Downloaded from www.ajronline.org by 117.253.221.134 on 11/10/15 from IP address 117.253.221.134. Copyright ARRS. For personal use only; all rights reserved
917
Factors Affecting the Development of Pneumothorax Associated with Thoracentesis
Vassilios Lawrence
This
2
M. Davis1’3 Grace Lee1’4 Cynthia UmaIi1 Robert Lew5 Richard S. Irwin6
ment
study
is a retrospective
of pneumothorax
survey
after
of the
thoracentesis.
variables
results
show
that
156:917-920,
AJR
August
27, 1990:
accepted
after revi-
16, 1991.
Department of Radiology, University of Massachusetts Medical Center, 55 Lake Ave. N., Worcester, MA 01655. 2 Present address:
Department
of
Radiology,
Beth Israel Hospital and Harvard Medical 330 Brookline Ave., Boston, MA 02215. reprint requests to V. Raptopoulos. Present address: Department Worcester Memorial Hospital, 119 Worcester, MA 01605. Johns more,
School, Address
Department of Pharmacology, University of Massachusetts Medical School, Worcester, MA 01655. 6 Division of Pulmonary Medicine and Critical Worcester,
MA 01655.
0361 -803X/91/i 565-0917 © American Roentgen Ray Society
the
a computer
developsearch
of
sonography-guided
thoracentesis
is complicated
by pneumo-
done with conventional techniques. of the smallest possible amount of
May 1991
In a recent randomized, prospective study from our institution, Grogan et al. [1] showed that, compared with other methods, sonography-guided thoracenteses were associated with significantly fewer complications, including pneumothorax. That study took place under specific circumstances, limited to presumed low-risk patients who were cooperative, spontaneously breathing, and had substantial amounts of free-flowing pleural fluid. The purpose of this study was to see if Grogan’s,findings
favor
apply
the development
to a broader
range
of pneumothorax
of patients
in patients
and
having
to see
if other
factors
thoracentesis.
of Medicine, Belmont St.,
Present address: Department of Radiology, Hopkins Hospital, 600 N. Wolfe St., BaltiMD 21205-2191.
Care, University of Massachusetts
influence
Other factors surveyed included the patients’ age, sex, underlying pulmonary disease, and overall clinical condition; the size of the effusion; the type of tap (diagnostic or therapeutic); the amount and type (exudate or transudate) of fluid acquired; and the size of the needles used. The technique used was the most significant single risk factor affecting the development of pneumothorax (18% for clinical vs 3% for sonographyguided thoracenteses). The incidence of pneumothorax decreased when a smaller amount of pleural fluid was aspirated (mean, 246 ml aspirated from patients who did not vs 472 ml from those who did develop pneumothorax) and when thin needles were used (4% pneumothorax with 20-gauge or smaller and 18% with larger than 20-gauge needles). The other factors surveyed did not influence the development of pneumothorax. Our
Received
may period,
hospital records identified 342 thoracenteses, of which 154 were done with conventional techniques by the clinical services, and 188 were done with sonographic guidance.
thorax significantly less often than is thoracentesis Use of the smallest possible needle and aspiration fluid will also result in fewer cases of pneumothorax.
sion November
that
In a 30-month
Medical Center,
Materials In this procedures
search 188
and study
Methods we included
by radiologists.
during,
and
after
The variables clinical condition; amount
all thoracenteses
recorded
in the computerized
and the records of the sonography section identified 342 pleural taps, of which 1 54 were
and
type
The patients’
The the
hospital
procedures
surveyed method (exudate
records were
and
the
listings
of hospital
of the Department of Radiology. performed by clinical physicians
radiographs
and
sonograms
done
The and before,
reviewed.
included patients’ age, sex, underlying pulmonary disease, and used; size of the effusion; type of tap (diagnostic or therapeutic); or transudate)
ages ranged
from
of fluid
18 to 94 years
acquired; (mean,
and needle size used. 62 years); 187 were men and
155
RAPTOPOULOS
918
were women.
They had no underlying disease (n = 103), or they had disease (n = 32), focal pulmonary disease such as pneumonia (n = 1 36), or pulmonary manifestations of cardiovascular disease such as pulmonary edema (n = 50). No data were available diffuse
Downloaded from www.ajronline.org by 117.253.221.134 on 11/10/15 from IP address 117.253.221.134. Copyright ARRS. For personal use only; all rights reserved
79)
=
The
or not
patients’
critical
were in an intensive One
hundred
fifty
by clinical
physicians.
clinically,
whereas
and
the
four
in the
groups, of the
condition
was
considered
critical
=
conventional
thoracenteses
were
performed
In 1 06 cases,
the pleural effusion was localized the effusion was localized by sonography
in 48,
performed
thoracenteses
guidance
clinical
242) depending on whether or not they care unit. No data were available on 21 patients. (n
thoracentesis
guided
were
the needle
later
done
department
by
or
at
was inserted
by
clinicians.
radiologists
the
Sonography-
with
sonographic
188). In both in training in over 90%
bedside
(n
by physicians
=
cases.
The amount who
of pleural
reviewed
the
fluid was estimated
patients’
chest
by three
radiographs
investigators
without
knowledge
of
the tapping method used or the outcome sensus based on the following guidelines
of the procedure. A conwas recorded: Grade 1 =
blunting
frontal
of the
costophrenic
fluid on the lateral midthorax; lateral
on
upright
decubitus
film
greater
than
on the
lateral
disagreed
1/3
angle
on the
view,
or layering
of
decubitus film of less than 5 mm, measured at the 2 = obliteration of less than 1/3 of the ipsilateral
Grade
diaphragm
chest of
of the
decubitus
or when
to
or
1 cm;
diaphragm, film.
the
radiograph,
5 mm
layering
Grade
or greater
Half
particular
of fluid
3
=
on
the
obliteration
of
than
1 cm
of layering
when
observers
grades
were
given
amount
was
judged
to fall between
grades. For statistical purposes, the effusions also were divided into Small = grades 0-1 .5 (n = 21 6); and Large = grades 2-3 (n = 126). The amount of fluid aspirated in milliliters was recorded in 307 cases. No data were available on 35 taps. Analysis of the pleural fluid revealed a transudate in 21 1 , an exudate in 23, and blood in 35 thoracenteses. No information was available about 73 pleural taps. The thoracenteses were characterized as diagnostic (n = 234) when stated as such and 1 00 ml or less of fluid was aspirated and as therapeutic
(n
as possible
for symptomatic
was
106)
=
aspirated.
The
when
the
intent
was
to aspirate
relief or when
needles
used
were
as much
fluid
more than 100 ml of fluid
classified
as small
(equal
to
or smaller
than 20-gauge, used in 1 94 taps) and large (larger than 20-gauge, used in 1 34 taps). The size of pneumothorax was estimated according to the method of Rhea
et al. [2].
The
pneumothoraces
were
then
graded
as follows:
grade
1 = small pneumothorax, estimated to be smaller than (n = 1 8); grade 2 = medium pneumothorax estimated between and 30% (n = 8); grade 3 = large pneumothorax, estimated greater than 30% (n 6). Grade 0 meant uncomplicated, that pneumothorax (n = 31 0). Critical pneumothorax was defined as a chest
tube
radiographs
was
inserted
for
treatment
(n
=
1 4).
The
number
10%
10% to be is, no when of
taken until resolution
or, if first, until discharge from the 115). Data were analyzed with a commercially available software program: CSS (Complete Statistical System, StatSoft, Inc., Tulsa, OK). For multiway frequency tables, we used the Spearman R t-test to assess the relationship between variables with a level of significance set at p .05. For confirmation, the standard Pearson x2 was used, and for small sample size, the Yates correction or the V2 test was used [3]. Because the discriminations obtained from both tests were hospital
similar,
AJR:156,
May 1991
Results
pulmonary
in 21 patients. (n
ET AL.
was
only
recorded
the
(n
Spearman
=
R t-test
results
are
reported.
We
used
Student’s t test and analysis of variance (ANOVA) to compare the averages of continuous variables among categories. We also used multiple linear regression analyses with dummy variates for categories of nominal variables to find regression models. In parallel, we used stepwise multiple logistic regression analysis to evaluate candidate predictors of the binary outcome, pneumothorax [4].
During
a 30-month
period,
342
formed: 1 54 with the clinical guided method. Pneumothorax in the
clinical
and
five
(3%)
thoracenteses
were
per-
and 1 88 with the sonographyoccurred in 32 (9%): 27(18%) in the
sonography-guided
group.
This difference was statistically significant (p < .0001) and, comparing the two methods, a sevenfold increased risk was estimated for the clinical taps. The occurrence of critical pneumothoraces (pneumothorax requiring chest tube) was significantly more frequent with the clinical than with the sonography-guided method (7% vs 2%, p < .009). Of the 154 clinical taps, the occurrence of pneumothorax was not different in the clinically localized and performed taps (1 9%), compared with the sonographically localized but clinically performed taps (15%). Age, sex, underlying pulmonary disease, critical clinical condition, and the type of pleural fluid had no significant effect on occurrence of pneumothorax after thoracentesis. The mean age of the patients who did not and of those who did develop pneumothorax was 62 and 66 years, respectively. Pneumothorax occurred in 1 0% of men and in 9% of women, in 8% of patients without and 1 1 % of those with underlying pulmonary disease. In the latter group, the incidence of pneumothorax was not different among the patients with focal (7%) or diffuse (1 2%) lung disease or pulmonary edema (8%). Similarly,
pneumothorax
did not
occur
more
commonly
when
taps were performed in critically sick patients, many under assisted respiration (9% for patients in the intensive care unit vs 1 0% for patients not in intensive care). Furthermore, no difference was seen in occurrence of pneumothorax when the pleural fluid was transudate (9%) or exudate (7%), and when exudate was or was not hemorrhage (6% and 9%, respectively). These five variables (age, sex, pulmonary disease, clinical condition, and type of pleural fluid) had no significant effect on the development of pneumothorax when considered separately in patients tapped with the clinical or the sonography-guided group. However, when the method used for each of these variables was considered, the differences favored sonographic guidance. The occurrence of pneumothorax after thoracentesis was not
affected
by
the
amount
of fluid
present
cavity, but was affected by the amount the size of needle used. In addition, patients grouped under each of these ated
with
significant
differences
of fluid
in the
pleural
aspirated
and
the method used in variables was associ-
in occurrence
of pneumotho-
rax, again favoring sonographic guidance. The estimated amounts of pleural effusion were comparable in patients who did not and did develop pneumothorax (mean grade of 1.48 vs 1 .73; range 0-3), and occurrence of pneumothorax was not significantly different on effusions estimated as small (i.e., fluid
obscuring
less
than
1/3
of the
diaphragm)
or large
(8%
vs 1 2%). However, pneumothorax developed more frequently in patients with either small or large effusions when the taps were performed with the clinical, rather than the sonographyguided method (1 8% vs 3% for small, p = .0001 ; and 14% vs 2% for large effusions, p .012). Occurrence of pneumothorax was significantly higher in =
Downloaded from www.ajronline.org by 117.253.221.134 on 11/10/15 from IP address 117.253.221.134. Copyright ARRS. For personal use only; all rights reserved
AJR:156,
May
PNEUMOTHORAX
1991
ASSOCIATED
therapeutic than in diagnostic taps (18% vs 6%, p = .001), and the mean volume of fluid aspirated was significantly greater in taps associated with than in those not associated with pneumothorax (472 ml vs 246 ml, p = .006). Compared with clinical thoracentesis, the sonography-guided method was associated with significantly lower occurrence of pneumothorax in both diagnostic (2% vs 1 3%, p .001 9), and therapeutic taps (4% vs 21 %, p = .048). This was also the case for “dry” or “near-dry taps,” that is, less than 5 ml of pleural fluid aspirated (3% in sonography-guided vs 1 7% in clinical taps, p .049). Use of a large needle (i.e., larger than 20-gauge) was associated with higher occurrence of pneumothorax than when a small needle was used (1 8% vs 4%, p .00019). This was also the case when the effusions were small (17% for large vs 4% for small needle, p .001 5) or large (1 6% for large vs 5% for small needle, p < .05), and when the thoracentesis was done with the sonography-guided (1 4% for large vs 2% for small needle, p < .005) but not with the clinical method (25% for large vs 1 6% for small needle). The increased risk from the clinical compared with the sonography-guided method is also demonstrated by the higher grade of pneumothorax (mean grade of .28 vs .04; range, 0-3, p < .0001) and the larger number of follow-up chest radiographs (mean number of 0.70 vs 0.06 additional studies per tap; p < .001) and chest tubes (mean number of 0.080 vs 0.005 tubes per tap; p .004), required with the clinical method. Once pneumothorax did develop, however, the method used did not have a different effect on the grade of pneumothorax (mean grade of 1 .6 for both methods; range, 1 -3) or the number of chest radiographs and chest tubes in taps performed with one or the other method (mean number of four vs two chest radiographs and mean number of 0.4 vs 0.2 chest tubes per pneumothorax occurring with the clinical or sonography-guided method). The increase in cost from follow-up chest radiographs and chest tubes alone was estimated to be an average of $1 28 for every thoracentesis performed with the clinical method, as opposed to $5 for every sonography-guided tap. Multiple regression analysis showed that the method of thoracentesis was the dominant significant factor (p .00757) for the development of pneumothorax. When tested with three other important variables, including the amount of effusion (small or large), type oftap (diagnostic or therapeutic), and size of needle (small or large), these factors were not significant after controlling for method. This was the case when the thoracentesis method was considered with any combination of these three variables. When the method was excluded from the analysis, both the type of tap (p .0434) and size of needle (p = .015) but not the effusion size were significant in the development of pneumothorax. This was the case when all possible combinations of these three variables were tested. Multivariate analysis by logistic regression confirmed these findings. =
=
=
=
=
=
Discussion
In this retrospective centesis, the incidence
survey of patients of pneumothorax
undergoing thorawas significantly
WITH
THORACENTESIS
919
influenced by the method used: sonographic guidance was superior to conventional pleural taps. These results are in agreement with those from the prospective, randomized study of Grogan et aI. [1], which included only patients at low nsk to develop or to be seriously affected from the development of pneumothorax. Also, the occurrence of pneumothorax in our study is similar to previous accounts in which the one or the other method was used [5-10]. The disadvantage of a retrospective study such as this is the difficulty in controlling for extraneous factors, but ethical considerations posed restrictions to the prospective, randomized design of Grogan et al. [1 1. As these two studies were done in the same institutions, at approximately the same time, many ofthe extraneous factors were similar. Viewed together, many of the weaknesses of the two studies are counterbalanced. For the development of pneumothorax, our results showed no contribution from age, sex, underlying pulmonary disease, critical clinical condition, or the type of fluid aspirated. Interestingly, the amount of effusion present did not affect the occurrence of pneumothorax. Conversely, pneumothorax occurred more frequently as the amount of pleural fluid aspirated increased. This may be due to rapid changes in pressure in the pleural space while fitting reexpansion of the underlying lung lags behind. As expected, larger needles were associated with higher incidence of pneumothorax. One subgroup of patients had clinical pleural taps performed after sonographic localization of the fluid, but not under sonographic guidance. In these taps, the occurrence of pneumothorax was not different from that in the rest of the clinical taps, although it was significantly higher than that in the sonography-guided taps. Kohan et al. [1 1 ] also found that the occurrence of pneumothorax was not reduced with remote sonographic localization. This may be related to a change in position of the fluid over time and change in the patient’s position, and the ultimate reliance on physical examination for the selection of the tap site. Furthermore, the occurrence of pneumothorax in our clinically performed taps was not different for small or for large effusions. Assuming a large effusion is easier to localize, one would expect a higher rate of pneumothorax in small effusions. These findings suggest that there may be no significant room for improvement of the conventional techniques, and that the advantage of sonographic guidance probably lies in the synchronous direct visualization of the pleural fluid during thoracentesis. Other technical differences, such as training and experience, also may play a role. Sonographers are used to percutaneous needle placements, and it may be easier to train and closely supervise a relatively small group of radiologists to perform a procedure in a consistent manner than to work with a much larger group of physicians in various clinical disciplines. Although Grogan et al. [1 ] tried to control this factor prospectively by providing similar instructions to radiology and medical personnel and by ensuring house physicians were competent to perform the procedure, the difference in occurrence of pneumothorax among the two groups remained significant. Our results show sonography-guided thoracentesis to be
Downloaded from www.ajronline.org by 117.253.221.134 on 11/10/15 from IP address 117.253.221.134. Copyright ARRS. For personal use only; all rights reserved
920
RAPTOPOULOS
safer than a conventional pleural tap, and it does not appear to increase the cost substantially. The extra expense from additional chest radiographs and required chest tubes alone was estimated at $1 23 per patient having conventional thoracentesis, which is almost comparable to the additional cost of chest sonography. Therefore, a strong case could be made in favor of sonography, proclaiming safety and comfort without significant increase in cost. In a busy clinical service, however, the logistics of such a routine may be substantial and, for low-risk patients, it is reasonable to wait for future studies to confirm our results. At a minimum, we recommend sonography-guided thoracentesis for patients whose clinical condition may be jeopardized by the development of pneumothorax, such as critically sick patients and patients with severe lung disease or who are on respiratory assistance. In addition, when single-needle passage does not provide sufficient fluid for diagnostic tests, we suggest that taps be repeated under sonographic guidance. Moreover, we advocate use of the smallest possible needle and aspiration of the smallest possible amount of fluid, even for therapeutic taps, irrespective of the method used, but especially when sonographic guidance is not used. ACKNOWLEDGMENT We
are grateful
for the
dedication
of our
radiology
residents.
ET
AL.
AJR:156,
May 1991
REFERENCES
1 . Grogan DR. Irwin RS, Channick R, et al. Complications associated with thoracentesis: A prospective, randomized study comparing three different methods. Arch Intern Med i990;1 50:873-877 2. Rhea JT, DeLuca SA, Greene RE. Determining the size of pneumothorax in the upright patient. Radiology i982;144:733-736 3. Rhoades HM, Overall JE. A sample size correction for Pearson chi-square in 2x2 contingency tables. Psycho! Bull 1982;9i :418-423 4. Dixon W, Brown M, Engelman L, et al. BMDP statisticalsoftware. Berkeley, CA: University of California Press, 1985 5. Collins TR, Sahn SA. Thoracocentesis: clinical value, complications, technical problems, and patient experience. Chest i987;91 :817-822 6. Seneff MG, Corwin RW, Gold LH, Irwin RS. Complications associated with thoracocentesis. Chest 1986:90: 97-i 00 7. Schroeder SA, Marton KI, Strom BL. Frequency and morbidity of invasive procedures: report of a pilot study from two teaching hospitals. Arch Intern Med 1978;138:i809-1811 8. Harnsberger HR, Lee TG, Mukuno DH. Rapid, inexpensive real-time directed thoracentesis. Radiology 1983:146:545-546 9. O’Moore PV, Mueller PR, Simeone JF, et al. Sonographic guidance in diagnostic and therapeutic interventions in the pleural space. AJR i987;149:i-5 10. Hirsch JH, Rogers JV, Mack LA. Real-time sonography of pleural opacities. AJR i98i;i36:297-301 11. Kohan JM, Poe RH, Israel RH, et al. Value of chest ultrasonography versus decubitus roentgenography for thoracentesis. Am Rev Respir Dis 1986:133:1124-1126