Conservative Versus Surgical Management of Chylothorax Boyd C. Marts, MD, Keith S. Naunheim, MD, Andrew C. Fiore, MD, D. Glenn Pennington, MD, St. Louis,Missouri Chylothorax is a potentially life-threatening disorder that has profound respiratory, nutritional, and immunologic consequences. Treatment for this problem is controversial, and the results have been variable. From 1985 to 1990, 29 patients (16 males, 13 females; mean age: 20.1 years; range: 5 days to 76.8 years) were diagnosed as having chylothorax (18 right, 6 left, 5 bilateral). Etiologies included surgical trauma (26) and nonsurgical trauma ( 3 ) . Initial conservative treatment consisted of tube thoracostomy drainage (mean duration: 13.3 days; range: 1 to 62 days; mean total volume: 4,030 mL) and dietary modification (low-fat diet, total parenteral nutrition). This resulted in resolution of the chylothorax in 23 patients (79% success), although 2 patients died of unrelated causes while hospitalized (myocardial infarction and cardiopulmonary arrest). Five adult patients and one infant (21%) required ligation of the thoracic duct, with resolution of the chylothorax in all six (100% success). Despite succeo~sful duct closure, one infant died of respiratory failure unrelated to the operation, and one adult died as the result of a cerebrovaseular accident 6 weeks postoperatively, yielding an operative mortality of 33% and an overall mortality of 14% (4 of 2 9 ) . Our experience demonstrates that initial treatment of chylothorax with thoracostomy drainage and dietary modification is successful in the majority of patients and is not associated with high morbidity or mortality rates. Surgical intervention for chylothoraces that fail to respond to initial conservative measures will be required in a minority of patients but appears to be associated with a higher risk of complications.
C
hylothorax is an infrequent but troublesome complication following surgical or traumatic insult to the chest. Management options include special enteral diets, parenteral nutrition, pleural sclerosis, and varying surgical procedures. These measures are successful in a high percentage of patients, but mortality remains significant. This study was undertaken to examine modern results of management for traumatic chylothorax and to attempt to identify risk factors for mortality. PATIENTS AND METHODS A computer search of discharge records was undertaken to identify patients labeled with an ICD-9 457.8 "chylothorax" code. The charts of these patients were then reviewed to identify those with chylothorax secondary to surgical or traumatic insult. Spontaneous chylothoraces and those secondary to malignancy with no history of preceding surgery or trauma were excluded. Inclusion in the study required confirmation of the diagnosis of chylothorax, including a description of a milky-appearing fluid, a pH between 7.4 and 7.8, triglyceride level greater than 110 mg/dL, fat globules seen on a Sudan III stain, or chylomicrons proven by electrophoresis. In addition, a specific gravity of greater than 1.012 or a high pleural fluid cell count with lymphocyte predominance were thought to provide confirmatory data. The medical records of these patients were reviewed, and the specific data were recorded, including age, gender, etiology of chylothorax, laterality of fluid, lowest serum albumin level during treatment, lowest absolute lymphocyte count during treatment, triglyceride level of the fluid, peak rate of drainage, the institution of total parenteral nutrition (TPN), and the complications of infection or the requirement for surgery. Statistical analysis for discrete variables was accomplished utilizing a X2 analysis or Fisher's exact test when appropriate. Continuous variables were analyzed using unpaired Student's t-tests; p values of less than 0.05 were considered significant.
RESULTS From January 1985 through July 1990, 29 patients were identified as having chylothorax following traumatic or surgical injury to the thoracic duct. There were 16 males (55%) and 13 females (45%), with a mean age of 20.1 years (range: 5 days to 76 years). Nine patients were adults, and the remainder were children 11 years of age or younger. The etiology of chylothorax in these patients is depictFrom the Departmentof Surgery, CardiothoracicDivision,St. Louis ed in Table L The most frequent cause of this complicaUniversityMedicalCenter,St. Louis,Missouri. tion was congenital heart surgery, especially the Fontan Requests for reprintsshouldbe addressedto KeithS. Naunheim, procedure. Esophagectomy for cancer was the next most MD, Department of Surgery,St. LouisUniversityMedical Center, frequent cause (17%), followed by trauma (10%) and 3635 Vistaat Grand Boulevard,St. Louis,Missouri63110-0250. Presentedat the 44th AnnualMeetingof the SouthwesternSurgi- miscellaneous thoracic procedures (14%). The chylocal Congress,Seottsdale,Arizona,April26-29, 1992. thorax was right-sided in 62% (18), left-sided in 21% (6), 532
THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER1992
MANAGEMENTOFCHYLOTHORAX
and bilateral in 17% (5). Initial conservative treatment consisted of tube thoracostomy drainage, which ranged in duration from 1 to 62 days and averaged 13.3 days. In Table II, the metabolic profile of patients during drainage is shown. The mean triglyceride level in the fluid tested was 308 4- 355 mg/dL (SD). The lowest serum albumin level recorded averaged 2.1 rag/alL, and the mean of the lowest absolute lymphocyte count was 856 cells/mm 3. After chest tube placement, 16 patients were begun on a low'fat diet, and the majority of these (9) were supplemented with medium-chain triglycerides. Chylothorax resolved in 14 of these 16 patients (88%) with this conservative therapy, but the remaining 2 patients underwent thoracic duct ligation. Four patients were treated with TPN alone, with resolution in two (50%) and surgery for the remaining patients. Finally, a low-fat diet supplemerited with TPN was utilized in nine patients, with resolution in seven (78%). There was no significant difference between these three regimens with regard to the incidence of resolution with conservative management (Figure I ). Infectious morbidity was encountered in eight patients (28%) and included pneumonia in six, urinary tract infection in three, and a viral upper respiratory tract infection in one. Overall hospital mortality in these 29 patients was 14% (4 of 29). Hospital deaths were due to sudden death in one infant with complex congenital heart disease and sudden cardiopulmonary arrest in a patient after decortication for empyema. The third patient who died had undergone an esophagectomy 6 weeks prior to her death, but her hospitalization had been complicated by a chylothorax, gastric outlet obstruction, and cervical anastomotic leak. The leak was healing when the patient suddenly experienced an episode of aspiration and severe hypoxic brain damage. The final patient who had a complex congenital heart disease developed pneumonia followed by sepsis and multi-organ failure. The 6 patients treated with thoracic duct ligation were compared with the 23 who were successfully treated via conservative therapy. The results (Table HI) reveal that surgery was undertaken in an older population who demonstrated higher peak chyle flow rates from their tube thoracostomies. The lowest serum albumin level in surgicaUy treated patients was significantly lower than that found in the conservatively treated group, and the mean absolute lymphocyte count tended to be lower, a trend that approached significance. Half of the surgically treated patients developed a significant postoperative infection compared with 22% of the conservatively treated patients. However, due to the small numbers, this did not reach significance. Similarly, there was no significant difference between the hospital mortality of the surgical group versus the conservatively treated group. The mean number of hospital days after the diagnosis of chylothotax was identical in the two groups. Univariate analysis was undertaken to identify predictors of hospital mortality in this patient population. There was no significant correlation between hospital mortality and age, gender, type of traumatic insult, laterality, lowest serum albumin level, lowest serum absolute lymphocyte count, triglyceride level, peak flow rate, utilization of TPN, occurrence of infection, or requirement for surgery.
TABLE I
Etiology of Chylothorax In 29 Patients Congenital heart surgery Modified Fontan procedure Atrial septal defect secundum Senning procedure Mustard procedure Artrioventricular Canal Atrial septal defect/patent ductus arteriosus Glen shunt Blalock-Taussig shunt
17 (59%) 9 1 1 1 1 1 1 2
Esophagectomy
5 (17%)
Trauma Motor vehicle accident Gunshot wound
3 (10%) 2 1
Miscellaneous thoracic Tracheoesophageal fistula Congenital left diaphragmatic hernia Right thoracic decortication and right lower Iobectomy T1-5 laminectomy and arteriovenous malformation resection
4 (14%) 1 1 1
Metabolic Profile
TABLE II
Mean • SD Lowest albumin level (mg/dL) Lowest lymphocyte level (/ram 3) Triglyceride level (mg/dL) Peak output (mL/kg/d)
2.1 • 856 • 308 • 26.4 •
0.9 503 355 25.4
Range 0.8-4.4 0-1,800 20-1,425 2-104
,,,~,Resolution
! Chylothorax 2 9 ~
\
Low fat + MCT 16 ~,~Surgery
TPN alone 4 4 ~
2 2
"Res~176 Surgery
Low fat + TPN
14
2 7
9~'Res~176 Surgery
2
Figure 1. Summary of outcomes stratified by methods of management. TPN = total parenteral nutrition; MCT = medium-chain trlglycerides.
TABLE III Comparison of Clinical Profiles Surgically Treated No. of patients 6 Age (y) 43.2 • 28.4 Peak flow (mL/kg/d) 54.6 _+40.2 Lowalbumin (mg/dL) 1.4 • 0.4 Absolute lymphocyte count 479 • 444 (/mm3) Infection 50% Hospital mortality 33% Hospital days 21.5 -+ 8.1
Conservatively Treated p Value 23 14.3 _ 23.8 19.1 -+ 13.6 2.6 • 0.9 961 -- 477
C
hylothorax is an infrequent but troublesome complication following surgical or traumatic insult to the chest. Management options include special enteral diets, parenteral nutrition, pleural sclerosis, and varying surgical procedures. These measures are successful in a high percentage of patients, but mortality remains significant. This study was undertaken to examine modern results of management for traumatic chylothorax and to attempt to identify risk factors for mortality. PATIENTS AND METHODS A computer search of discharge records was undertaken to identify patients labeled with an ICD-9 457.8 "chylothorax" code. The charts of these patients were then reviewed to identify those with chylothorax secondary to surgical or traumatic insult. Spontaneous chylothoraces and those secondary to malignancy with no history of preceding surgery or trauma were excluded. Inclusion in the study required confirmation of the diagnosis of chylothorax, including a description of a milky-appearing fluid, a pH between 7.4 and 7.8, triglyceride level greater than 110 mg/dL, fat globules seen on a Sudan III stain, or chylomicrons proven by electrophoresis. In addition, a specific gravity of greater than 1.012 or a high pleural fluid cell count with lymphocyte predominance were thought to provide confirmatory data. The medical records of these patients were reviewed, and the specific data were recorded, including age, gender, etiology of chylothorax, laterality of fluid, lowest serum albumin level during treatment, lowest absolute lymphocyte count during treatment, triglyceride level of the fluid, peak rate of drainage, the institution of total parenteral nutrition (TPN), and the complications of infection or the requirement for surgery. Statistical analysis for discrete variables was accomplished utilizing a X2 analysis or Fisher's exact test when appropriate. Continuous variables were analyzed using unpaired Student's t-tests; p values of less than 0.05 were considered significant.
RESULTS From January 1985 through July 1990, 29 patients were identified as having chylothorax following traumatic or surgical injury to the thoracic duct. There were 16 males (55%) and 13 females (45%), with a mean age of 20.1 years (range: 5 days to 76 years). Nine patients were adults, and the remainder were children 11 years of age or younger. The etiology of chylothorax in these patients is depictFrom the Departmentof Surgery, CardiothoracicDivision,St. Louis ed in Table L The most frequent cause of this complicaUniversityMedicalCenter,St. Louis,Missouri. tion was congenital heart surgery, especially the Fontan Requests for reprintsshouldbe addressedto KeithS. Naunheim, procedure. Esophagectomy for cancer was the next most MD, Department of Surgery,St. LouisUniversityMedical Center, frequent cause (17%), followed by trauma (10%) and 3635 Vistaat Grand Boulevard,St. Louis,Missouri63110-0250. Presentedat the 44th AnnualMeetingof the SouthwesternSurgi- miscellaneous thoracic procedures (14%). The chylocal Congress,Seottsdale,Arizona,April26-29, 1992. thorax was right-sided in 62% (18), left-sided in 21% (6), 532
THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER1992
MANAGEMENTOFCHYLOTHORAX
and bilateral in 17% (5). Initial conservative treatment consisted of tube thoracostomy drainage, which ranged in duration from 1 to 62 days and averaged 13.3 days. In Table II, the metabolic profile of patients during drainage is shown. The mean triglyceride level in the fluid tested was 308 4- 355 mg/dL (SD). The lowest serum albumin level recorded averaged 2.1 rag/alL, and the mean of the lowest absolute lymphocyte count was 856 cells/mm 3. After chest tube placement, 16 patients were begun on a low'fat diet, and the majority of these (9) were supplemented with medium-chain triglycerides. Chylothorax resolved in 14 of these 16 patients (88%) with this conservative therapy, but the remaining 2 patients underwent thoracic duct ligation. Four patients were treated with TPN alone, with resolution in two (50%) and surgery for the remaining patients. Finally, a low-fat diet supplemerited with TPN was utilized in nine patients, with resolution in seven (78%). There was no significant difference between these three regimens with regard to the incidence of resolution with conservative management (Figure I ). Infectious morbidity was encountered in eight patients (28%) and included pneumonia in six, urinary tract infection in three, and a viral upper respiratory tract infection in one. Overall hospital mortality in these 29 patients was 14% (4 of 29). Hospital deaths were due to sudden death in one infant with complex congenital heart disease and sudden cardiopulmonary arrest in a patient after decortication for empyema. The third patient who died had undergone an esophagectomy 6 weeks prior to her death, but her hospitalization had been complicated by a chylothorax, gastric outlet obstruction, and cervical anastomotic leak. The leak was healing when the patient suddenly experienced an episode of aspiration and severe hypoxic brain damage. The final patient who had a complex congenital heart disease developed pneumonia followed by sepsis and multi-organ failure. The 6 patients treated with thoracic duct ligation were compared with the 23 who were successfully treated via conservative therapy. The results (Table HI) reveal that surgery was undertaken in an older population who demonstrated higher peak chyle flow rates from their tube thoracostomies. The lowest serum albumin level in surgicaUy treated patients was significantly lower than that found in the conservatively treated group, and the mean absolute lymphocyte count tended to be lower, a trend that approached significance. Half of the surgically treated patients developed a significant postoperative infection compared with 22% of the conservatively treated patients. However, due to the small numbers, this did not reach significance. Similarly, there was no significant difference between the hospital mortality of the surgical group versus the conservatively treated group. The mean number of hospital days after the diagnosis of chylothotax was identical in the two groups. Univariate analysis was undertaken to identify predictors of hospital mortality in this patient population. There was no significant correlation between hospital mortality and age, gender, type of traumatic insult, laterality, lowest serum albumin level, lowest serum absolute lymphocyte count, triglyceride level, peak flow rate, utilization of TPN, occurrence of infection, or requirement for surgery.
TABLE I
Etiology of Chylothorax In 29 Patients Congenital heart surgery Modified Fontan procedure Atrial septal defect secundum Senning procedure Mustard procedure Artrioventricular Canal Atrial septal defect/patent ductus arteriosus Glen shunt Blalock-Taussig shunt
17 (59%) 9 1 1 1 1 1 1 2
Esophagectomy
5 (17%)
Trauma Motor vehicle accident Gunshot wound
3 (10%) 2 1
Miscellaneous thoracic Tracheoesophageal fistula Congenital left diaphragmatic hernia Right thoracic decortication and right lower Iobectomy T1-5 laminectomy and arteriovenous malformation resection
4 (14%) 1 1 1
Metabolic Profile
TABLE II
Mean • SD Lowest albumin level (mg/dL) Lowest lymphocyte level (/ram 3) Triglyceride level (mg/dL) Peak output (mL/kg/d)
2.1 • 856 • 308 • 26.4 •
0.9 503 355 25.4
Range 0.8-4.4 0-1,800 20-1,425 2-104
,,,~,Resolution
! Chylothorax 2 9 ~
\
Low fat + MCT 16 ~,~Surgery
TPN alone 4 4 ~
2 2
"Res~176 Surgery
Low fat + TPN
14
2 7
9~'Res~176 Surgery
2
Figure 1. Summary of outcomes stratified by methods of management. TPN = total parenteral nutrition; MCT = medium-chain trlglycerides.
TABLE III Comparison of Clinical Profiles Surgically Treated No. of patients 6 Age (y) 43.2 • 28.4 Peak flow (mL/kg/d) 54.6 _+40.2 Lowalbumin (mg/dL) 1.4 • 0.4 Absolute lymphocyte count 479 • 444 (/mm3) Infection 50% Hospital mortality 33% Hospital days 21.5 -+ 8.1
Conservatively Treated p Value 23 14.3 _ 23.8 19.1 -+ 13.6 2.6 • 0.9 961 -- 477
