ASAIO Journal 2014
Adult Circulatory Support
Safety of Parenteral Nutrition in Patients Receiving a Ventricular Assist Device Corey Scurlock,*† Sean P. Pinney,‡ Hung-Mo Lin,* Matthew Potenza,§ Aaron J. Weiss,† Neeha Zaidi,¶ Anelechi Anyanwu,† and Jeffrey I. Mechanick§
Patients with advanced heart failure and poor nutritional status are predisposed to higher rates of infection, bleeding, and mortality. We have increasingly used perioperative parenteral nutrition (PN) in ventricular assist device (VAD) patients and now report our initial experience. We performed a retrospective review of 43 consecutive patients who received implantable VADs from 2006 to 2009. We compared outcomes for patients receiving PN for >7 days perioperatively vs ≤7 days. In addition, we compared patients who received preoperative enteral nutrition (EN) with those who did not. Fourteen patients received perioperative PN in addition to EN for >7 days compared with 29 patients who received either PN for ≤7 days or EN alone. Univariate analysis showed no differences in infection, bleeding, thrombus, stroke, length of stay, or mortality. Multivariate stepwise regression including EN, preoperative PN, Interagency Registry for Mechanically Assisted Circulation score, age, gender, and VAD indication showed that only EN was associated with infection. Prolonged use of perioperative PN appears to be safe and well tolerated in patients undergoing VAD implantation. Preoperative EN, while increasing infection risk, seems to have no harmful effect on survival. ASAIO Journal 2014; 60:376–380.
found to be predictive of 90 day mortality (odds ratio [OR] of 5.7).1 This is consistent with findings that lean muscle loss is an independent risk factor for mortality in heart failure2,3 and that an energy debt greater than approximately 10,000 kcal is associated with complications in critically ill patients.4 Moreover, cardiac cachexia and gastrointestinal dysfunction frequently occur with advanced heart failure. This condition is cytokine-mediated (primarily involving tumor necrosis factor-α) and associated with splanchnic hypoperfusion resulting in anorexia and impaired nutrient absorption.5,6 Although inflammatory events driving cardiac cachexia are not thought to be nutritionally responsive, we assert that the improvement in a patient’s overall nutritional status before ventricular assist device (VAD) implantation could reduce the energy debt and consequent malnutrition-associated mortality. Because splanchnic hypoperfusion may initially worsen following VAD placement before it improves,7 oral nutrition or enteral nutrition (EN; tube feeds) may be inadequate and potentially hazardous by inducing intestinal ischemia.8 On the contrary, parenteral nutrition (PN) circumvents the intestine. The ability to deliver nutrients has a salutary effect on lean muscle loss, the extent of which depends on the degree of cytokine-mediated catabolism.9,10 Previous studies have demonstrated a benefit in using preoperative PN in patients with cardiac cachexia undergoing mitral valve repair surgery.11 Parenteral nutrition has also been used successfully to prevent the progression of cancer cachexia.12,13 However, there is a general reluctance to use PN in patients undergoing VAD implantation because of historical concerns primarily about infections. Nevertheless, the risk of infection in the present era may be greatly reduced with improved patient selection, increased utilization of smaller continuous-flow devices, and advances in operative and postoperative management such as intensive insulin therapy (IIT) and mandated adherence with central line–associated bacteremia (CLAB) protocols.14–16 In our center, we have applied aggressive nutrition support and IIT protocols (intensive metabolic support; IMS) since 2006.17 The primary purpose of this report was to retrospectively analyze our data to determine the safety of perioperative PN in this high-risk patient population. This is a first step before we can proceed with the design of potential prospective, interventional clinical trials of IMS in patients undergoing VAD placement.
Key Words: ventricular assist device, parenteral nutrition, heart failure, critical illness
Poor
nutritional status is a predictor of mortality in patients undergoing left ventricular assist device (LVAD) therapy; a serum albumin level below 3.3 g/dl, as a surrogate marker for malnutrition and metabolic stress, has been
From the *Department of Anesthesiology, Icahn School of Medicine at Mount Sinai, New York, New York; †Department of Cardiothoracic Surgery, Icahn School of Medicine at Mount Sinai, New York, New York; ‡Department of Medicine, Division of Cardiology, Icahn School of Medicine at Mount Sinai, New York, New York; §Department of Medicine, Division of Endocrinology, Diabetes, and Bone Disease, Icahn School of Medicine at Mount Sinai, New York, New York; and ¶Icahn School of Medicine at Mount Sinai, New York, New York. Submitted for consideration September 2013; accepted for publication in revised form February 2014. Disclosure: The authors have no conflicts of interest to report. Correspondence: Sean P. Pinney, MD, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1030, New York, NY 10029. Email: [email protected]. Copyright © 2014 by the American Society for Artificial Internal Organs DOI: 10.1097/MAT.0000000000000078
376
SAFETY OF LVAD PARENTERAL NUTRITION
Methods Patients All patients (n = 43) who received an implantable VAD between June 2006 and April 2009 were included in this study. Once a decision was made to implant a VAD, patients were transferred to a specialized cardiac care unit or telemetry unit, if not already admitted to these floors. In addition, the Metabolic Support Team was consulted for nutritional assessment, and a decision was made regarding continuation of preoperative oral nutrition only or adding EN, PN, or both. When possible, VAD implantation was delayed for 2 to 5 days to allow medical and nutritional optimization. All patients received prophylactic antibiotics for the first 48 hours after implantation, including vancomycin, levofloxacin, rifampin, and fluconazole, and were free of systemic infection at the time of surgery. All patients received the same protocol for central venous cannulation and line maintenance to reduce the risk of a CLAB. All VAD implants were performed by a single surgeon using standard techniques with the choice of VAD depending on various patient, logistic, and team preference factors. Nineteen of the VADs placed were investigational devices. We defined complications using the criteria of the HeartMate II investigators.18 We compared these parameters between the patients who received perioperative PN for ≤7 days and >7 days. Nutritional Interventions The criterion to use PN peri-VAD implantation was based on a definite or perceived inability to meet nutritional needs enterally, such as anorexia, infrequent or inadequate oral intake, refusal of or contraindication for nasogastric/enteric tube access, or presumed hypoperfusion-related gut dysfunction evidenced by pain, cramps, bloating, diarrhea, and so on. Enteral nutrition was not withheld on the basis of concomitant vasopressor or inotrope use. Nutritional needs to suppress starvation-related catabolism were defined as at least 20 kcal/kg/d.19 Parenteral nutrition was not provided preoperatively to patients if they were not initially evaluated by the Metabolic Support Team (typically because VAD implantation was required emergently or semiemergently) but was started immediately postoperatively if the above criteria were met. Patients who tolerated sufficient oral or EN immediately pre- or postoperatively to achieve at least 20 kcal/kg/d were not placed on PN postoperatively. Once commenced, PN was continued using a nutrition protocol with an increased target combined EN + PN daily caloric intake of 25–30 kcal/kg. Parenteral nutrition was discontinued once enteral intake alone was adequate (as determined by the Metabolic Support Team and intensivist team taking care of the patient postoperatively). Preoperatively, patients destined to receive VAD implantation who were less critically ill (i.e., Interagency Registry for Mechanically Assisted Circulation [INTERMACS] 3 or above)20 and supported with oral nutrition alone were additionally managed with subcutaneous (SQ) insulin to achieve target blood glucose (BG) levels consistent with American Association of Clinical Endocrinologists guidelines (7 days (N = 14)
Perioperative PN Exposure ≤7 days (N = 29)
p Value
53.70 ± 13.26 12 (85.71%) 27.95 ± 8.15 30.29 ± 15.45 32.15 ± 4.80 1.66 ± 0.68
53.29 ± 11.95 23 (79.31%) 29.05 ± 7.53 35.52 ± 19.64 31.38 ± 4.41 1.47 ± 0.38
0.92* 1.00† 0.66* 0.39* 0.61* 0.33*
48 [10–2,367]
66 [11–7,000]
1.00†
185.14 ± 74.89 3.23 ± 0.70 1.43 ± 0.52
201.00 ± 73.71 3.14 ± 0.73 1.58 ± 0.67
0.51* 0.70* 0.45* 1.00‡
4 (28.57%) 8 (57.14%) 2 (14.29%)
8 (27.59%) 17 (58.62%) 4 (13.79%)
1 (7.14%) 2 (14.29%)
3 (10.34%) 5 (17.24%)
10 (71.43%)
19 (65.52%)
1 (7.14%)
2 (6.90%)
1 (7.14%) 13 (92.86%)
4 (13.79%) 25 (86.21%)
1.00‡
1.00‡
Data are presented as mean ± SD, median [range] if the variable is not normally distributed, or n (%) if the variable is categorical. *Two-sample t-test. †Wilcoxon rank sums test. ‡Fisher’s exact test. §The category “Other” under VAD indication included myocarditis, allograft rejection, and postcardiotomy. ¶The category “Pulsatile” under VAD device type included HMXVE, Thoratec PVAD, Thoratec IVAD, and “Nonpulsatile” included HMII, VentrAssist, Jarvik, and CentriMag. BMI, body mass index; BUN, blood urea nitrogen; MI, myocardial infarction; PN, parenteral nutrition; VAD, ventricular assist device.
stroke, or thrombosis. Nor was there a statistically significant difference in infection incidence. Stepwise logistic regression was used to demonstrate association but not causation (Table 4). In this analysis, the use of preoperative EN was identified as the sole significant risk factor of infection. The use of PN, regardless of duration, did not appear to be associated with an increased risk of infection or as an effect modifier of EN. Figure 1 graphically illustrates the probability of infection plotted against the number of days of perioperative EN (total perioperative days on tube feeds). The bar on day 0 is the infection rate among those who did not receive any EN, which was approximately 20%. Noticeably, the risk increased to approximately 55% on day 3 and then remained unchanged. There is a statistically significant difference between the first bar (day 0) and the rest of the bars (days 3–5, days 6–10, day 11, day 12, day >12, p = 0.032) demonstrating that the infection rate increased from baseline once patients received tube feeds. However, no significant difference were seen between bars 2 through 6.
379
SAFETY OF LVAD PARENTERAL NUTRITION
Table 3. Univariate Analysis of Preoperative and Perioperative Outcomes, by Exposure to More than 7 Days of Perioperative PN Outcome
Yes (N = 14)
No (N = 29)
Odds Ratio
95% CI for OR
p Value
Days on PN, days Days on tube feeds Days in ICU Days in hospital Infection Death CVA Thrombus Bleed
9 [8–20] 12 [0–20] 16 [6–201] 47 [13–231] 8 (57.1%) 4 (28.6%) 3 (21.4%) 2 (14.3%) 5 (35.7%)
3 [0–7] 3 [0–7] 7 [2–139] 28 [2–163] 10 (34.5%) 8 (27.6%) 4 (13.8%) 2 (6.9%) 5 (17.2%)
— — — — 2.53 1.05 1.70 2.25 2.67
— — — — (0.69–9.36) (0.25–4.33) (0.33–8.93) (0.28–17.91) (0.62–11.45)
7 days) PN for these patients is relevant as historically physicians have avoided the use of PN in at-risk patients with a VAD because of concerns about hematogenous infection. The absence of increased morbidity despite liberal use of PN can be explained by several factors. Replacing essential nutrients in an undernourished patient in a safe format may make patients more immunocompetent and thus reduce infection risk. Catheter-related infection is of prime concern when assessing the safety of PN. However, with strict prevention and treatment routines, we did not find any correlation between such infections and major device-related infections. This suggests that there was no increased infective risk of PN when used with tight glycemic control and a proper CLAB prevention protocol, as has been previously suggested.14,15 Another concern of PN safety is the potential for coagulopathy and bleeding. This is because parenteral lipid emulsions can affect platelet function and clotting factors.22 Although no statistically significant differences were found between the two groups, the trend toward increased bleeding in the PN group could suggest that lipids could be infused at lower doses or infusion rates, or omitted altogether, as a precaution against postoperative bleeding events.
The patients who received preoperative EN had a higher rate of infection that was statistically significant without a concomitant increase in mortality. Although speculative, this higher rate of infection could be because of increased rates of pneumonia and aspirations; however, our study was not designed to look at either of these complications in its original data collection. Our results are similar to previous studies that have shown that although early EN is associated with increased rates of pneumonia, it is also protective against ICU and hospital-related mortality.23 In our study, this was most likely because of EN preventing the accumulation of an energy deficit, helping to maintain enterocyte integrity, and attenuating the hypermetabolic response of surgery. With statistical modeling, there appeared to be a bimodal increased probability of infection with the time spent on EN. Although the first peak of infection is likely related to aspiration pneumonia, we could not identify a clear mechanism for later infections; it may well be that late infections are not causally related but rather an indirect association with prolonged critical illness.
Table 4. Multivariate Stepwise Regression Analysis of Infection Risk Factors
Effect EN Preoperative PN Bridge LVAD Age Male INTERMACS ≥ 2
Odds Ratio Estimate
95% Wald Confidence Limits
p Value
7.575 1.349 0.973 1.012 1.920 0.997
1.380–41.589 0.239–7.628 0.143–6.634 0.947–1.081 0.267–13.814 0.210–4.734
0.0198 0.7348 0.9780 0.7280 0.5169 0.9975
EN, enteral nutrition; INTERMACS, Interagency Registry for Mechanically Assisted Circulation; LVAD, left ventricular assist device; PN, parenteral nutrition.
Figure 1. A statistical model of probability of infection as a function of days on enteral nutrition. The six bars in the table indicate the infection rate (%) at various time intervals (days; number of patients is in parentheses). There was a statistically significant difference between the first bar (day 0) and the rest of the bars (days 3–5, days 6–10, day 11, day 12, day >12, p = 0.032), showing that the infection rate increased from baseline once patients received tube feeds. No significant differences were seen between bars two through six. Error bars represent SD.
380 SCURLOCK et al. The strength of this retrospective study is that patients were treated according to real world, established heart failure, VAD, nutrition, and glycemic control parameters by the same team in a relatively standardized fashion. The decision to use perioperative PN was made jointly by the cardiologists, surgeons, and the intensivists in consultation with the Metabolic Support Team. This study is unique, in that we have liberally applied perioperative PN to VAD patients and in many cases maintained it for several days as opposed to the more traditional approach of postoperative underfeeding with EN commenced on the second or later postoperative day. Continuous PN confers consistent nutritional support as opposed to the frequent interruptions occurring with EN that contribute to underfeeding and glycemic variability. Although we do not have a control group, we believe that if a substantial increase in morbidity was associated with PN, then we should have observed at least a tendency toward an increase in morbidity in our initial experience. The weaknesses of this study include the nonrandomized and retrospective design, as well as the heterogeneity in the patients, nutritional strategies, and timing of therapy. In addition, the sample size was relatively small and therefore predisposes to type II statistical error. While we could have increased the sample size by further chart review, i.e., more than 3 years, this would introduce more heterogeneity into the study (because of older VAD models and less standardized perioperative management regimes) and render the results less relevant. For the same reasons, we felt it would not be appropriate to compare the study groups with historical controls who did not receive PN. Although this study demonstrated the safety of perioperative PN in VAD recipients, it was not designed to demonstrate a benefit. In addition, the small sample size may have limited the power of the study to show a benefit in mortality. Pooling data from other institutions is another way to increase sample size, but this would introduce more variables across teams and methods. Lastly, we included a variety of implantable pumps in this analysis, including first-generation pulsatile VADs and investigational VADs, which are no longer manufactured. This may have confounded our results. Conclusion Preoperative and prolonged perioperative PN (>7 days) did not confer increased risk in patients undergoing VAD implantation. Preoperative EN, while increasing the risk of infection, seems to have no harmful effect on mortality in VAD patients. Given its theoretical benefits, PN may be a useful adjunct to other perioperative strategies for optimizing outcomes of VAD implantation. Further studies are required to confirm these findings and to document the efficacy of this approach. References 1. Lietz K, Long JW, Kfoury AG, et al: Outcomes of left ventricular assist device implantation as destination therapy in the post-REMATCH era: Implications for patient selection. Circulation 116: 497–505, 2007. 2. Anker SD, Negassa A, Coats AJ, et al: Prognostic importance of weight loss in chronic heart failure and the effect of treatment
with angiotensin-converting-enzyme inhibitors: An observational study. Lancet 361: 1077–1083, 2003. 3. Anker SD, Ponikowski P, Varney S, et al: Wasting as independent risk factor for mortality in chronic heart failure. Lancet 349: 1050–1053, 1997. 4. Villet S, Chiolero RL, Bollmann MD, et al: Negative impact of hypocaloric feeding and energy balance on clinical outcome in ICU patients. Clin Nutr 24: 502–509, 2005. 5. Anker SD, Ponikowski PP, Clark AL, et al: Cytokines and neurohormones relating to body composition alterations in the wasting syndrome of chronic heart failure. Eur Heart J 20: 683–693, 1999. 6. Sandek A, Bauditz J, Swidsinski A, et al: Altered intestinal function in patients with chronic heart failure. J Am Coll Cardiol 50: 1561–1569, 2007. 7. el-Amir NG, Gardocki M, Levin HR, et al: Gastrointestinal consequences of left ventricular assist device placement. ASAIO J 42: 150–153, 1996. 8. Sandek A, Rauchhaus M, Anker SD, von Haehling S: The emerging role of the gut in chronic heart failure. Curr Opin Clin Nutr Metab Care 11: 632–639, 2008. 9. Shaw JH, Wildbore M, Wolfe RR: Whole body protein kinetics in severely septic patients. The response to glucose infusion and total parenteral nutrition. Ann Surg 205: 288–294, 1987. 10. Ishibashi N, Plank LD, Sando K, Hill GL: Optimal protein requirements during the first 2 weeks after the onset of critical illness. Crit Care Med 26: 1529–1535, 1998. 11. Paccagnella A, Calò MA, Caenaro G, et al: Cardiac cachexia: Preoperative and postoperative nutrition management. JPEN J Parenter Enteral Nutr 18: 409–416, 1994. 12. Lundholm K, Daneryd P, Bosaeus I, Körner U, Lindholm E: Palliative nutritional intervention in addition to cyclooxygenase and erythropoietin treatment for patients with malignant disease: Effects on survival, metabolism, and function. Cancer 100: 1967–1977, 2004. 13. Shang E, Weiss C, Post S, Kaehler G: The influence of early supplementation of parenteral nutrition on quality of life and body composition in patients with advanced cancer. JPEN J Parenter Enteral Nutr 30: 222–230, 2006. 14. van den Berghe G, Wouters P, Weekers F, et al: Intensive insulin therapy in critically ill patients. N Engl J Med 345: 1359–1367, 2001. 15. Pronovost P, Needham D, Berenholtz S, et al: An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 355: 2725–2732, 2006. 16. Lazzeri C, Bevilacqua S, Ciappi F, Pratesi C, Gensini GF, Romagnoli S: Glucose metabolism in cardiovascular surgery. HSR Proc Intensive Care Cardiovasc Anesth 2: 19–26, 2010. 17. Scurlock C, Raikhelkar J, Mechanick JI: Intensive metabolic support: evolution and revolution. Endocr Pract 14: 1047–1054, 2008. 18. Slaughter MS, Rogers JG, Milano CA, et al; HeartMate II Investigators: Advanced heart failure treated with c ontinuous-flow left ventricular assist device. N Engl J Med 361: 2241–2251, 2009. 19. Debaveye Y, Van den Berghe G: Risks and benefits of nutritional support during critical illness. Annu Rev Nutr 26: 513–538, 2006. 20. Kirklin JK, Naftel DC, Stevenson LW, et al: INTERMACS database for durable devices for circulatory support: First annual report. J Heart Lung Transplant 27: 1065–1072, 2008. 21. Rodbard HW, Blonde L, Braithwaite SS, et al; AACE Diabetes Mellitus Clinical Practice Guidelines Task Force: American Association of Clinical Endocrinologists medical guidelines for clinical practice for the management of diabetes mellitus. Endocr Pract 13(suppl 1): 1–68, 2007. 22. Duerksen DR, Papineau N: The prevalence of coagulation abnormalities in hospitalized patients receiving lipid-based parenteral nutrition. JPEN J Parenter Enteral Nutr 28: 30–33, 2004. 23. Artinian V, Krayem H, DiGiovine B: Effects of early enteral feeding on the outcome of critically ill mechanically ventilated medical patients. Chest 129: 960–967, 2006.
Adult Circulatory Support
Safety of Parenteral Nutrition in Patients Receiving a Ventricular Assist Device Corey Scurlock,*† Sean P. Pinney,‡ Hung-Mo Lin,* Matthew Potenza,§ Aaron J. Weiss,† Neeha Zaidi,¶ Anelechi Anyanwu,† and Jeffrey I. Mechanick§
Patients with advanced heart failure and poor nutritional status are predisposed to higher rates of infection, bleeding, and mortality. We have increasingly used perioperative parenteral nutrition (PN) in ventricular assist device (VAD) patients and now report our initial experience. We performed a retrospective review of 43 consecutive patients who received implantable VADs from 2006 to 2009. We compared outcomes for patients receiving PN for >7 days perioperatively vs ≤7 days. In addition, we compared patients who received preoperative enteral nutrition (EN) with those who did not. Fourteen patients received perioperative PN in addition to EN for >7 days compared with 29 patients who received either PN for ≤7 days or EN alone. Univariate analysis showed no differences in infection, bleeding, thrombus, stroke, length of stay, or mortality. Multivariate stepwise regression including EN, preoperative PN, Interagency Registry for Mechanically Assisted Circulation score, age, gender, and VAD indication showed that only EN was associated with infection. Prolonged use of perioperative PN appears to be safe and well tolerated in patients undergoing VAD implantation. Preoperative EN, while increasing infection risk, seems to have no harmful effect on survival. ASAIO Journal 2014; 60:376–380.
found to be predictive of 90 day mortality (odds ratio [OR] of 5.7).1 This is consistent with findings that lean muscle loss is an independent risk factor for mortality in heart failure2,3 and that an energy debt greater than approximately 10,000 kcal is associated with complications in critically ill patients.4 Moreover, cardiac cachexia and gastrointestinal dysfunction frequently occur with advanced heart failure. This condition is cytokine-mediated (primarily involving tumor necrosis factor-α) and associated with splanchnic hypoperfusion resulting in anorexia and impaired nutrient absorption.5,6 Although inflammatory events driving cardiac cachexia are not thought to be nutritionally responsive, we assert that the improvement in a patient’s overall nutritional status before ventricular assist device (VAD) implantation could reduce the energy debt and consequent malnutrition-associated mortality. Because splanchnic hypoperfusion may initially worsen following VAD placement before it improves,7 oral nutrition or enteral nutrition (EN; tube feeds) may be inadequate and potentially hazardous by inducing intestinal ischemia.8 On the contrary, parenteral nutrition (PN) circumvents the intestine. The ability to deliver nutrients has a salutary effect on lean muscle loss, the extent of which depends on the degree of cytokine-mediated catabolism.9,10 Previous studies have demonstrated a benefit in using preoperative PN in patients with cardiac cachexia undergoing mitral valve repair surgery.11 Parenteral nutrition has also been used successfully to prevent the progression of cancer cachexia.12,13 However, there is a general reluctance to use PN in patients undergoing VAD implantation because of historical concerns primarily about infections. Nevertheless, the risk of infection in the present era may be greatly reduced with improved patient selection, increased utilization of smaller continuous-flow devices, and advances in operative and postoperative management such as intensive insulin therapy (IIT) and mandated adherence with central line–associated bacteremia (CLAB) protocols.14–16 In our center, we have applied aggressive nutrition support and IIT protocols (intensive metabolic support; IMS) since 2006.17 The primary purpose of this report was to retrospectively analyze our data to determine the safety of perioperative PN in this high-risk patient population. This is a first step before we can proceed with the design of potential prospective, interventional clinical trials of IMS in patients undergoing VAD placement.
Key Words: ventricular assist device, parenteral nutrition, heart failure, critical illness
Poor
nutritional status is a predictor of mortality in patients undergoing left ventricular assist device (LVAD) therapy; a serum albumin level below 3.3 g/dl, as a surrogate marker for malnutrition and metabolic stress, has been
From the *Department of Anesthesiology, Icahn School of Medicine at Mount Sinai, New York, New York; †Department of Cardiothoracic Surgery, Icahn School of Medicine at Mount Sinai, New York, New York; ‡Department of Medicine, Division of Cardiology, Icahn School of Medicine at Mount Sinai, New York, New York; §Department of Medicine, Division of Endocrinology, Diabetes, and Bone Disease, Icahn School of Medicine at Mount Sinai, New York, New York; and ¶Icahn School of Medicine at Mount Sinai, New York, New York. Submitted for consideration September 2013; accepted for publication in revised form February 2014. Disclosure: The authors have no conflicts of interest to report. Correspondence: Sean P. Pinney, MD, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1030, New York, NY 10029. Email: [email protected]. Copyright © 2014 by the American Society for Artificial Internal Organs DOI: 10.1097/MAT.0000000000000078
376
SAFETY OF LVAD PARENTERAL NUTRITION
Methods Patients All patients (n = 43) who received an implantable VAD between June 2006 and April 2009 were included in this study. Once a decision was made to implant a VAD, patients were transferred to a specialized cardiac care unit or telemetry unit, if not already admitted to these floors. In addition, the Metabolic Support Team was consulted for nutritional assessment, and a decision was made regarding continuation of preoperative oral nutrition only or adding EN, PN, or both. When possible, VAD implantation was delayed for 2 to 5 days to allow medical and nutritional optimization. All patients received prophylactic antibiotics for the first 48 hours after implantation, including vancomycin, levofloxacin, rifampin, and fluconazole, and were free of systemic infection at the time of surgery. All patients received the same protocol for central venous cannulation and line maintenance to reduce the risk of a CLAB. All VAD implants were performed by a single surgeon using standard techniques with the choice of VAD depending on various patient, logistic, and team preference factors. Nineteen of the VADs placed were investigational devices. We defined complications using the criteria of the HeartMate II investigators.18 We compared these parameters between the patients who received perioperative PN for ≤7 days and >7 days. Nutritional Interventions The criterion to use PN peri-VAD implantation was based on a definite or perceived inability to meet nutritional needs enterally, such as anorexia, infrequent or inadequate oral intake, refusal of or contraindication for nasogastric/enteric tube access, or presumed hypoperfusion-related gut dysfunction evidenced by pain, cramps, bloating, diarrhea, and so on. Enteral nutrition was not withheld on the basis of concomitant vasopressor or inotrope use. Nutritional needs to suppress starvation-related catabolism were defined as at least 20 kcal/kg/d.19 Parenteral nutrition was not provided preoperatively to patients if they were not initially evaluated by the Metabolic Support Team (typically because VAD implantation was required emergently or semiemergently) but was started immediately postoperatively if the above criteria were met. Patients who tolerated sufficient oral or EN immediately pre- or postoperatively to achieve at least 20 kcal/kg/d were not placed on PN postoperatively. Once commenced, PN was continued using a nutrition protocol with an increased target combined EN + PN daily caloric intake of 25–30 kcal/kg. Parenteral nutrition was discontinued once enteral intake alone was adequate (as determined by the Metabolic Support Team and intensivist team taking care of the patient postoperatively). Preoperatively, patients destined to receive VAD implantation who were less critically ill (i.e., Interagency Registry for Mechanically Assisted Circulation [INTERMACS] 3 or above)20 and supported with oral nutrition alone were additionally managed with subcutaneous (SQ) insulin to achieve target blood glucose (BG) levels consistent with American Association of Clinical Endocrinologists guidelines (7 days (N = 14)
Perioperative PN Exposure ≤7 days (N = 29)
p Value
53.70 ± 13.26 12 (85.71%) 27.95 ± 8.15 30.29 ± 15.45 32.15 ± 4.80 1.66 ± 0.68
53.29 ± 11.95 23 (79.31%) 29.05 ± 7.53 35.52 ± 19.64 31.38 ± 4.41 1.47 ± 0.38
0.92* 1.00† 0.66* 0.39* 0.61* 0.33*
48 [10–2,367]
66 [11–7,000]
1.00†
185.14 ± 74.89 3.23 ± 0.70 1.43 ± 0.52
201.00 ± 73.71 3.14 ± 0.73 1.58 ± 0.67
0.51* 0.70* 0.45* 1.00‡
4 (28.57%) 8 (57.14%) 2 (14.29%)
8 (27.59%) 17 (58.62%) 4 (13.79%)
1 (7.14%) 2 (14.29%)
3 (10.34%) 5 (17.24%)
10 (71.43%)
19 (65.52%)
1 (7.14%)
2 (6.90%)
1 (7.14%) 13 (92.86%)
4 (13.79%) 25 (86.21%)
1.00‡
1.00‡
Data are presented as mean ± SD, median [range] if the variable is not normally distributed, or n (%) if the variable is categorical. *Two-sample t-test. †Wilcoxon rank sums test. ‡Fisher’s exact test. §The category “Other” under VAD indication included myocarditis, allograft rejection, and postcardiotomy. ¶The category “Pulsatile” under VAD device type included HMXVE, Thoratec PVAD, Thoratec IVAD, and “Nonpulsatile” included HMII, VentrAssist, Jarvik, and CentriMag. BMI, body mass index; BUN, blood urea nitrogen; MI, myocardial infarction; PN, parenteral nutrition; VAD, ventricular assist device.
stroke, or thrombosis. Nor was there a statistically significant difference in infection incidence. Stepwise logistic regression was used to demonstrate association but not causation (Table 4). In this analysis, the use of preoperative EN was identified as the sole significant risk factor of infection. The use of PN, regardless of duration, did not appear to be associated with an increased risk of infection or as an effect modifier of EN. Figure 1 graphically illustrates the probability of infection plotted against the number of days of perioperative EN (total perioperative days on tube feeds). The bar on day 0 is the infection rate among those who did not receive any EN, which was approximately 20%. Noticeably, the risk increased to approximately 55% on day 3 and then remained unchanged. There is a statistically significant difference between the first bar (day 0) and the rest of the bars (days 3–5, days 6–10, day 11, day 12, day >12, p = 0.032) demonstrating that the infection rate increased from baseline once patients received tube feeds. However, no significant difference were seen between bars 2 through 6.
379
SAFETY OF LVAD PARENTERAL NUTRITION
Table 3. Univariate Analysis of Preoperative and Perioperative Outcomes, by Exposure to More than 7 Days of Perioperative PN Outcome
Yes (N = 14)
No (N = 29)
Odds Ratio
95% CI for OR
p Value
Days on PN, days Days on tube feeds Days in ICU Days in hospital Infection Death CVA Thrombus Bleed
9 [8–20] 12 [0–20] 16 [6–201] 47 [13–231] 8 (57.1%) 4 (28.6%) 3 (21.4%) 2 (14.3%) 5 (35.7%)
3 [0–7] 3 [0–7] 7 [2–139] 28 [2–163] 10 (34.5%) 8 (27.6%) 4 (13.8%) 2 (6.9%) 5 (17.2%)
— — — — 2.53 1.05 1.70 2.25 2.67
— — — — (0.69–9.36) (0.25–4.33) (0.33–8.93) (0.28–17.91) (0.62–11.45)
7 days) PN for these patients is relevant as historically physicians have avoided the use of PN in at-risk patients with a VAD because of concerns about hematogenous infection. The absence of increased morbidity despite liberal use of PN can be explained by several factors. Replacing essential nutrients in an undernourished patient in a safe format may make patients more immunocompetent and thus reduce infection risk. Catheter-related infection is of prime concern when assessing the safety of PN. However, with strict prevention and treatment routines, we did not find any correlation between such infections and major device-related infections. This suggests that there was no increased infective risk of PN when used with tight glycemic control and a proper CLAB prevention protocol, as has been previously suggested.14,15 Another concern of PN safety is the potential for coagulopathy and bleeding. This is because parenteral lipid emulsions can affect platelet function and clotting factors.22 Although no statistically significant differences were found between the two groups, the trend toward increased bleeding in the PN group could suggest that lipids could be infused at lower doses or infusion rates, or omitted altogether, as a precaution against postoperative bleeding events.
The patients who received preoperative EN had a higher rate of infection that was statistically significant without a concomitant increase in mortality. Although speculative, this higher rate of infection could be because of increased rates of pneumonia and aspirations; however, our study was not designed to look at either of these complications in its original data collection. Our results are similar to previous studies that have shown that although early EN is associated with increased rates of pneumonia, it is also protective against ICU and hospital-related mortality.23 In our study, this was most likely because of EN preventing the accumulation of an energy deficit, helping to maintain enterocyte integrity, and attenuating the hypermetabolic response of surgery. With statistical modeling, there appeared to be a bimodal increased probability of infection with the time spent on EN. Although the first peak of infection is likely related to aspiration pneumonia, we could not identify a clear mechanism for later infections; it may well be that late infections are not causally related but rather an indirect association with prolonged critical illness.
Table 4. Multivariate Stepwise Regression Analysis of Infection Risk Factors
Effect EN Preoperative PN Bridge LVAD Age Male INTERMACS ≥ 2
Odds Ratio Estimate
95% Wald Confidence Limits
p Value
7.575 1.349 0.973 1.012 1.920 0.997
1.380–41.589 0.239–7.628 0.143–6.634 0.947–1.081 0.267–13.814 0.210–4.734
0.0198 0.7348 0.9780 0.7280 0.5169 0.9975
EN, enteral nutrition; INTERMACS, Interagency Registry for Mechanically Assisted Circulation; LVAD, left ventricular assist device; PN, parenteral nutrition.
Figure 1. A statistical model of probability of infection as a function of days on enteral nutrition. The six bars in the table indicate the infection rate (%) at various time intervals (days; number of patients is in parentheses). There was a statistically significant difference between the first bar (day 0) and the rest of the bars (days 3–5, days 6–10, day 11, day 12, day >12, p = 0.032), showing that the infection rate increased from baseline once patients received tube feeds. No significant differences were seen between bars two through six. Error bars represent SD.
380 SCURLOCK et al. The strength of this retrospective study is that patients were treated according to real world, established heart failure, VAD, nutrition, and glycemic control parameters by the same team in a relatively standardized fashion. The decision to use perioperative PN was made jointly by the cardiologists, surgeons, and the intensivists in consultation with the Metabolic Support Team. This study is unique, in that we have liberally applied perioperative PN to VAD patients and in many cases maintained it for several days as opposed to the more traditional approach of postoperative underfeeding with EN commenced on the second or later postoperative day. Continuous PN confers consistent nutritional support as opposed to the frequent interruptions occurring with EN that contribute to underfeeding and glycemic variability. Although we do not have a control group, we believe that if a substantial increase in morbidity was associated with PN, then we should have observed at least a tendency toward an increase in morbidity in our initial experience. The weaknesses of this study include the nonrandomized and retrospective design, as well as the heterogeneity in the patients, nutritional strategies, and timing of therapy. In addition, the sample size was relatively small and therefore predisposes to type II statistical error. While we could have increased the sample size by further chart review, i.e., more than 3 years, this would introduce more heterogeneity into the study (because of older VAD models and less standardized perioperative management regimes) and render the results less relevant. For the same reasons, we felt it would not be appropriate to compare the study groups with historical controls who did not receive PN. Although this study demonstrated the safety of perioperative PN in VAD recipients, it was not designed to demonstrate a benefit. In addition, the small sample size may have limited the power of the study to show a benefit in mortality. Pooling data from other institutions is another way to increase sample size, but this would introduce more variables across teams and methods. Lastly, we included a variety of implantable pumps in this analysis, including first-generation pulsatile VADs and investigational VADs, which are no longer manufactured. This may have confounded our results. Conclusion Preoperative and prolonged perioperative PN (>7 days) did not confer increased risk in patients undergoing VAD implantation. Preoperative EN, while increasing the risk of infection, seems to have no harmful effect on mortality in VAD patients. Given its theoretical benefits, PN may be a useful adjunct to other perioperative strategies for optimizing outcomes of VAD implantation. Further studies are required to confirm these findings and to document the efficacy of this approach. References 1. Lietz K, Long JW, Kfoury AG, et al: Outcomes of left ventricular assist device implantation as destination therapy in the post-REMATCH era: Implications for patient selection. Circulation 116: 497–505, 2007. 2. Anker SD, Negassa A, Coats AJ, et al: Prognostic importance of weight loss in chronic heart failure and the effect of treatment
with angiotensin-converting-enzyme inhibitors: An observational study. Lancet 361: 1077–1083, 2003. 3. Anker SD, Ponikowski P, Varney S, et al: Wasting as independent risk factor for mortality in chronic heart failure. Lancet 349: 1050–1053, 1997. 4. Villet S, Chiolero RL, Bollmann MD, et al: Negative impact of hypocaloric feeding and energy balance on clinical outcome in ICU patients. Clin Nutr 24: 502–509, 2005. 5. Anker SD, Ponikowski PP, Clark AL, et al: Cytokines and neurohormones relating to body composition alterations in the wasting syndrome of chronic heart failure. Eur Heart J 20: 683–693, 1999. 6. Sandek A, Bauditz J, Swidsinski A, et al: Altered intestinal function in patients with chronic heart failure. J Am Coll Cardiol 50: 1561–1569, 2007. 7. el-Amir NG, Gardocki M, Levin HR, et al: Gastrointestinal consequences of left ventricular assist device placement. ASAIO J 42: 150–153, 1996. 8. Sandek A, Rauchhaus M, Anker SD, von Haehling S: The emerging role of the gut in chronic heart failure. Curr Opin Clin Nutr Metab Care 11: 632–639, 2008. 9. Shaw JH, Wildbore M, Wolfe RR: Whole body protein kinetics in severely septic patients. The response to glucose infusion and total parenteral nutrition. Ann Surg 205: 288–294, 1987. 10. Ishibashi N, Plank LD, Sando K, Hill GL: Optimal protein requirements during the first 2 weeks after the onset of critical illness. Crit Care Med 26: 1529–1535, 1998. 11. Paccagnella A, Calò MA, Caenaro G, et al: Cardiac cachexia: Preoperative and postoperative nutrition management. JPEN J Parenter Enteral Nutr 18: 409–416, 1994. 12. Lundholm K, Daneryd P, Bosaeus I, Körner U, Lindholm E: Palliative nutritional intervention in addition to cyclooxygenase and erythropoietin treatment for patients with malignant disease: Effects on survival, metabolism, and function. Cancer 100: 1967–1977, 2004. 13. Shang E, Weiss C, Post S, Kaehler G: The influence of early supplementation of parenteral nutrition on quality of life and body composition in patients with advanced cancer. JPEN J Parenter Enteral Nutr 30: 222–230, 2006. 14. van den Berghe G, Wouters P, Weekers F, et al: Intensive insulin therapy in critically ill patients. N Engl J Med 345: 1359–1367, 2001. 15. Pronovost P, Needham D, Berenholtz S, et al: An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 355: 2725–2732, 2006. 16. Lazzeri C, Bevilacqua S, Ciappi F, Pratesi C, Gensini GF, Romagnoli S: Glucose metabolism in cardiovascular surgery. HSR Proc Intensive Care Cardiovasc Anesth 2: 19–26, 2010. 17. Scurlock C, Raikhelkar J, Mechanick JI: Intensive metabolic support: evolution and revolution. Endocr Pract 14: 1047–1054, 2008. 18. Slaughter MS, Rogers JG, Milano CA, et al; HeartMate II Investigators: Advanced heart failure treated with c ontinuous-flow left ventricular assist device. N Engl J Med 361: 2241–2251, 2009. 19. Debaveye Y, Van den Berghe G: Risks and benefits of nutritional support during critical illness. Annu Rev Nutr 26: 513–538, 2006. 20. Kirklin JK, Naftel DC, Stevenson LW, et al: INTERMACS database for durable devices for circulatory support: First annual report. J Heart Lung Transplant 27: 1065–1072, 2008. 21. Rodbard HW, Blonde L, Braithwaite SS, et al; AACE Diabetes Mellitus Clinical Practice Guidelines Task Force: American Association of Clinical Endocrinologists medical guidelines for clinical practice for the management of diabetes mellitus. Endocr Pract 13(suppl 1): 1–68, 2007. 22. Duerksen DR, Papineau N: The prevalence of coagulation abnormalities in hospitalized patients receiving lipid-based parenteral nutrition. JPEN J Parenter Enteral Nutr 28: 30–33, 2004. 23. Artinian V, Krayem H, DiGiovine B: Effects of early enteral feeding on the outcome of critically ill mechanically ventilated medical patients. Chest 129: 960–967, 2006.
