome.
DISCUSSION
Whereas several studies have compared the complications
associated with CVC according to the site of insertion (23–25),
there is a lack of data on the indications of such CVC inser-
tions in comparison with PVC. One may wonder whether this
is not the result of a strategy of “risk avoidance” as described
by Knottnerus and Dinant (26), who highlighted the fact that
investigators “might focus their energies on topics where the
methodological criteria of reviewers and editors can be most

Table 4. List and Number of Occurrences of Minor Complications in Each Allocation Group
Centrala Peripherala p

Patients without any complication, no. 52 42 0.33

At least one complication, no. patients
 Mechanical 54 58 0.38
 Infectious 49 68 0.006
 Thrombotic 2 5 0.27
Minor complications, no. of complications 201 248 0.06
 Mechanical 111 120 0.35
 Infectious 88 123 0.009
 Thrombotic 2 5 0.23
Initial group assignment (central or peripheral venous catheter).
a

Critical Care Medicine www.ccmjournal.org 2113

Ricard et al
Table 5. Complications According to the Common Terminology Classification for Adverse
Events Classification
Peripherala
All grades 1.54 (198/128)
Grades 1 and 2 1.41 (180/128)
Grades 3 and 4 0.13 (18/128)
Initial group assignment (central or peripheral venous catheter).
a
easily met, rather than studying real life clinical problems.”
Obviously, a study like ours is more difficult to conduct than a
comparison of different sites of CVC insertion.
This is the very first study to compare in a prospective ran-
domized fashion two venous access strategies in ICU patients.
Hundreds of millions of catheters are inserted every year in
ICU patients around the world. Although many studies have
exhaustively reported complications related to both CVC (14,
27, 28) and PVC, no randomized study has ever compared
peripheral and central venous accesses in ICU patients in terms
of a general venous-access management strategy. Gueuniaud et
al (29) reported a greater, albeit nonsignificant, proportion of
resuscitation in out-of-hospital cardiac arrests after peripher-
ally compared with centrally injected epinephrine, thus sug-
gesting that the peripheral route was at least as efficient as the
central one in this context. This better outcome might be the
result of faster peripheral than central venous access and thus
appropriate drug administration (vasopressors). Two observa-
tional studies have reported a greater prevalence of phlebitis
in ICU patients with short peripheral IV lines in comparison
with either peripherally inserted or centrally inserted central
catheters, but the observational design of the study precludes
any conclusion to be drawn from the data (16, 30).
Given this lack of objective data, we were therefore inter-
ested in answering the following question: in patients who
could be managed either with a PVC or a CVC, is it safer to use
one or the other, in terms of both insertion- and maintenance-
related complications?
Our first difficulty resided in defining the complications
and rating their degree of severity. A tension pneumothorax
is obviously a major mechanical complication associated
with CVC insertion but is there any counterpart for PVC?
Had we decided there was by definition no major mechanical
complication associated with PVC insertion, we would have
biased the study right from the start. We therefore classified a
priori a number of PVC complications as major ones.
The main criticism for this rating strategy lies in its subjec-
tivity because the potential harm rather than the actual conse-
quence of the complication is considered. We therefore decided
to complement this analysis by using another classification, the
CTCAEs (21), which is widely used in cancer clinical trials. In
our first classification, a pneumothorax was a major complica-
tion regardless of its size and clinical impact. The CTCAE-V3
enables to distinguish between a pneumothorax that requires
no intervention whatsoever (grade 1) and a life-threatening
tension pneumothorax requiring immediate drainage (grade
2114 www.ccmjournal.org September 2013 • Volume 41 • Number 9
Centrala p
0.89 (120/135) 0.0001
0.83 (112/135) 0.0001
0.06 (8/135) 0.13
4). Analysis of our results with this more objective classifica-
tion yielded similar findings.
The main result is that both strategies seem safe in a popu-
lation of critically ill patients who has no mandatory indica-
tion for CVC insertion. The severest complications were not
observed more frequently with either strategy. However, the fre-
quency of other complications, which were rated as major in this
study (although considered essentially of grade 1 or 2, i.e., need-
ing no major intervention and being not life threatening), was
significantly higher with PVCs than with CVCs. Complications
that led to this superiority of CVCs were mainly difficulties with
the insertion/maintenance of PVCs and erythema at their site
of insertion. Then, the major strength of this study is to provide
evidence-based data on which clinicians can rely to decide which
approach they will favor in a given patient. A raw interpretation
of our data would obviously favor the immediate insertion of
a central line in most if not all critically ill patients. Indeed, in
addition to the fact that fewer complications were observed in
the CVC group, it is worth noting that half of the patients ran-
domized to the PVC group ultimately required a CVC, and the
time spent to insert and ensure maintenance was shorter in the
CVC group. Repetitive insertions of PVC with the latter strategy
may decrease patients’ comfort, expose them to more complica-
tions than patients managed with a CVC, and finally delay inser-
tion of a CVC. On the opposite, one may conclude from our data
that insertion of a PVC in a patient in whom there is no techni-
cal difficulty and no need for repeated venous puncture and/or
change of the PVC is safe and obviates the need to systemati-
cally insert a CVC, which may allow an appreciable gain of time
in some critically ill patients needing urgent treatment (such as
antibiotics for severe sepsis) (31).
Limitations
Our study bears several limitations. For obvious reasons, it
was not blinded, and investigators were aware of the alloca-
tion group. However, complications were all strictly defined
a priori, thus limiting the subjectivity of the evaluation. The
use of ultrasound for CVC insertion was not routine in the
participating ICUs at the time of the study. One can hypoth-
esize that a number of hematoma or arterial punctures would
have been avoided in the CVC group and had an ultrasound
examination of systematically guided CVC insertion. This
would have, however, further increased the difference in com-
plication rates in favor of the CVC. Finally, maximum dose
and duration of vasopressor administration was not recorded.
It is, thus, not possible to say if complication results can be

explained by higher and longer vasopressor administration in
the PVC group. However, because inclusion criteria were simi-
lar in both groups and protocol mandated a crossover to a CVC
when vasopressor exceeded 2 mg/hr on a peripheral vein, we
believe this possibility is limited.
Taken together, our results provide for the first time data
on which to a base a decision regarding venous access in ICU
patients. In patients whom could be managed either with
a PVC or a CVC, a strategy based on systematic insertion of
CVCs is associated with fewer complications.
ACKNOWLEDGMENTS
We thank Dr. Marc Wysocki and Dr. Laurence Mier for helpful
suggestions for the design of the study, Dr. Diane Bouvry for
her help in collecting data, and all the ICU physicians who took
part in the Delphi questionnaire.
REFERENCES
1. Cook DJ, Sibbald WJ, Vincent JL, et al: Evidence based critical care
medicine; what is it and what can it do for us? Evidence Based
Medicine in Critical Care Group. Crit Care Med 1996; 24:334–337
2. Cook D, Guyatt G, Marshall J, et al: A comparison of sucralfate
and ranitidine for the prevention of upper gastrointestinal bleeding
in patients requiring mechanical ventilation. Canadian Critical Care
Trials Group. N Engl J Med 1998; 338:791–797
3. Drakulovic MB, Torres A, Bauer TT, et al: Supine body position as
a risk factor for nosocomial pneumonia in mechanically ventilated
patients: A randomized trial. Lancet 1999; 354:1851–1858
4. de Smet AM, Kluytmans JA, Cooper BS, et al: Decontamination of the
digestive tract and oropharynx in ICU patients. N Engl J Med 2009;
360:20–31
5. Bernard GR, Vincent JL, Laterre PF, et al; Recombinant human pro-
tein C Worldwide Evaluation in Severe Sepsis (PROWESS) study
group: Efficacy and safety of recombinant human activated protein C
for severe sepsis. N Engl J Med 2001; 344:699–709
6. Brochard L, Roudot-Thoraval F, Roupie E, et al: Tidal volume reduction
for prevention of ventilator-induced lung injury in acute respiratory dis-
tress syndrome. The Multicenter Trail Group on Tidal Volume reduc-
tion in ARDS. Am J Respir Crit Care Med 1998; 158:1831–1838
7. The Acute Respiratory Distress Syndrome Network: Ventilation with
lower tidal volumes as compared with traditional tidal volumes for
acute lung injury and the acute respiratory distress syndrome. N Engl
J Med 2000; 342:1301–1308.
8. Mancebo J, Fernández R, Blanch L, et al: A multicenter trial of pro-
longed prone ventilation in severe acute respiratory distress syn-
drome. Am J Respir Crit Care Med 2006; 173:1233–1239
9. Papazian L, Forel JM, Gacouin A, et al; ACURASYS Study
Investigators: Neuromuscular blockers in early acute respiratory dis-
tress syndrome. N Engl J Med 2010; 363:1107–1116
10. Vincent JL, Bihari DJ, Suter PM, et al: The prevalence of nosocomial
infection in intensive care units in Europe. Results of the European
Prevalence of Infection in Intensive Care (EPIC) Study. EPIC
International Advisory Committee. JAMA 1995; 274:639–644
11. Taylor RW, Palagiri AV: Central venous catheterization. Crit Care Med
2007; 35:1390–1396
Critical Care Medicine www.ccmjournal.org 2115
Clinical Investigations
12. Climo M, Diekema D, Warren DK, et al: Prevalence of the use of
central venous access devices within and outside of the intensive
care unit: Results of a survey among hospitals in the prevention
epicenter program of the Centers for Disease Control and Prevention.
Infect Control Hosp Epidemiol 2003; 24:942–945
13. Trick WE, Vernon MO, Welbel SF, et al: Unnecessary use of central
venous catheters: The need to look outside the intensive care unit.
Infect Control Hosp Epidemiol 2004; 25:266–268
14. McGee DC, Gould MK: Preventing complications of central venous
catheterization. N Engl J Med 2003; 348:1123–1133
15. Tagalakis V, Kahn SR, Libman M, et al: The epidemiology of periph-
eral vein infusion thrombophlebitis: A critical review. Am J Med 2002;
113:146–151
16. Monreal M, Quilez F, Rey-Joly C, et al: Infusion phlebitis in patients with
acute pneumonia: A prospective study. Chest 1999; 115:1576–1580
17. Maki DG, Kluger DM, Crnich CJ: The risk of bloodstream infection
in adults with different intravascular devices: A systematic review
of 200 published prospective studies. Mayo Clin Proc 2006;
81:1159–1171
18. O’Grady NP, Alexander M, Dellinger EP, et al: Guidelines for the
prevention of intravascular catheter-related infections. Centers for
Disease Control and Prevention. MMWR Recomm Rep 2002;
51(RR-10):1–29
19. Pronovost P, Needham D, Berenholtz S, et al: An intervention to
decrease catheter-related bloodstream infections in the ICU. N Engl
J Med 2006; 355:2725–2732
20. O’Grady NP, Alexander M, Burns LA, et al; Healthcare Infection
Control Practices Advisory Committee (HICPAC): Guidelines for the
prevention of intravascular catheter-related infections. Clin Infect Dis
2011; 52:e162–e193
21. Trotti A, Colevas AD, Setser A, et al: CTCAE v3.0: Development of
a comprehensive grading system for the adverse effects of cancer
treatment. Semin Radiat Oncol 2003; 13:176–181
22. Le Gall JR, Lemeshow S, Saulnier F: A new Simplified Acute
Physiology Score (SAPS II) based on a European/North American
multicenter study. JAMA 1993; 270:2957–2963
23. Timsit JF, Bruneel F, Cheval C, et al: Use of tunneled femoral catheters
to prevent catheter-related infection. A randomized, controlled trial.
Ann Intern Med 1999; 130:729–735
24. Timsit JF, Sebille V, Farkas JC, et al: Effect of subcutaneous tun-
neling on internal jugular catheter-related sepsis in critically ill
patients: A prospective randomized multicenter study. JAMA 1996;
276:1416–1420
25. Merrer J, De Jonghe B, Golliot F, et al; French Catheter Study Group
in Intensive Care: Complications of femoral and subclavian venous
catheterization in critically ill patients: A randomized controlled trial.
JAMA 2001; 286:700–707
26. Knottnerus JA, Dinant GJ: Medicine based evidence, a prerequisite
for evidence based medicine. BMJ 1997; 315:1109–1110
27. Polderman H, Girbes J: Central venous catheter use: Infectious com-
plications. Intensive Care Med 2002; 28:18–28
28. Polderman KH, Girbes AJ: Central venous catheter use. Part 1:
Mechanical complications. Intensive Care Med 2002; 28:1–17
29. Gueugniaud PY, Theurey O, Vaudelin T, et al: Peripheral versus cen-
tral intravenous lines in emergency cardiac care. Lancet 1987; 2:573
30. Giuffrida DJ, Bryan-Brown CW, Lumb PD, et al: Central vs peripheral
venous catheters in critically ill patients. Chest 1986; 90:806–809
31. Kumar A, Roberts D, Wood KE, et al: Duration of hypotension before
initiation of effective antimicrobial therapy is the critical determinant of
survival in human septic shock. Crit Care Med 2006; 34:1589–1596