| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2001;71:1779-1785
© 2001 The Society of Thoracic Surgeons
a Department of Surgery, University of Minnesota, Minneapolis, Minnesota, USA
Accepted for publication February 4, 2001.
Address reprint requests to Dr Jaklitsch, University of Minnesota, Cardiovascular and Thoracic Surgery, Box 495 UMHC, 420 Delaware St SE, Minneapolis, MN 55455
e-mail: mjaklitsch{at}partners.org
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. This is a retrospective review of 64 open-lung biopsies (58 patients) from 1976 to 1996. Open-lung biopsies were used to grade vasculopathy in 8 patients (12% of 64) with pulmonary hypertension and in 10 patients (16% of 64) with combined pulmonary hypertension and lung parenchymal disease. Forty-six biopsies (72%) were obtained to diagnose parenchymal disease. Comparisons were made between biopsies performed from 1976 to 1989 and from 1990 to 1996.
Results. In the period 1990 to 1996, there were significantly more infants (p = 0.03), comorbid disease (p = 0.009), extracorporeal membrane oxygenation support (p < 10-4), and ventilator dependence (p = 0.05) and significantly less immunocompromise (p = 0.04). A definitive diagnosis was made in 43 of 64 cases (67%) and altered workup in 63 of 64 cases (98%). No correlation existed between Heath-Edwards grade of microangiopathy and catheterization data. Definitive diagnosis was most strongly associated with a nonimmunocompromised patient (p < 10-4). Although only one death (1.5%) was related to open-lung biopsy, the procedure was associated with a 30% inhospital mortality rate and an 11% morbidity rate. Of the 19 deaths, 1 patient died from the procedure, 13 died from their diseases, and 5 had support withdrawn. Death was associated with preoperative ventilator dependence (p < 10-4) and extracorporeal membrane oxygenation (p = 0.007).
Conclusions. Pediatric open-lung biopsy commonly alters the diagnostic workup (98%). It is recommended for children who have been supported for 2 weeks by extracorporeal membrane oxygenation and for those with combined pulmonary hypertension and parenchymal lung disease. It is less useful in immunocompromised children.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
The utility of pediatric open-lung biopsy in current medical practice cannot be judged from the literature for several reasons. Most previous series have been in print for more than a decade and predate many current pediatric diagnostic techniques and treatment strategies. The range of expected operative mortality (0% to 62%) and the range of diagnostic utility (9% to 100%) previously reported have been too broad to assist in current selection criteria of patients who may benefit from the procedure [2, 10]. Additionally, the acuity of the pediatric intensive care unit patient and the number of immunocompromised children have apparently increased in the past decade. For these reasons, we theorized that the demographics, and possibly the outcomes, of pediatric open-lung biopsy have changed over the past 20 years.
To characterize the changing role of lung biopsy in children, we retrospectively examined the results by a single surgeon (J.E.F.) from 1976 to 1996 in a combined pediatric general thoracic and cardiac surgical service within a major tertiary referral medical center.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
All patients had open-lung biopsy performed under general endotracheal anesthesia with both lungs ventilated. Incision sites depended on the other surgical procedures to be done or on the location of the primary disease process. Patients with bilateral and equal interstitial disease underwent a right anterior minithoracotomy with a submammary incision. Specimens in full-term infants and children were obtained by stapling and excision; the staple line was reinforced with fine monofilament suture in areas of bleeding or air leak. Specimens in premature infants were obtained by amputating a free edge of lung tissue distal to a fine monofilament pursestring suture.
The specimens were divided in the operating room according to a written protocol. One large piece was sent in a sterile cup to the surgical pathology lab for hematoxylin and eosin stain, fungal stain, acid-fast bacillus stain, and special stains when appropriate. Two sterile pieces were placed into an anaerobic culture tube and two into an aerobic culture tube, and these were sent to the microbiology lab for aerobic culture, anaerobic culture, Gram stain, potassium hydroxide prep, fungal culture, and legionella culture. A fourth sterile specimen was sent to the Minnesota State Lab for acid-fast bacillus culture. Since 1983, a fifth piece was sent in a sterile culture tube to the virology lab for cytomegalovirus, herpes simplex virus, respiratory syncytial virus, adenovirus and parainfluenza cultures, and cytomegalovirus rapid antigen. Since 1985, a sixth specimen was sent in sterile saline to the cytology lab for cytomegalovirus inclusion stains and pneumocystis stains.
To examine demographic trends, this pediatric population was arbitrarily divided for comparison purposes into a group of patients undergoing biopsy before 1990 and a second group biopsied in 1990 or later.
Measurable endpoints included definitive diagnosis, usefulness of biopsy, and operative complications. Early mortality was defined as death within 30 days or the same hospitalization as the biopsy. A definitive diagnosis required the causative agent of the lung pathology to be identified by the pathologist, not just the observation of nonspecific inflammatory changes. Usefulness, however, was determined by either a change in therapy or the withholding of a planned therapy or diagnostic procedure in light of the results of the lung biopsy.
Infants were defined as being less than 1 year of age, neonates as less than 30 days of age, and prematurity as less than 36 weeks gestation. Patients were considered immunocompromised if they were currently receiving chemotherapy for a hematologic neoplasm, if they had a solid organ or bone marrow transplant and were taking immunosuppressive agents, or if they had an autoimmune disorder or a congenital immunodeficiency disease. A comorbid disease scale was applied to estimate the overall health of patients, which consisted of assigning an equal value of 1 to each of the following diseases: congenital heart disease, prematurity, bone marrow transplant, solid organ transplant, history of cancer, history of radiation therapy, previous pneumonectomy, glomerulonephritis, vasculitis, dermatomyositis, persistent fetal circulation, congenital immunodeficiencies, drowning, coarctation of the aorta, thrombotic disorder, sepsis, hematologic malignancy, hydrocephalus, and chromosomal abnormalities. Length of stay was defined as time from surgery for open-lung biopsy to same hospital discharge.
The Heath-Edwards (HE) grade for pulmonary microangiopathy [14] has been in use at this institution since the 1950s. Briefly, grade 1 is medial hypertrophy with no intimal reaction, grade 2 is medial hypertrophy with cellular lesions of the intima, grade 3 has fibrosis of the intima, and grade 4 includes plexiform and dilatation lesions. Available cardiac catheterization data were correlated with the HE grade of biopsies in patients with pulmonary hypertension.
Linear regression was used to correlate HE grade with cardiac catheterization data. Comparison between groups was made with Fisher’s exact test. The length-of-stay curves were sufficiently skewed to suggest a nonnormal distribution and are described using the nonparametric method of reporting quartiles.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
There were three main indications for open-lung biopsy in this group: 46 biopsies (72%) were to diagnose pulmonary parenchymal disease, 8 (12%) were to histologically grade the vasculopathy of pulmonary hypertension, and 10 (16%) were in patients with a combined clinical picture of pulmonary hypertension and simultaneous pulmonary parenchymal disease, with the severity of both unclear.
Midline sternotomies were used in 15% of cases (all combined with other operative procedures mandating that incision), left thoracotomies in 44%, and right thoracotomies in 41%. The left lower lobe was the site of biopsy in 19% of cases, the left upper lobe in 25%, the right lower lobe in 23%, the right middle lobe in 17%, and the right upper lobe in 16%. A chest tube was placed in 92% of patients. Fifty-two (81%) of the biopsies consisted of one sample of lung tissue, seven (11%) consisted of two samples, and five (8%) consisted of three or four samples of lung tissue. The number of samples and choice of lobe sampled were guided by the radiographic and visual appearance of lung pathology. Only 3 of the 64 biopsies (5%) included lung tissue from more than one lobe.
Twenty-five biopsies (39%) were performed on infants (less than 1 year of age). Thirty-three biopsies (52%) were in ventilator-dependent patients, and an extracorporeal membrane oxygenator was required to support eight (13%) of the patients. Of the 64 biopsies, 10 (16%) were performed in patients with no other illnesses, 36 (56%) in patients with one illness, 10 (16%) in patients with two illnesses, and 8 (13%) in patients with at least three other comorbid conditions. Eighteen biopsies (28%) were performed in immunocompromised patients: eight after a bone marrow transplant, four in patients with a hematologic malignancy and receiving chemotherapy, three in solid organ transplant patients, two in children with congenital immunosuppressive disorders, and one in a patient with an autoimmune disease.
Nineteen patients (30%) had congenital heart disease: three atrioventricular canal, four ventricular septal defects, one atrial septal defects, one atrial and ventricular septal defects, two transposition of the great vessels, one truncus, one double-outlet right ventricle, one common atrium and common ventricle, one total anomalous pulmonary venous connection, and four patent ductus. Of these 19, 7 patients had isolated pulmonary hypertension, an additional 9 had pulmonary hypertension as well as significant lung disease, and 3 patients had no evidence of pulmonary hypertension but underwent biopsy to diagnose their lung disease. Additionally, 2 other patients underwent lung biopsy to evaluate their pulmonary hypertension without congenital heart disease (a neonate with persistent fetal circulation and a 15-year-old child following thoracic radiation for a previous malignancy). Thus, 18 patients with pulmonary hypertension underwent lung biopsy.
Comparison between the 20 open-lung biopsies performed since 1990 and the 44 biopsies before 1990 revealed a difference in demographic indicators (Table 1). There were significantly more infants, more ventilator-dependent patients, more patients on extracorporeal membrane oxygenation (ECMO), and more patients with at least three other major illnesses in the 1990s. There were significantly fewer immunocompromised patients in the 1990s, and there was no statistical difference in the percentage of patients with congenital heart disease.
|
Univariate analysis with Fisher’s exact test showed no difference in the rate of definitive diagnoses in the years before 1990 compared with the years after 1990 (Table 2). Age, ventilator dependence, ECMO support, comorbid disease, and type of lung disease did not influence the ability to make a definitive diagnosis in open-lung biopsy. A significant difference in the ability to make a definitive diagnosis was found between the immunocompromised and the nonimmunocompromised patients. Also, open-lung biopsy to grade pulmonary hypertensive disease proved far more definitive than the ability to diagnose pulmonary parenchymal disease.
|
A correlation between mean pulmonary artery pressure and HE grade and between Rp/Rs and HE grade could not be shown (Fig 1). This implies that the open-lung biopsy may provide additional information to the cardiac catheterization data (ie, the presence or absence of fibrosis).
|
Of the 12 survivors, 8 patients with grade 1 or 2 HE changes underwent repair based on the biopsy results with no subsequent deaths, despite an average mean pulmonary artery pressure of 48 mm Hg within this group. One patient with grade 3 and 1 patient with grade 4 changes were not repaired. Two patients with grade 1 changes did not undergo repair at our institution.
Morbidity
Seven patients suffered a new morbid event after open-lung biopsy (11% overall operative morbidity). Four patients (6%) suffered sepsis syndrome within 72 hours postoperatively. All 4 of these patients had undergone open-lung biopsy for an infectious lung disease. Bacteremia occurred in an additional patient (2%) within 24 hours of operation. Two patients (3%) who had previously been ventilator independent required mechanical ventilation for 2 weeks postoperatively. Both of these patients had significant pulmonary hypertension. There were no cases of prolonged air leak. There were no cases of bleeding, even though 10 cases (16%) were performed in association with either ECMO or cardiopulmonary bypass.
Mortality
Nineteen patients died within 30 days of the same hospitalization, for an overall mortality of 30%. One patient died from the procedure, 13 died from their diseases, and 5 had support withdrawn.
Only one of the deaths seemed to be directly attributable to the open-lung biopsy procedure. This was a patient with severe pulmonary hypertension (Rp/Rs = 1.1; mean pulmonary artery pressure = 55 mm Hg), atrioventricular canal defect, and a patent ductus. She was admitted electively for ductus ligation and open-lung biopsy to evaluate the severity of pulmonary hypertensive disease, but she developed postoperative congestive heart and pulmonary failure and could not be weaned from the ventilator. This open-lung biopsy was done (1983) before the availability of nitric oxide, an agent that might have been beneficial in the treatment of this hypertensive crisis.
Thirteen patients (20%) died from the progression of their primary disease pathology or from significant comorbid disease. Specifically, 5 patients (8%) died of infection; 3 patients (5%) died of pulmonary failure; 2 (3%) died of leukemia; and 1 each (1.5% each) died from graft-versus-host disease, intraventricular hemorrhage on ECMO, and diffuse pulmonary hemorrhage 1 week after open-lung biopsy.
Univariate analysis failed to detect a difference between mortality before and after 1990 (Table 3). As expected, preoperative ventilator dependence was found to be an important predictor of death, and to a lesser degree the need for ECMO support before open-lung biopsy was correlated with subsequent death.
|
In the other 4 patients with irreversible processes, autopsies were obtained within 1 to 24 days following biopsy. All four autopsies confirmed the impression of irreversible lung damage noted on the open-lung biopsy. One neonate requiring ECMO support for 2 weeks had support withdrawn because of intracranial hemorrhage. The marked lung fibrosis and distal airway obliteration noted in one lobe at biopsy was noted in all lobes at autopsy. A 2-year-old boy developed pulmonary failure due to respiratory syncytial virus pneumonitis. Lung biopsy after 46 days of ECMO showed marked fibrosis of the intraalveolar septa. The inability to wean from ECMO after 7 weeks of support and the fibrosis of the lungs seen on lung biopsy initiated discussions to withdraw support. The patient did not survive once ECMO was withdrawn. Autopsy confirmed diffuse pulmonary fibrosis and alveolar destruction in all lobes. A 17-year-old patient developed respiratory syncytial virus pneumonitis after bone marrow transplant for acute nonlymphocytic leukemia. Lung biopsy showed severe fibrosis, and steroids were initiated. This patient died secondary to a stroke 24 days later. Autopsy confirmed severe pulmonary fibrosis in all lobes despite steroid therapy. A neonate required ECMO support for pulmonary hypertension of unclear source. He developed massive pulmonary hemorrhage after 4 days of ECMO. Lung biopsy showed thickened alveolar septum without capillary ingrowth consistent with alveolar-capillary dysplasia. Support was withdrawn 2 days later because of the massive pulmonary hemorrhage, persistent pulmonary hypertension with almost no pulmonary blood flow by echocardiography, and the supporting lung biopsy diagnosis of alveolar-capillary dysplasia. This diagnosis was confirmed on autopsy.
Among the 8 patients in this series supported by ECMO, 4 patients were believed to have an irreversible lesion at lung biopsy, and all 4 died. Four others were believed to have a reversible lesion, and 2 of these survived. One of these 4 died of intracranial hemorrhage, and the other was found to have irreversible fibrosis at autopsy 20 days later.
Length of stay
Length of hospital stay was primarily dictated by the severity of the primary disease pathology. The range of stay was from 1 day to 323 days. Median stay was 12 days. The 25th percentile was 6 days, and the 75th percentile was 23 days.
Neonates
Of the 25 infants, 9 were neonates (less than 1 month of age). Six of these 9 neonates (67%) died during the initial hospitalization, and 5 of these 9 required ECMO support. A definitive congenital cause of pulmonary failure was established in all 9 neonates by lung biopsy (two hyaline membrane disease, one Hammon-Rich syndrome, two venous occlusive disease, two alveolar capillary dysplasia, and two interstitial emphysema).
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
The demographics of the children referred for open-lung biopsy at this institution have changed within the past decade. There has been a decrease in the percentage of immunocompromised children, perhaps in recognition of the unfavorable risk to benefit ratio within this group. There has been a relative increase in the percentage of patients who are ventilator dependent, who are less than 1 year of age, and who have at least three other major illnesses. Some patients referred for open-lung biopsy have been so profoundly ill that they have required ECMO support. Despite this seeming decrease in the overall health of the most recently biopsied children, open-lung biopsy has continued to provide valuable information.
Previous literature is difficult to compare to our data because the definition of diagnostic utility has differed between authors. We tried to resolve the discrepancy by evaluating both the definitive diagnosis rate and the rate at which the biopsy led to a therapeutic or diagnostic change (which we called "usefulness"). Our 67% definitive diagnosis rate is comparable to previously reported rates in patients with parenchymal lung disease [2, 3, 7, 9]. Furthermore, our 98% usefulness rate is comparable to many previous reports of open-lung biopsies in children published between 10 and 20 years ago [1–7].
Similar to Doolin and colleagues [9], we found the effectiveness of lung biopsy in providing a definitive diagnosis for the immunocompromised child to be less than for the nonimmunocompromised. However, those authors claim that the biopsy was not helpful in any patient who subsequently died and was therefore useful in only 23% of cases. We disagree with that definition of usefulness and agree with Prober and colleagues [7] that the biopsy altered either therapy or a planned diagnostic workup in almost every patient. However, if one accepts Doolin and colleagues’ [9] definition of usefulness, then open-lung biopsy in this series of children was useful in only 44 of 64 patients (69%).
The discordance of the HE grade with pulmonary artery pressures and vascular resistances within our selected cohort is in contrast to the observations of Wilson and colleagues [13, 15]. This group found that HE grade reliably correlated with mean pulmonary artery pressure and Rp/Rs ratios in 59 patients with congenital heart disease and suggested that open-lung biopsy offered no new information over cardiac catheterization. Similar to other authors [16–18], however, we failed to find a correlation between pulmonary artery measurements at cardiac catheterization and HE grades of open-lung biopsies. Furthermore, 8 of our patients with favorable HE grades underwent successful repair in the face of poor prognostic cardiac catheterization data. This may reflect an increased pulmonary vasoreactivity in our patients with both congenital heart disease and lung disease. This group has not been specifically addressed in the literature before our study and may represent a distinct subgroup that particularly benefits from open-lung biopsy. We agree, however, with Wilson and colleagues [13, 15] that the risk of ventilator dependence and mortality is higher in this group of patients with pulmonary hypertensive disease, and therefore, lung biopsy should not be undertaken lightly. We did not utilize pulmonary wedge angiography at the time of cardiac catheterization in our patients, but some authors [15, 18, 19] have reported that this less invasive technique correctly predicts the HE grade 89% to 100% of the time in certain subgroups. These subgroups included 17 of the 41 patients (41%) examined by Wilson and colleagues [15] and 8 of the 19 patients with correlative lung biopsies in the group of Nihill and McNamara [18]. Routine use of this technique at the time of cardiac catheterization in patients with high pulmonary artery pressures may preclude the need for open-lung biopsy.
The retrospective nature of this study limits its direct applicability to future pediatric cohorts of open-lung biopsy candidates. The demographics of the children referred to us for biopsy have changed within the past decade, and observations within this cohort may not apply to future generations. Larger numbers of patients within this study might have improved the statistical ability to detect potential associations between preoperative risk factors and postoperative endpoints. This study has been able to predict the risk of death within the 30 days following lung biopsy in certain high-risk groups of patients without being able to predict how the mortality might be favorably influenced.
Ninety-five percent of our biopsies included tissue from a single lobe. Previous authors have emphasized the need to select from more than one lobe, because histologic changes are not always uniform, and to avoid the lingula and middle lobe, which are frequent sites of nonspecific scarring [20, 21]. These recommendations, however, are based on observations in adults with infiltrates that have been present on roentgenogram for more than a month. No data are available to suggest that similar precautions are necessary in the neonatal lung. In fact, two elegant studies have shown that when the pulmonary blood flow is uniform, biopsy tissue from a single lobe is representative of pulmonary hypertensive changes of the entire pulmonary vascular bed [22, 23]. Furthermore, the four autopsies in our group of patients in whom the magnitude of disease was considered irreversible confirmed the nonlocalized nature of these severe lung diseases in children.
Circumstantial evidence of irreversible lung disease was provided by open-lung biopsy in 6 patients. Our definition of irreversible in these cases involved markedly abnormal anatomy due to either scarring or agenesis of the distal pulmonary vascular bed. Declaring these changes to be irreversible and then withdrawing support can be viewed as a self-fulfilling prophecy. There is no study that we are aware of that documents the histologic changes of the healing lung supported by prolonged ECMO, and perhaps the degree of scarring in our biopsy samples was not truly irreversible. It has not been our practice to withdraw support based solely on a lung biopsy specimen. We do believe, however, that open-lung biopsy can be very useful in the patient who fails to improve after prolonged support on ECMO. Based on our data, we would recommend open-lung biopsy after 2 weeks of ECMO support without improvement. The histology and cultures can guide therapy if therapy is available. We have found that a profoundly abnormal biopsy in the face of a seemingly futile clinical course can facilitate discussions of prognosis.
Diagnostic surgical procedures are rare. The use of open-lung biopsy has conformed to new roles as the care of the pediatric intensive care unit patient has advanced. Subgroups of children who may benefit from open-lung biopsy have emerged, including children on ECMO support for 2 weeks and children with both pulmonary hypertension and parenchymal lung disease. Open-lung biopsy in the immunocompromised patient has a limited role. The emergence of nitric oxide as a therapy for reversible pulmonary hypertension may increase the need for open-lung biopsies to stratify patients according to the microscopic pathology of the lungs. Open-lung biopsy remains a useful technique to guide therapy in selected children.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Kanngiesser, F. Liewald, G. Halter, and L. Sunder-Plassmann Thoracic surgery in children Eur. J. Cardiothorac. Surg., July 1, 2005; 28(1): 50 - 55. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D Inwald, K Brown, F Gensini, M Malone, and A Goldman Open lung biopsy in neonatal and paediatric patients referred for extracorporeal membrane oxygenation (ECMO) Thorax, April 1, 2004; 59(4): 328 - 333. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Mavroudis and R. M. Sade The Southern Thoracic Surgical Association 50th anniversary celebration: the impact of STSA pediatric cardiothoracic surgery manuscripts on surgical practice Ann. Thorac. Surg., November 1, 2003; 76(90050): S47 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Gorenflo Open lung biopsy in patients with congenital heart disease. Ann. Thorac. Surg., April 1, 2002; 73(4): 1357 - 1358. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
