66

Prévia do material em texto

Should We Still Order Chest X-rays in the ICU?
V. Ioos, A. Galbois, and B. Guidet
J.-L. Vincent (ed.), Annual Update in Intensive Care and Emergency Medicine 2011
DOI 10.1007/978-3-642-18081-1, ˇ Springer Science+Business Media LLC 2011
Introduction
Among the investigations performed daily in the intensive care unit (ICU), bed-
side chest x-rays (CXRs) are trivialized. However, they are a source of discomfort
and irradiation for the patients, and carry a potential risk of accidental removal
of devices (catheters, tubes) and microbial dissemination, all resulting in addi-
tional cost for the community. It is, therefore, essential to assess current practices
in order to establish recommendations for prescription of CXRs and to determine
whether it is possible to reduce the number of CXRs performed during an ICU
stay without impairing quality of care. We will discuss alternatives to CXRs in
specific situations (such as placement of feeding tubes and central venous lines).
We will also consider bedside CXR prescription strategies, without considering
medico-economic aspects such as the possible savings resulting from a reduction
in the number of CXRs performed in the ICU. Indeed, the real cost of performing
a bedside CXR is unknown because it integrates multiple parameters, such as cost
of consumables, depreciation of equipment, working time of the x-ray technician
and logistical costs.
Heterogeneity of Prescription Practices
In a preliminary, retrospective study (unpublished data), we compared the num-
ber of CXRs ordered in 19 medical ICUs of the Assistance Publique–Hôpitaux de
Paris (APHP) group of hospitals during the year 2004. Two databases were used:
RADOS, which collects the records of all medical imaging investigations (includ-
ing bedside CXRs) performed in each hospital, and the medico-administrative
database (Disease Related Groups database, DRG), which describes hospital stays
in terms of length of stay, dates of admission and discharge, and diagnosis using
the ICD-10. The RADOS and DRG databases were matched and each bedside CXR
performed in the ICU was retrieved.
Differences among the ICUs were measured through direct standardization.
The ratio comparing the number of bedside CXRs per day and per ICU stay in a
given ICU to the mean number of bedside CXRs in all the 19 ICUs was called the
“Radiology Performance Index” (RPI). We adjusted this RPI for several variables
including case-mix, mechanical ventilation and ICU length of stay. The main
characteristics of the 8975 analyzed ICU stays were: Mean length of stay 6.6 days
(95 % confidence interval [CI] 3.9–12.8); percentage of mechanically ventilated
patients 51.4 % (95 % CI 30.9–66.9) of the ICU stays; mean number of bedside
753
XVII
CXRs performed per day 1.1 (95 % CI 0.4–2). The distribution of RPI showed
great variability in the number of CXR prescriptions. There was still considerable
heterogeneity after adjustment for case-mix (0.31 to 1.54), mechanical ventilation
(0.31 to 1.53) and length of stay (0.41 to 1.58). One explanation for these varia-
tions could have been differences in indications according to medical teams and
patient diagnoses. To explore this possibility, we designed a study to interview
senior ICU physicians on their opinion about whether a systematic CXR should
be ordered in a set of usual ICU clinical situations.
Searching for a Consensus on Indications for CXRs in ICUs
Indications for ordering CXRs in ICUs have been poorly studied in a systematic
way. Research has mainly focused on prescribing strategies (routine versus on-
demand) [1–8], rather than the precise clinical context, excluding invasive proce-
dures [9–11] which are easier to study.
Taking into account the heterogeneity of the constraints of each hospital in
terms of logistics (human resources, physical structure of the hospital, technology
of equipments) and practices (prescription habits), we asked ICU physicians of
the Cub-Réa network (34 medical, surgical or mixed ICUs in teaching and non-
teaching hospitals in the Paris area) to participate in a search for a consensus
about indications for CXR prescription using the Delphi method [12]. We used a
structured questionnaire to which 95 ICU physicians (rate of answer = 50 %)
agreed to answer (82 completed the full study). Their opinion on the systematic
prescription of CXRs in various clinical scenarios was collected during iterative
sequences of interrogation using a dedicated Web application. The questionnaire
included 29 questions relative to the placement of medical devices and their sur-
veillance, as well as various clinical situations (Table 1). During each round of
interrogation, the respondent was asked to answer each item using a 9-point
scale, in which 1 point indicated that routine CXR was considered to be always
unnecessary and 9 points indicated that routine CXR was considered to be always
necessary. Each doctor received detailed feedback of the answers (of all the par-
ticipants) from the previous phase and could change his/her opinion in later
rounds. The process of interrogation was interrupted when there was a stability
in the answers. The distributions of answers observed for each question were
analyzed in terms of convergence of opinion and strength of the consensus
obtained.
We observed a strong consensus – that is a low variability in the answers – for
10 questions which represented widely accepted reasonable attitudes (Table 2).
However, concerning whether or not to perform daily routine CXRs in patients
on mechanical ventilation, a consensus was not reached. In certain very specific
clinical situations (acute asthma, acute respiratory distress syndrome [ARDS]),
systematic daily prescription was widely recommended, while in others (neuro-
logical or toxic coma with healthy lungs, for example) the answers of the doctors
did not converge. These results demonstrate the importance of the clinical con-
text in the decision of prescription and the difficulty in making general recom-
mendations that do not take into account the heterogeneity of the clinical scenar-
ios.
754 V. Ioos, A. Galbois, and B. Guidet
XVII
Table 1. Item wording of the Delphi survey [12]
item Question
01* After an intubation?
02* After tracheotomy?
03* After placement of a sub-clavian central venous catheter?
04* After placement of an internal jugular central venous catheter?
05* After placement of a pulmonary artery catheter?
06* After placement of temporary transvenous pacing leads?
07* After placement of a nasogastric tube for enteral nutrition?
08* After placement of a nasogastric tube (for another reason than enteral nutrition), through
which gastric secretions could be aspirated?
09* After placement of a ballasted nasogastric tube (risk of pneumothorax)?
10* After placement of a chest tube?
11# Presence of an endotracheal tube?
12# Non-invasive ventilation with PaO2/FiO2 < 220 mmHg?
13# Non-invasive ventilation with PaO2/FiO2 > 220 mmHg?
14# Presence of a nasogastric tube?
15# Pulmonary artery catheter?
16# Presence of a catheter in the superior vena cava system?
17# Presence of a chest tube?
18# Presence of temporary transvenous pacing leads?
19# Invasive mechanical ventilation for ARDS?
20# Invasive mechanical ventilation for hemodynamic pulmonary edema (cardiogenic, kidney
failure)?
21# Non-invasive ventilation for hemodynamic pulmonary edema (cardiogenic, kidney failure)?
22# Hemodynamic instability in a patient under invasive mechanical ventilation?
23# Invasive mechanical ventilation for status asthmaticus?
24# Invasive mechanical ventilation for acute respiratory failure in a severely immunocompro-
mised patient?
25# Non-invasive ventilation for acute respiratory failure in a severely immunocompromised
patient?
26# Invasive mechanical ventilation for acute-on-chronic respiratory failure?
27# Non-invasive ventilation for acute-on-chronic respiratory?
28# Invasive mechanical ventilation for neurologic or toxic coma with normal respiratory function?
29 Is it justified tosystematically realize a CXR before extubation?
* item began with “Should a routine CXR be obtained (within 1h) .....”
# item began with “Should a routine CXR be obtained daily for ....”
Reduction in the Number of CXRs Ordered in Patients on Mechanical
Ventilation
The American College of Radiology recommends routine daily CXRs for mechan-
ically ventilated patients, and use of additional CXRs if necessary [13]. This strat-
egy is, however, controversial [1, 6, 7, 12, 14, 15]: Some clinicians support it,
whereas others advocate on-demand prescription of CXRs only when warranted
by the patient’s clinical status. Our Delphi study revealed that the recommenda-
tions for ordering CXRs in the ICU were not followed and that indications for
prescription varied from one department to another, in particular for ICU
patients on mechanical ventilation.
Routine CXR has two main advantages. First, some potentially life-threatening
situations that might otherwise be missed can be discovered and treated. Second,
Should We Still Order Chest X-rays in the ICU? 755
XVII
Table 2. Description of the 80 % central answers (n = 66) of the Delphi survey (the 10 % of the
answers located at each end of the range were eliminated) for the situations listed in Table 1 [12]
Question Summary of answers Indication for system-
atic prescription of
radiography
Strength of the
consensusmean median min max
1 8.64 9 7 9 YES HIGH
2 7.64 8 5 9 YES MODERATE
3 8.97 9 8 9 YES HIGH
4 8.67 9 8 9 YES HIGH
5 8.91 9 8 9 YES HIGH
6 8.65 9 7 9 YES HIGH
7 6.65 6 3 9 UNCERTAINTY LOW
8 3.02 3 1 5 NO MODERATE
9 8.05 8 5 9 YES MODERATE
10 8.95 9 8 9 YES HIGH
11 5.33 5 2 8 UNCERTAINTY LOW
12 7.26 7 5 9 YES MODERATE
13 4.82 5 3 7 UNCERTAINTY MODERATE
14 1.79 2 1 3 NO HIGH
15 6.36 6 3 9 UNCERTAINTY LOW
16 1.53 2 1 2 NO HIGH
17 6.65 7 4 9 YES MODERATE
18 4.79 5 2 7 YES MODERATE
19 8.73 9 8 9 YES HIGH
20 7.38 7 5 9 YES MODERATE
21 6.30 6 4 8 YES MODERATE
22 8.14 8 6 9 YES MODERATE
23 8.83 9 8 9 YES HIGH
24 8.42 9 6 9 YES MODERATE
25 7.74 8 6 9 YES MODERATE
26 6.09 6 3 9 YES LOW
27 4.80 5 2 7 UNCERTAINTY MODERATE
28 4.77 5 2 9 UNCERTAINTY LOW
29 2.67 3 1 6 NO MODERATE
scheduling CXRs during morning rounds might be more efficient from a logisti-
cal point of view. In contrast, the on-demand strategy might avoid unnecessary
radiation exposure and provides substantial cost savings [16] but an increased
number of CXRs may be needed during the rest of the day to compensate for
those not done in the morning.
In a randomized study, patients managed with an on-demand strategy had
fewer CXRs during the ICU stay (4.4 vs 6.8) [6]. Compared with routine imaging,
on-demand CXRs had more images with new findings (53 % vs 33 %) and with
new findings requiring intervention (27 % vs 13 %). Another prospective con-
trolled study showed that abandoning the routine strategy decreased the volume
of CXRs ordered by 35 % [3]. Routine CXRs identified new or progressive major
findings in only 4.4 % and changed management in only 1.9 %. In a single center
randomized study, delayed diagnoses occurred in only 0.7 % of the on-demand
group and all findings were minor [1]. In another study, after an on-demand
strategy was implemented, the number of CXRs per patient-day decreased from
1.1 to 0.6 [4]. None of these studies identified complications associated with the
756 V. Ioos, A. Galbois, and B. Guidet
XVII
on-demand approach, such as increased duration of mechanical ventilation,
length of stay, or mortality [1, 3, 4, 6]. Finally, a meta-analysis selected eight stud-
ies comparing on demand and daily routine strategies, including a total of 7078
patients [17]. In a pooled analysis, no difference in ICU mortality, ICU length of
stay or duration of mechanical ventilation was found between the on-demand and
daily routine groups. However, this meta-analysis included only two small size
randomized controlled trials, both single center studies. No study has, therefore,
had sufficient power to totally convince ICU physicians to abandon daily routine
CXRs.
In the RARE study [18], we evaluated two strategies for prescribing morning
CXRs in ICU patients on mechanical ventilation using a cluster-randomized,
open-label crossover design. In the ‘routine strategy’, CXRs were performed daily
in patients on mechanical ventilation, irrespective of their clinical status. In the
‘on-demand strategy’, CXRs were performed in the morning if warranted by the
clinical examination and the analysis of biological parameters. We randomly
assigned 21 ICUs (medical, surgical or medico-surgical) in 18 hospitals (teaching
and non-teaching) to use the ‘routine’ or ‘on-demand’ strategy during the first of
two treatment periods. ICUs used the alternative strategy in the second period.
The primary outcome measure was the mean number of CXRs per patient-day of
mechanical ventilation. Secondary outcome measures where related to the quality
and safety of care (days of mechanical ventilation, ICU length of stay and ICU
mortality, number of unscheduled CXRs performed) as well as diagnostic and
therapeutic value of the CXRs performed within each strategy. Finally, distur-
bances in the organization of the ICU and medical imaging departments were
also analyzed.
During the study period, 424 patients had 4607 routine CXRs (mean per
patient-day of mechanical ventilation 1.09, 95 % CI 1.05–1.14), and 425 had 3148
on-demand chest radiographs (mean 0.75, 95 % CI 0.67–0.83), which corre-
sponded to a reduction of 32 % (95 % CI 25–38) with the on-demand strategy
(p < 0.0001). Duration of mechanical ventilation as well as ICU length of stay and
28-day mortality did not differ between the groups (Table 3). The difference in the
total number of routine and on-demand chest radiographs was not significant
Table 3. Numbers of CXRs, length of stay, duration of mechanical ventilation, and mortality rates in the
RARE study [18]
Strategy of CXR prescription
Routine
(424 patients)
On-demand
(425 patients)
CXRs performed
Total number
With new finding
With new finding leading to therapeutic intervention
Mean number per patient and per day
4607
688
728
1.11
3148
618
729
0.77
Length of mechanical ventilation (days)
Total number of days
LOS: mean, median [5th and 95th percentile]
4172
9.82, 7 [2;30]
4226
9.94, 7 [2;30]
Mortality in ICU (dead/alive, %) 131/293, 30.9 % 136/289, 32.0 %
Should We Still Order Chest X-rays in the ICU? 757
XVII
Fig. 1. Distribution of the CXRs
performed during the day in the
RARE study [18]. The morning
reduction in the number of CXRs
was not compensated for by a
concomitant increase in CXRs
ordered during the rest of the
day.
when the analysis was restricted to CXRs with new findings that led or contrib-
uted to diagnostic procedures or therapeutic interventions. Finally, there was no
increase in the number of unscheduled CXRs performed in the afternoon and in
the night in the on-demand strategy, and, therefore, no disruption in the organi-
zation of the medical imaging department (Fig. 1).
This study strongly suggests that routine daily CXRs in the ICU patient on
mechanical ventilation should be abandoned [19]. This restrictive strategy is in
line with previous studies that had some methodological flaws [17]. The main
limit to its broad application lies in the fact that French ICUs are closed units and
the results may not be applicable to open ICUs, an organizational model which is
found in other countries [20].
Alternatives to CXR
Alternatives to CXR to Ensure Correct Placement of Enteral Feeding Tube
Accidental placement of the enteral feeding tube in the tracheobronchial tract can
lead to potentially lethal complications. Tracheal intubation does not always pre-
vent this misplacement [21]. Ensuring correct enteral feeding tube position is
therefore of paramount importance for patients in ICU. Confirmation of the ade-
quate placement of an enteral feeding tube by epigastric auscultation after air
injection through the feeding tube is not a reliable test when used alone [22–24].Some studies have suggested testing the pH of an aspirate obtained from the
enteral feeding tube to ensure proper placement, but this test can be inconclusive
in patients with small bore feeding tubes or in those receiving acid suppression
therapies [22]. Therefore, most of the guidelines recommend confirmation of
enteral feeding tube placement by CXR before starting enteral nutrition [24, 25].
We studied two interesting alternatives to CXR.
Bedside ultrasonography is a non-invasive procedure that is increasingly used
in the ICU by non-radiologist physicians who can obtain reliable results after a
short period of training [26, 27]. Using a probe of 2 to 5 MHz, we showed that
758 V. Ioos, A. Galbois, and B. Guidet
XVII
Fig. 2. Assessment of intraga-
stric position of a small bore
enteral feeding tube by ultraso-
nography. The probe is placed in
the middle epigastric area and
oriented towards the left upper
abdominal quadrant to visualize
the gastric area. The small bore
feeding tube appears as two
parallel hyperechogenic lines.
ultrasonography enables small-bore enteral feeding tubes to be visualized in the
digestive tract with a sensitivity of 97 % and to assess whether it is properly
placed in the stomach within 5 minutes (Fig. 2) [28]. If the enteral feeding tube is
not immediately visible by ultrasound, injection of 5 ml normal saline mixed with
5 ml air into the tube increases the sensitivity. This technique is more rapid than
conventional radiography, is radiation free, and can be taught to ICU physicians
during a short training period [28]. Radiography could then be performed only
in the rare cases of ultrasonography failure, due to gas interposition, for example.
Capnography is often used to assess expiratory CO2. However, it is possible to
connect the capnography device to the enteral feeding tube via the tip of an
endotracheal tube and to assess the correct placement of the enteral feeding tube
by the absence of CO2. The feeding tube must be inserted to a depth of 30 cm
from the nostril and should not get coiled in the pharynx. When the enteral feed-
ing tube is accidentally inserted in the respiratory tract, the capnograph displays
a normal capnogram. When the enteral feeding tube is inserted in the esophagus,
the capnograph displays a “purging warning” [29]. It has to be noted that enteral
feeding tube permeability is essential for CO2 detection. In our ICU, we ensure
this permeability by removing the guidewire, insufflating then exsufflating with
air using a 50 ml syringe, before connecting the capnography device. We use a
colorimetric capnography device after 30 cm insertion and then we complete the
insertion until 50 cm from the nostril. Finally, in order to check that the enteral
feeding tube is not coiled in the esophagus after its complete insertion, nurses
perform epigastric auscultation. Radiography is required only in cases of incon-
clusive epigastric auscultation (10.1 % of cases). We showed that this local proto-
col combining colorimetric capnography and epigastric auscultation had perfect
specificity for confirming enteral feeding tube placement in stomach, improved
nurse organization of care, saved time, and decreased costs [30, 31]. Another
advantage of this procedure is that accidental tracheobronchial insertion is
detected after 30 cm insertion, which prevents any risk of pneumothorax or
hydrothorax, which are rare but potentially fatal complications of feeding tube
misplacement that are detected but not prevented by post-procedural radiogra-
phy.
Should We Still Order Chest X-rays in the ICU? 759
XVII
Alternative to CXR to Diagnose and Monitor Pneumothorax
Chest computed tomography (CT-scan) is the gold standard for diagnosing pneu-
mothorax. Conversely, CXRs miss 30 % to 72 % of pneumothoraces due to their
anterior location [32], of which half can be tension pneumothoraces. Pleural
ultrasonography has greater sensitivity than CXR for pneumothorax diagnosis in
patients in ICUs [33] or trauma centers [32, 34] and after pleural biopsy [35, 36].
Lichtenstein et al. reported that ultrasonography detected all pneumothoraces in
ICU patients, including those not identified by CXR [33]. Ultrasound diagnosis of
pneumothorax relies on three signs: Abolition of lung sliding, the A-line sign, and
the lung point. The sensitivity and specificity of the abolition of lung sliding for
the diagnosis of pneumothorax are 100 % and 91 % respectively (Figs. 3–5) [37].
The presence of horizontal linear artefacts at regular intervals below the pleural
line (A-lines) is part of the ultrasound semiology of normal lungs and pneumo-
thoraces (Fig. 3). In contrast, vertical linear artefacts arising from the pleural line,
i.e., B-lines or comet-tail artefacts, are observed when alveolar-interstitial syn-
drome is present, as well as in the last two intercostal spaces in 27 % of healthy
subjects (Fig. 6) [38]. The A-line sign is defined as the presence of A-lines without
B-lines (Fig. 3) and has a sensitivity of 100 % and a specificity of 60 % for the
diagnosis of pneumothorax. The presence of B-lines rules out a diagnosis of
pneumothorax [39]. The lung point is detected while the probe is stationary:
There is lung sliding during inspiration (when the lung contacts the wall), which
disappears during expiration (when the lung is not in contact with the wall). The
sensitivity of this sign for diagnosis of pneumothorax is 66 % while its specificity
is 100 % [40]. After drainage, ultrasonography is better than CXR for detecting
residual pneumothoraces, with 39 % not identified by CXR [27]. After drainage of
primary spontaneous pneumothoraces, performance of ultrasonography is excel-
lent [27]. After drainage of non-primary spontaneous pneumothorax, the predic-
tive positive value of ultrasonography was 100 % if a lung point was present.
However, it decreased to 90 % in the absence of a lung point. Many factors have
been reported to be responsible for lung sliding abolition, including atelectasis,
ARDS, selective intubation, and pleural symphysis, among others [37, 38, 41].
Fig. 3. Pleural ultrasonography
in two-dimensional mode. The
pleural line is seen between two
ribs. Lung sliding is abolished
when both the parietal and vis-
ceral pleura do not slide while
the patient is breathing. The
A-line sign is the presence of
linear horizontal artefacts at reg-
ular intervals below the pleural
line (A-lines) without B-lines.
The A-line sign is part of the
ultrasound semiology of the
normal lung and of pneumotho-
rax. From [27] with permission
760 V. Ioos, A. Galbois, and B. Guidet
XVII
Fig. 4. Assessment of lung slid-
ing on pleural ultrasonography
in time-motion mode in a
patient without pneumothorax.
Lung sliding generates a granu-
lar pattern under the pleural
line. Subcutaneous tissue over
the pleural line does not move
while the patient is breathing,
generating horizontal lines. From
[27] with permission
Fig. 5. Detection of abolition of
lung sliding on pleural ultraso-
nography in time-motion mode
in a patient with pneumothorax.
While the patient is breathing,
the (normal) granular pattern
under the pleural line is
replaced by horizontal lines,
indicating abolition of lung slid-
ing. From [27] with permission
Fig. 6. Detection of B-lines on
pleural ultrasonography in two-
dimensional mode. The presence
of vertical linear artefacts arising
from the pleural line (B-lines or
comet-tail artefacts) rules out
pneumothorax in this patient
with interstitial syndrome. From
[27] with permission
Should We Still Order Chest X-rays in the ICU? 761
XVII
Exclusive use of ultrasonography for follow-up of non-primary spontaneous
pneumothorax seems possible, but the physician must be aware that in the
absence of a lung point, a diagnosis of pneumothorax should not be made if other
causes of lung sliding abolition have not been ruled out. We recommend per-
forming a CT-scan if doubt persists, especially if new chest tube insertion is being
considered.
These excellent performances make pleuralultrasonography more than an
alternative to CXR and it should be considered as the ‘bedside gold-standard’ to
diagnose and monitor pneumothorax. Moreover, ultrasonography gives faster
results than CXR and can be performed competently by naı̈ve physicians after a
brief training session [27, 42].
Alternative to CXR after Central Venous Catheter (CVC) Insertion
Radiography is usually performed after CVC insertion in the superior vena caval
system to assess the correct placement of the CVC tip and to ensure the absence
of pneumothorax. CVC tip misplacement occurs in 5–6 % of cases after catheteri-
zation of the subclavian or internal jugular vein, and pneumothorax occurs in
1.5–3.1 % and 0.1–0.2 %, respectively [43].
Clinical evaluation of the patient to predict the absence of complications after
CVC insertion via the subclavian vein or internal jugular vein was very accurate
in a study by Gray et al. [44]. However, Gladwin et al. showed that the clinical
impression of the operator (based on the number of needle passes, difficulty
establishing access, operator experience, poor anatomical landmarks, number of
previous catheter placements, resistance to wire or catheter advancement, resis-
tance to aspiration of blood or flushing of the catheter ports, sensations in the
ear, chest, or arm, and development of signs or symptoms suggestive of pneumo-
thorax) had poor sensitivity and specificity for predicting a complication (44 %
and 55 %, respectively) [45]. Gladwin and colleagues concluded that post-proce-
dural CXR remains necessary because clinical factors alone cannot reliably iden-
tify tip misplacement.
However, as detailed above, numerous pneumothoraces can be missed by bed-
side CXR whereas ultrasonography showed excellent sensitivity and specificity for
the diagnosis of pneumothorax within a few minutes. We compared post-proce-
dural ultrasonography to CXR after insertion of 85 central venous catheters (70
subclavian and 15 internal jugular) [46]. Ten misplacements and one pneumotho-
rax occurred. Ultrasonic examination feasibility was 99.6 %. The only pneumo-
thorax and all misplacements except one were diagnosed by ultrasound. Taking
into consideration signs of misplacement and pneumothorax, ultrasonic exami-
nation did not give any false positive results. Moreover, ultrasound guidance
increased the success rate of CVC insertion, saved time and decreased the compli-
cation rate [47]. Considering these results, it seems logical to use the same ultra-
sonographic device to assess both the adequate position of the CVC and the
absence of pneumothorax after the procedure. The only limit of ultrasonography
in this indication is the lack of visualization of the azygos, internal thoracic and
cardiophrenic veins and an inconstant visualization of the superior vena cava.
Thus, it may be proposed that the absence of misplacement and pneumothorax
could be assessed using ultrasonography and CXR limited to patients in whom
ultrasonographic analysis is incomplete.
762 V. Ioos, A. Galbois, and B. Guidet
XVII
Conclusion
We have convincingly showed that bedside CXR could be avoided in most cir-
cumstances. This is true for mechanically ventilated patients and to insure proper
placement of devices, such as feeding tubes and central venous catheters. This
new restrictive policy for ordering bedside CXRs implies that the patient’s clinical
status is assessed at least once a day before ordering a CXR; CXRs should never
replace clinical evaluation of the patient but should rather be prescribed on the
basis of a clinical suspicion. As a consequence, ICU organization may have to be
modified in order to allow implementation of such a prescribing strategy and a
reduction in the number of CXRs ordered.
Ultrasonography is a very good alternative to CXR. For example, ultrasonogra-
phy is more accurate than CXR for detecting pneumothorax. However, short
training courses must be organized in order to achieve a basic level of compe-
tency for every physician working in the ICU.
A policy of reducing the number of CXRs has many advantages (patient com-
fort, better organization of the radiology department, cost reduction) and should
be widely implemented in the ICU.
References
1. Clec’h C, Simon P, Hamdi A, et al (2008) Are daily routine chest radiographs useful in
critically ill, mechanically ventilated patients? A randomized study. Intensive Care Med.34:
264–270
2. Kappert U, Stehr SN, Matschke K (2008) Effect of eliminating daily routine chest radio-
graphs on on-demand radiograph practice in post-cardiothoracic surgery patients. J Tho-
rac Cardiovasc Surg 135: 467–468
3. Hendrikse KA, Gratama JW, Hove W, Rommes JH, Schultz MJ, Spronk PE (2007) Low
value of routine chest radiographs in a mixed medical-surgical ICU. Chest 132: 823–828
4. Graat ME, Kroner A, Spronk PE, et al (2007) Elimination of daily routine chest ra-
diographs in a mixed medical-surgical intensive care unit. Intensive Care Med 33: 639–
644
5. Bhagwanjee S, Muckart DJ (1996) Routine daily chest radiography is not indicated for
ventilated patients in a surgical ICU. Intensive Care Med 22: 1335–1338
6. Krivopal M, Shlobin OA, Schwartzstein RM (2003) Utility of daily routine portable chest
radiographs in mechanically ventilated patients in the medical ICU. Chest 123: 1607–1614
7. Silverstein DS, Livingston DH, Elcavage J, Kovar L, Kelly KM (1993) The utility of routine
daily chest radiography in the surgical intensive care unit. J Trauma 35: 643–646
8. Hall JB, White SR, Karrison T (1991) Efficacy of daily routine chest radiographs in intu-
bated, mechanically ventilated patients. Crit Care Med 19: 689–693
9. Rao PS, Abid Q, Khan KJ, et al (1997) Evaluation of routine postoperative chest X-rays in
the management of the cardiac surgical patient. Eur J Cardiothorac Surg 12: 724–729
10. Houghton D, Cohn S, Schell V, Cohn K, Varon A (2002) Routine daily chest radiography
in patients with pulmonary artery catheters. Am J Crit Care 11: 261–265
11. Haddad SH, Aldawood AS, Arabi YM (2007) The diagnostic yield and clinical impact of
a chest X-ray after percutaneous dilatational tracheostomy: a prospective cohort study.
Anaesth Intensive Care 35: 393–397
12. Hejblum G, Ioos V, Vibert JF, et al (2008) A web-based Delphi study on the indications of
chest radiographs for patients in ICUs. Chest 133: 1107–1112
13. American College of Radiology (2006) Appropriateness criteria: routine chest radiograph.
Available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/app_crite-
ria/pdf/ExpertPanelonThoracicImaging/RoutineChestRadiographDoc7.aspx, accessed Oct
11, 2010
14. Graat ME, Hendrikse KA, Spronk PE, Korevaar JC, Stoker J, Schultz MJ (2006) Chest radi-
Should We Still Order Chest X-rays in the ICU? 763
XVII
ography practice in critically ill patients: a postal survey in the Netherlands. BMC Med
Imaging 6: 8
15. Graat ME, Stoker J, Vroom MB, Schultz MJ (2005) Can we abandon daily routine chest
radiography in intensive care patients? J Intensive Care Med 20: 238–246
16. Price MB, Grant MJ, Welkie K (1999) Financial impact of elimination of routine chest
radiographs in a pediatric intensive care unit. Crit Care Med 27: 1588–1593
17. Oba Y, Zaza T (2010) Abandoning daily routine chest radiography in the intensive care
unit: meta-analysis. Radiology 255: 386–395
18. Hejblum G, Chalumeau-Lemoine L, Ioos V, et al (2009) Comparison of routine and on-
demand prescription of chest radiographs in mechanically ventilated adults: a multicen-
tre, cluster-randomised, two-period crossover study. Lancet 374: 1687–1693
19. Hejblum G, Guidet B (2010) Evidence-based data for abandoning unselective daily chest
radiographs in Intensive Care Units. Radiology 256: 1013–1014
20. Siegel MD, Rubinowitz AN (2009) Routine daily vs on-demand chest radiographs in inten-
sive care. Lancet 374: 1656–1658
21. Woodall BH, Winfield DF, Bisset GS 3rd (1987) Inadvertent tracheobronchial placement of
feeding tubes. Radiology 165: 727–729
22. Kunis K (2007)Confirmation of nasogastric tube placement. Am J Crit Care 16: 19
23. Stroud M, Duncan H, Nightingale J (2003) Guidelines for enteral feeding in adult hospital
patients. Gut 52 Suppl 7: vii1-vii12
24. Metheny NA (2006) Preventing respiratory complications of tube feedings: evidence-based
practice. Am J Crit Care 15: 360–369
25. Jolliet P, Pichard C, Biolo G, et al (1998) Enteral nutrition in intensive care patients: a
practical approach. Working Group on Nutrition and Metabolism, ESICM. European Soci-
ety of Intensive Care Medicine. Intensive Care Med 24: 848–859
26. Chalumeau-Lemoine L, Baudel JL, Das V, et al (2009) Results of short-term training of
naive physicians in focused general ultrasonography in an intensive-care unit. Intensive
Care Med 35: 1767–1771
27. Galbois A, Ait-Oufella H, Baudel JL, et al (2010) Pleural ultrasound compared with chest
radiographic detection of pneumothorax resolution after drainage. Chest 138: 648–655
28. Vigneau C, Baudel JL, Guidet B, Offenstadt G, Maury E (2005) Sonography as an alterna-
tive to radiography for nasogastric feeding tube location. Intensive Care Med 31:
1570–1572
29. Burns SM, Carpenter R, Truwit JD (2001) Report on the development of a procedure to
prevent placement of feeding tubes into the lungs using end-tidal CO2 measurements.
Crit Care Med 29: 936–939
30. Meyer P, Henry M, Maury E, Baudel JL, Guidet B, Offenstadt G (2009) Colorimetric cap-
nography to ensure correct nasogastric tube position. J Crit Care 24: 231–235
31. Galbois A, Vitry P, Ait-Oufella H, et al (2011) Colorimetric capnography, a new procedure
to ensure correct feeding tube placement in the intensive care unit: An evaluation of a
local protocol. J Crit Care (in press)
32. Soldati G, Testa A, Pignataro G, et al (2006) The ultrasonographic deep sulcus sign in
traumatic pneumothorax. Ultrasound Med Biol 32: 1157–1163
33. Lichtenstein DA, Meziere G, Lascols N, et al (2005) Ultrasound diagnosis of occult pneu-
mothorax. Crit Care Med 33: 1231–1238
34. Soldati G, Testa A, Sher S, Pignataro G, La Sala M, Silveri NG (2008) Occult traumatic
pneumothorax: diagnostic accuracy of lung ultrasonography in the emergency depart-
ment. Chest 133: 204–211
35. Chung MJ, Goo JM, Im JG, Cho JM, Cho SB, Kim SJ (2005) Value of high-resolution ultra-
sound in detecting a pneumothorax. Eur Radiol 15: 930–935
36. Sartori S, Tombesi P, Trevisani L, Nielsen I, Tassinari D, Abbasciano V (2007) Accuracy of
transthoracic sonography in detection of pneumothorax after sonographically guided
lung biopsy: prospective comparison with chest radiography. AJR Am J Roentgenol 188:
37–41
37. Lichtenstein DA, Menu Y (1995) A bedside ultrasound sign ruling out pneumothorax in
the critically ill. Lung sliding. Chest 108: 1345–1348
38. Lichtenstein D, Meziere G, Biderman P, Gepner A, Barre O (1997) The comet-tail artifact.
764 V. Ioos, A. Galbois, and B. Guidet
XVII
An ultrasound sign of alveolar-interstitial syndrome. Am J Respir Crit Care Med 156:
1640–1646
39. Lichtenstein D, Meziere G, Biderman P, Gepner A (1999) The comet-tail artifact: an ultra-
sound sign ruling out pneumothorax. Intensive Care Med 25: 383–388
40. Lichtenstein D, Meziere G, Biderman P, Gepner A (2000) The ”lung point”: an ultrasound
sign specific to pneumothorax. Intensive Care Med 26: 1434–1440
41. Blaivas M, Tsung JW (2008) Point-of-care sonographic detection of left endobronchial
main stem intubation and obstruction versus endotracheal intubation. J Ultrasound Med
27: 785–789
42. Noble VE, Lamhaut L, Capp R, et al (2009) Evaluation of a thoracic ultrasound training
module for the detection of pneumothorax and pulmonary edema by prehospital physi-
cian care providers. BMC Med Educ 9: 3
43. McGee WT, Ackerman BL, Rouben LR, Prasad VM, Bandi V, Mallory DL (1993) Accurate
placement of central venous catheters: a prospective, randomized, multicenter trial. Crit
Care Med 21: 1118–1123
44. Gray P, Sullivan G, Ostryzniuk P, McEwen TA, Rigby M, Roberts DE (1992) Value of post-
procedural chest radiographs in the adult intensive care unit. Crit Care Med 20:
1513–1518
45. Gladwin MT, Slonim A, Landucci DL, Gutierrez DC, Cunnion RE (1999) Cannulation of
the internal jugular vein: is postprocedural chest radiography always necessary? Crit Care
Med 27: 1819–1823
46. Maury E, Guglielminotti J, Alzieu M, Guidet B, Offenstadt G (2001) Ultrasonic examina-
tion: an alternative to chest radiography after central venous catheter insertion? Am J
Respir Crit Care Med 164: 403–405
47. Hind D, Calvert N, McWilliams R, et al (2003) Ultrasonic locating devices for central
venous cannulation: meta-analysis. BMJ 327: 361
Should We Still Order Chest X-rays in the ICU? 765
XVII
	Should We Still Order Chest X-rays in the ICU?
	Introduction
	Heterogeneity of Prescription Practices
	Searching for a Consensus on Indications for CXRs in ICUs
	Reduction in the Number of CXRs Ordered in Patients on Mechanical Ventilation
	Alternatives to CXR
	Alternatives to CXR to Ensure Correct Placement of Enteral Feeding Tube
	Alternative to CXR to Diagnose and Monitor Pneumothorax
	Alternative to CXR after Central Venous Catheter (CVC) Insertion
	Conclusion
	References