Complications of parenteral nutrition

Complications of parenteral nutrition

Complications of Parenteral Nutrition Bfuce M. woh, MD, Davis, California March A. Ryrkr, RN, USN, Davis, California Reid A. Nbhkawa, Fiband, Dav...

1023KB Sizes 5 Downloads 110 Views

Complications of Parenteral Nutrition

Bfuce M. woh,

MD, Davis, California

March A. Ryrkr, RN, USN,

Davis, California

Reid A. Nbhkawa, Fiband, Davis, California Char& H. HaWed, MD, lkis,


Edan F. Schmidt, MD, Davis, California

Soon after the techniques for provision of parenteral nutrition solutions through central venous catheters were established by Dudrick and associates [I], a variety of complications of this procedure were identified [2].~These complications arise as the result of placement or maintenance of the intravenous catheter required for the infusion or from excessive or insufficient provision of one or more nutrients. Seemingly, every complication possible from placement of a central venous catheter and subsequent infusion of total parenteral nutrition (TPN) solutions has been described [3]. This study reports the incidence of specific complications of TPN at the University of California, Davis Medical Center (UCDMC) from 1981 through 1985. Factors involved in the evolution of rates of specific complications and the potential impact on complication rates by the quality assurance program at UCDMC are also reported.

A continuous prospective study of all complications of TPN at UCDMC was instituted in 1981. All possible complications were reviewed monthly in order to determine whether a given event represented a complication of TPN and to take corrective action if indicated. All patients receiving TPN were evaluated and followed by the nutrition support service. Direct observation rather than a retrospective review of medical records was used to identify possible complications. Responsibility for the decision to treat with TPN, catheter placement, and phyFrom the Dqwbwnts of Surgery and Medicine, University of California, Davis, Sacramento, California. Supported in part by Clinical Nutrition Research Unit r&ant P30 AM35747-01 funded by the National Institutes of Health. Bethesda. Maryland. Requests for reprints should be addressed to Bruce M. Wolfe, MD, Depwtment of Surgmy, University of Callfomia. Davis, 4301 X Street, Sacramento, California 95817. Presented at the 57th Annual Meeting of the Pacific Coast Surgical Assoclatlon, Maui. Hawaii, February 16-19, 1986.

Volume 152, July 1986

sicians’ orders were retained by the patients’ primary physicians. All patients except infants who received TPN by central venous catheter infusion were included in this study. A complication was considered to have occurred if the patient experienced obvious morbidity, mortality, or both; an event known to be deleterious, despite a lack of demonstrable morbidity; or premature loss of the central venous catheter. In the final 6 months of the study, an effort was made to classify the severity of the impact of the complication on the patient. Complications were classified as being (1) secondary to central venous catheter placement, (2) responsible for induction of systemic sepsis, (3) mechanical, or (4) metabolic. All definite arterial injuries were considered to be complications even if no direct treatment was required. Catheter malposition was regarded as a complication only if the entire catheter required replacement or if morbidity resulted from TPN infusion through the malpositioned catheter. A diagnosis of catheter-related sepsis required a positive blood culture and clinical indication that the catheter was the most likely source of sepsis. Information considered in the review of each case included the organism cultured, blood culture site, semiquantitative culture of the subcutaneous and intravascular catheter segments [4], response to catheter removal, and alternative explanations for the positive blood culture. Mechanical complications refer to premature loss of the catheter, venous thrombosis, air embolism secondary to tubing disruption, or other mechanical mishap. A diagnosis of venous thrombosis was based on physical findings of swelling, venous congestion, or both; venography; or evidence of the TPN solution leaking around the skin and catheter junction. Metabolic complications satisfied the following criteria: the abnormality was (1) known to occur with TPN, (2) known to be detrimental, (3) appeared after TPN was initiated and could reasonably be attributed to the TPN solution itself, and (4) when applicable, was observed to subside after appropriate adjustment of the TPN. Examples of specific laboratory abnormalities considered complications included a chloride concentration of more than 115 mEq/liter associated with acidosis, a serum glucose


Wolfe et al


Summary of Complkatlom

ot Total Parenteral Nutrltlon

CathederPlacement Year



1981 1982 1983 1984 1985

284 298 358 368 341

3 11 30 27 24





sepsis %


1.1 3.7 8.4 7.3 7

8 13 13 41 34



3i 11:1 10

3: 24 40





level of more than 400 or leas than 60 mg/dl, and a phosphate concentration of less than 1.5 mg/dl. Tests for statistical significance of observed differences were carried out by chi-square analysis. ResUtts

A total of 1,647 patients received central venous TPN at our institution from 1981 through 1985, of whom 93 percent were adults (Table I). The mean duration of TPN was 18.9 days, and the total number of patient-days of TPN was 31,112. There were 509 complications in these patients. Two thousand four hundred forty-six central venous catheters were placed by the following techniques: 1,438 by new venapuncture (59 percent), 881 by exchange of an existing catheter using a guide wire (36 percent), and 127 by operative placement of a cuffed silicone rubber catheter (5 percent). From 1981 through 1983, 1.2 catheters per patient were placed, whereas from 1984 through 1985,l.g catheters per patient were placed (p
FAechanlcal %


20 9.1 9.2 0.5 11.7 9

MhbOllC n


88 15 20 17 7

24 5.1 5.6 4.0 2.1



Pneumothorax occurred in 18 patients, 2 of which were tension pneumothoraces. One, in a patient requiring mechanical ventilation who suffered a cardiac arrest before its resolution, proved fatal. Other catheter placement complications that were life threatening included air embolism in one patient and puncture of the endotracheal tube cuff in one patient. Complications of catheter placement using a guide wire included an unsuccessful procedure that resulted in loss of the catheter in 10 patients and induction of arrhythmia that necessitated treatment and catheter withdrawal in four patients. The annual percentage of patients with nosocomial sepsis associated with TPN is shown in Table I. Annual sepsis rates of 2.1 percent in 1981 and 4.4 percent in 1982 compare favorably with rates reported from other centers [I ,5]. The abrupt increase of TPN-associated nosocomial sepsis (11.1 percent in 1984 and 10 percent in 1985) is statistically significant (p

cal infections were S. epidermidis, an established pathogen [6]. Fungi were isolated in 24 percent of cases (94 percent of which were of the Candida species), whereas gram-negative rods were identified in 23 percent of cases. The impact of a given episode of TPN-induced sepsis is often difficult to determine without an analysis of matched patients without sepsis [ 7j. Specific examples of obvious major morbidity were as follows: Sepsis secondary to Candida tropicalis was attributed to the TPN catheter in a woman with intestinal obstruction and fistulas. The catheter was removed and signs and symptoms of sepsis resolved without further treatment. Four months later, the patient was found to have destruction of portions of the lumbar spine. Candida tropicalis was isolated from the aspirate of the second and third lumbar interspace. Candida sepsis developed in a second patient who required mechanical ventilation, fluid resuscitation, and general critical care support, including a full course of amphotericin B. A third patient was undergoing aggressive immunosuppressive therapy for rejection of a liver transplant when Candida sepsis from the TPN line occurred. The patient died from complications of fungal sepsis and graft rejection. Mechanical complications were common during the initial year of study (20 percent) and decreased to 9 to 12 percent for the remaining 4 years. Inadvertent removal of the catheter was the most frequent mechanical complication, generally due to nursing mishaps. An intensive in-service program for the nursing staff resulted in substantial reduction, but not elimination of this problem. Major venous thrombosis was also relatively common (44 patients). No case of clinically manifested pulmonary embolus was identified among these patients. Permanent inability to gain central venous access at the site of thrombosis was the most frequent long-term morbidity secondary to TPN catheter-related venous thrombosis. Other mechanical complications leading to loss of the catheter included clotting of the line and catheter damage. Air embolism as a mechanical complication resulting from tubing disconnection was rare (two patients), but resulted in major morbidity in one patient, who suffered cerebral injury. A second lifethreatening mechanical complication was hydrothorax secondary to erosion of the TPN catheter through the vein wall. Prompt recognition and tube thoracostomy prevented the development of tension hydrothorax [8]. Metabolic complications were frequent in 1981, with hyperchloremic metabolic acidosis predominating. Correction of this problem was accomplished by alteration of the standard electrolyte additive protocol from a chloride to acetate ratio of 7~1to 4:l and finally to a ratio of 2.5:1. The frequency of metabolic complications has decreased steadiVolume 152, July 1996



of Parenteral Nutrition



Figure 1. Sqsis by catheter type, 1984 through 1985. After the lntroducthmofmultkmen catheters for 7PN lnfuskn4the rata of TPN-br&cedaepsk3 was tianged for sbfgtelumen catheter9 (3.2percunf) but was i46percentfor multIlumencatheters and 13.6 pwceM fof Swan-Ganz catheters No aepsfs occurred among hospitalized patients with cuffed sllieone rubber catheters Such catheters were used only In stab& pat/ants bag trabmd for home TPN or patknts sbeady on homa 7PN. This pat/ent populathm is therefore not comparab/e to the patlents treated with other 8lngIe or muMlumencatheters.

ly and has not been a major problem; however, one fatal and two life threatening complications have occurred. A 13 month old infant on permanent home TPN was admitted with catheter sepsis. The catheter was removed and a new catheter placed under general anesthesia. Postoperatively, the same nutrient formulation was infused that had been used over the previous several weeks. Hyperglycemic hyperosmolar coma with a serum glucose level of 1,980 mg/dl developed over the next several hours, producing cardiovascular instability and ultimately death. The life threatening metabolic complications consisted of acute congestive heart failure secondary to rapid infusion of normal saline solution immediately after catheter placement in one patient, and hypercarbia and acute respiratory failure due to a high rate of glucose infusion that required endotracheal intubation and mechanical ventilation in a patient with marginal pulmonary function [9]. Hepatic dysfunction was judged to have occurred in one patient primarily due to the TPN infusion. Among the aforementioned complications, four are considered to have been either primarily responsible for death or the major contributing factor at the time of death: tension pneumothorax, subclavian artery laceration, Candida sepsis, and hyperglycemic hyperosmolar coma (each occurring in one patient). Thus, the TPN-associated mortality in this series was 4 of 1,647 patients or 0.2 percent. During the final 6 months of the study, an effort was made to classify the impact of the complication on the patient’s course. A review of the treatment required and resultant morbidity of the 51 complications during this interval showed 7 major complications among the 174 patients. Thus, 29 percent of 95

Wolfe et al

patients receiving TPN experienced a complication, and 4 percent experienced what was considered a major complication. These included four patients with persistent Candida sepsis who required a full course of amphotericin B, bacterial sepsis in one patient that resulted in the loss of the only available access site to the superior vena cava, tube thoracostomy for pneumothorax in one patient with severe respiratory failure, and one patient with psychiatric illness who refused for several days to return to the operating room for correction of a malpositioned Hickman catheter. No TPN-associated mortality was identified during this 6 month interval.

Comments Despite a plethora of reports regarding complications from TPN, complacency among physicians and nurses sometimes occurs due to the frequent administration of TPN to hospitalized patients, protocols for all aspects of TPN delivery that make the establishment of venous access and order writing routine, and the tendency for the majority of TPN complications to have seemingly limited adverse effects on patients. This study demonstrates that complications of TPN are common and that life threatening or fatal complications, although uncommon, occur on a regular basis. It is apparent that many of the complications reported herein would not have had a major impact on otherwise healthy patients. TPN is generally provided only to patients with complications or who are in a compromised state; thus, events such as partial pneumothorax or transient septicemia assume more importance than usual. Continued reliance on glucose as a major component of TPN solutions necessitates central venous access. The frequency of complications of catheter insertion can be minimized by utilization of experienced personnel, such as a nutrition support team, and by the correction of hypovolemia, proper positioning, and other precautions [IO]. Nevertheless, this complication rate cannot be decreased to zero. Reported rates of complications have varied between our annual rates of 1.1 to 8.4 percent [1,11]. The relatively high proportion of tension pneumothoraces seen in our series (11 percent) resulted from the large number of patients who required mechanical ventilation at the time of catheter placement. Materials for immediate tube thoracostomy must always be available. A normal appearing chest radiograph does not ensure a lack of a complication of catheter placement such as mediastinal catheter position or arterial injury [12]. The fatal case of subclavian arterial injury in our series occurred despite successful catheter placement, and a subsequent chest radiograph showed satisfactory catheter position with no evidence of hemothorax or other complications. A false sense of security contributed to a delay in the definitive diagnosis of the arterial injury. Although 18 98

percent of total complications were secondary to catheter placement, 50 percent of the fatal complications resulted from injuries that occurred during placement. . The data presented in this study appear to demonstrate an increase in the sepsis rate as the result of the alteration of our protocol to allow the infusion of TPN solutions through multilumen venous or Swan-Ganz catheters. Sicker patients with difficult vascular access or unstable cardiovascular function are more prone to the development of septic complications for a variety of reasons. In order to attribute the increased sepsis rate in 1984 to 1985 to the patient population treated, it would be necessary to demonstrate a change in the population between 1981 to 1983 and 1984 to 1985. There is no basis to suggest that the increased TPN-associated sepsis rates in 1984 and 1985 at our institution occurred due to a change in patient population. The protocol change was the only major change made between 1983 and 1984. Essentially identical data was recently reported by Pemberton et al [13]. Identification of an increased sepsis rate resulting from multilumen catheter use, however, does not necessarily establish a basis for abandonment of these catheters for TPN infusion. A separate study is required to determine if total patient morbidity is decreased by placement of an additional central venous catheter dedicated to TPN when a multilumen or Swan-Ganz catheter is also in place. There is no evidence that the addition of TPN infusion to a multilumen catheter increases the incidence of sepsis in that catheter, especially since hypertonic TPN solutions are a poor culture medium [14]. A decreased requirement for venous access procedures, thereby decreasing placement complications, might justify the use of multilumen central venous catheters. Unfortunately, the number of catheters per patient in this study increased abruptly from 1.2 catheters per patient in 1981 to 1983 to 1.9 in 1984, and the increase continued in 1985. It appears that the introduction of the multilumen catheters has increased the TPN sepsis rate and increased both the number of catheters and placement procedures needed per patient. The reasons for the frequent induction of sepsis by multilumen catheters were not established by this study. Historically, dedication of a single central venous catheter for TPN has been regarded as crucial in the prevention of nosocomial TPN-associated sepsis [15]. An additional factor not evaluated by the present study that could increase the sepsis rate is that larger, stiffer catheters produce more injury to vein walls, resulting in associated thrombosis and sepsis [16]. The determination of the severity of nosocomial bacteremia is difficult in many critically ill patients. Patients with nosocomial bacteremia have a mortality rate 3.8 to 14 times greater than that of similarly matched nonbacteremic patients, with an average The Amrkan

Joumal of Surgery

Complications of Parenterai Nutrition

increase in hospitalization of 17 to 19 days, costing in excess of $4,000 [7,17]. The role of specific antimicrobial therapy combined with catheter removal has not been clearly established in the past [I ,5]. The occurrence of Candida osteomyelitis in one patient in our series resulting in permanent disability supports the recommendation of Edwards [18] that Candida sepsis usually be treated with amphotericin B in addition to catheter removal. Our present recommendation is for routine antimicrobial treatment of both bacterial and fungal catheter-induced sepsis. Exchange of the central venous catheter by use of a guide wire for evaluation of suspected catheter sepsis has been safe and efficacious in our hands [19]. Several studies have addressed the issue of venous thrombosis in TPN patients. Although the routine addition of heparin to TPN solutions has been advocated, more recent studies have not demonstrated any efficacy for routine heparin administration among hospitalized TPN patients [20,21]. Utilization of silicone rubber or polyurethane catheters with hydromer coating rather than polyvinyl chloride or polyethylene catheters may effectively decrease the incidence of venous thrombosis without the addition of heparin. Metabolic complications have shown a steady decrease in frequency, but nevertheless accounted for major morbidity and mortality. Our fatal case of hyperglycemic hyperosmolar coma emphasizes the need for constant monitoring of urine and serum glucose concentrations. Such monitoring, with appropriate responses to glucose intolerance, should prevent this complication [2,3]. Alterations of glucose metabolism induced by surgical or septic stress have the potential to result in rapid and marked hyperglycemia [22]. The value of quality of care reviews are demonstrated by this study. Corrective actions are unlikely to be vigorously pursued if the occurrence of ongoing complications is not demonstrated within a given hospital. At our institution, revisions of the standard protocol resulted in gradual but definite decreases in the incidence of metabolic complications. The demonstration of a high rate of nursing errors with resultant mechanical complications formed the basis for a vigorous educational program in the hospital. Finally, progress was made in decreasing the use of multilumen catheters for TPN during the last 6 months of 1985. We conclude that complications of TPN are frequent and may be severe. Quality assurance mechanisms for identification of these complications are necessary and form the basis for establishment of appropriate protocols. mry All hospitalized patients except infants (a total of 1,647 patients) who received central venous TPN solutions at UCDMC from 1981through 1985 were Volume 152, Jufy 1999

studied to determine the incidence of complications from the use of TPN. A complication was considered to have occurred if the patient experienced (1) obvious morbidity, mortality, or both; (2) an event known to be deleterious, despite a lack of demonstrable morbidity; or (3) premature loss of the central venous catheter. Complications related to catheter placement occurred in 5.7 percent of patients, sepsis in 6.5 percent, mechanical complications in 9 percent, and metabolic complications in 7.7 percent. The incidence of induction of sepsis increased during 1984 to 1985 due to the introduction of multilumen central venous catheters. The most frequent catheter placement complications were hemorrhage and pneumothorax. Major venous thrombosis and nursing mishaps were the most common mechanical complications. Metabolic complications were infrequent and were generally not severe after adjustment of the protocol in late 1981. Four patients (0.2 percent) died from TPN-associated complications: a child on home TPN who underwent a catheter change and in whom hyperosmolar hyperglycemic coma developed, a patient with end-stage chronic obstructive pulmonary disease in whom tension pneumothorax occurred, a patient who died from complications of subclavian artery laceration, and a patient who died from Candida septicemia. Complications of TPN are frequent and may be severe. Quality assurance mechanisms for identification of these complications are necessary and should form the basis for the establishment of appropriate protocols.

References 1. Dudrick SJ, Macfadyen BV, Van Buren CT, et al. Parenterai hyperalimentation: metabolic problems and solutions. Ann Surg 1972;178:259-64. 2. Ryan JA, Abel RM, Abbott WM, et al. Catheter complications in total parenteral nutrition. N Engl J Med 1974;290:75761. 3. Ryan JA. Complications of total parenteral nutrition. In: Fischer JE, ed. Total parenteral nutrition. Boston: Little, Brown, 197655-100. 4. Maki DG, W&e CE, Sarafin HW. A semiquantitatlve culture method for identifying intravenous catheter-related infeci tion. N Engl J h&d 1977;296:1305-9. 5. Sanders RA, Sheldon GF. Septic complications of total parenteral nutrition. Am J Surg 1976;132:214-9. 6. Forse A, Dixon C, Bernard K, et al. Staphylococcus epklermidis: an important pathogen. Surgery 1979;86:507-14. 7. Rose R, Hunting KJ, Townsend TR. Wenzel RP. Morbklity/mortalky and economics of hospital-acquired blood stream infections: a controlled study. South Med J 1977;70: 12679. 6. Chute E, Cerra F. Late development of hydrothorax and hydromediastlnum in patients with central venous catheters. Crit Care Med 1962;10:866-9. 9. Askanazi J, Elwyn DH, Silverberg PA, et al. Respiratory distress secondary to a high carbohydrate load. Surgery 1980;87:5%6-9. 10. Nehme AE. Nutritlonal support of the hospitalized patient. JAMA 1980;243:1906~6. 11. Padberg FT, Rugglero J, Blackburn GL, Bistrian BR. Central 97

Wolfe et al

12. 13.


15. 16. 17.

18. 19.




venous catheterization for parenterai nutrition. Ann Surg 1981;193:284-70. Langston CS. The aberrant central venous catheter and its complications. Radiology 1971;100:55-9. Pemberton LB. Lyman B, Lander V, et al. Triple versus sinaie lumen catheter sepsis during TPN in surgical or criticaii~iii patients. Arch Surg 1988; 121:591-4. Gebari SM, Reinhardt GF, Greeniee HB. Muitipiicatlon of nosocomiai pathogens in intravenous feeding solutions. Appi Microbioi 1973;28:874-9. Williams WW. Infection control during parenterai nutrition therapy. JPEN 1985;9:735-48. Stiiiman RM, Fawzi S, Garci L, Sawyer PN. Etiology of catheter associated sepsis. Arch Surg 1977;112:1497-9. Spengler RF, Greenough WB, Stoiiey PD. A descriptive study of nosocomiai bacteremias at the Johns Hopkfns Hospital. Johns Hopkins Med Bull 1978;142:77-84. EdwardsJE Jr. Disseminated candidosis. infect Dis Newsletter 1985;4:81-4. Ryder M, Wolfe BM, Flynn N. Safety and efficacy of catheter exchange for replacement and/or evaluation of sepsis in central venous catheters. Presented at the ASPENCilnicai Congress, Dallas, Texas, February 1988. Kudsk KA. Powell C. Mirtailo JM. et al. Heoarin does not reduce catheter sepsis during total parenterai nutrition. JPEN 1985;9:348-9. Macoviak JA, Melnik KG, Rombeau JL, et al. The effect of low dose heparin on the prevention of venous thrombosis In patients receiving short-term parenterai nutrition. Curr Surg 1984;41:98-100. Wolfe R, O’Donnell T, Stone M, et al. investigation of factors determining the optimal glucose infusion rate in total parenterai nutrition. Metabolism 1980;29:892-900.

John A. Ryan, Jr. (Seattle, WA): Dr. Wolfe, you and your colleagues have provided information about complications of TPN in 1,647 patients, the largest series ever reported. The most striking finding was the recent increase in catheter sepsis associated with a new policy of using multilumen central venous catheters for TPN. Evidence has accumulated that the most important avenues for sepsis are the ingrowth of skin organisms along the outside of the catheter and breaks in a closed system that contaminate the inside of the catheter. Organisms normally found on the skin of the hospitalized patients and on the hands of hospital personnel are most commonly cultured in catheter sepsis. Despite the high incidence of anatomic infections in patients who receive TPN, hematogenous spread to the catheter from septic sites is a rare occurrence. Likewise, solution contamination rarely occurs, since refrigerated nutrient solutions are poor media for bacterial and fungal growth. Several institutions have demonstrated that the rate of catheter sepsis decreased with enforcement of a protocol for catheter care. Catheter sepsis has a direct relationship with infections at the skin entry site and with positive skin cultures. Catheter sepsis also bears a direct relationship to the number of attempted catheterizations. Skin damage by multiple needle punctures enhances the opportunity for low-grade skin infection that can lead to an infected catheter. The duration of catheter use is not related to the incidence of catheter sepsis. The incidence of suspected catheter sepsis is three times the incidence of confirmed catheter sepsis. This knowledge has allowed surgeons to exchange catheters suspected of sepsis over a guide wire, culturing the old catheter for a final confirmation. A


quality assurance committee involved with the design and enforcement of a protocol for total parenteral nutrition can keep the incidence of complications to an acceptable rate. This study details that a systematic ongoing review of all complications identified not only an increased incidence of catheter sepsis but, more importantly, found the culprit, the multilumen catheter. These large, stiff thrombogenic catheters, which by design cause multiple breaks in a closed system, led to an incidence of sepsis of 14.2 percent. Total parenteral nutrition should not be delivered in an institution unless a quality assurance mechanism is in place. In the current malpractice environment, institutions that do not abide by this recommendation will be in a disadvantageous position when significant complications occur. Eric W. Fonkalsrud (Los Angeles, CA): In our experience with over 800 children receiving TPN during the past 10 years, we have observed 27 cases of vena caval thrombosis in which the entire cava, either the superior vena cava or the inferior vena cava, was completely occluded. This eventually led to death in five patients. Thrombosis of the superior vena cava had at least a five times greater incidence of morbidity and mortality than did thrombosis of the inferior vena cava and, consequently, we have used the saphenous vein as the major access route for infusion of TPN solutions into the inferior vena cava in infants and young children. The use of 25 to 30 percent glucose solutions appears to be particularly conducive to endothelial injury and sclerosis in young children. Dr. Wolfe, have you observed vena caval thrombosis in many of your patients and, if so, how were they treated? Have you used urokinase or streptokinase to lyse thrombi? Previous experimental studies have shown that the infusion of concentrated glucose solutions directly into the right atrium or into the pulmonary artery led to a high incidence of pulmonary complications that simulated emboli. Dr. Wolfe, have you observed pulmonary obstruction in your patients who had catheters inserted into the right atrium or into the pulmonary artery? Cholelithiasis has occurred in over 5 percent of children receiving long-term TPN in our experience, particularly those in whom only very small volumes of oral feedings could be administered. Dr. Wolfe, what has been your experience with TPN-induced cholelithiasis? Henry A. Pitt (Baltimore, MD): Dr. Wolfe, you have reported a very low incidence of metabolic complications. Dr. Fonkalsrud and several colleagues at UCLA have been interested in some of the TPN-related metabolic problems, including gallbladder, liver, and bone disease, and we presented data on the high incidence and severity of TPN-associated gallbladder disease 2 years ago. Gallbladder problems and other metabolic abnormalities OCcur mainly in patients on long-term TPN. What percentage of your patients received TPN for more than 6 weeks? How were metabolic complications defined in this study? Finally, what was the incidence of TPN-associated gallbladder, liver, and bone disease? Donald B. McConnell (Portland, OR): With the wider use of the triple lumen catheters in Portland, we have noticed increased catheter sepsis as well as mortality and have concluded that more specific indications for use of these triple lumen lines are needed. Physician or nursing

The Amrkrn

Joumal ol Surgery

Complications of Parenteral Nutrition

convenience is not enough to justify their use. In addition, there should be development of a softer triple lumen line. Steatohepatitis is an insidious metabolic complication in our patients receiving long-term TPN and is associated with increases in serum glutamic pyruvate transaminase and alkaline phosphatase levels as well as jaundice in a few. Dr. Wolfe, have you encountered this complication, and do you have a protocol for its management? Bruce M. Wolfe (closing): The use of the triple lumen catheter is not entirely resolved, because many patients require additional central venous access for the purpose of hemodynamic monitoring and administration of cardioactive or vasoactive drugs. A separate protocol is necessary to determine whether total patient morbidity is decreased by putting an additional central venous catheter in place for the purpose of having a dedicated line for the TPN. With regard to the Swan-Ganz catheter, it is crucial that the TPN be administered through the central venous port and that the position of the catheter be constantly confirmed, because catheters are often repositioned to determine the wedge pressure. When this is done, the various ports on the Swan-Ganz catheter move, and when hypertonic solutions are infused into a smaller vein, it promptly results in injury to the vein wall and thrombosis. Dr. Fonkalsrud, only 7 percent of our patients were children. This presumably accounts for many of the differences that we have seen as opposed to the UCLA experience, such as thrombosis of the superior vena cava, more

Book Reviews (cont’d) expected, the sections on abdominal healing and control of adhesions are splendid. The pioneer work of the senior author in the understanding of development and clinical control of fibrous adhesions is summarized in elegant fashion, Selection of sutures and dressings is also covered well and without redundancy. A highlight of the text is a complete bibliography at the end of every chapter. I am not aware of a single significant reference that was omitted. The weaknesses of the book are not significant. The chapter on management of burns concisely presents the general consensus most surgeons already hold for proper management of thermal injuries. I do not agree with all of these concepts and will, therefore, list as a minor deficit failure to present competing views. Overemphasis on the use of mesh grafts is an example of such a topic. There are more photographs in several chapters than are needed to enhance the text, some of which will not add anything to a surgeon’s knowledge of wound healing. The publishers have done a nice job producing a clean, easy to read volume, but it is overpriced, as most medical texts are today. Overall, this book is one of the most clearly presented



often in our permanent, long-term patients. Our study was limited to inpatient complications of total parenteral nutrition. We have generally treated acute TPN-related thrombosis by removal of the catheter and have proceeded with anticoagulation only if the patient’s clinical condition would permit it and if the patient had apparent morbidity in terms of swelling or associated sepsis. We have used streptokinase in one of our patients receiving long-term TPN at home whose survival depends on our capacity to maintain permanent central venous access. Otherwise, we have not used this more hazardous and costly treatment for venous thrombosis. Dr. Pitt, the issue of metabolic complications, particularly with regard to the biliary tract and gallbladder, is difficult to resolve. We have limited ourselves to those complications that could clearly be attributed to the TPN itself in this study. Complications such as acalculous cholecystitis are prone to occur in the critically ill population whether they are given parenteral nutrition or not. We have seen bone disease, but this is a long-term complication rather than one that is seen in the hospital. Dr. McConnell, the stiffness of the triple lumen catheter is part of the reason why it induces more sepsis, although we have not confirmed a higher thrombosis rate with this catheter. Pericardial tamponade is a fatal complication when it occurs, and it is important that these catheters not be placed low in the right atrium unless they are soft Silastic catheters. Our studies indicate that hepatic steatosis is a function of the severity of illness rather than the provision of TPN as long as caloric excess is avoided.

and engaging texts available today. It is literally crammed with useful information presented in a refreshingly direct and simple style. I recommend it to students and accomplished surgeons without reservation. It is a very useful addition to the connective tissue biology and clinical surgery literature alike. Erie E. Peacock, Jr., MD Chapel Hill, North Carolina Practical Fiberoptic Bronchoscopy. 2nd Edition. By Kenkichi Oho and Ryuta Amemiya (translated by J.P. Barron). Tokyo: Kgaku-Shoin, 1984.218 pages. The excellence of the first edition of this book is admirably present in this second edition. Although only 4 years have passed since publication of the first edition, Oho and Amemiya have successfully updated the important technical developments in fiberoptic bronchoscopy instrumentation. More importantly, an entire section is devoted to their evaluation of technique, and indications and limitations of laser methods now recognized as a major development in the field. Their description of bronchscopic instrumentation and of the normal anatomy of the bronchus remains unsurpassed. The photographic presenta-

on page 109.