Lactic acidosis after cardiac surgery is an ominous

Transcription

1 Lactic Acidosis After Cardiac Surgery Is Associated With Polymorhisms in Tumor Necrosis Factor and Interleukin 10 Genes Thomas Ryan, FFARCSI, Joanna Balding, BA, Mod (Genetics), Eilis M. McGovern, FRCSI, John Hinchion, FRCSI, Wendy Livingstone, PhD, Zeb Chughtai, FRCSI, and Owen P. Smith, FRCPI Deartments of Anaesthesia, Hereditary Coagulation Disorders, and Cardiothoracic Surgery, St. James s Hosital, Dublin, Ireland Background. Lactic acidosis after cardiac surgery is a manifestation of excess cytokine roduction. Cytokinerelated genetic olymorhisms account for variability in cytokine resonse and may redisose to the develoment of lactic acidosis after cardiac surgery. Methods. Routine ostoerative cardiac surgery atients were studied. Lactic acid levels were greater than 4 mmol/l in study atients and less than 4 mmol/l in controls. Polymerase chain reaction-based techniques were used to examine carriage of tumor necrosis factor (TNF- ), TNF G 308A, and interleukin 10 (IL-10) G 1082A alleles. Results. Demograhic characteristics and details of surgery were similar for 30 control and 21 study atients. Lactic acidosis after cardiac surgery is an ominous event and is associated with excess mortality among atients in shock [1]. In ediatric cardiac surgery lactic acidosis is a owerful redictor of outcome with greater lactic acid levels associated with adverse outcome including excess mortality [2]. Lactic acidosis in cardiac surgical atients is a manifestation of systemic inflammation and excess roinflammatory cytokine roduction [3]. Systemic inflammation after cardiac surgery may also resent as the low systemic vascular resistance syndrome with hyotension and high cardiac index and is always accomanied by lactic acidosis [3]. Lactic acid accumulation in shocked cardiac surgical atients is a net result of excess lactic acid roduction with unchanged utilization and is not related to altered carbohydrate metabolism [4]. This rocess may be a direct metabolic effect of tumor necrosis factor (TNF), as lactic acid roduction is increased by TNF-mediated inhibition of yruvate dehydrogenase [5]. Interindividual variation in TNF roduction in atients with sesis has been linked to olymorhisms in the TNF- gene [6]. Polymorhisms in the TNF- romoter gene are associated with excess mortality in setic shock [7]. Interleukin 10 (IL-10) is a otent antiinflammatory cytokine that inhibits TNF roduction. Polymorhisms in IL-10 romoter genes are associated with variation in Acceted for ublication Feb 17, Address rerint requests to Dr Ryan, Deartment of Anaesthesia, St. James s Hosital, James St, Dublin 8, Ireland; ryants@iol.ie. Lactic acid levels after intensive care admission changed over time and were related to both TNF- and IL-10 G 1082A olymorhisms. All 4 study atients homozygous for TNF- 1 and carrying an IL A allele develoed lactic acidosis ( 0.02). There was no relation between the rate of einehrine infusion or duration of cardioulmonary byass and lactic acid levels. Conclusions. Genetic factors have a role in the develoment of lactic acidosis after cardiac surgery. (Ann Thorac Surg 2002;73: ) 2002 by The Society of Thoracic Surgeons IL-10 roduction [8]. It is lausible that these cytokine genomic olymorhisms modulate cytokine roduction and systemic inflammation in cardiac surgical atients and that the occurrence and severity of lactic acidosis in these atients is influenced by the resence of TNF and IL-10 genetic olymorhism. We conducted a study to test this hyothesis. Patients and Methods Consenting atients scheduled for routine cardiac surgery with normal reoerative ventricular function and uneventful surgical rocedures were recruited over a 6-month eriod. Patients with a history of heatic disease were excluded. Patients were admitted to a dedicated cardiac surgical intensive care unit after surgery with care determined by the referring cardiac surgical service. Patient care was not modified for the urose of the study. Cardioulmonary byass was erformed using an oen system rimed with 1,500 ml of Hartmann s solution with hearin-coated circuits and roller ums. Suction systems were controlled. Patient demograhic characteristics and history of myocardial infarction, hyertension, congestive heart failure, vascular surgery, and diabetes were collected. The nature of the surgical rocedure, duration of cardioulmonary byass and the minimum temerature on cardioulmonary byass (CPB), the first and last arterial blood gases on cardioulmonary byass, the blood flows and hemoglobin concentrations on cardioulmonary byass that corresonded with these blood gases, and the 2002 by The Society of Thoracic Surgeons /02/$22.00 Published by Elsevier Science Inc PII S (02)

2 1906 RYAN ET AL Ann Thorac Surg LACTIC ACIDOSIS AFTER CARDIAC SURGERY 2002;73: least blood flow on cardioulmonary byass were documented. A case-control study was erformed with study atients having arterial lactic acid level in excess of 4 mmol/l at any time in the first 24 hours after cardiac surgery and a control grou with lactic acid level never greater than 4 mmol/l in the first 24 hours after surgery. Lactic acid levels, hemodynamics, inotroic requirement, arterial blood gases were recorded at the end of cardioulmonary byass, on arrival in the intensive care, and 6, 12, and 24 hours later. The duration of ostoerative mechanical ventilation and the blood loss in the first 12 hours after surgery were recorded. Genetic analysis was erformed by a erson who was unaware of grou allocation and lactic acid levels. DNA was extracted from whole blood using overnight roteinase K (1 mg/ml) cell lysis at 37 C in the resence of 0.5% sodium dodecyl sulfate followed by extraction with henol/chloroform and reciitation with ethanol. Polymerase chain reaction (PCR) amlification of all olymorhic sites was erformed in a 50 L total volume. The standard reaction mix consisted of Taq DNA Polymerase buffer with MgCl 2 (Promega; 50 mmol/l KCl, 10 mmol/l Tris-HCl [H 9.0], 0.1% Triton X-100, and 1.5 mmol/l MgCl 2 ), 0.4 U of DNA Taq olymerase, 2 L of genomic DNA, 4% dimethyl sulfoxide (DMSO), 30 mol/l each of deoxyribonucleoside trihoshates, and 0.2 mol/l each of sense rimer and antisense rimer (Aendix 1). The cycling variables for each assay are listed in Aendix 2, along with any changes to the standard PCR reaction mix. Restriction enzymes used for each assay are listed in Aendix 2. The IL-6, TNF-, IL , and IL PCR roducts were digested with the aroriate enzyme overnight at 37 C. The TNF- PCR roduct was digested for 3 hours at 37 C, and the IL-1 PCR roduct was digested for 12 hours at 65 C. Restriction digest roducts were run in the aroriate ercentage of agarose gel containing 1.6 g/ml ethidium bromide. Continuous variables were analyzed with Student s t test and analysis of variance (ANOVA). The 2 test and Fisher s exact test were used to comare categorical variables. The relation between lactic acid levels and individual olymorhisms was analyzed at each time oint using Student s t test or ANOVA where aroriate. Where association between an individual olymorhism and lactic acid levels was detected on such a univariate test, then the interaction between olymorhism and change in lactic acid levels with time was analyzed by multivariate ANOVA (MANOVA) with reeated measures. The institutional ethics committee aroved this study in March Results There were 30 control atients and 21 atients in the study grou. Age, gender distribution, body surface area, reoerative chronic disease states, and the nature of surgical rocedure were similar in both grous (Aendix 3). The minimum temerature on cardioulmonary byass was similar in study and control grous. Cardioulmonary byass time was longer in study atients; however, this difference did not reach statistical significance. Arterial artial ressure of carbon dioxide was greater in the study grou at the end of cardioulmonary byass ( Ka versus Ka, 0.05). Otherwise arterial blood gases, hemoglobin, and circuit flow rates at the beginning and termination of cardioulmonary byass were similar in the two grous. Postoerative blood loss was similar in the two grous. The duration of mechanical ventilation was greater in the study grou (Table 1). Lactic acid levels were significantly greater in the study grou at all times in the first 24 hours, with the greatest difference seen 6 hours after intensive care admission (Table 2). MANOVA for reeat measures of lactic acid over time with atient grou as a factor found that lactic acid levels changed significantly over time ( ) and found an interaction between atient grou and time ( ), indicating that atient grouing affected the temoral change in lactic acid levels. Although the duration of cardioulmonary byass was longer in the study grou this duration did not correlate with lactic acid levels. Lactic acid levels at 6 hours after intensive care admission correlated with the duration of mechanical ventilation (lactic acid level at 6 hours ventilation hours; 0.04, R ). On univariate testing 6 hours after intensive care admission lactic acid levels were significantly higher in atients homozygous for the TNF- 1 allele and significantly lower in atients homozygous for the IL G allele (Tables 3 and 4). On MANOVA with TNF- allele and IL G allele as factors and analyzing reeat measurement of lactic acid at 1 and 6 hours (time) after intensive care admission, there was a significant association between TNF- allele and lactic acid level ( 0.03), between IL G allele and lactic acid level ( 0.03), and in lactic acid level change with time ( ); the interaction between time and TNF- allele was also significant ( ) as was the interaction between time and IL G allele ( 0.01). Patient grouing was not associated with the distribution of any individual cytokine olymorhism allele; however, all 4 atients homozygous for the TNF- 1 allele and who carried the IL A allele were in the study grou ( 0.02). One other atient who was homozygous for the TNF- 1 allele who did not carry the IL A allele had normal lactic acid levels. There was no association between TNF G-308A alleles and lactic acid levels. However, only 2 atients were homozygous for the TNF-308A allele. One of these carried the IL A allele and develoed marked lactic acidosis. The other did not carry the IL A allele and had normal lactic acid levels. There was no association between IL , IL-6 174, and IL olymorhisms and lactic acid levels. There was no relation between blood ressure, central venous ressure, and genotye. There was no significant relation between genotye and ostoerative blood loss. There was no association between lactic acid levels 6 hours after intensive care admission and infusion rate of einehrine (lactic acid level mmol/l einehrine g/kg er minute, 0.4).

3 Ann Thorac Surg RYAN ET AL 2002;73: LACTIC ACIDOSIS AFTER CARDIAC SURGERY 1907 Table 1. Patient Demograhics and Oerative Details Comment Control Grou Study Grou Number Age (years) NS Sex (male) NS Body surface area (m 2 ) NS Hemoglobin (g/dl) NS Urea (mmol/l) NS Creatinine (mmol/l) NS Albumin (g/dl) NS Hyertension 9 6 NS Myocardial infarction 6 4 NS Diabetes mellitus 2 2 NS CABG Valve 1 4 CABG/valve 2 1 CPB time (min) Minimum CPB NS temerature (C ) Blood loss 24 hours (ml) NS Postoerative ventilation time (hours) All values are quoted as a frequency for categorical values and as a mean standard error for continuous variables. CABG coronary artery byass graft; CPB cardioulmonary byass; In this study routine cardiac surgical atients dislayed two distinct temoral atterns of lactic acidosis. The duration of cardioulmonary byass may have accounted for some of this difference yet the occurrence of ostoerative lactic acidosis and the change in lactic acid levels over time was associated with the carriage of secific TNF and IL-10 alleles. A genotye was identified which was always associated with lactic acidosis, yet not all atients with lactic acidosis had this genotye. Excess lactic acid accumulation after CPB has been attributed to slanchnic hyoerfusion with reerfusion in the initial hours following surgery and it is not inconceivable that visceral regional hyoerfusion might occur on a frequent basis during cardioulmonary byass. However, Haisjackl and colleagues [9] measuring slanchnic blood flow with indocyanine green and using Table 2. Lactic Acid Levels (mmol/l) in Control and Study Patients After Surgery Control Grou Study Grou Number End cardioulmonary byass Intensive care hour Intensive care hour Intensive care hour Intensive care hour Table 3. Lactic Acid Levels (mmol/l) After Cardiac Surgery in Relation to TNF B Genotye Genotye TNFB1/B1 TNFB1/B2 TNFB2/B2 Number End cardioulmonary NS byass Intensive care hour NS Intensive care hour Intensive care hour NS Intensive care hour NS TNF tumor necrosis factor; gastric tonometry for mucosal H measurement found no evidence of slanchnic hyoerfusion. Indeed ost- CPB slanchnic erfusion and lactic acid levels were both increased comared with re-cpb levels, suggesting that slanchnic lactic acid roduction after cardiac surgery is related to systemic inflammation. Cremer and associates [3] investigated the low systemic vascular syndrome after CPB and found that atients with low systemic vascular resistance had greatly increased levels of TNF and that this excess TNF was always associated with a concomitant lactic acidosis. Thus lactic acid roduction after CPB can be a manifestation of TNF-mediated systemic inflammation rather than hyoerfusion. The association between inflammatory cytokines and excess lactic acid roduction is well recognized in sesis related organ failure. In this setting excess roduction of lactic acid arallels excess inflammatory cytokine roduction in organs that are failing [10]. Vary and colleagues [5] in an animal model linked TNF with inhibition of yruvate dehydrogenase and excess lactic acid roduction. Using a rat model of sesis, they observed that anti-tnf antibody reversed both TNF-mediated yruvate dehydrogenase inhibition and excess lactate roduction. Einehrine and other otent -adrenergic agonists may cause lactic acidosis [11]. The mechanism for this is unclear but a hyothesis suggests that couling of membrane bound Na/K ATPases and anaerobic glycolytic enzymes may be resonsible. Lactic acidosis has been reorted with einehrine administration after cardioulmonary byass [12]. Totaro and Raer [12] reorted that 6 of 18 atients who received einehrine after cardioulmonary byass develoed lactic acidosis whereas none of 17 atients in a noreinehrine grou had lactic acidosis. The study did not determine why only a third of atients develoed acidosis in the einehrine grou. As this study did not include atients with lactic acidosis who did not require inotroic suort, the occurrence of lactic acidosis in such atients was not investigated. TNF- and TNF- are similar comounds and both are active at TNF recetors [13]. TNF- is rimarily roduced by activated monocytes and TNF- by activated lymhocytes. The B1 allele of the TNF- olymorhism was first associated with excess TNF- roduction by Messer and colleagues [14] and has been associated with greater severity of colitis by Koss and associategs [15]. Many olymorhisms in the TNF- romoter gene have been described

4 1908 RYAN ET AL Ann Thorac Surg LACTIC ACIDOSIS AFTER CARDIAC SURGERY 2002;73: Table 4. Lactic Acid Levels (mmol/l) After Cardiac Surgery in Relation to Interleukin 10 Genotye Genotye IL GG IL GA or AA Number End cardioulmonary NS byass Intensive care hour NS Intensive care hour Intensive care hour NS Intensive care hour NS and of these, the functional significance and disease association of the TNF-G 308A olymorhism is best documented. This TNF- olymorhism is associated with enhanced gene transcrition [16], a 3.75-fold increase in mortality with setic shock [7], and a sevenfold excess mortality from cerebral malaria [17]. Thus TNF- and TNF- olymorhisms are functional and modulate inflammation. The TNF-B1 allele occurs with greater frequency than the TNF 308A allele and as a consequence it is easier to investigate in a small study such as this. The resent study was too small to determine the effects of the TNF-G 308A olymorhism. IL-10 is a otent antiinflammatory cytokine that inhibits TNF roduction. The A allele of a olymorhism at osition 1082 in the IL-10 gene romoter region is associated with lower IL-10 roduction [5]. In inflammatory bowel disease carriage of the IL A allele is associated with greater severity of disease [15]. Thus a genotye associated with excess TNF- and a decrease in IL-10 could be characterized as roinflammatory. We observed that all atients with this genotye had excess lactic acid. If one considered an additional atient homozygous for TNF 308A and carrying the IL A alleles in this roinflammatory genotye then the association would be more rominent. This study demonstrated that a combination of cytokine genetic olymorhisms interact to romote lactic acidosis. It is ossible that atients with this roinflammatory genotye may develo systemic inflammation and lactic acidosis after a lower threshold stimulus. However only 20% of atients in the study grou had the identified genotye. Thus the roinflammatory genotye identified was sufficient but not necessary to initiate systemic inflammation. Further study is required to determine whether alternate genetic factors are associated with lactic acidosis in cardiac surgical atients. It is likely that the etiology underlying the generation of systemic inflammation after cardioulmonary byass is multifactorial. Surgical factors such as rolonged cardioulmonary byass and temerature reduction during cardioulmonary byass likely act as a trigger to reciitate systemic inflammation. We excluded atients with rolonged, comlex, or eventful surgery and thus the effects of these factors were minimized. Further study is required to determine the relative imortance of genetic and surgical factors in the occurrence of lactic acidosis after cardiac surgery. The small size of this study and a focus on low-risk atients excluded any ossibility of relating genotye and outcome measures. More extensive study for an association between genotye and outcome will follow this study. Having genetic information on a atient s inflammatory resonse before surgery may have distinct benefits. Beside the basic science interest concerning the role and interaction of the inflammatory mediators, there could be very ractical clinical benefits as those atients with genetic redisosition to high cytokine release may benefit from antiinflammatory mediator strategies. This study was funded by a grant from the Royal City of Dublin Hosital Fund. References 1. Davies AR, Bellomo R, Raman JS, Gutteridge GA, Buxton BF. High lactate redicts failure of intra aortic balloon uming after cardiac surgery. Ann Thorac Surg 2001;71: Duke T, Butt W, South M, Karl T. Early markers of adverse events in children after cardiac oerations. J Thorac cardiovasc Surg 1997;114: Cremer J, Martin M, Redl H, et al. Systemic inflammatory resonse after cardiac oerations. Ann Thorac Surg 1996;61: Chiolero RL, Revelly JP, Leverve X, et al. Effects of cardiogenic shock on lactate and glucose metabolism after heart surgery. Crit Care Med 2000;28: Vary T, Hazen S, Maish G, Cooney R. TNF binding rotein revents hyerlactaemia and inactivation of PDH comlex in skeletal muscle during sesis. J Surg Res 1998;80: Stuber F, Petersen M, Bokelmann F, Schade U. A genomic olymorhism within the tumor necrosis factor locus influences lasma tumor necrosis factor concentrations and outcome of atients with severe sesis. Crit Care Med 1996;24: Mira JP, Cariou A, Grall F, et al. Association of TNF2, atnf-alha romoter olymorhism, with setic shock suscetibility, and mortality: a multicenter study. JAMA 1999; 282: Turner D, Williams D, Sankaran D, Lazarus M, Sinnott PJ, Hutchinson IV. An investigation of olymorhism in the interleukin-10 romoter gene. Eur J Immunogenet 1997;24: Haisjackl M, Birnbaum J, Redlin M, et al. Slanchnic oxygen transort and lactate metabolism during normothermic cardioulmonary byass in humans. Anaesth Anal 1998;86: Douzinas EE, Tsidemiadou PD, Pitaridis MT, et al. The regional roduction of cytokines and lactate in sesis related multile organ failure. Am J Res Crit Care Med 1997;155: James JH, Luchette FA, McCarter FD, Fischer JF. Lactate is an unreliable indicator of tissue hyoxia in injury or sesis. Lancet 1999;354: Totaro RJ, Raer RF. Einehrine induced lactic acidosis following cardioulmonary byass. Crit Care Med 1997;25: Bazzoni F, Beutler B. The tumor necrosis factor ligand and recetor family. NEJM 1996;334: Messer G, Sengler U, Jung MC, et al. Polymorhic structure of the tumor necrosis factor (TNF) locus: an NcoI olymorhism in the first intron of the human TNF- gene correlates with a variant amino acid in osition 26 and a reduced level of TNF- roduction. J Ex Med 1991;173: Koss K, Satsangi J, Fanning GC, Welsh KI, Jewell DP. Cytokine (TNF alha, LT alha and IL-10) olymorhism in inflammatory bowel diseases and normal controls: differential effects on roduction and allele frequencies. Genes Immun 2000;1: Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW. Effects of a olymorhism in the human tumor necrosis factor alha romoter on transcritional activation. Proc Natl Acad Sci 1997;94: McGuire W, Hill AV, Allso CE, Greenwood BM, Kwiatkowski D. Variation in the TNF alha romoter region associated with suscetibility to cerebral malaria. Nature 1994;371:

5 Ann Thorac Surg RYAN ET AL 2002;73: LACTIC ACIDOSIS AFTER CARDIAC SURGERY 1909 Aendix 1 Polymerase Chain Reaction Primer Sequences With Position of 5 Base, Restriction Enzymes, and Cut Sites Primer Sequence (5-3 ) Position of 5 Base Restriction Enzyme Restriction Sites Variant Constant IL sense ATGACTTCAGCTTTACTCTT 324 Hs 92 II IL antisense ATAAATCTTTGTTGGAGGGT 81 TNFB sense CCCTCCTGCACCTGCTGCCTGG 112 Hinf I TNFB antisense AGAGGGG TGGATGCTTGGGTTC 833 TNF- 308 sense GGAGGCAATAGGTTTTGAGGGCCAT 334 Nco I TNF- 308 antisense CTGTCTCGGT TTCTTCTCCATGGCG 140 IL sense TCTGAAGAAGTCCTGATGTC 1248 Mnl I , 1191 IL antisense CTCTTACCTATCCCTACTTCC IL sense GACTCCAGCCACAGAAGCTTA 967 Rsa I , 842, 834 IL antisense ATATCCTCAAAGTTCCCAAGC 531 IL sense GTGTTGTCATCAGACTTTGACCGTA 3816 Taq I IL antisense GAGAGCTTTCAGTTCATATCGACCA 4073 IL interleukin; TNF tumor necrosis factor. Aendix 2 Polymerase Chain Reaction Cycling Parameters, Changes to Standard Reaction Mix, and Percentage Agarose Gel Used in Each Assay Polymorhism Cycles Denaturation Primer Annealing Elongation Changes to Standard Reaction Mix IL C, 1 min 58 C, 1 min 72 C, 90 sec 3% TNFB C, 30 sec 68 C, 30 sec 74 C, 42 sec 1 M of each rimer 2% TNF C, 1 min 65 C, 1 min 72 C, 1 min 2% IL C, 1 min 58 C, 1 min 72 C, 1 min No DMSO 4% IL C, 1 min 64 C, 1 min 72 C, 1 min 1 M of each rimer 3% IL C, 1 min 65 C, 1 min 72 C, 1 min 3% Agarose Gel IL interleukin; TNF tumor necrosis factor. Aendix 3 Arterial Blood Gases and Blood Flow Rate on Cardioulmonary Byass Control Grou Study Grou Initial H NS PCO 2 (kpa) NS PO 2 (kpa) NS Bicarbonate (mmol/l) NS Base excess NS Hemoglobin (g/dl) NS Flow L/m 2 BSA NS Minimal flow L/m 2 BSA NS Final H NS PCO 2 (kpa) PO 2 (kpa) NS Bicarbonate (mmol/l) NS Base excess NS Hemoglobin (g/dl) NS Flow L/m 2 BSA NS BSA Body surface area in square meters;