nejmoa1500964.pdf

8/18/2019 nejmoa1500964.pdf

1/10

T he n e w e n g l a n d j o u r n a l o f medicine

n engl j med 373;13 nejm.org September 24, 20151220

The authors’ affiliations are listed in theAppendix. Address reprint requests to Dr.Parienti at the Department of Biostatisticsand Clinical Research, CHU de Caen, Ave.de la Côte de Nacre, Caen 14000, France,or at [email protected].

* A complete list of investigators in the3SITES study is provided in the Supple-mentary Appendix, available at NEJM.org.

Drs. Mermel and du Cheyron contributedequally to this article.

This article was updated on September 24,2015, at NEJM.org.

N Engl J Med 2015;373:1220-9.

DOI: 10.1056/NEJMoa1500964

Copyright © 2015 Massachusetts Medical Society.

BACKGROUND

Three anatomical sites are commonly used to insert central venous catheters, butinsertion at each site has the potential for major complications.

METHODS

In this multicenter trial, we randomly assigned nontunneled central venous catheter-ization in patients in the adult intensive care unit (ICU) to the subclavian, jugular,or femoral vein (in a 1:1:1 ratio if all three insertion sites were suitable [three-choice scheme] and in a 1:1 ratio if two sites were suitable [two-choice scheme]).The primary outcome measure was a composite of catheter-related bloodstream

infection and symptomatic deep-vein thrombosis.

RESULTS

A total of 3471 catheters were inserted in 3027 patients. In the three-choice com-parison, there were 8, 20, and 22 primary outcome events in the subclavian, jugu-lar, and femoral groups, respectively (1.5, 3.6, and 4.6 per 1000 catheter-days;P = 0.02). In pairwise comparisons, the risk of the primary outcome was signifi-cantly higher in the femoral group than in the subclavian group (hazard ratio, 3.5;95% confidence interval [CI], 1.5 to 7.8; P = 0.003) and in the jugular group thanin the subclavian group (hazard ratio, 2.1; 95% CI, 1.0 to 4.3; P = 0.04), whereasthe risk in the femoral group was similar to that in the jugular group (hazard ratio,1.3; 95% CI, 0.8 to 2.1; P = 0.30). In the three-choice comparison, pneumothorax

requiring chest-tube insertion occurred in association with 13 (1.5%) of the sub-clavian-vein insertions and 4 (0.5%) of the jugular-vein insertions.

CONCLUSIONS

In this trial, subclavian-vein catheterization was associated with a lower risk ofbloodstream infection and symptomatic thrombosis and a higher risk of pneumo-thorax than jugular-vein or femoral-vein catheterization. (Funded by the HospitalProgram for Clinical Research, French Ministry of Health; ClinicalTrials.gov num-ber, NCT01479153.)

A B S T R A CT

Intravascular Complications of Central

Venous Catheterization by Insertion Site Jean-Jacques Parienti, M.D., Ph.D., Nicolas Mongardon, M.D.,

Bruno Mégarbane, M.D., Ph.D., Jean-Paul Mira, M.D., Ph.D.,Pierre Kalfon, M.D., Ph.D., Antoine Gros, M.D., Sophie Marqué, M.D.,Marie Thuong, M.D., Véronique Pottier, M.D., Michel Ramakers, M.D.,

Benoît Savary, M.D., Amélie Seguin, M.D., Xavier Valette, M.D.,Nicolas Terzi, M.D., Ph.D., Bertrand Sauneuf, M.D.,

Vincent Cattoir, Pharm.D., Ph.D., Leonard A. Mermel, D.O.,and Damien du Cheyron, M.D., Ph.D., for the 3SITES Study Group*

Original Article

8/18/2019 nejmoa1500964.pdf

2/10

n engl j med 373;13 nejm.org September 24, 2015 1221

Complications of Cathe terization by Insertion Site

Subclavian, jugular, and femoral

central venous catheterization are associated with infectious, thrombotic, and mechani-

cal complications.1 Catheter-related bloodstreaminfection has a significant effect on morbidity,mortality, and health care costs.2-4 The risk ofshort-term catheter-related bloodstream infec-

tion is influenced mainly by extraluminal micro-bial colonization of the insertion site,5 and suchcolonization is also associated with thrombo-sis.6,7 Although the importance of catheter-relateddeep-vein thrombosis has been debated,1 allthromboses have the potential to embolize. Inaddition, catheter-related deep-vein thrombosis7-9 and pulmonary embolism10 may remain undiag-nosed in critically ill patients undergoing me-chanical ventilation.11

We conducted the 3SITES multicenter studyto evaluate the risk of catheter-related blood-stream infection or symptomatic catheter-relateddeep-vein thrombosis in adult patients who hadbeen admitted to an intensive care unit (ICU).On the basis of our previous meta-analysis,12 wehypothesized that the risk of these major com-plications would differ according to the site ofcatheter insertion.

Methods

Study Design and Oversight

The 3SITES study was a multicenter randomized,controlled trial conducted in four university-affiliated hospitals and five general hospitals,representing 10 ICUs, in France from December2011 through June 2014. The study was supportedby funds from the French Ministry of HealthProgramme Hospitalier de Recherche CliniqueNational to the Délégation de la Recherche Cli-nique et de l’Innovation of the Caen UniversityHospital. The first author designed the study.CareFusion provided chlorhexidine products freeof charge; no other commercial entity contributed

to this trial. The research ethics committee atCaen University approved the study protocol (avail-able with the full text of this article at NEJM.org)for all the participating centers. The first authoranalyzed the data and vouches for the accuracyand completeness of the reported data and forthe fidelity of the study to the protocol.

Patients and Randomization

Patients 18 years of age or older were eligible forthe study if they were admitted to the ICU, re-

quired nontunneled central venous vascular ac-cess through a new venipuncture, and wereconsidered by the physician inserting the cathe-ter to be suitable candidates for venous catheter-ization in at least two of the following threesites: the subclavian veins, the jugular veins, orthe femoral veins. The determination that a ve-

nous access site was suitable (or usable) forcatheterization was based on clinician judgment.More than one catheter per patient could be in-cluded in the trial. Written informed consent was obtained from all participants or from theirproxies in cases of impaired decision-makingcapacity at the time of enrollment.

If all three venous access sites (subclavian, jugular, and femoral) were considered suitablefor catheter placement, the catheterization site was assigned in a 1:1:1 randomization scheme(three-choice scheme). If one of the three sites was not suitable on both the left and right sidesof the body, the catheterization site was assignedin a 1:1 randomization scheme for the other twosites (two-choice scheme), an approach termed“selective exclusion.”13 If only one site was suit-able, that catheterization procedure was not in-cluded in the study. Randomization was strati-fied according to ICU and according to the useof antibiotic therapy versus no use of antibiotictherapy 14; it was implemented by means of acentralized 24-hour, web-based or telephone

interactive computerized response system (EOL,MedSharing), with the use of permuted-blockrandomization with varying block sizes.

Trial Procedures

All participating ICUs followed the French HauteAutorité de Santé checklist 15 and U.S. guidelinesfor preventing catheter-related infections.16 Resi-dents or staff physicians who had performed atleast 50 previous procedures inserted the cathe-ters or supervised the catheterization in the ICU.Maximal sterile barrier precautions were used,

including surgical hand antisepsis,17

sterile gloves,surgical long-sleeved gowns, caps, and masks.Patients were covered by sterile drapes. Antisep-tics, dressing, and catheter products are listedaccording to participating ICU in Table S1 in theSupplementary Appendix, available at NEJM.org.None of the study catheters were antiseptic-impregnated, antibiotic-impregnated, or tunneled.

Catheterization was achieved by means of theSeldinger technique with the use of anatomicallandmarks or ultrasonographic guidance. After

8/18/2019 nejmoa1500964.pdf

3/10


T h e n e w e n g l a n d j o u r n a l o f medicine

jugular and subclavian catheterizations, chestradiography was used to confirm the position ofthe catheter tip inside the superior vena cava andto assess for pneumothorax. Catheters were notused for routine blood sampling or renal re-placement therapy.

Decisions to remove catheters were made in-

dependently by the physicians caring for each pa-tient. After aseptic removal, the catheter tips weresent for quantitative culture.18 Peripheral bloodfor culture was systematically drawn at the timeof catheter removal (details are provided in theSupplementary Appendix). Patients dischargedfrom the ICU with the catheter in place hadblood drawn for culture simultaneously from aperipheral vein and the central venous catheterto determine the differential time to positivity.

Within 2 days after removal of the catheter,compression ultrasonography was performedat the insertion site to confirm symptomaticcatheter-related deep-vein thrombosis and to de-tect asymptomatic deep-vein thrombosis. Casesof symptomatic deep-vein thrombosis and casesof asymptomatic deep-vein thrombosis were com-bined for some analyses and referred to as totaldeep-vein thrombosis. Data on asymptomaticdeep-vein thrombosis were missing for all pa-tients who died or were discharged from the ICU with the catheter in place.

External, independent clinical monitors vali-

dated a randomly selected 12% of the data andall primary and secondary outcomes. Patients were followed until ICU discharge or death.

Outcomes

The primary outcome was the incidence of majorcatheter-related complications from the time ofcatheter insertion to 48 hours after catheter re-moval; major complications were defined as thecomposite of catheter-related bloodstream in-fection (Medical Dictionary for Regulatory Activities [MedDRA], version 17, code 10064687, grade 3

or higher) and symptomatic deep-vein thrombo-sis (MedDRA, version 17, code 10062169, grade 3or higher), whichever occurred first (see the defi-nitions in Table S2 in the Supplementary Appen-dix). Key secondary outcomes included the timeto catheter-tip colonization and time to totaldeep-vein thrombosis after catheter removal.

A diagnosis of catheter-related bloodstreaminfection required catheter-tip colonization withthe same phenotypic microorganism isolatedfrom a peripheral blood culture.16 For a diagnosis

of catheter-related bloodstream infection with apotential skin contaminant, two separate periph-eral-blood cultures had to grow the same micro-organism that colonized the catheter tip. Colo-nization of the catheter tip was defined as 1000or more colony-forming units per milliliter.18 Anadjudication committee that was unaware of the

study-group assignments reviewed all suspectedcases of catheter-related bloodstream infection.

If a patient had signs or symptoms of catheter-related deep-vein thrombosis, compression ultra-sonography was used to confirm the diagnosis.The ultrasonography-confirmed diagnosis servedas the deep-vein thrombosis component of theprimary outcome.

The secondary safety outcome was the rateof major mechanical complications (grade 3 orhigher) during insertion of the central venouscatheter and follow-up. Mechanical complications were defined in accordance with the modifiedNational Cancer Institute Common TerminologyCriteria for Adverse Events, version 4.0 (defini-tions are provided in Table S2 in the Supplemen-tary Appendix), with the modif ication that pneu-mothorax requiring chest-tube insertion wasclassified as grade 3 instead of grade 2.

Statistical Analysis

Our sample-size estimation is described in theSupplementary Appendix. We estimated that a

total sample of 3333 catheters was required,given our initial assumptions about the inci-dence of catheter-related complications.

The statistical unit of analysis was the cath-eter. Analyses followed the intention-to-treatprinciple. A per-protocol sensitivity analysis ex-cluded catheters that were not inserted in theallocated site and side of the body because offailure to gain vascular access. For cases in which catheter-tip culture data and blood-culturedata were missing, follow-up was censored atcatheter removal or at ICU discharge with the

catheter, whichever occurred first. Completecase and multiple-imputation sensitivity analy-ses were also performed (described in the Sup-plementary Appendix).

The incidence of the primary outcome wascompared among the three insertion sites in thethree-choice comparison with the use of theoverall log-rank test. Pairwise comparisons wereconducted (combining insertion-site groups fromamong the catheters that were randomized in thethree-choice scheme with the relevant groups from

8/18/2019 nejmoa1500964.pdf

4/10



the relevant two-choice scheme [Fig. 1]) with theuse of a Cox model that included catheter site,stratification variables, and design variables (inclu-sion in the three-choice or the relevant two-choicescheme).13 A robust sandwich covariance estimate

was used to account for a possible clusteringeffect of multiple catheters per patient. A sub-sample sensitivity analysis that included onerandomly selected catheter per patient was per-formed. The proportionality assumption was con-firmed visually and tested by including the site asa time-dependent covariate in the Cox model.

The intention-to-treat secondary safety outcome was analyzed by means of a random-intercept lo-gistic regression. Planned subgroup analyses of theprimary outcome were conducted by testing the

interaction term between each pairwise compari-son and the use of alcoholic chlorhexidine for cuta-neous antisepsis,19-21 antibiotic treatment,14 antico-agulation, body-mass index (the weight in kilogramsdivided by the square of the height in meters)

greater than 28,9

and selective site exclusion.13

Sim-ilarly, subgroup analyses for the secondary safe-ty outcome were conducted according to whetherultrasonography was used to guide insertion.22

We used SAS software, version 9.4 (SAS in-stitute), for the statistical analyses. The Holm–Bonferroni method23 was used to account formultiple testing of the primary outcome in thetwo superiority pairwise comparisons. Therefore,a P value of less than 0.025 was considered to in-dicate statistical signif icance for the lower P value,

Figure 1. Screening and Randomization of Catheter Insertions.Pairwise comparisons combine insertion-site groups from among the catheters that were randomized in the three-

choice scheme with the relevant groups from the relevant two-choice scheme. For example, the subclavian insertion-site group that was included in the subclavian-versus-femoral pairwise comparison included 878 catheters: the 843

catheters assigned in the three-choice scheme (i.e., with all three sites available) plus the 35 catheters in patientsfor whom the jugular site was excluded.

3471 Underwent randomization2532 Were assigned to 1:1:1939 Were assigned to 1:1

7559 Catheter insertions were screened

4088 Were excluded becauseonly one site was available

1016 Were assigned to subclavian site843 Had all sites available138 Had femoral site excluded35 Had jugular site excluded

1284 Were assigned to jugular site845 Had all sites available300 Had subclavian site excluded139 Had femoral site excluded

1171 Were assigned to femoral site844 Had all sites available296 Had subclavian site excluded31 Had jugular site excluded

866 Were inserted in assigned site51 Were inserted in the femoral site96 Were inserted in the jugular site3 Were inserted in the contralateral

subclavian site

1174 Were inserted in assigned site29 Were inserted in the subclavian site61 Were inserted in the femoral site20 Were inserted in the contralateral

jugular site

1114 Were inserted in assigned site4 Were inserted in the subclavian site

51 Were inserted in the jugular site2 Were inserted in the contralateral

femoral site

843 Were included in the intention-to-treat three-choice comparison

878 Were included in the intention-to-treat pairwise subclavian-versus-femoral comparison

981 Were included in the intention-to-treat pairwise subclavian-versus-jugular comparison


875 Were included in the intention-to-treat pairwise femoral-versus-subclavian comparison

1140 Were included in the intention-to-treat pairwise femoral-versus-jugular comparison


984 Were included in the intention-to-treat pairwise jugular-versus-subclavian comparison

1145 Were included in the intention-to-treat pairwise jugular-versus-femoral comparison

8/18/2019 nejmoa1500964.pdf

5/10



and a P value of less than 0.05 was considered toindicate significance for the higher P value.

Results

Baseline and Procedural Characteristics

and Follow-up

A total of 3027 patients were included in the study.A total of 3471 catheters (1284 jugular, 1171 femo-ral, and 1016 subclavian) were included, of which2532 (72.9%) were randomly assigned in the three-choice scheme (845 jugular, 844 femoral, and843 subclavian) (Fig. 1). The reasons for excludingone of the three sites are provided in the Table S3in the Supplementary Appendix. Catheters wereinserted in the randomly assigned site and side in3154 cases (90.9%) overall, including 866 cases(85.2%) assigned to the subclavian site, 1174cases (91.4%) assigned to the jugular site, and 1114cases (95.1%) assigned to the femoral site (Fig. 1).

The characteristics of the patients at baselineaccording to the site of catheter insertion were well balanced between the groups within thethree-choice comparison and the three pairwisecomparisons (Table 1). Catheter-related and pro-cedural characteristics are shown in Table 2. Theuse of anatomical landmarks was more frequentin the subclavian and femoral groups than in the jugular group. Catheterization was performedmore quickly in the femoral group than in either of

the other two groups. Alcohol-based products werethe predominant cutaneous antiseptics used forcleaning the catheter insertion site; the frequencyof the use of chlorhexidine-containing products was similar among the different insertion sites.

The median duration of catheter use was 5 daysfor each of the three insert ion sites (Table 2). Nopatients were lost to follow-up. Catheter-tip cul-tures and peripheral-blood cultures were missingin 101 cases (2.9%). Data on asymptomatic deep- vein thrombosis were missing in 2049 cases (59.0%).

Catheter-Related Infection and Symptomatic

Deep-Vein Thrombosis

In the three-choice comparison, there were 50nonduplicate primary outcome events (i.e., eventsthat did not occur in the same catheter), and theirincidence differed according to the randomlyassigned site of catheter insertion, with 8 eventsin the subclavian group, 20 events in the jugulargroup, and 22 events in the femoral group (1.5,3.6, and 4.6 per 1000 catheter-days, respectively; T

a b l e 1 .

C h a r a c t e r i s t i c s o f t h e P a

t i e n t s a t B a s e l i n e . *

C h a r a c t e r i s t i c

T h r e e - C h o i c e C o m p a r i s o

n

P a i r w i s e C o m p a r i s o n †



J u g u l a r

( N = 8 4 5 )

F e m o r a l

( N = 8 4 4 )

S

u b c l a v i a n

( N = 8 4 3 )

F e m o r a l

( N = 8 7 5 )

S u b c l a v i a n

( N = 8 7 8 )

J u g u l a r

( N = 9 8 4 )

S u b c l a v i a n

( N = 9 8 1 )

F e m o

r a l

( N = 1 1 4 0 )

J u g u l a r

( N = 1 1 4 5 )

A g e

6 3 . 0 ± 1 6 . 3

6 3 . 5 ± 1 5 . 3

6 2 . 9 ± 1 5 . 9

6 3 . 3 ± 1 5 . 5

6 2 . 8 ± 1 5 . 8

6 3 . 0 ± 1 6 . 1

6 2 . 5 ± 1 5 . 8

6 3 . 0 ± 1 5 . 1

6 3 . 1 ± 1 6 . 1

M a l e — n o . ( % )

5 3 1 ( 6 2 . 8 )

5 4 5 ( 6 4 . 6 )

5 4 3 ( 6 4 . 4 )

5 7 2 ( 6 5 . 4 )

5 7 0 ( 6 4 . 9 )

6 2 1 ( 6 3 . 1 )

6 3 5 ( 6 4 . 7 )

7 4 8 ( 6 5 . 6 )

7 2 1 ( 6 3 . 0 )

S A P S I I ‡

5 7 . 1 ± 1 9 . 4

5 6 . 1 ± 1 8 . 9

5 5 . 1 ± 2 0 . 2

5 6 . 1 ± 1 8 . 9

5 5 . 2 ± 2 0 . 1

5 6 . 9 ± 1 9 . 6

5 4 . 8 ± 2 0 . 2

5 7 . 3 ± 1 9 . 7

5 7 . 7 ± 2 0 . 3

B o d y - m a s s i n d e x

2 6 . 1 ± 5 . 7

2 5 . 8 ± 5 . 5

2 5 . 9 ± 5 . 3

2 5 . 9 ± 5 . 5

2 5 . 9 ± 5 . 3

2 6 . 2 ± 5 . 8

2 6 . 4 ± 5 . 8

2 5 . 9 ±

5 . 5

2 6 . 1 ± 5 . 7

D i a b e t e s m e l l i t u s — n o . ( % )

1 7 3 ( 2 0 . 5 )

1 7 5 ( 2 0 . 7 )

1 5 1 ( 1 7 . 9 )

1 8 2 ( 2 0 . 8 )

1 6 1 ( 1 8 . 3 )

2 1 4 ( 2 1 . 7 )

1 8 8 ( 1 9 . 2 )

2 3 8 ( 2 0 . 9 )

2 3 6 ( 2 0 . 6 )

C a n c e r — n o . ( % )

7 6 ( 9 . 0 )

9 6 ( 1 1 . 4 )

7 4 ( 8 . 8 )

9 7 ( 1 1 . 1 )

7 7 ( 8 . 8 )

8 2 ( 8 . 3 )

8 3 ( 8 . 5 )

1 4 0 ( 1 2 . 3 )

1 2 0 ( 1 0 . 5 )

A I D S — n o . ( % )

1 2 ( 1 . 4 )

8 ( 0 . 9 )

1 3 ( 1 . 5 )

8 ( 0 . 9 )

1 5 ( 1 . 7 )

1 2 ( 1 . 2 )

1 3 ( 1 . 3 )

1 2 ( 1 . 1 )

1 5 ( 1 . 3 )

N e u t r o p h i l c o u n t

8/18/2019 nejmoa1500964.pdf

6/10

8/18/2019 nejmoa1500964.pdf

7/10



P = 0.02) (Fig. 2). Corresponding Kaplan–Meiercurves showed a constant increase in risk foreach insertion-site group (Fig. S1 and S2 in theSupplementary Appendix). In pairwise compari-sons for the primary outcome in groups fromthe three-choice and two-choice schemes com-

bined (Table 3), the risk of the primary outcome was significantly higher in the femoral groupthan in the subclavian group (hazard ratio, 3.5;95% conf idence interval [CI], 1.5 to 7.8; P = 0.003)and in the jugular group than in the subclaviangroup (hazard ratio, 2.1; 95% CI, 1.0 to 4.3;P = 0.04), whereas the risk in the femoral group was similar to that in the jugular group (hazardratio, 1.3; 95% CI, 0.8 to 2.1; P = 0.30).

In a sensitivity analysis that included onerandomly selected catheter per patient, the results were consistent with those in the primary analy-

sis (Table S4 in the Supplementary Appendix).Differences between the subclavian group andthe other two groups were larger in the per-protocol sensitivity analysis (Table S5 in the Supple-mentary Appendix), because half of the catheter-related bloodstream infections in the subclaviangroup occurred in catheters that were, in fact,inserted elsewhere. The results of complete-caseand multiple-imputation sensitivity analyses wereconsistent with those in the primary analysis(Table S6 in the Supplementary Appendix).

The results for the secondary outcomes ofcatheter-tip colonization and total deep-veinthrombosis also favored the subclavian group(Table 3). Kaplan–Meier curves of these data areshown in Figures S3 and S4 in the Supplemen-tary Appendix, respectively. The causative patho-gens identified in each case of catheter-related

bloodstream infection and catheter-tip coloniza-tion are shown in Table S7 in the SupplementaryAppendix. Among the 171 blood samples drawnfor culture to determine the differential time topositivity in patients discharged from the ICU with their central venous catheter in place, theone catheter-related bloodstream infection iden-tified with the use of this method was subse-quently confirmed in a catheter-tip culture.

Mechanical Complications

The frequency of major mechanical complicationsin the three-choice comparison (Fig. 2) differedaccording to insertion-site group (P = 0.047), with18 events in the subclavian group, 12 events inthe jugular group, and 6 events in the femoralgroup. Pneumothorax accounted for 13 events inthe subclavian group and 4 events in the jugulargroup. In the pairwise comparisons (Table 3),there were significantly fewer mechanical com-plications in the femoral group than in the sub-clavian group (odds ratio, 0.3; 95% CI, 0.1 to 0.8;P = 0.03), but there were no significant differ-

ences in the other pairwise comparisons.

Subgroup Analyses

None of the preplanned subgroup analysesshowed a significant interaction for the primaryoutcome (Table S9 in the Supplementary Appen-dix). With regard to major mechanical complica-tions, there was a significant interaction betweenthe use of ultrasonography and the comparisonbetween the femoral group and the jugular group(P = 0.007), as well as a nonsignif icant trend foran interaction between the use of ultrasonogra-

phy and the comparison between the femoralgroup and the subclavian group (P = 0.07); thedifferences between the groups in these twocomparisons were larger when ultrasonography was not used to guide catheter insertion (TableS9 in the Supplementary Appendix).

Discussion

In this randomized, controlled trial, catheter-ization of the subclavian vein was associated

Figure 2. Complications in the Three-Choice Comparison, According to

Insertion-Site Group.

The primary end point (the composite of symptomatic deep-vein thrombosisand bloodstream infection) differed significantly among the insertion-sitegroups (P = 0.02 by the log-rank test), as did the principal safety secondaryend point (mechanical complications) (P = 0.047 by the chi-square test).

P e r c e n

t a g e o f C a t h e t e r s

w i t h

C o m p l i c a t i o n

4

2

3

1

0Subclavian(N=843)

Jugular (N=845)

Femoral(N=844)

Mechanical (grade ≥3)Symptomatic deep-vein

thrombosisBloodstream infection

18 (2.1%) 4 (0.5%)

4 (0.5%)

12 (1.4%) 8 (0.9%)

12 (1.4%)

6 (0.7%)12 (1.4%)

10 (1.2%)

8/18/2019 nejmoa1500964.pdf

8/10



T a b l e 3 .

I n t e n t i o n - t o - T r e a t P a i r w i s e C o m p a r i s o n s f o r t h e T r i a l O u t c o m e s .

O u t c o m e

F e m o r a l v e r s u s S u b c l a v i a n

J u g u l a r v e r s u s S u b c l a v i a n

F e m o r a l v e r s u s J u g u l a r

F e m o r a l

S u b c l a v i a n

H a z a r d R a t i o

( 9 5 % C I ) *

P V a l u e

J u g u l a r

S u b c l a v i a n

H a z a

r d R a t i o

( 9 5 % C I ) *

P V a l u e

F e m o r a l

J u g u l a r

H a z a r d R a t i o

( 9 5 % C I ) *

P V a l u e

n u m b e r

n u m b e r

n u m b e r

C a t h e t e r s

8 7 5

8 7 8

—

—

9 8 4

9 8 1

—

—

1 1 4 0

1 1 4 5

—

—

C a t h e t e r - d a y s

5 1 9 8

5 7 3 9

—

—

6 5 7 3

6 6 5 1

—

—

6 6 5 8

7 4 2 7

—

—

P r i m a r y c o m p o s i t e o u t c o m e †

2 5

8

3 . 5 ( 1 . 5 – 7 . 8 )

0 . 0 0 3

2 3

1 1

2 . 1 ( 1

. 0 – 4 . 3 )

0 . 0 4

3 3

3 0

1 . 3 ( 0 . 8 – 2 . 1 )

0 . 3 0

B l o o d s t r e a m i n f e c t i o n

1 1

4

3 . 4 ( 1 . 0 – 1 1 . 1 )

0 . 0 4 8

1 3

6

2 . 3 ( 0

. 8 – 6 . 2 )

0 . 1 1

1 5

2 1

0 . 9 ( 0 . 5 – 1 . 8 )

0 . 8 1

S y m p t o m a t i c d e e p - v e i n

t h r o m b o s i s

1 5

5

3 . 4 ( 1 . 2 – 9 . 3 )

0 . 0 2

1 0

6

1 . 8 ( 0

. 6 – 4 . 9 )

0 . 2 9

2 0

9

2 . 4 ( 1 . 1 – 5 . 4 )

0 . 0 4

S e c o n d a r y o u t c o m e

C a t h e t e r - t i p c o l o n i z a t i o n

1 0 7

3 9

3 . 4 ( 2 . 4 – 5 . 0 )

< 0 . 0 0 1

1 0 4

4 2

2 . 5 ( 1

. 7 – 3 . 5 )

< 0 . 0 0 1

1 4 5

1 2 1

1 . 6 ( 1 . 2 – 2 . 0 )

0 . 0 0 3

D e e p - v e i n t h r o m b o s i s ‡

4 6

1 9

3 . 0 ( 1 . 7 – 5 . 3 )

< 0 . 0 0 1

6 9

2 0

3 . 1 ( 1

. 9 – 5 . 0 )

< 0 . 0 0 1

6 2

7 8

0 . 9 ( 0 . 7 – 1 . 3 )

0 . 6 8

M a j o r m e c h a n i c a l c o m p l i c a -

t i o n s

6

1 8

0 . 3 ( 0 . 1 – 0 . 8 ) §

0 . 0 3

1 2

2 2

0 . 5 ( 0

. 3 – 1 . 1 ) §

0 . 0 9

7

1 3

0 . 5 ( 0 . 2 – 1 . 4 ) §

0 . 1 9

A r t e r i a l i n j u r y

4

1

—

—

2

2

—

—

4

2

—

—

H e m a t o m a

0

1

—

—

4

1

—

—

0

4

—

—

P n e u m o t h o r a x

N A

1 3

—

—

4

1 4

—

—

N A

4

—

—

M i s c e l l a n e o u s ¶

2

3

—

—

2

5

—

—

3

3

—

—

* V a l u e s i n t h i s c o l u m n a r e h a z a r d

r a t i o s u n l e s s o t h e r w i s e i n d i c a t e d . A l l h a z a r d r a t i o s w e r e a d j u s t e d f o r s t r a t i f i c a t i o n v a r i a b l e s ( i n t e n s i v e c a r e u n i t a n d a n t i b i o t i c u s e ) a n d d e s i g n v a r i -

a b l e s ( t h r e e - c h o i c e v s . t w o - c h o i c

e s c h e m e ) . T h e c o n f i d e n c e i n t e r v a l s f o r h a

z a r d r a t i o s a r e r o b u s t c o n f i d e n c e i n t e r v a l s

t a k i n g i n t o a c c o u n t m u l t i p l e c a t h e t e r s p e r

p a t i e n t . N A d e n o t e s n o t

a p p l i c a b l e .

† S o m e c a t h e t e r s h a d t w o e v e n t s .

‡ T h e n u m b e r s o f u l t r a s o n o g r a p h i c e v a l u a t i o n s i n t h e t h r e e p a i r w i s e c o m p a r

i s o n s w e r e 7 4 4 ( 3 9 8 i n t h e f e m o r a l g r o u p a n d 3 4 6 i n t h e s u b c l a v i a n g r o u p ) , 7 8 6 ( 3 9 9

i n t h e j u g u l a r g r o u p a n d

3 8 7 i n t h e s u b c l a v i a n g r o u p ) , a n

d 9 5 5 ( 4 9 7 i n t h e f e m o r a l g r o u p a n d 4 5 8 i n t h e j u g u l a r g r o u p ) .

§ T h e v a l u e f o r m a j o r m e c h a n i c a l c o m p l i c a t i o n s i s a n o d d s r a t i o c o m p u t e d b

y r a n d o m - i n t e r c e p t l o g i s t i c r e g r e s s i o n , r a t h e r t h a n a h a z a r d r a t i o .

¶ T h e s p e c i f i c m i s c e l l a n e o u s m e c h a n i c a l c o m p l i c a t i o n s f o r e a c h p a i r w i s e c o m

p a r i s o n a r e l i s t e d i n T a b l e S 8 i n t h e S u p p

l e m e n t a r y A p p e n d i x .

8/18/2019 nejmoa1500964.pdf

9/10



with a reduced risk of the combined outcomeof catheter-related bloodstream infection andsymptomatic deep-vein thrombosis. This was truein a comparison with femoral-vein catheterization,as others have suggested,8 but also in a com-parison with jugular-vein catheterization. Thesefindings are consistent with the Centers for Dis-

ease Control and Prevention guideline for preventing intravascular catheter-related infections,in which the recommendation is to “use a sub-clavian site, rather than a jugular or a femoralsite, in adult patients.”16 However, subclavian- vein catheterization was associated with an in-creased risk of mechanical complications.

The low incidence of catheter-related blood-stream infection in the ICUs in our study is con-sistent with data from other ICUs.24 Moreover,the differences in the incidences of catheter-related bloodstream infection and symptomaticdeep-vein thrombosis according to insertion siteare consistent with the differences found incatheter-tip colonization and total deep-veinthrombosis. Of note, the incidence of total deep- vein thrombosis should be interpreted with cau-tion, because more than half of the insertedcatheters had missing data for this secondaryoutcome, entirely because of missing data forasymptomatic patients.

There are probably several factors contributingto our findings. The subcutaneous course of the

subclavian catheter before entry into the vein isgenerally longer than for the other two types.The subclavian insertion site has the lowestbacterial bioburden25,26 and is relatively protectedagainst dressing disruption.27 Finally, subclaviancatheters are associated with less thrombosis.7,8

The overall risk of mechanical, infectious, andthrombotic complications of grade 3 or higher was similar among the three insert ion sites(Fig. 2), which suggests that an ideal site forcentral venous catheter insertion does not exist when all types of complications are considered

to be of equal concern. However, the expectedduration of catheterization is important, becausethe cumulative risk of infectious and thromboticcomplications increases with increasing catheterexposure, whereas the risk of mechanical com-plications does not. Furthermore, the mechani-cal complications associated with subclaviancatheter insertion can be limited by ultrasono-

graphic guidance28,29 and physician experience withthe procedure.30 Pneumothorax, which accountedfor most of the difference in mechanical compli-cations among insertion sites in our study, can bediagnosed promptly and treated immediately. Thismay not be the case for catheter-related blood-stream infection or deep-vein thrombosis. Deci-

sions regarding the choice of insertion site shouldtherefore be considered on a case-by-case basis.

A number of limitations of this trial shouldbe considered. The use of ultrasonographic guid-ance during catheter insertion was not random-ized. This may have influenced the risk of me-chanical and infectious complications found inthis study, although the reduction in catheterinfection risk associated with the use of ultra-sonography that was found in one randomizedstudy 22 was not confirmed in a subsequent largeobservational study.31 Daily chlorhexidine bath-ing32,33 and chlorhexidine-impregnated dressings34 were not used. Whether these measures inf lu-ence the difference in infectious risk betweeninsertion sites is unknown. Last, we did not studythe use of peripherally inserted central venouscatheters. Peripherally inserted central venouscatheters have been associated with a risk ofinfection similar to that associated with central venous catheters among patients in the ICU and with a higher risk of thrombosis.35,36

In conclusion, in the 3SITES study, we found

that catheterization of the subclavian vein was as-sociated with a lower risk of the composite outcomeof catheter-related bloodstream infection and symp-tomatic deep-vein thrombosis than that associated with catheterization of either the jugular vein orfemoral vein. However, subclavian-vein catheter-ization was associated with a higher risk of me-chanical complications, primarily pneumothorax.

Supported by a grant (PHRC-N 2010, 06-03) from the HospitalProgram for Clinical Research, French Ministry of Health.

Disclosure forms provided by the authors are available withthe full text of this art icle at NEJM.org.

We thank Cynthia T. Crosby, Ph.D. (Department of Clinical

Research, CareFusion, San Diego, CA), for providing alcoholicchlorhexidine (ChloraPrep, CareFusion) at no cost; Caroline Sabin,Ph.D. (Royal Free Campus, London), who reviewed an earlier version of the manuscript; Christian Brun-Buisson, M.D., Ph.D.(Hôpital Henri Mondor, Créteil), and Jean-François Timsit, M.D.,Ph.D. (Hôpital Bichat, Paris), for reviewing the initial protocoland providing supporting comments; and the nursing and med-ical staf f members of all the part icipating centers for their dedi-cation and the study participants and their family members, without whom this study could not have been completed.

Appendix

The authors’ affiliations are as follows: the Departments of Biostatistics and Clinical Research (J-J.P.), Infectious Diseases (J.-J.P.),Surgical Intensive Care (V.P.), Medical Intensive Care (A.S., X.V., N.T., B. Sauneuf, D.C.), and Microbiology (V.C.), Centre Hospitalier

8/18/2019 nejmoa1500964.pdf

10/10



Universitaire (CHU) Caen, INSERM Unité Mixte de Recherche Scientifique 1075 COMETE (N.T.) and EA4655 Risques Microbiens(J.-J.P., V.C., D.C.), Université de Caen Normandie, Caen, Department of Medical Intensive Care, CHU Cochin (N.M., J.-P.M.), andDepartment of Medical and Toxicologic Intensive Care, CHU Lariboisière (B.M.), Paris, Department of Anesthesiology and SurgicalIntensive Care, CHU Mondor, Créteil (N.M.), Department of Intensive Care Medicine, Centre Hospitalier Général, Chartres (P.K.), De-partment of Intensive Care Medicine, Centre Hospitalier Général, Versailles (A.G.), Department of Intensive Care Medicine, Centre Hospit-alier Général, Corbeil–Essonnes (S.M.), Department of Intensive Care Medicine, Centre Hospitalier Général, Pontoise (S.M.), and Depart-ment of Intensive Care Medicine, Centre Hospitalier Général, Saint-Lô (M.R., B. Savary) — all in France; and the Rhode Island HospitalDepartment of Medicine, Division of Infectious Diseases, Warren Alpert Medical School of Brown University, Providence (L.A.M.).

References

1. McGee DC, Gould MK. Preventingcomplications of central venous catheter-ization. N Engl J Med 2003;348:1123-33.2. Mermel LA. Prevention of intravascu-lar catheter-related infections. Ann InternMed 2000;132:391-402. [Erratum, Ann In-tern Med 2000:133:395.]3. Siempos II, Kopterides P, Tsangaris I,Dimopoulou I, Armaganidis AE. Impact ofcatheter-related bloodstream infections onthe mortality of critically ill patients: a meta-analysis. Crit Care Med 2009;37:2283-9.4. Zimlichman E, Henderson D, Tamir O,et al. Health care-associated infections:a meta-analysis of costs and financial im-pact on the US health care system. JAMAIntern Med 2013;173:2039-46.5. Mermel LA. What is the predominantsource of intravascular catheter infections?Clin Infect Dis 2011;52:211-2.6. Raad II, Luna M, Khali l SA, Costerton JW, Lam C, Bodey GP. The relationshipbetween the thrombotic and infectiouscomplications of central venous catheters. JAMA 1994;271:1014-6.7. Timsit JF, Farkas JC, Boyer JM, et al.Central vein catheter-related thrombosisin intensive care patients: incidence, risksfactors, and relationship with catheter-related sepsis. Chest 1998;114:207-13.

8. Merrer J, De Jonghe B, Golliot F, et al.Complications of femoral and subclavian venous catheterization in crit ically il l pa-tients: a randomized controlled trial. JAMA2001;286:700-7.9. Parienti J-J, Thirion M, Mégarbane B,et al. Femoral vs jugular venous catheter-ization and risk of nosocomial events inadults requiring acute renal replacementtherapy: a randomized controlled trial. JAMA 2008;299:2413-22.10. Minet C, Lugosi M, Savoye PY, et al.Pulmonary embolism in mechanically ventilated patients requiring computedtomography: prevalence, risk factors, andoutcome. Crit Care Med 2012;40:3202-8.

11. Cook DJ, Crowther MA. Thrombopro-phylaxis in the intensive care unit: focus onmedical-surgical patients. Crit Care Med2010;38:Suppl:S76-S82.12. Parienti J-J, du Cheyron D, Timsit J-F,et al. Meta-analysis of subclavian insertionand nontunneled central venous catheter-associated infection risk reduction incritically ill adults. Crit Care Med 2012;40:1627-34.13. Law MG, Emery S. Selective exclusionof treatment arms in multi-arm random-ized clinical trials. Stat Med 2003;22:19-30.14. Souweine B, Heng AE, Aumeran C,

et al. Do antibiotics administered at thetime of central venous catheter removalinterfere with the evaluation of coloniza-tion? Intensive Care Med 2008;34:286-91.15. French Haute Autorité de Santé. Check-list: pose d’un catheter veineux central(CVC) ou autre dispositif vasculaire (DV)(http://www.has-sante.fr/portail/upload/docs/application/pdf/2011-01/11_01_check-list-cvc-dv.pdf).16. O’Grady NP, Alexander M, Burns LA,et al. Guidelines for the prevention of in-travascular catheter-related infections. Am J Infect Control 2011;39:Suppl 1:S1-S34.17. Parienti JJ, Thibon P, Heller R, et al.Hand-rubbing with an aqueous alcoholicsolution vs traditional surgical hand-scrubbing and 30-day surgical site infec-tion rates: a randomized equivalence study. JAMA 2002;288:722-7.18. Brun-Buisson C, Abrouk F, Legrand P,Huet Y, Larabi S, Rapin M. Diagnosis ofcentral venous catheter-related sepsis:critical level of quantitat ive tip cultures.Arch Intern Med 1987;147:873-7.19. Chaiyakunapruk N, Veenstra DL, Lip-sky BA, Saint S. Chlorhexidine compared with povidone-iodine solution for vascu-lar catheter-site care: a meta-analysis.

Ann Intern Med 2002;136:792-801.20. Parienti J-J, du Cheyron D, RamakersM, et al. Alcoholic povidone-iodine to pre- vent central venous catheter colonization:a randomized unit-crossover study. CritCare Med 2004;32:708-13.21. Maiwald M, Chan ESY. The forgottenrole of alcohol: a systematic review andmeta-analysis of the clinical efficacy andperceived role of chlorhexidine in skinantisepsis. PLoS One 2012;7(9):e44277.22. Karakitsos D, Labropoulos N, DeGroot E, et al. Real-time ultrasound-guidedcatheterisation of the internal jugular vein:a prospective comparison with the land-mark technique in critical care patients.

Crit Care 2006;10:R162.23. Aickin M, Gensler H. Adjusting formultiple testing when reporting researchresults: the Bonferroni vs Holm methods.Am J Public Health 1996;86:726-8.24. Vital signs: central line-associatedblood stream infections — United States,2001, 2008, and 2009. MMWR Morb Mor-tal Wkly Rep 2011;60:243-8.25. Moro ML, Viganò EF, Cozzi Lepri A.Risk factors for central venous catheter-related infections in surgical and inten-sive care units. Infect Control Hosp Epi-demiol 1994;15:253-64.

26. Kwakman PH, Müller MC, Binnekade JM, et al. Medical-grade honey does notreduce skin colonization at central venouscatheter-insertion sites of critically ill pa-tients: a randomized controlled trial. CritCare 2012;16:R214.27. Timsit J-F, Bouadma L, Ruckly S, et al.Dressing disruption is a major risk factorfor catheter-related infections. Crit CareMed 2012;40:1707-14.28. Fragou M, Gravvanis A, Dimitriou V,et al. Real-time ultrasound-guided sub-clavian vein cannulation versus the land-mark method in critical care patients:a prospective randomized study. Crit CareMed 2011;39:1607-12.29. O’Leary R, Ahmed SM, McLure H, et al.Ultrasound-guided infraclavicular axil lary vein cannulation: a useful alternative tothe internal jugular vein. Br J Anaesth2012;109:762-8.30. Eisen LA, Narasimhan M, Berger JS,Mayo PH, Rosen MJ, Schneider RF. Me-chanical complications of central venouscatheters. J Intensive Care Med 2006;21:40-6.31. Cartier V, Haenny A, Inan C, Walder B,Zingg W. No association between ultra-sound-guided insertion of central venouscatheters and bloodstream infection: a pro-

spective observational study. J Hosp Infect2014;87:103-8.32. Climo MW, Yokoe DS, Warren DK,et al. Effect of daily chlorhexidine bathingon hospital-acquired infection. N Engl JMed 2013;368:533-42.33. Noto MJ, Domenico HJ, Byrne DW,et al. Chlorhexidine bathing and healthcare-associated infections: a randomizedclinical trial. JAMA 2015;313:369-78.34. Timsit J-F, Schwebel C, Bouadma L,et al. Chlorhexidine-impregnated spongesand less frequent dressing changes forprevention of catheter-related infectionsin critically ill adults: a randomized con-trolled trial. JAMA 2009;301:1231-41.

35. Chopra V, O’Horo JC, Rogers MAM,Maki DG, Safdar N. The risk of blood-stream infection associated with periph-erally inserted central catheters com-pared with central venous catheters inadults: a systematic review and meta-analysis. Infect Control Hosp Epidemiol2013;34:908-18.36. Chopra V, Anand S, Hickner A, et al.Risk of venous thromboembolism associ-ated with peripherally inserted centralcatheters: a systematic review and meta-analysis. Lancet 2013;382:311-25.Copyright © 2015 Massachusetts Medical Society.

nejmoa1500964.pdf

Documents

Transcript of nejmoa1500964.pdf