Sonography of iatrogenic pneumothorax in pediatric patients

Abstract

Pneumothorax is defined as the presence of air in the pleural cavity. The incidence of iatrogenic pneumothorax in the pediatric population is 0.3–0.48 in 1000 patients. A conventional chest X-ray, in some cases supplemented with chest computed tomography, is a typical imaging examination used to confirm the diagnosis of pneumothorax. Within the last years, the relevance of transthoracic lung ultrasound in the diagnostic process of this disease entity has greatly increased. This is confirmed by the opinion of a group of experts in ultrasound lung imaging in patients in a life-threatening condition, who strongly recommend a transthoracic ultrasound examination for the diagnosis of pneumothorax in such patients. These data constituted the basis for initiating the prospective studies on the application of this method in pneumothorax diagnosis in patients of pediatric hematology and oncology wards.

Aim

The aim of the study was to present the possibility of using the transthoracic lung ultrasound in the diagnostic process of pneumothorax in pediatric patients, with particular attention paid to its iatrogenic form. The article discusses sonographic criteria for pneumothorax diagnosis in pediatric patients, including the sensitivity and specificity of the method, in relation to conventional chest X-ray.

Material and methods

The prospective studies included a group of patients treated in the Clinic of Pediatrics, Pediatric Hematology, Oncology and Endocrinology of the Academic Clinical Centre (Medical University of Gdańsk, Poland) in whom a central venous catheter was placed in the subclavian veins. The studies lasted for one year – from 1 July 2011 to 30 June 2012. The examined group comprised 63 patients – 25 girls (39.7%) and 38 boys (60.3%) aged from 1 to 17. The analysis included the results of 115 ultrasound examinations conducted in this group.

Results

In t he examined group with suspected or diagnosed neoplasm, iatrogenic pneumothorax was identified in 4 out of 63 patients (6.3%). In all cases, it was a consequence of procedures connected with the introduction of a central catheter. In the examined material, the sensitivity, specificity and predictive values (positive and negative) of the transthoracic lung ultrasound equaled 100%.

Conclusions

The transthoracic ultrasound examination is a highly effective method in diagnosing iatrogenic pneumothorax in pediatric patients. However, to confirm this conclusion, multicenter studies, among considerably larger populations, are necessary.

Introduction

Pneumothorax, defined as the presence of air in the pleural cavity, occurs rarely in pediatric patients. There are a dozen or so publications discussing the incidence of pneumothorax in adults but information about children can be found only in single papers. According to the British data based on the Healthcare Cost and Utilization Project (HCUP) report, in 1997, iatrogenic pneumothorax was diagnosed in 0.48 per 1000 patients aged 0–17. This value was similar to the incidence in other age groups (0.42 per 1000 patients aged 18–44 and 0.43 in those aged 45–64; a slightly higher incidence – 0.74 – was noted in the group of patients above the age of 65)⁽¹⁾. According to Miller's data, in 2000 in Great Britain, iatrogenic pneumothorax was diagnosed in 0.3 per 1000 patients aged 0–18⁽²⁾. In the pediatric population, one of the main groups at risk of developing pneumothorax is the group of neonates in whom the most common cause of this condition is barotrauma. However, in older pediatric groups and in adult patients, more significant etiological factors are invasive diagnostic and therapeutic procedures, such as central venous access (CVA) or thoracocentesis⁽³⁾.

Transthoracic lung ultrasound in the diagnosis of pneumothorax

A conventional chest X-ray, which in some cases is supplemented with computed tomography (chest CT), is a typical imaging examination used to confirm the diagnosis of pneumothorax. These diagnostic methods are used both in pediatrics as well as in adult patients and CT scan is commonly accepted as the gold standard in pneumothorax diagnosis. In the last years, the relevance of transthoracic lung ultrasound (TLUS) has increased. This primarily refers to adult patients in a life-threatening condition, patients with trauma to the chest and those who underwent diagnostic or therapeutic procedures within the thoracic cavity. This increased relevance is confirmed by the published opinion of a group of experts in ultrasound imaging (US) of the lungs in patients in a life-threatening condition, who strongly recommend TLUS for the diagnosis of pneumothorax in the aforementioned patient groups⁽⁴⁾. Moreover, in the metaanalysis of the published data concerning the diagnosis of pneumothorax with the use of TLUS, Ding et al. (2011) demonstrated great usefulness of this imaging modality by concluding that its sensitivity and specificity in adult patients is non-inferior to conventional chest X-ray⁽⁵⁾.

These data constituted the basis for initiating the prospective studies on the application of TLUS in the diagnostic process of pneumothorax in patients of pediatric hematology and oncology wards. The main cause of pneumothorax in this pediatric group is a complication resulting from the insertion of a central venous catheter to the subclavian vein. This group of patients was therefore enrolled into the prospective study. Another argument for initiating these studies was the lack of similar data concerning the pediatric population in available literature.

Material and methods

The prospective studies included a group of patients treated in the Clinic of Pediatrics, Pediatric Hematology, Oncology and Endocrinology of the Academic Clinical Centre (Medical University of Gdańsk, Poland) in whom a central venous catheter was placed in the subclavian vein. The studies were conducted for one year – from 1 July 2011 to 30 June 2012. The study population comprised 63 patients, including 25 girls (39.7%) and 38 boys (60.3%) aged from 1 to 17. The mean age in the examined group was 7.8 years (standard deviation – 5.0 years). The analysis included the results of 115 US examinations performed in these patients. The procedures of inserting central catheters were conducted in two centers: Academic Clinical Centre and Nicolaus Copernicus Pomeranian Trauma Centre of the Regional Specialized Hospital in Gdańsk (Poland). The procedures were performed by experienced anesthesiologists in a general anesthesia via the percutaneous access according to anatomic markers. In 40 children (63.5%), these procedures were conducted after an initial sonographic assessment of the thoracic vessels, which included anatomical conditions, diameter of the subclavian and brachiocephalic veins as well as blood flow in these vessels evaluated in color or power Doppler examinations.

Radiological diagnosis

The placement of a central venous catheter was performed in procedure rooms with the possibility to monitor the correctness of the procedure and its potential complications with the use of radioscopy. After the procedure, the placement of the catheter and the image of the lungs were documented in the form of a conventional chest X-ray performed in the supine position. Another chest X-ray was performed in those patients who, based on the lung US examination, were diagnosed with pneumothorax. These examinations were conducted at the patients’ bedside or in the Department of Radiology of the Medical University of Gdańsk (Poland) depending on the clinical needs and in accordance with the principles of medical art, bearing in mind well-being of the patients. The examinations were used in further therapeutic procedures.

Ultrasound imaging

The US examination of the lungs and venous vessels was performed according to the previously prepared protocol. All initial transthoracic US examinations of the lungs as well as the assessment of the jugular and subclavian veins, which were performed prior to the placement of the central catheter (40/115 examinations which constitutes 34.8% of all scans), took place in the Laboratory of Ultrasound Imaging and the Clinic of Pediatric Biopsy, Hematology and Oncology at the Medical University of Gdańsk. Follow-up examinations after the placement of venous catheters were performed at the patients’ bedside or in the Laboratory of Ultrasound Imaging and the Clinic of Pediatric Biopsy, Hematology and Oncology at the Medical University of Gdańsk. 25/75 scans (33.3%) were bedside examinations. The indications for such examinations were each time determined by the physician directly responsible for a given patient's treatment. The examinations were performed at the day of the procedure – within 1–2 hours after its conclusion if CVA was conducted in the Academic Clinical Centre, or directly after the patient's return from the Trauma Centre – on the next day following catheter placement.

TLUS – examination technique

TLUS examinations were performed with the use of two US systems: GE Logiq 7 (GE Medical Systems, Waukesha, WI, USA) and Philips iU 22 (Philips Medical Systems, Bothell, WA, USA) with convex probes with the frequency of 3.5–5.0 MHz and linear transducers with the frequency of 8.0–12.0 MHz. The preset settings, designed for abdominal examinations, were used and modified during each examination according to current needs. All TLUS examinations were performed in accordance with the standards of the Polish Ultrasound Society, which had been created by the author of this paper⁽⁶⁾.

Each examination was documented in the digital form in standard formats (WMV or AVI). Due to the dynamic character of the majority of assessed parameters, a printed image would be of limited relevance, or completely irrelevant, for documenting the diagnosed changes.

TLUS examinations were performed with the patients in the supine position. The image of the lungs was assessed in the accessible intercostal spaces on the anterior and lateral surfaces of the thorax. Each examination was conducted with both convex and linear probes, placing them longitudinally and transversely and using the intercostal spaces as acoustic windows. During each examination, the US presentation of the lungs was compared to the image of the contralateral side of the chest with the transducer placed at the same level. The assessment of the lungs and pleura in one intercostal space was made during at least three respiratory cycles. The total duration of TLUS did not exceed 10 minutes. As with the CT diagnosis, the ALARA principle applied (as low as reasonably achievable).

TLUS – sonographic criteria and signs used in the diagnosis of pneumothorax

The images obtained during the examinations were analyzed in terms of the presence of the sonographic criteria that either confirm or rule out pneumothorax. In this study, the criteria used for adult patients were applied with certain modifications for pediatric patients, which are discussed below.

Sonographic criteria to rule out pneumothorax

The sonographic criteria based on which pneumothorax was ruled out included:

• presence of pleural line slide (lung sliding sign)⁽⁷⁾;

• presence of B-line sonographic artefacts⁽⁸⁾, presence of Z-line and I-line sonographic artefacts (author's own modification);

• detection of the sign referred to as lung pulse⁽⁹⁾.

Sonographic criteria to confirm pneumothorax

The sonographic criteria based on which pneumothorax was confirmed included:

• presence of changes referred to as A’-profile (A prime) according to the criteria created by Daniel Lichtenstein in the diagnostic “BLUE protocol”⁽⁸⁾;

• detection of the sign referred to as lung point⁽¹⁰⁾;

• detection of the sign referred to as double lung point⁽¹¹⁾.

The sonographic criteria and signs used for pneumothorax diagnosis are defined in the following way:

• Lung sliding sign

  It is a dynamic ultrasound artefact that appears on a so-called pleural line which physiologically consists of the parietal pleura, pulmonary pleura and a slight amount of fluid between these membranes. The pleural line is visible on the ultrasound monitor as a strongly hyperechoic line which moves upwards and downwards as the air volume in the lungs changes. The movement of this hyperechoic line is referred to as the lung sliding sign. Due to its dynamic character, the real-time assessment is necessary⁽⁷⁾. If it is difficult to visualize the sliding pleural line, power/color Doppler sign is used which appears during the examination of the pleural line in power or color Doppler modes⁽¹²⁾. The printed documentation of the lung sliding sign should be performed in the M-mode option⁽¹³⁾.

• Presence of the B-line, Z-line and I-line sonographic artefacts

  This group of artefacts observed in TLUS appears on the pleural line. Thus, these artefacts may be visualized only if the line itself is visible. In pneumothorax, the pleural line is not visible and these artefacts are not present either. The B-line artefacts are vertical hyperechoic lines which appear on the pleural line, spread along the entire length of the monitor and move clockwise together with the movements of the pulmonary pleura. It is believed that these artefacts are generated by the slight amount of fluid under the pulmonary pleura, usually in the interlobular septa (air-water artefact). Despite not entirely explained etiopathogenesis, the majority of authors consider them reverberation artefacts^(14–16). The Z-line artefacts are very similar to the B-line ones. It is claimed that they resemble a comet tail; they are observed as vertical, hyperechoic lines which begin on the pleural line. The difference from the B-line artefacts lies in their length – the Z-line artefacts end at approximately 1/3–1/2 of the monitor's length and do not reach its edges. By contrast with the B-line artefacts, they do not “erase” the A-line artefacts when visualized with the use of convex or sector transducers. The practical r elevance of t hese a rtefacts h as not b een unequivocally determined but they were initially named with the last letter of the alphabet, thus suggesting clinical irrelevance⁽¹⁵⁾. The manner of their formation has not been sufficiently explained. As with the two types of artefacts mentioned above, the I-line artefacts are comet tail artefacts. They appear on the pleural line, have 1–3 cm in length and do not erase A-line artefacts when convex or sector transducers are used. They are currently believed to be diagnostically irrelevant⁽¹³⁾.

  Despite the fact that only the B-line artefacts are unequivocally relevant in the diagnosis of adult patients, the sole existence of the Z-line and I-line artefacts appears to be a justifiable argument for their usage in pneumothorax diagnosis. Another reason for using these two types of artefacts in pediatric examinations is the fact that, when compared to adults, examinations of children are more frequently performed with the use of linear high-frequency probes (10–15 MHz). This way, we obtain a more accurate image of the pleura and these artefacts are observed more frequently and are more clearly visible than in imaging with convex or sector probes. Therefore, the presence or absence of these artefacts was included in the diagnostic criteria of pneumothorax (the author's own modification).

• Lung pulse

  This term denotes a dynamic sonographic sign which is the effect of the transmission of the heart beat to the pleural line by the pulmonary consolidation area adjacent to the pericardium. The presence of this sign definitively rules out pneumothorax⁽⁹⁾.

• Detection of changes referred to as A’-profile

  This sign corresponds to the presence of air in the pleural c avity. I t i s c haracterized b y t he p resence o f t he A-line artefacts only, and the absence of the pleural line sliding. It differs from the A-profile (reflecting normal lung presentation) in the lack of the lung sliding sign. In normal conditions, the A-line artefacts – horizontal, averagely hyperechoic lines, parallel to the pleural line and repeatable in the same distance within the edges of the monitor – appear on the pleural line. They are reverberation artefacts and result from multiple reflections of the ultrasound wave on the border between air in the alveoli located under the pulmonary pleura and the chest wall⁽¹⁷⁾. In the case of pneumothorax, the manner of A-line artefact formation is precisely the same but the border on which they appear is different – it is the parietal pleura and air in the pleural cavity. Being the basic artefacts of the normal lung image in TLUS, paradoxically, the A-line artefacts also constitute one of the prime signs of pneumothorax. The visualization of the profile which is characterized by the presence of the A-line artefacts, the absence of B-line, Z-line and I-line artefacts as well as no lung sliding sign confirms the diagnosis of pneumothorax⁽¹⁸⁾.

• Lung point

  This term denotes a sonographic sign which marks the border between the pneumothorax and normally aerated lung⁽¹⁰⁾. It is the effect of the collapsed lung moving closer to the chest wall due to a slight increase of its volume on inspiration. In such a situation, the monitor shows a border between the normal lung image (A-profile) and the area referred to as the A’-profile (only the A-line artefacts are visible without the sign of lung sliding).

  The visualization of the lung point is also possible in the M-mode examination. In this case, two bordering sonographic signs may be observed: the seashore sign, as the normal lung image, and the stratosphere sign which confirms pneumothorax⁽¹⁹⁾. The seashore sign is characterized by the lack of movement above the pleural line in the form of several parallel, horizontal lines and a homogeneous granular pattern of reflections which appears below the pleural line as a result of the movement of structures located below the pleural line according to the breathing cycle. The stratosphere sign (or “barcode sign”) consists of numerous horizontal lines with invisible pleural line. It is detected when the lung sliding sign is negative and corresponds to a complete absence of the pleural movement in the examined region.

• Double lung point

  As with the lung point, the double lung point marks the border between pneumothorax and normally aerated lung but it is observed only when pneumothorax is of a small size⁽¹¹⁾. It is best visible with a transverse placement of a linear probe. The image consists of a centrally located A-profile area and the lung point that is visible on its both sides (author's own observations).

Assessment of the size of pneumothorax in TLUS

In order to assess the size of pneumothorax, the author of this study used generally accepted principles which had been created for adult patients. The marker which allows for the determination of the size of pneumothorax is the lung point. Depending on its localization in the chest, it is possible to roughly estimate the size of pneumothorax. It is believed that if its presence is detected on the anterior chest wall, pneumothorax is small. The localization of the lung point on the lateral chest wall up to the level of the midaxillary line suggests the diagnosis of moderate pneumothorax. If, however, it is located below this line, substantial pneumothorax is identified⁽²⁰⁾. The impossibility to visualize the lung point remains an unsolved issue. In such situations in adult patients, further diagnosis with chest CT is necessary⁽⁴⁾.

Statistical analysis

The author prepared standard descriptive statistics of the analyzed parameters. In order to compare the diagnostic value of the tested methods, the sensitivity, specificity, positive and predictive values as well as McNemar's Chi-square value were calculated. Due to a low number of positive results, the exact form of the test was used. The agreement of results obtained by means of the two studied diagnostic methods was assessed and Cohen's kappa coefficient was calculated. The value of this coefficient was interpreted according to Landis and Koch's classification, i.e. the values <0 denote the lack of agreement, 0–0.20 indicate slight agreement, 0.21–0.40 – fair agreement, 0.41–0.60 – moderate agreement, 0.61–0.80 – substantial agreement and 0.81–1 – almost perfect agreement. The assumed level of statistical significance equaled α = 0.05. The statistical analysis was conducted with the use of Stata 12 statistical system by StataCorp LP (USA, 2011).

Results

In the analyzed group with suspected or diagnosed neoplasm, iatrogenic pneumothorax was identified in 4 out of 63 patients (6.3%). In all cases, it was a consequence of procedures connected with the insertion of a central venous catheter. The detailed data are listed in tab. 1.

Tab. 1.

Characteristics of patients with diagnosed iatrogenic pneumothorax

Patient	Gender	Age (years)	Diagnosis	Side
1	M	15	ALL	Right
2	Ż	6	DRESS	Left
3	M	9	ALL/BMT	Left
4	M	10	ALL	Left

ALL– acute lymphoblastic leukemia; BMT– bone marrow transplantation; DRESS– drug reaction (or skin rash) with eosinophilia and systemic symptoms.

In 3 patients (N = 63), TLUS was the first imaging examination in which pneumothorax was diagnosed and in one case, it was a subsequent imaging examination which confirmed the presence of pneumothorax following a previously conducted chest CT. In this patient, the diagnosis on the basis of TLUS was established before the attending physician obtained the CT results. Among the patients with TLUS-based diagnosis of pneumothorax, in 2 cases, its presence was confirmed by chest X-ray and a chest drain was conducted. In one case (patient 2), chest X-ray was not performed due to the small size of pneumothorax which was monitored by TLUS only. The normal lung image was observed on the fourth day of monitoring. Table 2 presents the comparison of US- and X-ray-based lung examinations.

Tab. 2.

Comparison of TLUS and X-ray results

Method	Pneumothorax	Normal image
TLUS performed after central catheter implantation	4/63	59/63
Chest X-ray performed after central catheter implantation	1/63	62/63
Chest X-ray performed after TLUS examination during which pneumothorax was diagnosed	2/2	0/2

In one case, TLUS was a bedside examination (patient 1) and in 3 cases, it took place in the US laboratory. The time interval between the TLUS examination and the conclusion of the procedures connected with central catheter insertion equaled 1–2 hours in patients who remained at the Academic Clinical Centre. In 2 patients who had central venous access performed in the Trauma Centre the time interval was 30 and 36 hours. The mean time within which pneumothorax was diagnosed by means of TLUS was analyzed on the basis of the examination sequence stored on the hard disk of the US system. The time that lapsed from the first recorded sequence to the moment at which sonographic signs typical of pneumothorax appeared on the monitor was analyzed. It did not exceed 25 seconds in the study population. The total duration of the TLUS examination in the group of patients with diagnosed iatrogenic pneumothorax, analyzed from the first to the last sequence, did not exceed 3 minutes. The US presentation of detected changes in the context of the US diagnostic criteria for pneumothorax is presented in tab. 3.

Tab. 3.

Comparison of the normal image and pneumothorax in the study population, including sonographic signs and artefacts that confirm or rule out pneumothorax

	N– number of cases
	Patients with normal image	Patients with pneumothorax
Lung sliding sign	59	0
B-line artefact	5	0
Z-line or I-line artefacts	59	0
A’-profile	0	4
Lung point	0	2
Double lung point	0	2
Lung pulse	0	0

In the study population with iatrogenic pneumothorax diagnosed by means of TLUS, all typical criteria confirming pneumothorax were observed: the lack of lung sliding sign, changes referred to as A’-profile and the lung point. Based on the localisation of the last above-mentioned criterion, the presence of large pneumothorax was suggested in 2 patients and required interventional treatment (patient 1 and patient 3). In further two cases (patient 2 and patient 4), slight pneumothorax was diagnosed – limited to the anterior aspect of the chest with a visible double lung point. These lesions were observed on the side at which the CVA procedures were conducted. In one patient, it was the right side and in 3 – the left side. Nevertheless, no differences in the observed diagnostic criteria were found. In the group of patients with the normal lung image in TLUS, none of the criteria confirming pneumothorax were observed whereas all criteria that ruled out pneumothorax were present.

Statistical analysis – results

Statistical methods were used to assess the diagnostic value of TLUS and chest X-ray in the analyzed group – the results are presented in tab. 4.

Tab. 4.

Diagnostic value of TLUS and X-ray (values in the parentheses refer to the 95% confidence interval)

	TLUS	X-ray
Sensitivity	100% (39,8–100)	25% (0,6–80,6)
Specificity	100% (93,9–100)	100% (93,9–100)
Positive predictive value	100% (39,8–100)	100% (2,5–100)
Negative predictive value	100% (93,9–100)	92,5% (86,5–99)

In the studied material, the sensitivity, specificity and predictive values (positive and negative) for the TLUS examination equaled 100%. The width of the 95% confidence interval for specificity and negative predictive value was acceptable. However, the range of 95% confidence interval for sensitivity and positive predictive value was very wide. This results from a relatively low number of instances of pneumothorax in the study population.

As for the X-ray examination, comparable or slightly lower specificity and negative predictive values were observed. The width of the 95% confidence interval was acceptable. The sensitivity equaled 25% and positive predictive value – 100%. As with the TLUS, the width of the 95% confidence interval for the sensitivity and positive predictive value was too large to enable precise estimation of their value in the population and allow for their comparison in both methods.

Despite a considerable difference in sensitivity (100% for TLUS and 25% for X-ray), the result of McNemar's test, for which p value = 0.250 (tab. 5), does not allow to claim that one method is superior to the other. However, it should be borne in mind that due to a low number of positive results, there is a risk of making a type II error.

Tab. 5.

Statistical assessment of the agreement of TLUS and X-ray results (95% confidence interval is provided in the parentheses)

	TLUS vs. X-ray
McNemar's test (exact)	P = 0,250
Observed agreement	95,2%
Anticipated agreement	92,3%
Cohen's kappa coefficient	0,38(0,19 − 0,51)

The agreement of the diagnoses and the value of kappa coefficient for both tested methods are presented in tab. 5. The agreement of the diagnoses in the study population equaled 95.2%. However, due to the low number of positive results, the anticipated agreement should equal 92.3%. This is the reason for a low kappa coefficient (0.38) which is a corrected and more reliable measure of agreement of the diagnoses established with the use of the compared diagnostic methods.

Discussion

The first physician who proposed the term pneumothorax for the presence of air in the pleural cavity was Itard. It took place in 1803 when he was a student of Laennec, the inventor of the stethoscope⁽²¹⁾, who described the clinical symptoms of pneumothorax in 1819⁽²²⁾. As of today, they are still valid and still constitute the basis for interview and physical examination in the course diagnosing this pathology. Despite well-known clinical symptoms of pneumothorax and its inclusion in a differential diagnosis in pediatric patients with dyspnea, the detection of such changes in children frequently surprises clinicians. This confirms the thesis that pneumothorax is diagnosed much more frequently in adults than in the pediatric population. The difference in the incidence of spontaneous pneumothorax between these age groups is substantial. Despite the availability of only scarce publications concerning the epidemiology of this disease entity in the pediatric population, the data from American centers allow for a conclusion that in the adult population, the incidence of pneumothorax is approximately 100 times greater than in pediatric population^{(23, 24)}.

The “gold standard” in medical imaging of pneumothorax is chest CT and in daily practice, the first imaging examination is a conventional chest X-ray.

A question needs to be asked whether the development of a new diagnostic method, i.e. transthoracic lung ultrasound, is necessary in the situation when the estimated incidence of pneumothorax in the pediatric population is so low, the X-ray examination is so commonly accessible and, additionally, when CT is becoming more and more commonly accessible. The answer to this question appears to be very easy. However, when only iatrogenic forms of pneumothorax and cases diagnosed in intensive care units are taken into account in the incidence analysis, the frequency with which pneumothorax occurs in the pediatric and adult populations will be completely different than in the general incidence of this disease. It occurs that the incidence of iatrogenic pneumothorax in both these populations is similar and estimated at 0.48/1000 patients aged 0–17 and 0.43/1000 patients aged 18–64⁽¹⁾.

The same refers to patients in intensive care wards, in whom the incidence of pneumothorax equals 3–4%^{(25, 26)} and is similar in both populations when the pediatric population includes the group of neonates⁽²⁷⁾. Taking these epidemiological data into account, it should be stated that with such a small number of the instances of pneumothorax in the general pediatric population, the introduction of a new imaging technique is groundless. However, taking into consideration patients in pediatric intensive care units or pediatric wards of similar profiles, the transportation of whom for the X-ray examination may be prevented for numerous reasons, as well as a group of pediatric patients after interventional procedures within the chest, the introduction of a method which allows for a safe, fast and effective diagnosis of pneumothorax in daily practice appears to be necessary.

The application of ultrasounds in pneumothorax diagnosis is not a new idea. The first publication connected with their usage in humans was published 25 years ago when Wernecke et al. described the basic criteria for the diagnosis of pneumothorax⁽²⁸⁾ based on a previous experiment of Rantanen who, a year before, had presented the US image of equine pneumothorax⁽²⁹⁾. The ultrasound assessment of pneumothorax was practically not conducted until the second half of the 1990s when the first Lichtenstein's publications appeared. Not only did his papers systematize the sonographic signs and artefacts, which are currently used as the basis for pneumothorax US diagnosis, but also greatly influenced the development of lung sonography in patients in a life-threatening condition. Earlier than that – in 1990, Targhetta et al. were the first authors to describe the US presentation of pneumothorax in a group of 28 patients and compared it with a control group of 100 patients⁽³⁰⁾. Moreover, in 1992, they were the first to suggest that TLUS should be used for diagnosing iatrogenic pneumothorax in patients directly after an aspiration biopsy of a pulmonary lesion⁽³¹⁾.

At present, the issue of lung biopsies with ultrasound-guidance of the procedure is relatively widely discussed in the literature, but still, only single publications refer to the usage of TLUS for diagnosing the pulmonary consequences of CVA procedures^{(32, 33)}. In two most frequently quoted publications, the number of cases of iatrogenic pneumothorax following central catheter placement in the subclavian veins, which are diagnosed by means of TLUS, is very low and equals merely 37 cases (Maury – 1 patient, Kreuter – 36 patients). In the available literature, no publication was found concerning the application of TLUS in the diagnosis of pulmonary complications after central catheter implantation in pediatric patients.

In the presented group of patients, the sensitivity and specificity of TLUS in the diagnosis of pulmonary complications of the CVA procedure equaled 100%. Kreuter et al. report the same values with respect to adult patients⁽³³⁾. What is more, in the analyzed pediatric group, the sensitivity of TLUS was slightly higher than that of a conventional chest X-ray when conducted directly after catheter placement. Perhaps, the difference in time from the catheter placement to the examinations by means of both methods affected the obtained results. Perhaps, if the X-ray examinations had been performed again after 1–2 hours after the procedure, the results would be comparable. Pneumothorax caused by interventional procedures is generally manifested within the first 3 hours after the procedures. This may be a probable reason for a normal lung image in the chest X-ray conducted directly after the procedure in 3 patients even though one of them suffered from a substantial pneumothorax (patient 1). In one case (patient 2), TLUS helped to diagnose slight pneumothorax, probably, so-called radiologically silent pneumothorax. In this case, the supine position of the patient during the examination might have influenced such radiological findings. The X-ray examination performed with the patient in such a position may affect the visualisation of slight pneumothorax to a considerable degree⁽³⁴⁾. Furthermore, it should be emphasized that according to the data of other authors, some instances of pneumothorax remain undiagnosed with the use of the chest X-ray irrespective of the patient's position^{(35, 36)}.

In the 4 presented cases of iatrogenic pneumothorax diagnosed with the use of sonography, its size was correctly estimated based on the lung point localization. In 3 cases, it corresponded to the size of pneumothorax assessed in radiological examinations: in X-ray (2 cases) and in CT (1 case). In one case (patient 2), further X-ray examination was not performed due to the small size of pneumothorax – it was monitored with TLUS only. The patient 1 is the first case that demonstrates the possibility to diagnose iatrogenic pneumothorax by means of the bedside TLUS examination.

The study presented herein has several significant limitations which include: a low number of pediatric patients, the lack of the chest C T a s a reference imaging method in pneumothorax, the time of performing the first TLUS, which should have been performed directly after the procedure, and the time of performing the chest X-ray examination, which should have been performed 1–3 hours a fter the procedure for the most optimal results. Nonetheless, in the available literature, this is the first publication on the application of T LUS for the diagnosis of iatrogenic pneumothorax in pediatric patients and the second one in terms of the number of presented patients. Unfortunately, in our center, ultrasound guidance of CVA procedures is performed only when the placement of a catheter is difficult, at the request of the physician that performs the procedure. In the analyzed period of time, such situations occurred twice and in both cases ultrasound-guided CVA procedures proceeded with no complications. It appears that a potential l imitation of the study, i.e. the l ack of routinely conducted CT as the reference method in iatrogenic pneumothorax diagnosis, is of no relevance in the case of the pediatric population.

Conclusions

1. TLUS is a highly effective method in the diagnostic process of iatrogenic pneumothorax in pediatric patients.

2. To confirm th is c onclusion, m ulticenter s tudies among considerably larger populations should be conducted.

The article was prepared on the basis of the author's habilitation thesis and includes its numerous fragments also published in the “Annales Academiae Medicae Gedanensis” 2013, v. XLIII, supplement 10.

Conflict of interest

Author does not report any financial or personal links with other persons or organizations, which might affect negatively the content of this publication and/or claim authorship rights to this publication.