Download Left Superior Vena Cava Persistence in Patients Undergoing

Left Superior Vena Cava Persistence in
Patients Undergoing Pacemaker or
Cardioverter-Defibrillator Implantation*
A 10-Year Experience
Mauro Biffi, MD; Giuseppe Boriani, MD; Lorenzo Frabetti, MD;
Gabriele Bronzetti, MD; and Angelo Branzi, MD
Objective: The persistence of a left superior vena cava (LSVC) has been observed in 0.3% of the
general population as established by autopsy. In the adult population, it is an important anatomic
finding if a left superior approach to the heart is considered. The aim of the study was to evaluate
the prevalence of a LSVC in patients undergoing pacemaker (PM) and cardioverter-defibrillator
(CD) implantation.
Design: We observed the prevalence of LSVC during a 10-year period; each patient undergoing
PM or transvenous CD implantation received a left cephalic/left subclavian venous approach to
the heart. With this technique, LSVC persistence is easily diagnosed during lead placement.
Results: A total of 1,139 patients consecutively underwent PM implantation during 10 years: 4
patients had persistent LSCV (0.34%). Among 115 patients undergoing CD implantation, 2
patients with LSVC (1.7%) were observed. Overall LSVC persistence was found in 6 of 1,254
patients (0.47%). Two patients, one of whom had no right superior vena cava (RSVC), received a
left-sided PM, whereas two other patients received right-sided devices. Both CD patients
received a left-sided active-can device: the first patient with a right-sided lead tunneled to the left
pectoral pocket, as a result of poor catheter handling through the LSVC and coronary sinus, and
the second patient with a screw-in lead from LSVC. Long-term follow-up of these patients
(average ⴞ SD, 41 ⴞ 26 months) revealed absence of lead dislodgment and appropriate device
function regardless of lead implantation site.
Conclusions: Persistence of LSVC in adults undergoing PM/CD implantation is similar to that of
the general population (0.47% in our study). The left-sided implant can be achieved by stylet
shaping and by use of active fixation leads in most patients, with a reliable outcome at short term
in addition to appropriate device performance at follow-up. Assessment of the RSVC is advisable
when planning a right-sided implantation, since a minority of patients lacks this vessel.
(CHEST 2001; 120:139 –144)
Key words: left superior vena cava persistence; pacemaker/defibrillator lead
Abbreviations: AV ⫽ atrioventricular; CD ⫽ cardioverter-defibrillator; LSVC ⫽ left superior vena cava;
PM ⫽ pacemaker; RSVC ⫽ right superior vena cava; SSS ⫽ sick sinus syndrome; VT ⫽ ventricular tachycardia
persistence of a left superior vena cava
T he(LSVC)
is due to an abnormal development of
the sinus venosus in the early stages of fetal life. In
the 4-mm embryo, this structure consists of three
distinct parts, namely, the right horn, the transverse
part, and the left horn, and collects three pairs of
veins (the omphalomesenteric veins, the umbilical
*From the Institute of Cardiology, University of Bologna, Bologna, Italy.
Manuscript received May 9, 2000; revision accepted January 10,
2001.
Correspondence to: Mauro Biffi, MD, Institute of Cardiology,
University of Bologna, Via Massarenti 9, 40138 Bologna, Italy
veins, and the common cardinal veins). Due to the
rightward direction of blood flow, the right horn
undergoes preferential growth and ultimately constitutes the intercaval (nonmuscular) part of the right
atrium.1,2
The transverse part and the proximal left horn of
the sinus venosus invaginate from the left atrium and
form the coronary sinus. The distal left horn and the
left cardinal vein obliterate, and form the so-called
ligament of Marshall1,2 in the adult subject. In the
case of patency of the left cardinal vein, its drainage
reaches the right atrium through the coronary sinus,
which becomes enlarged to variable degrees. In 92%
of cases, drainage occurs in the right atrium; in the
CHEST / 120 / 1 / JULY, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
139
remainder of cases, drainage occurs in the left atrium
because of failure to form the coronary sinus.3
Persistence of LSVC has been reported to occur in
approximately 0.3% of the general population in a
single, large report (⬎ 4,000) of unselected autopsies.4 Its prevalence is in fact much higher in patients
with congenital cardiac abnormalities compared to
the general population, ranging from 2.8 to 4.3%.5– 8
In addition, about 10% of these subjects with congenital cardiac abnormalities do not have a right
superior vena cava (RSVC).4,9
When isolated, LSVC persistence is usually not
recognized until a left superior approach to the heart
is required, when it becomes a relevant anatomic
finding. In fact, it can complicate the positioning of
left-sided pacemaker (PM) and cardioverter-defibrillator (CD) leads,10 –13 the placement of central venous lines for therapeutic purposes and hemodynamic monitoring,12,14 and cardiopulmonary bypass
in patients undergoing cardiac/thoracic surgery procedures.12
Isolated cases of LSVC persistence in patients
undergoing PM or CD implantation have been
reported in literature10 –13,15–19; the most common
problems related to the unusual anatomic access to
the heart were reaching a convenient pacing site and
ensuring stable lead placement.10 –12 This background activated our center to prospectively investigate the prevalence of LSVC persistence in adults
undergoing either PM or CD implantation over a
10-year period, and to evaluate the relative benefits
of the different approaches adopted by the distinct
physicians to overcome implant-related difficulties.
Materials and Methods
The study was prospectively conducted from October 1989 to
September 1999 on all patients undergoing PM implantation;
from January 1992 to September 1999, patients undergoing
transvenous CD implantation were also taken into consideration.
Patients with congenital thoracic-abdominal abnormalities were
not considered due to the likelihood of associated LSVC persistence5– 8; this would have increased the prevalence in the patients
with bradyarrhythmias and tachyarrhythmias. Patients requiring
generator replacement and patients with known thrombosis of
the left subclavian vein were not considered in the study.
From October 1989, all device implantations were consecutively carried out using the left-sided approach. In case of LSVC
persistence, the guidewire (or the lead in case of cephalic access)
enters the subclavian vein and does not cross the column, but lays
parallel to it before entering the coronary sinus posteriorly to the
left atrium. After entering the right atrium, the passage into the
tricuspid orifice is accomplished from the posterior to the
anterior and then leftwards, almost on the same horizontal plane,
and requires skilful shaping of the stylet depending on right-heart
size and geometry. The access to the right atrial appendage goes
downwards, then upwards, and then anteriorly.
The physicians defined different steps to overcome implantrelated difficulties: hand shaping the stylet in predefined fashion,
using active fixation leads, and moving to a right-sided implant
after right peripheral venography by nonionic medium (iomeprol,
20 mL) or echocardiography. The practical outcome of the
different policies was assessed by lead performance, total fluoroscopy time, and total procedure time.
Statistical Analysis
Comparisons of total fluoroscopy times and total procedure
times were performed using unpaired t tests.
Results
During 10 years, 1,147 consecutive patients without congenital abnormalities underwent PM implantation for commonly accepted indications; 8 patients
were occasionally discovered to have asymptomatic
left subclavian thrombosis (angiographically documented during implantation) and were hence excluded from analysis due to the impossibility of
ruling out LSVC persistence. Of the other 1,139
patients, 627 patients were male (55%) and 512 were
female (45%); 4 of 1,139 patients had patency of
LSVC (0.34%). In the period from 1992 to 1999, 115
consecutive patients (84 male [73%] and 31 female
patients [27%]) underwent transvenous CD implantation; of these patients, LSVC persistence was
observed in 2 patients (1.7%). Overall, LSVC persistence was found in 6 of 1,254 patients (0.47%)
patients, all of whom were male.
The first patient (aged 57 years) had sick sinus
syndrome (SSS) with episodic sinus arrest; LSVC
persistence was recognized during lead advance by
the left cephalic vein. An elective right-sided implant
was performed after visualization of the RSVC by
contrast venography (Table 1). Over 94 months of
follow-up, this patient was asymptomatic with ⬍ 5%
pacing by rate hysteresis.
The second patient (aged 60 years) had SSS.
LSVC was recognized while advancing a standard
bipolar ventricular lead through the left cephalic
vein. Placement in the right ventricular apex was
accomplished by a hand-shaped stylet consisting of a
distal rounded “L” following a proximal leftward
sharp bending (Fig 1, middle, B; Fig 2). Placement
into the right appendage was successful with a
rounded L-shaped stylet (Fig 1, top, A) and an active
fixation lead (Fig 2). A normal RSVC entering the
right atrium was visualized by contrast venography
after implantation.
The third patient (aged 58 years) had SSS and
mitral valve prolapse; LSVC was diagnosed after
subclavian vein puncture by anomalous wire course.
This patient was treated by shaping stylets in the
same way as was the second patient (Fig 1) but
received active fixation leads in both chambers. This
patient had no RSVC at venography performed at
the end of implantation.
140
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
Clinical Investigations
Table 1—Main Parameters at Implantation and at Last Follow-up of the Implanted Devices*
Patient No./
Arrhythmia
Type
1/SSS
2/SSS
3/SSS
4/AVB
5/VT
6/VT
A/V Sensing
Threshold, mV
A/V Pacing
Threshold, mA
A/V Impedance, ⍀
Device
Type
Lead
Side
Shortterm
Longterm
Shortterm
Longterm
Shortterm
Longterm
Defibrillation
Threshold, J
Fluoroscopy
Time, s
Procedure
Time, min
PM
VVI
PM
DDD
PM
DDD
PM
DDD
CD
CD
Right
—/17.2
—/⬎ 8
—/0.5
—/1.2
—/523
—/614
—
370
190
Left
3.5/18.5
3.5/⬎ 15
0.8/0.6
1.6/1.1
530/610
718/677
—
460
150
Left
3.4/16.8
3/14.3
1.1/0.8
1.7/1.5
422/472
380/563
—
500
140
Right
2.8/5.8
2.5/—
1.4 /0.4
2.1/0.8
350/1,070
440/1,184
—
550
200
Right
Left
—/12.7
—/20.9
—/⬎ 9
—/18
—/1.0
—/0.7
—/1.6
—/1.9
—/546
—/870
—/593
—/825
15
12
1,380
620
220
190
*A/V ⫽ atrium/ventricle; AVB ⫽ atrioventricular block; DDD ⫽ double-chamber pacemaker; VVI ⫽ single-chamber ventricular pacemaker.
The fourth patient (aged 65 years) had complete
right bundle-branch block, second-degree atrioventricular (AV) block, and episodic complete heart
block following an old inferior myocardial infarction.
LSVC was diagnosed at subclavian vein puncture.
The physician moved to a right-sided implantation
after visualization of the RSVC by contrast venography. This patient was PM- dependent 6 months after
implantation.
Both CD patients had dilated cardiomyopathy due
to coronary artery disease and previous myocardial
infarction.
The fifth patient (aged 58 years) had been rescued
five times from monomorphic ventricular tachycardia
(VT) while screening for heart transplantation. CD
implantation was planned as a bridge to heart transplantation. LSVC was diagnosed following subclavian
puncture: passage through the right cardiac chambers
was difficult, and the lead neither reached the right
Figure 1. Drawing of a hand-shaped stylet for lead placement at
the right atrial (A) and right ventricular (B) site through the
coronary sinus.
ventricular apex nor demonstrated stable or reliable
sensing. Then, the physician changed to another implantation strategy. The RSVC was visualized by echocardiography at the bedside in the operating room. The
patient received a left pectoral active-can device (housing can takes active part into the defibrillation process)
but with a right-sided lead tunneled to the left pectoral
pocket. This patient eventually underwent heart transplantation 6 months later, having been rescued once
more from fast VT. This patient is not considered in
long-term follow-up.
The sixth patient (aged 64 years) had right bundlebranch block and sustained monomorphic VT unresponsive to drug therapy. He received a pectoral
left-sided active-can device, with an active-fixation,
single-coil lead passing through the LSVC and coronary sinus. Placement in the right ventricular apex
(Fig 3) was accomplished by a hand-shaped stylet
(Fig 1, middle, B). Right-sided venography performed during implantation demonstrated absence
of the RSVC.
In this experience, total fluoroscopy time
(mean ⫾ SD) was 526 ⫾ 83 s for left-sided implants
and 766 ⫾ 538 s for right-sided implants, respectively (p ⫽ not significant). Total procedure time was
160 ⫾ 26 min for left-sided implants and 203 ⫾ 15
min for right-sided implants, respectively (p ⫽ 0.07).
Active fixation leads were used in two of two patients
at the atrial site, and in two of three patients at the
ventricular site in those with left-sided implantation.
Absence of the RSVC was observed in two of six
patients (33%).
Five patients were followed up for 41 ⫾ 32
months on average; absence of lead dislodgment and
appropriate function either for PM or CD patients
regardless of lead implantation site were observed
(Table 1). Follow-up times for leads placed through
LSVC and coronary sinus were 46 months, 26
CHEST / 120 / 1 / JULY, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
141
Figure 3. Radiographic views of lead placement into the right
ventricle in a patient with VT, LSVC persistence, and absent
RSVC: anteroposterior view (top, A), right anterior oblique
(middle, B), and left anterior oblique (bottom, C).
Figure 2. Radiographic views of lead placement into the right
atrial appendage and right ventricle in a patient with SSS:
anteroposterior view (top, A), right anterior oblique (middle, B),
and left anterior oblique (bottom, C).
months, and 20 months, respectively, whereas they
were 94 months and 18 months for those placed
through the RSVC. Lead integrity was 100%, as
assessed by telemetry system check.
Discussion
Persistence of LSVC in adult life is normal in
rabbits and some other mammals, but it is a rare
abnormality in man. Besides being associated with
congenital diseases,5– 8 its most relevant clinical implication is the association with disturbances of
cardiac impulse formation and conduction.20 The
ontogenetic development of the sinus node, the AV
node, and the His bundle may be heavily influenced
by the lack of regression of the left cardinal vein,21–23
since these structures are located at the junctions of
the right and left cardinal veins with the sinus
venosus. Persistence of LSVC and/or anomalies of
RSVC have been described to alter the location and
also the histologic organization of both the sinus
142
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
Clinical Investigations
node and the AV junction,24 causing small and poorly
formed sinus node, fetal dispersion of the AV node
and His bundle within the central fibrous body, small
diameter of the His bundle, and poor arterial supply
to either the AV node or the sinus node.25,26 This
pathologic substrate may predispose patients to arrhythmias and also to sudden death.25,26
LSVC Prevalence
The prevalence of LSVC in general population is
about 0.3%, as established by an autopsy study4;
however, to our knowledge, the prevalence in patients with symptomatic bradyarrhythmias requiring
permanent cardiac pacing has never been investigated before. In our study, we observed a similar,
although slightly higher, prevalence of LSVC persistence compared to general population as assessed by
unselected autopsies4; this may be viewed as lower
than what expected a priori, given the association
with abnormalities of impulse formation and conduction.20 –26 In our study, RSVC was absent in 33% of
patients, compared to an average 10% in other
reports.4,7,9 This difference may be due to the low
prevalence of RSVC absence in each study. Indeed,
this anatomic finding is of key importance from a
surgical point of view (in this situation, the left
vascular axis constitutes the only superior access to
the heart) and implies a nontransvenous implantation in case of failure from the LSVC.
The finding of a LSVC complicating placement of
left-sided PM or CD systems has been reported as a
sporadic observation in literature, mainly in isolated
patients. Of a total of 661 patients undergoing
VVI-mode PM system implantation, Zerbe et al16
reported 4 patients with LSVC persistence, but the
observation was retrospective and no systematic
attempt to assess the prevalence of this abnormality
by consecutive left-sided approach to each PM implantation was made. The customary left-sided approach adopted in our center allowed easy diagnosis
of LSVC persistence and prevented missing “falsenegatives,” thus giving a true estimate of its prevalence in this cardiologic population.
Practical Implications
Different techniques have been used to obtain a
reliable function of the implanted system in these
patients,10 –13,15–18 minimizing exposure to radiation
and risk of lead dislodgment.10 –12 Many authors
found helpful ways to shape stylets as to enter the
tricuspid valve or to reach the right atrial appendage15–17; active fixation leads were used to ensure
lead stability,17,19 although this was not mandatory in
all the cases.
In our experience, lead placement by the left
approach through the coronary sinus was feasible
with reliable results. We suggest use of active fixation
leads in the right appendage, guided by a large-curve
L-shaped stylet (Fig 1, top, A). Careful shaping of
the stylet (Fig 1, middle, B) allowed an easy access to
the right ventricular apex in three of four patients in
whom it was attempted. A standard ventricular lead
was used in one patient with success; however, active
fixation leads may be used when diagnosis is made
before lead insertion.
In patients with poor handling through the coronary sinus, a right approach is recommended after
visualization of a right vena cava entering the right
atrium by echocardiography27 or contrast venography; the absence of RSVC would suggest an epicardial implantation. In our experience, successful implantation by the left side did not require
significantly longer fluoroscopy time compared to a
right-sided implantation except for the first CD
patient. Crossing over to a right-sided approach
requires a longer procedure time, echocardiography
or contrast venography to ensure RSVC presence,
and thus may cause greater discomfort to the patients. In fact, the two patients who crossed over to
the right approach in “elective” manner (no attempt
from the left side after diagnosis) had longer procedure times (190 min and 200 min, respectively)
compared to average procedure time of the leftsided approach (160 ⫾ 26 min). We therefore suggest that implantation be accomplished left sided
when diagnosis is made intraoperatively, unless catheter handling appears difficult or fluoroscopy time
exceeds a defined maximum.
In patients undergoing CD implantation, devices
with outer shell taking part into the defibrillation
process (termed active can or hot can, depending on
different manufacturers) have lower defibrillation
current requirements when placement is left sided
compared to right-sided placement.28,29 Current CD
generation is capable of achieving acceptable defibrillation thresholds also when placed right pectoral,30 so that tunneling the lead to a left subclavian
pocket is unnecessary in several cases. A left-sided
implantation may be mandatory in the minority of
patients with high defibrillation thresholds from the
right-sided configuration to improve the effectiveness of heart defibrillation. We suggest trying a
transvenous left-sided approach first, and considering a right venous access in case of failure to reach a
convenient site.
Relevant implications of our observation may apply to the fields of critical care, anesthesiology,
general and thoracic surgery, oncology, and hematology when central lines for monitoring or therapeutic purposes are required12,14 or when permanent
catheters for drug delivery are implanted. Awareness
CHEST / 120 / 1 / JULY, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
143
of this venous anomaly may obviate unnecessary
catheter removal and troublesome placement of a
new one when arterial puncture is suspected by
imaging techniques, but not obvious at bedside
verification tests. In addition, assessment of the
RSVC may be very important in these patients, since
its absence may represent a major obstacle in providing care of critical patients.14
In conclusion, persistence of LSVC in patients
undergoing PM/CD implantation is similar to that in
the general population (0.47% in our study). Specifically, shaped stylets and active fixation leads are
helpful tools to overcome technical difficulties in
ordinary procedure times. Right-sided implantation
is easily feasible once the presence of a RSVC has
been confirmed either by echocardiography or by
peripheral venography, since a substantial minority
of these patients (two of six patients in our study)
lacks a right-sided superior venous access to the right
atrium.
References
1 Los JA. The development of the human pulmonary vein and
the coronary sinus in the human embryo [thesis]. Leiden, The
Netherlands: University of Leiden, 1958
2 Van Mierop LHS. Embryology of the heart. In: Netter FH,
ed. The CIBA collection of medical illustrations (vol 5).
Summit, NJ: CIBA Pharmaceutical, 1969; 112–130
3 Meadows WR, Sharp JT. Persistent left superior vena cava
draining into the left atrium without arterial oxygen unsaturation. Am J Cardiol 1965; 16:273–279
4 Geissler W, Albert M. Persistierende linke ober Hohlvene
und Mitralstenose. Z Gesamte Inn Med 1956; 11:865
5 Loogen F, Rippert R. Anomalien der grosse Körper und
Lungenvenen. Z Kreisslauforsch 1958; 47:677
6 Mantini E, Grondin CM, Lillehei CW, et al. Congenital
anomalies involving the coronary sinus. Circulation 1966;
33:317–327
7 Wood P. Diseases of the heart and circulation. 2nd ed.
Philadelphia, PA: JB Lippincott, 1956; 457
8 Campbell M, Deuchar DC. The left-sided superior vena cava.
Br Heart J 1954; 16:423
9 Karnegis JN, Wang Y, Winchell P, et al. Persistent left
superior vena cava, fibrous remnant of the right superior vena
cava and ventricular septal defect. Am J Cardiol 1964;
14:573–577
10 Krukal JC. Transvenous pacemaker failure due to anomalous
venous return to the heart. Chest 1971; 59:458 – 461
11 Garcia L, Levine R, Kosowsky W, et al. Persistent left
superior vena cava complicating pacemaker insertion. Chest
1972; 61:396 –397
12 Rubenfire M, Evangelista J, Wajszczuk WJ, et al. Implication
of a persistent left superior vena cava in transvenous pacemaker therapy and cardiac hemodynamic monitoring. Chest
1974; 65:145–147
13 Robbens EJ, Ruiter JH. Atrial pacing via unilateral persistent
left superior vena cava. PACE 1986; 9:594 –596
14 Mooney DP, Snyder CL, Holder TM. An absent right and
persistent left superior vena cava in an infant requiring
extracorporeal membrane oxygenation. J Pediatr Surg 1993;
28:1633–1634
15 Dirix LY, Kersschot IE, Fierens H, et al. Implantation of a
dual chamber pacemaker in a patient with persistent left
superior vena cava. Pacing Clin Electrophysiol 1988; 11:343–
345
16 Zerbe F, Bornakowski J, Sarnowski W. Pacemaker electrode
implantation in patients with left superior vena cava. Br
Heart J 1992; 67:65– 66
17 Rusk PA, Bexton JM, Mc Comb JM. Persistent left sided and
absent right sided superior vena cava complicating pacemaker
insertion. Heart 1996; 75:413
18 Brooks R, Jackson G, Mc Govern BA, et al. Transvenous
cardioverter-defibrillator implantation via persistent left superior vena cava. Am Heart J 1995; 129:195–197
19 Hellestrand KJ, Ward DE, Bexton RS, et al. The use of active
fixation electrodes for permanent endocardial pacing via a left
superior vena cava. Pacing Clin Electrophysiol 1982; 5:180 –
184
20 Camm AJ, Dymond D, Spurrell RAJ. Sinus node dysfunction
associated with the absence of right superior vena cava. Br
Heart J 1979; 41:504 –507
21 Patten BM. The development of the sinoventricular conduction system. Univ Mich Med Bull 1956; 22:1
22 James TN. Cardiac conduction system: fetal and postnatal
development. Am J Cardiol 1970; 25:213–226
23 Van Mierop LHS, Patterson PR, Reynolds RW. Two cases of
congenital asplenia with isomerism of the cardiac atria and
the sinoatrial nodes. Am J Cardiol 1964; 13:407
24 Anderson RH, Latham RA. The cellular architecture of the
human atrioventricular node, with a note on its morphology in
the presence of a left superior vena cava. J Anat 1971;
109:443– 455
25 James TN, Marshall TK. Persistent fetal dispersion of the
atrioventricular node and His bundle within the central
fibrous body. Circulation 1976; 53:1026 –1034
26 James TN, Marshall TK, Edwards JE. Cardiac electrical
instability in the presence of a left superior vena cava.
Circulation 1976; 54:689 – 697
27 Snider AR, Ports TA, Silverman RH. Venous anomalies of the
coronary sinus: detection by M-mode, two-dimensional and
contrast echocardiography. Circulation 1979; 60:721–724
28 Roberts PR, Allen S, Betts T, et al. Increased defibrillation
threshold with right-sided active pectoral can. J Interv Card
Electrophysiol 2000; 4:245–249
29 Epstein AE, Kay GN, Plumb VJ, et al. Elevated defibrillation
threshold when right-sided venous access is used for nonthoracotomy implantable defibrillator lead implantation: the
Endotak investigators. J Cardiovasc Electrophysiol 1995;
6:979 –986
30 Flaker GC, Tummala R, Wilson J. Comparing of right- and
left-sided pectoral implantation parameters with the Jewel
active can cardiodefibrillator: the World Wide Jewel Investigators. Pacing Clin Electrophysiol 1998; 21:447– 451
144
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21964/ on 05/04/2017
Clinical Investigations