- Email: [email protected]
Fatal Postoperative Pulmonary Edema
Fatal Postoperative Pulmonary Edema* : Pathogenesis and Literature Review Allen I. Arieff Chest 1999;115;1371-1377 DOI 10.1378/chest.115.5.1371 The online version of this article, along with updated information and services can be found online on the World Wide Web at: http://chestjournal.chestpubs.org/content/115/5/1371.full.html
Chest is the official journal of the American College of Chest Physicians. It has been published monthly since 1935. Copyright1999by the American College of Chest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights reserved. No part of this article or PDF may be reproduced or distributed without the prior written permission of the copyright holder. (http://chestjournal.chestpubs.org/site/misc/reprints.xhtml) ISSN:0012-3692
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
Fatal Postoperative Pulmonary Edema* Pathogenesis and Literature Review Allen I. Arieff, MD
Study objectives: Pulmonary edema is a known postoperative complication, but the clinical manifestations and danger levels for fluid administration are not known. We studied (1) 13 postoperative patients (11 adult, 2 pediatric) who developed fatal pulmonary edema, and (2) one contemporaneous year of inpatient operations at two university teaching hospitals to determine the clinical manifestations, causes, epidemiology, and guidelines for fluid administration. Design: Retrospective analysis of 13 patients with fatal postoperative pulmonary edema and one contemporaneous year of major inpatient surgery. Patients and methods: Thirteen patients had net fluid retention of at least 67 mL/kg in the initial 24 postoperative hours and developed pulmonary edema. Ten were generally healthy while three had serious associated medical conditions. Measurements and results: There was no measurement, laboratory value, or clinical finding predictive of impending pulmonary edema. The most common clinical manifestation following the onset of pulmonary edema was cardiorespiratory arrest (n 5 8). Patients had metabolic acidosis (pH 5 7.15 6 .33), hypoxia (PO2 5 45 6 18 mm Hg), and normal electrolytes. The diagnosis of pulmonary edema was established by chest radiograph and confirmed by autopsy and pulmonary artery pressure (21 6 4 mm Hg). The mean net fluid retention was 7.0 6 4.5 L (90 6 36 mL/kg/d) and exceeded 67 mL/kg/d in all patients. Autopsy revealed pulmonary edema with no other cause of death. Among 8,195 major operations, 7.6% developed pulmonary edema with a mortality of 11.9%. Extrapolation to the 8.2 million annual major surgeries in the United States yields a projection of 8,000 to 74,000 deaths. Conclusions: Pulmonary edema can occur within the initial 36 postoperative hours when net fluid retention exceeds 67 mL/kg/d. There are no known predictive warning signs and cardiorespiratory arrest is the most frequent clinical presentation. The monitoring systems currently in use neither detect nor predict impending pulmonary edema, and as yet, there are no known panic values for excessive fluid administration or retention. (CHEST 1999; 115:1371–1377) Key words: acute renal failure; cardiorespiratory arrest; hypoxia; metabolic acidosis; postoperative complications; pulmonary edema Abbreviation: TURP 5 transurethral resection of the prostate
edema is a reported postoperative P ulmonary complication, but the etiology and mortality are 1
not known.2–5 Although the usual clinical manifestations of pulmonary edema have been described For editorial comment see page 1224
extensively, the presentation of postoperative pulmonary edema is not known. There are several possible *From the Department of Medicine, University of California School of Medicine, San Francisco, CA. Supported by a grant RO1 AG 08575-01A2 from the National Institute on Aging, Department of Health and Human Services, Bethesda, MD. Manuscript received June 29, 1998; revision accepted August 4, 1998. Correspondence to: Allen I. Arieff, MD, Department of Medicine, University of California School of Medicine, 299 South St, Sausalito, CA 94965
etiologic factors that may lead to postoperative pulmonary edema. Available evidence suggests that the disorder will be most common in patients who have preexisting heart disease. Fluid overload with high hydrostatic pressures resulting in left ventricular dysfunction would probably be the most common sequence1,6 and some of these patients would likely have suffered myocardial infarction5,7 or renal failure. Other groups of patients with postoperative pulmonary edema would be those with neurogenic (noncardiogenic) pulmonary edema secondary to postoperative hyponatremic encephalopathy,8 head trauma,9 pheochromocytoma,10 or other causes.11,12 ARDS is often an important differential consideration.13–15 One frequent reason for administration of large volumes of postoperative fluids is to compensate for apparent fluid losses for which there is no clinically CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1371
accessible method to accurately evaluate. Such loss includes “third space” losses, evaporative losses, blood loss, and insensible loss. Replacement is generally based on clinical approximations.16,17 Currently, there is little information available concerning the maximum postoperative volume of fluid that can be administered safely. In a postoperative patient who does not have serious cardiovascular, hepatic, or renal disorders, the quantity of fluid necessary to induce pulmonary edema would vary according to such factors as age; body weight; tissue turgor; cardiovascular, pulmonary, and renal function; plasma vasopressin levels; plasma proteins; and the volume of the “third space.”3 The differential diagnosis of postoperative pulmonary congestion includes hyponatremic encephalopathy, pulmonary capillary leak syndromes (ARDS), pneumonia, sepsis, and volume overload,18 –20 as well as other noncardogenic causes of pulmonary edema.4,8,10,11 Furthermore, it is unclear from the literature what the usual clinical manifestations of postoperative pulmonary edema are, or if it can result in death or serious morbidity in patients without heart disease or renal failure.18,21,22 In the current study, we present 13 fatal cases of postoperative pulmonary edema in generally healthy individuals in whom the pulmonary edema was secondary to excessive fluid administration.
Materials and Methods There were 13 study patients who, from 1991 to 1996, had postoperative pulmonary edema documented by clinical criteria and characteristic findings on chest radiograph. The diagnosis was confirmed by autopsy in nine patients and by pulmonary artery catheterization in four. Demographic features are shown in Table 1. There were four men and nine women, and the average age (6 SD) was 38 6 21 years. The operations are shown in Table 1. All but one operation (patient 9) were elective. All patients were seen in consultation at university teaching hospitals (n 5 9) or affiliated community hospitals (n 5 4). From a review of the literature, a net fluid retention of 2.2 L/d can lead to pulmonary edema,1,3,16 so patients were considered for study only if their net fluid retention was at least 2.2 L during the initial 24 postoperative hours. In four patients, the pulmonary capillary wedge pressure was measured by means of a pulmonary artery catheter. In the “Results” and “Discussion” sections, “cholecystectomy” refers to an open abdominal procedure rather than a laparoscopic one.3 The medical condition leading to hospitalization is shown in Table 1. Three patients had severe associated medical conditions: multiple traumatic injuries; acute pancreatitis; acute glomerulonephritis. However, all 13 patients were in stable condition and admitted to the hospital for elective surgical procedures, including tonsillectomy, knee-joint replacement, endometrial ablation, and sacral laminectomy. Only one patient had known heart disease (aortic valvular insufficiency), and 12 of 13 had normal results of heart examination on preoperative physical examination, chest radiograph, and ECG. Every patient received IV fluids preoperatively, during surgery, 1372
and postoperatively. The intake and output values are shown in Table 2 and are the totals from admission to the hospital to the time pulmonary edema was initially diagnosed. The fluid retention index is the net fluid retention for the initial 24 postoperative hours in milliliters per kilogram per day. The output values were extracted from the inpatient records and represent the sum of urine, estimated blood loss, and tube drainage. IV fluid intake was also extracted from the inpatient records. The predominant intraoperative IV fluid was Ringers lactate (130 mmol/L of sodium), others being blood, blood products, or 154 mmol/L NaCl. No patient received IV hypotonic fluid intraoperatively. Three patients had endoscopic surgery and received parenteral intraoperative hypotonic irrigating fluids by absorption through the cauterized operative bed, either the uterus or prostate.23 The irrigating fluids were either 200 mmol/L glycine or 165 mmol/L sorbitol. In the three patients who had endoscopic surgery, the intake was determined as the sum of IV fluids plus irrigating fluids, while output was the sum of urine volume, estimated blood loss, and collected drainage of irrigating fluid from the operative field. The operative drainage was collected in stainless steel vessels. Tube drainage was collected in a polyethylene bag by gravity and the volume was measured by means of the gradations on the bags. Three patients had indwelling Foley catheters postoperatively, while in the other 10 patients, urine was collected by spontaneous voiding. The postoperative fluids consisted of either Ringer’s lactate, 154 mmol/L NaCl, or 77 mmol/L NaCl. Pulmonary edema occurred within three postoperative days (range of 3 to 66 h). Epidemiologic Studies The records of major inpatient surgery for one contemporaneous year (1993) from two university teaching hospitals were evaluated for (1) the incidence and mortality of pulmonary edema, and (2) the incidence of other major postoperative complications in patients with pulmonary edema (pulmonary embolus, pneumonia, acute renal failure, GI bleeding, stroke, myocardial infarction, and hyponatremia). We examined only major surgical procedures; diagnostic procedures and minor surgery were not tabulated.24 A computer search of hospital records for the 8,195 hospital inpatients operated on in 1993 was done using a database (SAS; Cary, NC) to determine which surgical patients had postoperative pulmonary edema and their outcome.25 The incidence of seven other postoperative complications—pulmonary embolus, pneumonia, acute renal failure, GI bleeding, stroke, myocardial infarction, and hyponatremia (plasma sodium , 130 mmol/L)—were also tabulated. Outcome was determined only as death or discharge from the hospital within 30 days of the episode of pulmonary edema.5 Patients dying with pulmonary edema plus one of the other seven postoperative complications were subtracted from the total deaths with pulmonary edema to estimate the number of patients who died from pulmonary edema only. In addition to the information on patients with pulmonary edema, the overall postoperative mortality was determined.
Results Postoperative Patients With Fatal Pulmonary Edema All results are presented as mean 6 SD. In eight patients, the initial manifestation of pulmonary edema was cardiopulmonary arrest requiring emergent intubation (Table 1). One patient had intraopClinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
Table 1—Patients With Pulmonary Edema Patient/Sex/Age, yr/Weight, kg
Operations
Other Medical and Surgical Conditions
1/M/58/88.6
TURP
Prostatic hypertrophy
2/F/52/52.4
Aortic valve replacement
Aortic valve insufficiency
3/M/58/59
Open cholecystectomy
Cholelithiasis, gangrenous gall bladder
4/M/8/36.8
Myringotomy
5/F/63/68.2
Replacement of both knee joints
Otitis media, upper respiratory tract infection Osteoarthritis of knees
6/F/20/63.6
Open cholecystectomy
Acute pancreatitis, acute renal failure
7/F/67/51.4
Sacral laminectomy
Herniated sacral disk
8/M/31/74.8
Bronchoscopy, placement of dialysis graft Exploratory laparotomy, splenectomy
Acute glomerulonephritis, acute renal failure Motor vehicle accident, multiple traumatic injuries; ruptured spleen, fractures of ribs and arm Otitis media
9/F/19/63.2
10/F/3/15
11/F/39/60.5
Tonsillectomy, adenoidectomy, myringotomy Endometrial ablation
12/F/39/69.3
Exploratory laparotomy
Gastroenteritis, dehydration
13/F/31/54.3
Endometrial ablation
Dysfunctional uterine bleeding
Menorrhagia
Clinical Presentation Intraoperative chest pain Acute florid pulmonary edema (rales, dyspnea, pink frothy sputum) Unwitnessed cardiopulmonary arrest Acute florid pulmonary edema, shock Unwitnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest Witnessed cardiopulmonary arrest Acute renal failure Combative, acute florid pulmonary edema (rales, dyspnea, pink frothy sputum) Unwitnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest Witnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest
Mean 6 SD 9/F/38/58 4/M/21/18
Pulmonary Edema Diagnosed by* Chest radiograph, autopsy Chest radiograph, autopsy, PCWP 5 22 mm Hg
Autopsy
Chest radiograph, autopsy Chest radiograph, PCWP 5 18 mm Hg Chest radiograph, PCWP 5 19 mm Hg Chest radiograph
Chest radiograph, autopsy PCWP 5 23 mm Hg, chest radiograph, autopsy
Chest radiograph, autopsy
Chest radiograph, autopsy
Chest radiograph
Chest radiograph, autopsy
PCWP (mm Hg) 5 20 6 2
*PCWP 5 pulmonary capillary wedge pressure; M 5 male; F 5 female.
erative chest pain, three had acute florid pulmonary edema, and one had acute renal failure. The diagnosis of acute pulmonary edema was established by a diagnostic chest radiograph accompanied by typical physical findings of bilateral rales, a third heart sound, and copious pink frothy sputum in three patients (Table 1). Two patients died before the diagnosis of pulmonary edema was established and in these patients, the diagnosis was established at autopsy. The five patients not presenting with cardiorespiratory arrest were intubated and had a chest radiograph that demonstrated findings of diffuse pulmonary edema. At the time of respiratory arrest or when pulmo-
nary edema was first diagnosed (n 5 11), arterial blood gases revealed a mixed metabolic-hypercapnic acidosis (Table 2), with pH 5 7.15 6 0.33, Pco2 5 58 6 37 mm Hg, and bicarbonate 5 18.5 6 4.8 mmol/L. All patients had hypoxemia, with Po2 5 45 6 18 mm Hg. These values are similar to those previously reported in patients with cardiogenic pulmonary edema.26 IV Fluids The average amount of IV fluid administered for the initial 27 postoperative hours was 9.9 6 6.0 L (171 6 103 mL/kg) (Table 2). The mean patient CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1373
Table 2—Laboratory Values
Mean 6 SD
Plasma Na, mmol/L
Plasma K, mmol/L
135 6
4.2 0.8
Time After Surgery When Pulmonary Edema Noted, h Mean
27
6 SD
20
Plasma Cl, mmol/L
Plasma Bicarbonate, mmol/L
Arterial pH
Arterial Po2, mm Hg
Arterial Pco2, mm Hg
Urea, mmol/L
18.5 4.8
7.15 0.33
45 18
58 37
11.3 17.7
Fluid Output, L
Fluid Retention, L
Fluid Retention Index, mL/kg/d
2.88 4.36
7.04 4.52
90 36
103 7
Fluid Intake, L 9.92 6.03
weight was 58 6 18 kg and using available tables, the calculated total body water was 29.8 6 8.9 L.27 The average net fluid retention was 7.04 6 4.5 L over the initial 27 6 20 postoperative hours, a quantity of retained fluid equal to 24% of the total body water or 12.2% of body weight. The fluid retention index (Table 2) was 90 6 36 mL/kg/d, almost three times the “normal” value (Table 3) of 38 mL/kg/d. The net fluid retention corresponded to an increase of 24% of total body water or 12.2% of body weight (Table 2). Thus, the quantity of administered fluid was clearly excessive.1,17,28,29 The reasons for the excessive quantities of administered fluid were not readily apparent. Recording of intake and output was ordered in all 13 patients. The orders were carried out, but in all patients, totals could be ascertained only by tedious searching of multiple sources: operative records, nurses notes, graphic records, physicians orders, and billing records. However, the excessive net fluid intake that was charted was not appreciated by the physicians caring for any of the patients until after pulmonary edema was diagnosed. In no instance was any physician notified that his or her patient had been administered or retained an excessive amount of fluid. Postoperative weights were ordered in all patients, but they were not actually obtained in any patient before the onset of pulmonary edema. Eight patients suffered cardiopulmo-
Right Lung Weight, g
Creatinine, mg/dL 1.2 1.1
Left Lung Weight, g
906 154
726 171
nary arrest, and in all eight, the net excess of administered fluid was not appreciated until after the arrest. In eight patients, the diagnosis of pulmonary edema was confirmed at autopsy (Table 1). All eight had pulmonary edema, with no other cause of death. In particular, fat emboli were not present and none had acute myocardial infarction. In these eight patients, lung weights were 906 6 154 g (right) and 726 6 171 g (left), more than twice the expected normal values.30 Epidemiologic Studies of Postoperative Pulmonary Edema During 1993, there were 8,195 major inpatient operations performed at two university teaching hospitals. The overall operative and postoperative mortality from all causes was 3.88% (302 patients). There were a total of 621 cases of postoperative pulmonary edema, an incidence of 7.6%. Among the 621 cases of pulmonary edema, 74 patients (11.9%) died. Those patients with pulmonary edema were then evaluated for the occurrence of other comorbid conditions that would be expected either to increase the chances of developing pulmonary edema or increase its mortality. The conditions were as follows: acute myocardial infarction, acute renal failure, acute GI bleeding, stroke, pneumonia, pulmonary
Table 3—Fluid Losses During and After Surgery
Type of Fluid Loss Blood loss Operative evaporation Insensible loss Third space loss Urine output Tube drainage
1374
Method of Quantitative Evaluation Clinical estimation Clinical estimation Clinical estimation Clinical estimation Actual measurement Actual measurement
Usual Range, mL/h
Typical Estimated Values for a Major Procedure Such as Open Cholecystectomy, mL
Varies according to operation 150 25 200 70 Varies according to operation Total losses
250 400 100 1,000 250 200 2,200
Clinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
embolus, and hyponatremia (plasma sodium , 130 mmol/L). When the patients with pulmonary edema plus one other comorbid postoperative complication were subtracted from the total, there were 204 patients remaining who had pulmonary edema without any of the other comorbid conditions (2.6%), of whom 8 (3.9%) died. In the United States in 1993, there were 8.2 million major inpatient surgical procedures.24 If the pulmonary edema figures were extrapolated from those 8.2 million operations, there would be a projected 622,000 annual cases of postoperative pulmonary edema in the United States, with 74,000 deaths. Even if all patients with comorbid factors are subtracted, there would still be projected 213,200 annual cases of postoperative pulmonary edema that had no other postoperative complications, totaling 8,315 deaths. Thus, the projected range of deaths from postoperative pulmonary edema would be 8,300 to 74,000 per year.
Discussion These data demonstrate that in generally healthy individuals undergoing elective surgery, pulmonary edema may be the initial clinical manifestation of fluid overload. If the pulmonary edema is not recognized and treated promptly, the patients may not survive. Although most textbooks of surgery recognize that excessive postoperative fluid administration can cause pulmonary edema, essentially no guidelines are available as to the quantity of fluid that may lead to such a complication.17 There are many reasons for administration of large quantities of both intraoperative and postoperative fluids.3 Some of the more common reasons are as follows: (1) hypotension following induction of anesthesia; (2) fluid sequestration in the alleged “third space” during surgery; (3) maintainance of BP after traumatic injury;4,13 (4) excessive blood loss; and (5) postoperative fever with increased insensible loss. Intraoperatively, hypotension may develop after the onset of anesthesia, requiring fluids and vasopressor therapy. In patients with cardiac disease, the afterload reduction secondary to the effects of anesthesia on sympathetic tone may improve myocardial performance transiently and allow the administration of large quantities of IV fluids. However, the return of vascular resistance to normal as the anesthetic recedes can result in congestive heart failure and pulmonary edema due to the relative fluid overload.16 The net fluid retention in these patients was . 7 L/d (Table 2). There are a number of different modalities and formulas for the calculation of postoperative fluid requirements in adults17,28,31 (Table 3). These include maintenance requirements esti-
mated at 2,500 mL/d, 30 to 35 mL/kg/d, or 1 mL/calorie/d. In addition, blood losses and tube drainage (nasogastric, chest tube, Penrose drain) are replaced and 1 to 3 L of extra fluid may be administered after a major complicated procedure in order to make up for estimated evaporative and third space losses. Operative fluid losses by evaporation are usually estimated at 150 mL/h. Using any of the aforementioned criteria, these 13 patients should have received from 2.5 to 5.5 L in the initial 36 postoperative hours. Based on any of the commonly used formulas for determining postoperative fluid requirements, the quantity administered (9.9 6 6 L; Table 2) was excessive. In general, most textbooks recommend that the patients’ actual losses be assessed at the bedside, and fluid administration should be based on estimated maintenance requirements plus estimated losses.17,29,32 A gain in fluid above 20% of total body water has been classified as fluid overload.1 Assuming that the patients’ net fluid retention was associated with a corresponding gain in body weight, the mean weight gain was 12.2%, a quantity that is by itself associated with a mortality of 19%.1 The “third space” is an area in the body that is occupied by fluids but is not in equilibrium with the bloodstream and thus hemodynamically inactive. Examples of the third space include burns, traumatized operative bed, and traumatically injured tissues. There is no bedside method to measure third space losses, and replacement is usually based on clinical approximation. However, the postoperative sequestered extracellular fluid has been estimated to vary from an imperceptible minimum to as much as 3 L.17 Ranges of 5 to 17 mL/kg/h of surgery after cholecystectomy or similar procedure during the initial postoperative day are typical estimated replacement values, with 500 to 1,000 mL/d being about average.31,32 Fluids in the third space may return to the vascular compartment several days after surgery because of the changes in physiology resulting in intravascular reaccumulation, which may result in pulmonary edema. In addition, operative blood loss is usually estimated by the operating surgeon to be 15 to 40% less than the isotopically measured blood loss from that patient.17 Although insensible loss is typically about 600 mL/d for an adult, postoperative hyperventilation, fever, and hypermetabolism may increase this substantially (1,500/d).17 However, in the postoperative patient who is not eating, there is ordinarily no net insensible loss,16 and this fact is often neglected when considering replacement, resulting in additional fluid excess. Additional surgical fluid losses that are difficult to calculate are the following: additional insensible fluid losses by evaporation from the exposed peritoneal (or pleural) surfaces; and evaporative losses by virtue of operaCHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1375
tive drapes and from the administration of dry inhaled gases. Actual data are limited, but during laparotomy, the loss is about 150 mL/h or about 300 mL for an uncomplicated cholecystectomy.33–35 Most protocols for postoperative fluid therapy recommend about 30 to 35 mL/kg/d plus estimated losses. For a 65-kg individual following cholecystectomy, the total for the initial postoperative day might thus be calculated as 2,275 mL (35 mL/kg maintenance) 1 600 mL (insensible loss) 1 1,000 mL (third space loss) 1 300 mL (evaporative loss) 5 4.2 L. Virtually all protocols for postoperative fluid management recommend frequent monitoring of the patient’s volume status, utilizing such readily available clinical parameters as BP, heart rate, urine output, auscultation of the lungs, changes in body weight, and calculation of net fluid retention.17,28,31,32 If fluid retention is noted to be excessive, then the rate of administration can be decreased. In the current series, values for heart rate, BP, pulse, and urine output were generally recorded in the nurses notes or graphic sheets. However, not a single patient had a postoperative weight recorded before pulmonary edema was diagnosed and after discharge from the recovery room, auscultation of the lungs by a physician was never recorded in any patient. These observations suggest that among physicians caring for these patients, the index of suspicion concerning possible consequences of excessive fluid administration was not high. Since to our knowledge there are no established guidelines for “panic values” or their equivalent for fluid retention, the nursing staff did not alert physicians as to possible excessive fluid retention. When postoperative pulmonary edema has occurred in generally healthy individuals, the mortality has often been in excess of 50%, as occurred in the present study.1,3 The parameters that are included in the physician’s calculation of replacement plus maintenance fluids include estimations of parameters shown in Table 3. All of the five parameters not only represent “estimates,” but it is generally agreed that the estimates vary widely among different physicians.17 It would appear to be important that “panic” values for fluid retention be established for different categories of surgical procedures. Postoperative pulmonary edema has been a reported complication in generally healthy postoperative subjects receiving known amounts of IV fluids.16 Among 161 postoperative patients, 27% of those retaining 2.2 L per 24 h developed pulmonary edema.16 However, since a number of these patients also had postoperative hyponatremic encephalopathy, the pulmonary edema may have been neurogenic in some cases. In the present study, neurogenic pulmonary edema secondary to postoperative hy1376
ponatremic encephalopathy should not have occurred, because the plasma sodium level was normal (135 6 6 mmol/L). The incidence of postoperative pulmonary edema is uncertain. In generally elderly, mostly male patients and those with heart disease, the incidence was 2.3 to 4.8%, less than the 7.6% in the current study.5,21,22 In 14,707 generally low-risk female patients undergoing hysteroscopic hysterectomy, the incidence was only 0.2%.36 In none of these studies were mortality figures available.5,21,22,36 There may have been patients in the current study who had undocumented silent postoperative myocardial ischemia.7 Major causes of perioperative morbidity and mortality are myocardial infarction and ischemia.18,21,22 In response, there is now ever more sophisticated perioperative monitoring of cardiac function.7 However, fluid monitoring is generally carried out in an unsophisticated manner, and this appears to have contributed to the mortality from pulmonary edema (Table 3). In three patients, pulmonary edema occurred following endoscopic procedures, in which the operative field was irrigated with copious quantities of fluid. Although the staff attempted to evaluate intake and output, such a task is technically difficult. Monitoring of fluid retention during transurethral resection of the prostate (TURP) can be accomplished by the ethanol method or monitoring of body weight,37,38 and a device is available for fluid monitoring during hysteroscopic hysterectomy,39 but such procedures are not in wide use. The most widely accepted method of monitoring intravascular volume status is by pulmonary artery catheter, but this procedure has been criticized recently because of associated morbidity.40,41 The incidence of pulmonary edema following hysteroscopic hysterectomy (endometrial ablation) is , 1% and the number of deaths is not known.36,42 The incidence of pulmonary edema following TURP is unknown, but there are at least five reported cases with one death.43– 45 In 25 patients undergoing TURP, the volume of irrigating fluid absorbed (6 SD) was found to be 2.1 6 0.9 L.38 In five patients undergoing hysteroscopic hysterectomy who developed fluid overload, the volume of irrigating fluid absorbed was 5 to 7 L.23 Retention of such volumes of fluid can lead to pulmonary edema in otherwise healthy subjects16 (Table 2), emphasizing the necessity for monitoring net fluid retention during such procedures.
References 1 Lowell JA, Schifferdecker C, Driscoll DF, et al. Postoperative fluid overload: not a benign problem. Crit Care Med 1990; 18:728 –733 Clinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
2 van Hoozen BE, van Hoozen CM, Alberton TE. Pulmonary considerations and complications in the neurosurgical patient: pulmonary edema. In: Youmans JR, ed. Neurological surgery. Philadelphia, PA: WB Saunders, 1996; 624 – 626 3 Roth E, Lax LC, Maloney JV. Ringer’s lactate solution and extracellular fluid volume in the surgical patient: a critical analysis. Ann Surg 1969; 169:149 –164 4 Finn JC, Rosenthal MH. Pulmonary edema in trauma and critically ill patients. Semin Anesth 1989; 8:265–274 5 Khuri SF, Daley J, Henderson W, et al. The National Veterans Administration surgical risk study: risk adjustment for the comparative assessment of the quality of surgical care. J Am Coll Surg 1995; 180:519 –531 6 Katz AM. Cardiomyopathy of overload. N Engl J Med 1990; 322:100 –110 7 Mangano DT, Browner WS, Hollenberg M, et al. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery. N Engl J Med 1990; 323:1781–1788 8 Ayus JC, Arieff AI. Pulmonary complications of hyponatremic encephalopathy: noncardiogenic pulmonary edema and hypercapnic respiratory failure. Chest 1995; 107:517–521 9 Becker DP, Gade GF, Young HF, et al. Diagnosis and treatment of head injury in adults: fluids and electrolytes. In: Youmans JR, ed. Neurological surgery. Philadelphia, PA: WB Saunders, 1990; 2113–2115 10 Suga K, Tsukamoto K, Nishigauchi K, et al. Iodine-123MIBG imaging in pheochromocytoma with cardiomyopathy and pulmonary edema. J Nucl Med 1996; 37:1361–1364 11 Malik AB. Mechanisms of neurogenic pulmonary edema. Circ Res 1985; 57:1–18 12 Frim DM, Wollman L, Evans AB, et al. Acute pulmonary edema after low-level air embolism during craniotomy. J Neurosurg 1996; 85:937–940 13 Falk JL, Rackow EC, Weil MH. Colloid and crystalloid fluid resuscitation. In: Shoemaker WC, Ayres S, Grenvik A, eds. Textbook of critical care. Philadelphia, PA: WB Saunders, 1989; 1055–1073 14 Baumgartner F, Omari B, Lee J, et al. Survival after trauma pneumonectomy: the pathophysiologic balance of shock resuscitation with right heart failure. Am Surg 1996; 62:967–972 15 Christenson JT, Aeberhard JM, Badel P. Adult respiratory distress syndrome after cardiac surgery. Cardiovasc Surg 1996; 4:15–21 16 Rosenthal MH, Arieff AI. Fluid and electrolyte therapy in critically ill patients and those who are pre-, post- or intraoperative. In: Arieff AI, DeFronzo RA, eds. Fluid, electrolyte and acid-base disorders. New York, NY: Churchill Livingstone, 1995; 597– 632 17 Shires GT, Shires GTI, Lowry SF. Fluid, electrolyte, and nutritional management of the surgical patient. In: Schwartz SI, Shires GT, Spencer FC, eds. Principles of surgery. New York, NY: McGraw-Hill, 1994; 61–93 18 Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990; 72:153–184 19 Bishop MH, Jorgens J, Shoemaker WC, et al. The relationship between ARDS, pulmonary infiltration, fluid balance, and hemodynamics in critically ill surgical patients. Am Surg 1991; 57:785–792 20 Connors AF. Invasive measurements are required for assessing hemodynamic status in critically ill patients. Am Rev Respir Dis 1988; 138:1070 –1075 21 Rao TLK, Jacobs KH, El-Etr AA. Reinfarction following anesthesia in patients with myocardial infarction. Anesthesiology 1983; 59:499 –505 22 Goldman L, Caldera DL, Nussbaum SR. Multifactorial index
23 24
25 26 27
28
29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45
of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977; 297:845– 850 Ayus JC, Arieff AI. Glycine-induced hypo-osmolar hyponatremia. Arch Intern Med 1997; 157:223–226 National Health Survey. Vital and health statistics: detailed diagnoses and procedures, National Hospital Discharge Survey, 1995. Hyattsville, MD: National Center for Health Statistics, 1997; 146 SAS Institute I. SAS language reference, version 6, 1st ed. Cary, NC: SAS Institute, 1990; 1290 Aberman A, Fulop M. The metabolic and respiratory acidosis of acute pulmonary edema. Ann Intern Med 1972; 76:173–184 Park R, Guisado R, Arieff AI. Nutrient deficiencies in man and animals: water. In: Rechcigl M Jr, ed. CRC handbook of nutrition and food: section E, nutritional disorders. West Palm Beach, FL: CRC Press, 1978; 363– 400 Darby JM, Nelson PB. Fluid, electrolyte, and acid-base balance in neurosurgical intensive care. In: Andrews BT, ed. Neurosurgical intensive care. New York, NY: McGraw-Hill, 1993; 133–162 Drucker RP. Fluid therapy. In: Davis JH, Drucker RP, Foster RS, et al, eds. Clinical surgery. Washington, DC: CV Mosby, 1987; 847– 851 Whimster WF, Macfarlane AJ. Normal lung weights in a white population. Am Rev Respir Dis 1974; 110:478 – 483 Tonnesen AS. Crystalloids and colloids: third space losses. In: Miller RD, ed. Anesthesia. New York, NY: Churchill Livingstone, 1990; 1439 –1465 Proctor HJ, Rutledge R. Fluid and electrolyte management. In: Hardy JD, Kukora JS, Pass HI, eds. Hardy’s textbook of surgery. Philadelphia, PA: JB Lippincott, 1988; 15–29 Virtue RW, LeVine DS, Aikawa JK. Fluid shifts during the surgical period: RISA and 35S determinations following glucose, saline or lactate infusion. Ann Surg 1966; 163:523–528 Giesecke AH, Egbert LD. Perioperative fluid therapy— crystalloids. In: Giesecke AH, Egbert LD, eds. Anesthesia. New York, NY: Churchill-Livingstone, 1986; 1313–1328 Roberts JP, Roberts JD, Skinner C. Extracellular fluid deficit following operation and its correction with Ringer’s lactate: a reassessment. Ann Surg 1985; 202:1– 8 Baggish MS, Sze EH. Endometrial ablation: a series of 568 patients treated over an 11 year period. Am J Obstet Gynecol 1996; 174:908 –913 Lyon RP, St Lezin M, Thomas C, et al. Monitoring of body weight during transurethral resection of the prostate: preliminary report. J Endourol 1994; 8:161–163 Hahn RG, Ekengren JC. Patterns of irrigating fluid absorption during transurethral resection of the prostate as indicated by ethanol. J Urol 1993; 149:502–506 Hawe JA, Chien PFW, Martin D, et al. The validity of continuous automated fluid monitoring during endometrial surgery— luxury or necessity? Br J Obstet Gynaecol 1998; 105:797–801 Weil MH. The assault on the Swan-Ganz catheter. Chest 1998; 113:1379 –1386 Kefalides P. Pulmonary artery catheters on trial. Ann Intern Med 1998; 128:161–162 Hulka JF, Peterson HA, Phillips JM, et al. Operative hysteroscopy: American Association of Gynecologic Laparoscopists’ 1993 membership survey. J Am Assoc Gynecol Laparosc 1995; 2:131–132 Drinker HR, Shields T, Grayhack JT, et al. Simulated transurethral resection in the dog: early signs and optimal treatment. J Urol 1963; 89:595– 602 Bulkley GJ, O’Connor VJ, Sokol JK. Overhydration during transurethral prostatic resection. JAMA 1954; 161:1042 Aasheim GM. Hyponatraemia during transurethral surgery. Can Anaesth Soc J 1973; 20:274 –280 CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1377
Fatal Postoperative Pulmonary Edema* : Pathogenesis and Literature Review Allen I. Arieff Chest 1999;115; 1371-1377 DOI 10.1378/chest.115.5.1371 This information is current as of May 14, 2012 Updated Information & Services Updated Information and services can be found at: http://chestjournal.chestpubs.org/content/115/5/1371.full.html References This article cites 32 articles, 7 of which can be accessed free at: http://chestjournal.chestpubs.org/content/115/5/1371.full.html#ref-list-1 Cited Bys This article has been cited by 14 HighWire-hosted articles: http://chestjournal.chestpubs.org/content/115/5/1371.full.html#related-urls Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.chestpubs.org/site/misc/reprints.xhtml Reprints Information about ordering reprints can be found online: http://www.chestpubs.org/site/misc/reprints.xhtml Citation Alerts Receive free e-mail alerts when new articles cite this article. To sign up, select the "Services" link to the right of the online article. Images in PowerPoint format Figures that appear in CHEST articles can be downloaded for teaching purposes in PowerPoint slide format. See any online figure for directions.
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
Chest is the official journal of the American College of Chest Physicians. It has been published monthly since 1935. Copyright1999by the American College of Chest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights reserved. No part of this article or PDF may be reproduced or distributed without the prior written permission of the copyright holder. (http://chestjournal.chestpubs.org/site/misc/reprints.xhtml) ISSN:0012-3692
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
Fatal Postoperative Pulmonary Edema* Pathogenesis and Literature Review Allen I. Arieff, MD
Study objectives: Pulmonary edema is a known postoperative complication, but the clinical manifestations and danger levels for fluid administration are not known. We studied (1) 13 postoperative patients (11 adult, 2 pediatric) who developed fatal pulmonary edema, and (2) one contemporaneous year of inpatient operations at two university teaching hospitals to determine the clinical manifestations, causes, epidemiology, and guidelines for fluid administration. Design: Retrospective analysis of 13 patients with fatal postoperative pulmonary edema and one contemporaneous year of major inpatient surgery. Patients and methods: Thirteen patients had net fluid retention of at least 67 mL/kg in the initial 24 postoperative hours and developed pulmonary edema. Ten were generally healthy while three had serious associated medical conditions. Measurements and results: There was no measurement, laboratory value, or clinical finding predictive of impending pulmonary edema. The most common clinical manifestation following the onset of pulmonary edema was cardiorespiratory arrest (n 5 8). Patients had metabolic acidosis (pH 5 7.15 6 .33), hypoxia (PO2 5 45 6 18 mm Hg), and normal electrolytes. The diagnosis of pulmonary edema was established by chest radiograph and confirmed by autopsy and pulmonary artery pressure (21 6 4 mm Hg). The mean net fluid retention was 7.0 6 4.5 L (90 6 36 mL/kg/d) and exceeded 67 mL/kg/d in all patients. Autopsy revealed pulmonary edema with no other cause of death. Among 8,195 major operations, 7.6% developed pulmonary edema with a mortality of 11.9%. Extrapolation to the 8.2 million annual major surgeries in the United States yields a projection of 8,000 to 74,000 deaths. Conclusions: Pulmonary edema can occur within the initial 36 postoperative hours when net fluid retention exceeds 67 mL/kg/d. There are no known predictive warning signs and cardiorespiratory arrest is the most frequent clinical presentation. The monitoring systems currently in use neither detect nor predict impending pulmonary edema, and as yet, there are no known panic values for excessive fluid administration or retention. (CHEST 1999; 115:1371–1377) Key words: acute renal failure; cardiorespiratory arrest; hypoxia; metabolic acidosis; postoperative complications; pulmonary edema Abbreviation: TURP 5 transurethral resection of the prostate
edema is a reported postoperative P ulmonary complication, but the etiology and mortality are 1
not known.2–5 Although the usual clinical manifestations of pulmonary edema have been described For editorial comment see page 1224
extensively, the presentation of postoperative pulmonary edema is not known. There are several possible *From the Department of Medicine, University of California School of Medicine, San Francisco, CA. Supported by a grant RO1 AG 08575-01A2 from the National Institute on Aging, Department of Health and Human Services, Bethesda, MD. Manuscript received June 29, 1998; revision accepted August 4, 1998. Correspondence to: Allen I. Arieff, MD, Department of Medicine, University of California School of Medicine, 299 South St, Sausalito, CA 94965
etiologic factors that may lead to postoperative pulmonary edema. Available evidence suggests that the disorder will be most common in patients who have preexisting heart disease. Fluid overload with high hydrostatic pressures resulting in left ventricular dysfunction would probably be the most common sequence1,6 and some of these patients would likely have suffered myocardial infarction5,7 or renal failure. Other groups of patients with postoperative pulmonary edema would be those with neurogenic (noncardiogenic) pulmonary edema secondary to postoperative hyponatremic encephalopathy,8 head trauma,9 pheochromocytoma,10 or other causes.11,12 ARDS is often an important differential consideration.13–15 One frequent reason for administration of large volumes of postoperative fluids is to compensate for apparent fluid losses for which there is no clinically CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1371
accessible method to accurately evaluate. Such loss includes “third space” losses, evaporative losses, blood loss, and insensible loss. Replacement is generally based on clinical approximations.16,17 Currently, there is little information available concerning the maximum postoperative volume of fluid that can be administered safely. In a postoperative patient who does not have serious cardiovascular, hepatic, or renal disorders, the quantity of fluid necessary to induce pulmonary edema would vary according to such factors as age; body weight; tissue turgor; cardiovascular, pulmonary, and renal function; plasma vasopressin levels; plasma proteins; and the volume of the “third space.”3 The differential diagnosis of postoperative pulmonary congestion includes hyponatremic encephalopathy, pulmonary capillary leak syndromes (ARDS), pneumonia, sepsis, and volume overload,18 –20 as well as other noncardogenic causes of pulmonary edema.4,8,10,11 Furthermore, it is unclear from the literature what the usual clinical manifestations of postoperative pulmonary edema are, or if it can result in death or serious morbidity in patients without heart disease or renal failure.18,21,22 In the current study, we present 13 fatal cases of postoperative pulmonary edema in generally healthy individuals in whom the pulmonary edema was secondary to excessive fluid administration.
Materials and Methods There were 13 study patients who, from 1991 to 1996, had postoperative pulmonary edema documented by clinical criteria and characteristic findings on chest radiograph. The diagnosis was confirmed by autopsy in nine patients and by pulmonary artery catheterization in four. Demographic features are shown in Table 1. There were four men and nine women, and the average age (6 SD) was 38 6 21 years. The operations are shown in Table 1. All but one operation (patient 9) were elective. All patients were seen in consultation at university teaching hospitals (n 5 9) or affiliated community hospitals (n 5 4). From a review of the literature, a net fluid retention of 2.2 L/d can lead to pulmonary edema,1,3,16 so patients were considered for study only if their net fluid retention was at least 2.2 L during the initial 24 postoperative hours. In four patients, the pulmonary capillary wedge pressure was measured by means of a pulmonary artery catheter. In the “Results” and “Discussion” sections, “cholecystectomy” refers to an open abdominal procedure rather than a laparoscopic one.3 The medical condition leading to hospitalization is shown in Table 1. Three patients had severe associated medical conditions: multiple traumatic injuries; acute pancreatitis; acute glomerulonephritis. However, all 13 patients were in stable condition and admitted to the hospital for elective surgical procedures, including tonsillectomy, knee-joint replacement, endometrial ablation, and sacral laminectomy. Only one patient had known heart disease (aortic valvular insufficiency), and 12 of 13 had normal results of heart examination on preoperative physical examination, chest radiograph, and ECG. Every patient received IV fluids preoperatively, during surgery, 1372
and postoperatively. The intake and output values are shown in Table 2 and are the totals from admission to the hospital to the time pulmonary edema was initially diagnosed. The fluid retention index is the net fluid retention for the initial 24 postoperative hours in milliliters per kilogram per day. The output values were extracted from the inpatient records and represent the sum of urine, estimated blood loss, and tube drainage. IV fluid intake was also extracted from the inpatient records. The predominant intraoperative IV fluid was Ringers lactate (130 mmol/L of sodium), others being blood, blood products, or 154 mmol/L NaCl. No patient received IV hypotonic fluid intraoperatively. Three patients had endoscopic surgery and received parenteral intraoperative hypotonic irrigating fluids by absorption through the cauterized operative bed, either the uterus or prostate.23 The irrigating fluids were either 200 mmol/L glycine or 165 mmol/L sorbitol. In the three patients who had endoscopic surgery, the intake was determined as the sum of IV fluids plus irrigating fluids, while output was the sum of urine volume, estimated blood loss, and collected drainage of irrigating fluid from the operative field. The operative drainage was collected in stainless steel vessels. Tube drainage was collected in a polyethylene bag by gravity and the volume was measured by means of the gradations on the bags. Three patients had indwelling Foley catheters postoperatively, while in the other 10 patients, urine was collected by spontaneous voiding. The postoperative fluids consisted of either Ringer’s lactate, 154 mmol/L NaCl, or 77 mmol/L NaCl. Pulmonary edema occurred within three postoperative days (range of 3 to 66 h). Epidemiologic Studies The records of major inpatient surgery for one contemporaneous year (1993) from two university teaching hospitals were evaluated for (1) the incidence and mortality of pulmonary edema, and (2) the incidence of other major postoperative complications in patients with pulmonary edema (pulmonary embolus, pneumonia, acute renal failure, GI bleeding, stroke, myocardial infarction, and hyponatremia). We examined only major surgical procedures; diagnostic procedures and minor surgery were not tabulated.24 A computer search of hospital records for the 8,195 hospital inpatients operated on in 1993 was done using a database (SAS; Cary, NC) to determine which surgical patients had postoperative pulmonary edema and their outcome.25 The incidence of seven other postoperative complications—pulmonary embolus, pneumonia, acute renal failure, GI bleeding, stroke, myocardial infarction, and hyponatremia (plasma sodium , 130 mmol/L)—were also tabulated. Outcome was determined only as death or discharge from the hospital within 30 days of the episode of pulmonary edema.5 Patients dying with pulmonary edema plus one of the other seven postoperative complications were subtracted from the total deaths with pulmonary edema to estimate the number of patients who died from pulmonary edema only. In addition to the information on patients with pulmonary edema, the overall postoperative mortality was determined.
Results Postoperative Patients With Fatal Pulmonary Edema All results are presented as mean 6 SD. In eight patients, the initial manifestation of pulmonary edema was cardiopulmonary arrest requiring emergent intubation (Table 1). One patient had intraopClinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
Table 1—Patients With Pulmonary Edema Patient/Sex/Age, yr/Weight, kg
Operations
Other Medical and Surgical Conditions
1/M/58/88.6
TURP
Prostatic hypertrophy
2/F/52/52.4
Aortic valve replacement
Aortic valve insufficiency
3/M/58/59
Open cholecystectomy
Cholelithiasis, gangrenous gall bladder
4/M/8/36.8
Myringotomy
5/F/63/68.2
Replacement of both knee joints
Otitis media, upper respiratory tract infection Osteoarthritis of knees
6/F/20/63.6
Open cholecystectomy
Acute pancreatitis, acute renal failure
7/F/67/51.4
Sacral laminectomy
Herniated sacral disk
8/M/31/74.8
Bronchoscopy, placement of dialysis graft Exploratory laparotomy, splenectomy
Acute glomerulonephritis, acute renal failure Motor vehicle accident, multiple traumatic injuries; ruptured spleen, fractures of ribs and arm Otitis media
9/F/19/63.2
10/F/3/15
11/F/39/60.5
Tonsillectomy, adenoidectomy, myringotomy Endometrial ablation
12/F/39/69.3
Exploratory laparotomy
Gastroenteritis, dehydration
13/F/31/54.3
Endometrial ablation
Dysfunctional uterine bleeding
Menorrhagia
Clinical Presentation Intraoperative chest pain Acute florid pulmonary edema (rales, dyspnea, pink frothy sputum) Unwitnessed cardiopulmonary arrest Acute florid pulmonary edema, shock Unwitnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest Witnessed cardiopulmonary arrest Acute renal failure Combative, acute florid pulmonary edema (rales, dyspnea, pink frothy sputum) Unwitnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest Witnessed cardiopulmonary arrest Unwitnessed cardiopulmonary arrest
Mean 6 SD 9/F/38/58 4/M/21/18
Pulmonary Edema Diagnosed by* Chest radiograph, autopsy Chest radiograph, autopsy, PCWP 5 22 mm Hg
Autopsy
Chest radiograph, autopsy Chest radiograph, PCWP 5 18 mm Hg Chest radiograph, PCWP 5 19 mm Hg Chest radiograph
Chest radiograph, autopsy PCWP 5 23 mm Hg, chest radiograph, autopsy
Chest radiograph, autopsy
Chest radiograph, autopsy
Chest radiograph
Chest radiograph, autopsy
PCWP (mm Hg) 5 20 6 2
*PCWP 5 pulmonary capillary wedge pressure; M 5 male; F 5 female.
erative chest pain, three had acute florid pulmonary edema, and one had acute renal failure. The diagnosis of acute pulmonary edema was established by a diagnostic chest radiograph accompanied by typical physical findings of bilateral rales, a third heart sound, and copious pink frothy sputum in three patients (Table 1). Two patients died before the diagnosis of pulmonary edema was established and in these patients, the diagnosis was established at autopsy. The five patients not presenting with cardiorespiratory arrest were intubated and had a chest radiograph that demonstrated findings of diffuse pulmonary edema. At the time of respiratory arrest or when pulmo-
nary edema was first diagnosed (n 5 11), arterial blood gases revealed a mixed metabolic-hypercapnic acidosis (Table 2), with pH 5 7.15 6 0.33, Pco2 5 58 6 37 mm Hg, and bicarbonate 5 18.5 6 4.8 mmol/L. All patients had hypoxemia, with Po2 5 45 6 18 mm Hg. These values are similar to those previously reported in patients with cardiogenic pulmonary edema.26 IV Fluids The average amount of IV fluid administered for the initial 27 postoperative hours was 9.9 6 6.0 L (171 6 103 mL/kg) (Table 2). The mean patient CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1373
Table 2—Laboratory Values
Mean 6 SD
Plasma Na, mmol/L
Plasma K, mmol/L
135 6
4.2 0.8
Time After Surgery When Pulmonary Edema Noted, h Mean
27
6 SD
20
Plasma Cl, mmol/L
Plasma Bicarbonate, mmol/L
Arterial pH
Arterial Po2, mm Hg
Arterial Pco2, mm Hg
Urea, mmol/L
18.5 4.8
7.15 0.33
45 18
58 37
11.3 17.7
Fluid Output, L
Fluid Retention, L
Fluid Retention Index, mL/kg/d
2.88 4.36
7.04 4.52
90 36
103 7
Fluid Intake, L 9.92 6.03
weight was 58 6 18 kg and using available tables, the calculated total body water was 29.8 6 8.9 L.27 The average net fluid retention was 7.04 6 4.5 L over the initial 27 6 20 postoperative hours, a quantity of retained fluid equal to 24% of the total body water or 12.2% of body weight. The fluid retention index (Table 2) was 90 6 36 mL/kg/d, almost three times the “normal” value (Table 3) of 38 mL/kg/d. The net fluid retention corresponded to an increase of 24% of total body water or 12.2% of body weight (Table 2). Thus, the quantity of administered fluid was clearly excessive.1,17,28,29 The reasons for the excessive quantities of administered fluid were not readily apparent. Recording of intake and output was ordered in all 13 patients. The orders were carried out, but in all patients, totals could be ascertained only by tedious searching of multiple sources: operative records, nurses notes, graphic records, physicians orders, and billing records. However, the excessive net fluid intake that was charted was not appreciated by the physicians caring for any of the patients until after pulmonary edema was diagnosed. In no instance was any physician notified that his or her patient had been administered or retained an excessive amount of fluid. Postoperative weights were ordered in all patients, but they were not actually obtained in any patient before the onset of pulmonary edema. Eight patients suffered cardiopulmo-
Right Lung Weight, g
Creatinine, mg/dL 1.2 1.1
Left Lung Weight, g
906 154
726 171
nary arrest, and in all eight, the net excess of administered fluid was not appreciated until after the arrest. In eight patients, the diagnosis of pulmonary edema was confirmed at autopsy (Table 1). All eight had pulmonary edema, with no other cause of death. In particular, fat emboli were not present and none had acute myocardial infarction. In these eight patients, lung weights were 906 6 154 g (right) and 726 6 171 g (left), more than twice the expected normal values.30 Epidemiologic Studies of Postoperative Pulmonary Edema During 1993, there were 8,195 major inpatient operations performed at two university teaching hospitals. The overall operative and postoperative mortality from all causes was 3.88% (302 patients). There were a total of 621 cases of postoperative pulmonary edema, an incidence of 7.6%. Among the 621 cases of pulmonary edema, 74 patients (11.9%) died. Those patients with pulmonary edema were then evaluated for the occurrence of other comorbid conditions that would be expected either to increase the chances of developing pulmonary edema or increase its mortality. The conditions were as follows: acute myocardial infarction, acute renal failure, acute GI bleeding, stroke, pneumonia, pulmonary
Table 3—Fluid Losses During and After Surgery
Type of Fluid Loss Blood loss Operative evaporation Insensible loss Third space loss Urine output Tube drainage
1374
Method of Quantitative Evaluation Clinical estimation Clinical estimation Clinical estimation Clinical estimation Actual measurement Actual measurement
Usual Range, mL/h
Typical Estimated Values for a Major Procedure Such as Open Cholecystectomy, mL
Varies according to operation 150 25 200 70 Varies according to operation Total losses
250 400 100 1,000 250 200 2,200
Clinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
embolus, and hyponatremia (plasma sodium , 130 mmol/L). When the patients with pulmonary edema plus one other comorbid postoperative complication were subtracted from the total, there were 204 patients remaining who had pulmonary edema without any of the other comorbid conditions (2.6%), of whom 8 (3.9%) died. In the United States in 1993, there were 8.2 million major inpatient surgical procedures.24 If the pulmonary edema figures were extrapolated from those 8.2 million operations, there would be a projected 622,000 annual cases of postoperative pulmonary edema in the United States, with 74,000 deaths. Even if all patients with comorbid factors are subtracted, there would still be projected 213,200 annual cases of postoperative pulmonary edema that had no other postoperative complications, totaling 8,315 deaths. Thus, the projected range of deaths from postoperative pulmonary edema would be 8,300 to 74,000 per year.
Discussion These data demonstrate that in generally healthy individuals undergoing elective surgery, pulmonary edema may be the initial clinical manifestation of fluid overload. If the pulmonary edema is not recognized and treated promptly, the patients may not survive. Although most textbooks of surgery recognize that excessive postoperative fluid administration can cause pulmonary edema, essentially no guidelines are available as to the quantity of fluid that may lead to such a complication.17 There are many reasons for administration of large quantities of both intraoperative and postoperative fluids.3 Some of the more common reasons are as follows: (1) hypotension following induction of anesthesia; (2) fluid sequestration in the alleged “third space” during surgery; (3) maintainance of BP after traumatic injury;4,13 (4) excessive blood loss; and (5) postoperative fever with increased insensible loss. Intraoperatively, hypotension may develop after the onset of anesthesia, requiring fluids and vasopressor therapy. In patients with cardiac disease, the afterload reduction secondary to the effects of anesthesia on sympathetic tone may improve myocardial performance transiently and allow the administration of large quantities of IV fluids. However, the return of vascular resistance to normal as the anesthetic recedes can result in congestive heart failure and pulmonary edema due to the relative fluid overload.16 The net fluid retention in these patients was . 7 L/d (Table 2). There are a number of different modalities and formulas for the calculation of postoperative fluid requirements in adults17,28,31 (Table 3). These include maintenance requirements esti-
mated at 2,500 mL/d, 30 to 35 mL/kg/d, or 1 mL/calorie/d. In addition, blood losses and tube drainage (nasogastric, chest tube, Penrose drain) are replaced and 1 to 3 L of extra fluid may be administered after a major complicated procedure in order to make up for estimated evaporative and third space losses. Operative fluid losses by evaporation are usually estimated at 150 mL/h. Using any of the aforementioned criteria, these 13 patients should have received from 2.5 to 5.5 L in the initial 36 postoperative hours. Based on any of the commonly used formulas for determining postoperative fluid requirements, the quantity administered (9.9 6 6 L; Table 2) was excessive. In general, most textbooks recommend that the patients’ actual losses be assessed at the bedside, and fluid administration should be based on estimated maintenance requirements plus estimated losses.17,29,32 A gain in fluid above 20% of total body water has been classified as fluid overload.1 Assuming that the patients’ net fluid retention was associated with a corresponding gain in body weight, the mean weight gain was 12.2%, a quantity that is by itself associated with a mortality of 19%.1 The “third space” is an area in the body that is occupied by fluids but is not in equilibrium with the bloodstream and thus hemodynamically inactive. Examples of the third space include burns, traumatized operative bed, and traumatically injured tissues. There is no bedside method to measure third space losses, and replacement is usually based on clinical approximation. However, the postoperative sequestered extracellular fluid has been estimated to vary from an imperceptible minimum to as much as 3 L.17 Ranges of 5 to 17 mL/kg/h of surgery after cholecystectomy or similar procedure during the initial postoperative day are typical estimated replacement values, with 500 to 1,000 mL/d being about average.31,32 Fluids in the third space may return to the vascular compartment several days after surgery because of the changes in physiology resulting in intravascular reaccumulation, which may result in pulmonary edema. In addition, operative blood loss is usually estimated by the operating surgeon to be 15 to 40% less than the isotopically measured blood loss from that patient.17 Although insensible loss is typically about 600 mL/d for an adult, postoperative hyperventilation, fever, and hypermetabolism may increase this substantially (1,500/d).17 However, in the postoperative patient who is not eating, there is ordinarily no net insensible loss,16 and this fact is often neglected when considering replacement, resulting in additional fluid excess. Additional surgical fluid losses that are difficult to calculate are the following: additional insensible fluid losses by evaporation from the exposed peritoneal (or pleural) surfaces; and evaporative losses by virtue of operaCHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1375
tive drapes and from the administration of dry inhaled gases. Actual data are limited, but during laparotomy, the loss is about 150 mL/h or about 300 mL for an uncomplicated cholecystectomy.33–35 Most protocols for postoperative fluid therapy recommend about 30 to 35 mL/kg/d plus estimated losses. For a 65-kg individual following cholecystectomy, the total for the initial postoperative day might thus be calculated as 2,275 mL (35 mL/kg maintenance) 1 600 mL (insensible loss) 1 1,000 mL (third space loss) 1 300 mL (evaporative loss) 5 4.2 L. Virtually all protocols for postoperative fluid management recommend frequent monitoring of the patient’s volume status, utilizing such readily available clinical parameters as BP, heart rate, urine output, auscultation of the lungs, changes in body weight, and calculation of net fluid retention.17,28,31,32 If fluid retention is noted to be excessive, then the rate of administration can be decreased. In the current series, values for heart rate, BP, pulse, and urine output were generally recorded in the nurses notes or graphic sheets. However, not a single patient had a postoperative weight recorded before pulmonary edema was diagnosed and after discharge from the recovery room, auscultation of the lungs by a physician was never recorded in any patient. These observations suggest that among physicians caring for these patients, the index of suspicion concerning possible consequences of excessive fluid administration was not high. Since to our knowledge there are no established guidelines for “panic values” or their equivalent for fluid retention, the nursing staff did not alert physicians as to possible excessive fluid retention. When postoperative pulmonary edema has occurred in generally healthy individuals, the mortality has often been in excess of 50%, as occurred in the present study.1,3 The parameters that are included in the physician’s calculation of replacement plus maintenance fluids include estimations of parameters shown in Table 3. All of the five parameters not only represent “estimates,” but it is generally agreed that the estimates vary widely among different physicians.17 It would appear to be important that “panic” values for fluid retention be established for different categories of surgical procedures. Postoperative pulmonary edema has been a reported complication in generally healthy postoperative subjects receiving known amounts of IV fluids.16 Among 161 postoperative patients, 27% of those retaining 2.2 L per 24 h developed pulmonary edema.16 However, since a number of these patients also had postoperative hyponatremic encephalopathy, the pulmonary edema may have been neurogenic in some cases. In the present study, neurogenic pulmonary edema secondary to postoperative hy1376
ponatremic encephalopathy should not have occurred, because the plasma sodium level was normal (135 6 6 mmol/L). The incidence of postoperative pulmonary edema is uncertain. In generally elderly, mostly male patients and those with heart disease, the incidence was 2.3 to 4.8%, less than the 7.6% in the current study.5,21,22 In 14,707 generally low-risk female patients undergoing hysteroscopic hysterectomy, the incidence was only 0.2%.36 In none of these studies were mortality figures available.5,21,22,36 There may have been patients in the current study who had undocumented silent postoperative myocardial ischemia.7 Major causes of perioperative morbidity and mortality are myocardial infarction and ischemia.18,21,22 In response, there is now ever more sophisticated perioperative monitoring of cardiac function.7 However, fluid monitoring is generally carried out in an unsophisticated manner, and this appears to have contributed to the mortality from pulmonary edema (Table 3). In three patients, pulmonary edema occurred following endoscopic procedures, in which the operative field was irrigated with copious quantities of fluid. Although the staff attempted to evaluate intake and output, such a task is technically difficult. Monitoring of fluid retention during transurethral resection of the prostate (TURP) can be accomplished by the ethanol method or monitoring of body weight,37,38 and a device is available for fluid monitoring during hysteroscopic hysterectomy,39 but such procedures are not in wide use. The most widely accepted method of monitoring intravascular volume status is by pulmonary artery catheter, but this procedure has been criticized recently because of associated morbidity.40,41 The incidence of pulmonary edema following hysteroscopic hysterectomy (endometrial ablation) is , 1% and the number of deaths is not known.36,42 The incidence of pulmonary edema following TURP is unknown, but there are at least five reported cases with one death.43– 45 In 25 patients undergoing TURP, the volume of irrigating fluid absorbed (6 SD) was found to be 2.1 6 0.9 L.38 In five patients undergoing hysteroscopic hysterectomy who developed fluid overload, the volume of irrigating fluid absorbed was 5 to 7 L.23 Retention of such volumes of fluid can lead to pulmonary edema in otherwise healthy subjects16 (Table 2), emphasizing the necessity for monitoring net fluid retention during such procedures.
References 1 Lowell JA, Schifferdecker C, Driscoll DF, et al. Postoperative fluid overload: not a benign problem. Crit Care Med 1990; 18:728 –733 Clinical Investigations in Critical Care
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
2 van Hoozen BE, van Hoozen CM, Alberton TE. Pulmonary considerations and complications in the neurosurgical patient: pulmonary edema. In: Youmans JR, ed. Neurological surgery. Philadelphia, PA: WB Saunders, 1996; 624 – 626 3 Roth E, Lax LC, Maloney JV. Ringer’s lactate solution and extracellular fluid volume in the surgical patient: a critical analysis. Ann Surg 1969; 169:149 –164 4 Finn JC, Rosenthal MH. Pulmonary edema in trauma and critically ill patients. Semin Anesth 1989; 8:265–274 5 Khuri SF, Daley J, Henderson W, et al. The National Veterans Administration surgical risk study: risk adjustment for the comparative assessment of the quality of surgical care. J Am Coll Surg 1995; 180:519 –531 6 Katz AM. Cardiomyopathy of overload. N Engl J Med 1990; 322:100 –110 7 Mangano DT, Browner WS, Hollenberg M, et al. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery. N Engl J Med 1990; 323:1781–1788 8 Ayus JC, Arieff AI. Pulmonary complications of hyponatremic encephalopathy: noncardiogenic pulmonary edema and hypercapnic respiratory failure. Chest 1995; 107:517–521 9 Becker DP, Gade GF, Young HF, et al. Diagnosis and treatment of head injury in adults: fluids and electrolytes. In: Youmans JR, ed. Neurological surgery. Philadelphia, PA: WB Saunders, 1990; 2113–2115 10 Suga K, Tsukamoto K, Nishigauchi K, et al. Iodine-123MIBG imaging in pheochromocytoma with cardiomyopathy and pulmonary edema. J Nucl Med 1996; 37:1361–1364 11 Malik AB. Mechanisms of neurogenic pulmonary edema. Circ Res 1985; 57:1–18 12 Frim DM, Wollman L, Evans AB, et al. Acute pulmonary edema after low-level air embolism during craniotomy. J Neurosurg 1996; 85:937–940 13 Falk JL, Rackow EC, Weil MH. Colloid and crystalloid fluid resuscitation. In: Shoemaker WC, Ayres S, Grenvik A, eds. Textbook of critical care. Philadelphia, PA: WB Saunders, 1989; 1055–1073 14 Baumgartner F, Omari B, Lee J, et al. Survival after trauma pneumonectomy: the pathophysiologic balance of shock resuscitation with right heart failure. Am Surg 1996; 62:967–972 15 Christenson JT, Aeberhard JM, Badel P. Adult respiratory distress syndrome after cardiac surgery. Cardiovasc Surg 1996; 4:15–21 16 Rosenthal MH, Arieff AI. Fluid and electrolyte therapy in critically ill patients and those who are pre-, post- or intraoperative. In: Arieff AI, DeFronzo RA, eds. Fluid, electrolyte and acid-base disorders. New York, NY: Churchill Livingstone, 1995; 597– 632 17 Shires GT, Shires GTI, Lowry SF. Fluid, electrolyte, and nutritional management of the surgical patient. In: Schwartz SI, Shires GT, Spencer FC, eds. Principles of surgery. New York, NY: McGraw-Hill, 1994; 61–93 18 Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990; 72:153–184 19 Bishop MH, Jorgens J, Shoemaker WC, et al. The relationship between ARDS, pulmonary infiltration, fluid balance, and hemodynamics in critically ill surgical patients. Am Surg 1991; 57:785–792 20 Connors AF. Invasive measurements are required for assessing hemodynamic status in critically ill patients. Am Rev Respir Dis 1988; 138:1070 –1075 21 Rao TLK, Jacobs KH, El-Etr AA. Reinfarction following anesthesia in patients with myocardial infarction. Anesthesiology 1983; 59:499 –505 22 Goldman L, Caldera DL, Nussbaum SR. Multifactorial index
23 24
25 26 27
28
29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45
of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977; 297:845– 850 Ayus JC, Arieff AI. Glycine-induced hypo-osmolar hyponatremia. Arch Intern Med 1997; 157:223–226 National Health Survey. Vital and health statistics: detailed diagnoses and procedures, National Hospital Discharge Survey, 1995. Hyattsville, MD: National Center for Health Statistics, 1997; 146 SAS Institute I. SAS language reference, version 6, 1st ed. Cary, NC: SAS Institute, 1990; 1290 Aberman A, Fulop M. The metabolic and respiratory acidosis of acute pulmonary edema. Ann Intern Med 1972; 76:173–184 Park R, Guisado R, Arieff AI. Nutrient deficiencies in man and animals: water. In: Rechcigl M Jr, ed. CRC handbook of nutrition and food: section E, nutritional disorders. West Palm Beach, FL: CRC Press, 1978; 363– 400 Darby JM, Nelson PB. Fluid, electrolyte, and acid-base balance in neurosurgical intensive care. In: Andrews BT, ed. Neurosurgical intensive care. New York, NY: McGraw-Hill, 1993; 133–162 Drucker RP. Fluid therapy. In: Davis JH, Drucker RP, Foster RS, et al, eds. Clinical surgery. Washington, DC: CV Mosby, 1987; 847– 851 Whimster WF, Macfarlane AJ. Normal lung weights in a white population. Am Rev Respir Dis 1974; 110:478 – 483 Tonnesen AS. Crystalloids and colloids: third space losses. In: Miller RD, ed. Anesthesia. New York, NY: Churchill Livingstone, 1990; 1439 –1465 Proctor HJ, Rutledge R. Fluid and electrolyte management. In: Hardy JD, Kukora JS, Pass HI, eds. Hardy’s textbook of surgery. Philadelphia, PA: JB Lippincott, 1988; 15–29 Virtue RW, LeVine DS, Aikawa JK. Fluid shifts during the surgical period: RISA and 35S determinations following glucose, saline or lactate infusion. Ann Surg 1966; 163:523–528 Giesecke AH, Egbert LD. Perioperative fluid therapy— crystalloids. In: Giesecke AH, Egbert LD, eds. Anesthesia. New York, NY: Churchill-Livingstone, 1986; 1313–1328 Roberts JP, Roberts JD, Skinner C. Extracellular fluid deficit following operation and its correction with Ringer’s lactate: a reassessment. Ann Surg 1985; 202:1– 8 Baggish MS, Sze EH. Endometrial ablation: a series of 568 patients treated over an 11 year period. Am J Obstet Gynecol 1996; 174:908 –913 Lyon RP, St Lezin M, Thomas C, et al. Monitoring of body weight during transurethral resection of the prostate: preliminary report. J Endourol 1994; 8:161–163 Hahn RG, Ekengren JC. Patterns of irrigating fluid absorption during transurethral resection of the prostate as indicated by ethanol. J Urol 1993; 149:502–506 Hawe JA, Chien PFW, Martin D, et al. The validity of continuous automated fluid monitoring during endometrial surgery— luxury or necessity? Br J Obstet Gynaecol 1998; 105:797–801 Weil MH. The assault on the Swan-Ganz catheter. Chest 1998; 113:1379 –1386 Kefalides P. Pulmonary artery catheters on trial. Ann Intern Med 1998; 128:161–162 Hulka JF, Peterson HA, Phillips JM, et al. Operative hysteroscopy: American Association of Gynecologic Laparoscopists’ 1993 membership survey. J Am Assoc Gynecol Laparosc 1995; 2:131–132 Drinker HR, Shields T, Grayhack JT, et al. Simulated transurethral resection in the dog: early signs and optimal treatment. J Urol 1963; 89:595– 602 Bulkley GJ, O’Connor VJ, Sokol JK. Overhydration during transurethral prostatic resection. JAMA 1954; 161:1042 Aasheim GM. Hyponatraemia during transurethral surgery. Can Anaesth Soc J 1973; 20:274 –280 CHEST / 115 / 5 / MAY, 1999
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians
1377
Fatal Postoperative Pulmonary Edema* : Pathogenesis and Literature Review Allen I. Arieff Chest 1999;115; 1371-1377 DOI 10.1378/chest.115.5.1371 This information is current as of May 14, 2012 Updated Information & Services Updated Information and services can be found at: http://chestjournal.chestpubs.org/content/115/5/1371.full.html References This article cites 32 articles, 7 of which can be accessed free at: http://chestjournal.chestpubs.org/content/115/5/1371.full.html#ref-list-1 Cited Bys This article has been cited by 14 HighWire-hosted articles: http://chestjournal.chestpubs.org/content/115/5/1371.full.html#related-urls Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.chestpubs.org/site/misc/reprints.xhtml Reprints Information about ordering reprints can be found online: http://www.chestpubs.org/site/misc/reprints.xhtml Citation Alerts Receive free e-mail alerts when new articles cite this article. To sign up, select the "Services" link to the right of the online article. Images in PowerPoint format Figures that appear in CHEST articles can be downloaded for teaching purposes in PowerPoint slide format. See any online figure for directions.
Downloaded from chestjournal.chestpubs.org at University of Illinois at Chicago Library on May 14, 2012 © 1999 American College of Chest Physicians