Impact of Renal Function Before Mechanical Circulatory Support on Posttransplant Renal Outcomes

Transcription

1 Impact of Renal Function Before Mechanical Circulatory Support on Posttransplant Renal Outcomes Madhurmeet Singh, DO, Michael Shullo, PharmD, Robert L. Kormos, MD, Kathleen Lockard, RN, Rachelle Zomak, BSN, Marc A. Simon, MD, Christian Bermudez, MD, Jay Bhama, MD, Dennis McNamara, MD, Yoshiya Toyoda, MD, and Jeffrey J. Teuteberg, MD Cardiovascular Institute, Department of Pharmacy and Therapeutics, and Heart Lung and Esophageal Institute, University of Pittsburgh, Pittsburgh, Pennsylvania Background. Renal dysfunction is common before mechanical circulatory support (MCS). Mechanical circulatory support frequently improves renal function, but the impact of pre-mcs renal dysfunction on renal function after cardiac transplantation (CTX) is unknown. Methods. Patients with MCS from January 1995 until April 2008 at a single center were included if their MCS duration was at least 60 days and they underwent successful CTX. Patients were followed for 1 year after CTX. Results. A total of 116 patients were included in the study. Mechanical circulatory support was biventricular assist device in 28% and left ventricular assist device in 72% (continuous flow left ventricular assist device, 14%). Mean duration of MCS was 124 days. Patients were grouped according to tertiles of pre-mcs creatinine clearance (CrCl): group 1, CrCl less than 45 ml/min; group 2, CrCl between 45 and 65 ml/min inclusive; and group 3, CrCl more than 65 ml/min. Group 3 had the best renal outcomes both after MCS and 1 year after CTX. Regardless of group, patients who had a CrCl of at least 60 ml/min before CTX had similar 1-year posttransplant CrCl (55 versus 53 versus 56 ml/min for groups 1 through 3, respectively; not significantly different). However, the ability to achieve this level of renal function after MCS was less likely in those with the worst renal function before the initiation of MCS (53% versus 74% versus 90% for groups 1 through 3, respectively; p 0.001). Conclusions. The use of MCS leads to improvements in renal function in patients after MCS. However, the renal outcomes after CTX seem to be more dependent on the level of renal function achieved during MCS than on the level of renal function before MCS. (Ann Thorac Surg 2011;91: ) 2011 by The Society of Thoracic Surgeons End-stage heart failure is frequently accompanied by renal insufficiency. Elevated right atrial pressures, low cardiac output, an adverse neurohormonal milieu, and high-dose diuretics contribute to the renal dysfunction in this population [1, 2]. With the institution of mechanical circulatory support (MCS), these processes are reversed, which, in turn, leads to substantial renal recovery [3, 4]. Most patients who receive MCS are implanted as a bridge to cardiac transplantation (CTX). However, the long-term need for immunosuppression after CTX results in renal dysfunction, much of which can be attributed to calcineurin inhibitors (CNI) [5 7]. Renal toxicity from CNI leads to end-stage renal disease in upward of 20% of patients by 10 years [8]. Furthermore, those with renal insufficiency after CTX have significantly higher rates of morbidity and mortality [9, 10]. Despite improvements in renal function reported during MCS, it is unknown whether the preexisting renal Accepted for publication Oct 18, Address correspondence to Dr Teuteberg, Scaife Hall, Ste 550, 200 Lothrop St, Pittsburgh, PA 15213; teutebergjj@upmc.edu. dysfunction before MCS would predispose patients to renal dysfunction after CTX. In this study we sought to characterize not only how the level of renal dysfunction before MCS affects the magnitude of improvement in renal function after MCS but also the durability of those improvements after CTX. Material and Methods This was a single-center retrospective review of patients who underwent placement of long-term MCS, were supported at least 60 days, subsequently underwent transplantation, and had tacrolimus as their initial CNI. The cutoff of 60 days of support was chosen to allow for a reasonable period of renal recovery after MCS. All patients had MCS and CTX between January 1, 1995, and April 1, 2008, and were followed until April 1, Patients were excluded if they had a temporary device, were supported with a right ventricular assist device alone, or required permanent dialysis after MCS. Data were drawn from the Transplant Patient Management System database at the University of Pittsburgh, which prospectively collects data on all recipients of MCS and 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg SINGH ET AL 2011;91: MCS AND RENAL FUNCTION AFTER TRANSPLANT CTX at the University of Pittsburgh Medical Center. This study was approved by the University of Pittsburgh s Institutional Review Board, and all patients signed informed consent before being included in the database. Patients were grouped according to tertiles of preimplant creatinine clearance (CrCl, in ml/min per 1.73 m 2 )as calculated by the amdrd formula: group 1, CrCl less than 45 ml/min; group 2, CrCl between 45 and 65 ml/min inclusive; and group 3, CrCl greater than 65 ml/min. The primary outcome was change in CrCl at the end of MCS. Baseline characteristics were compared using Kruskal- Wallis or analysis of variance (ANOVA) for continuous variables and 2 or Fisher s exact test for categorical variables. For assessment of CrCl and tacrolimus levels with time by tertile, a multiway ANOVA was performed, and only if the overall model was significant for changes across time or tertile were changes assessed across these variables. For continuous variables, a Dunn or Bonferroni correction was used for multiple comparisons if the one-way ANOVA was significant. The CrCl and tacrolimus levels at each period (exclusive of the baseline and pretransplantation CrCl) were calculated from the mean of each patient s CrCl or immunosuppressive levels during the period of interest. Thus an individual s 1-month CrCl is the mean of all the CrCl measurements from day 14 (the end of the prior period) to day 30. Baseline CrCl was the measurement immediately before MCS and the pretransplantation value was the last measurement before CTX. Statistical analysis was performed with Stata 9.2 (College Station, TX). Results 1349 There were a total of 116 patients who met inclusion criteria and had a mean age of 50 years; 83% were male and 49% had an ischemic cardiomyopathy. There were no significant differences in the patient s demographics among the three groups (Table 1). Patients with the lowest tertile of renal function had a trend toward a greater utilization of biventricular assist device support; otherwise there were no significant differences among the groups in terms of the type or duration of MCS. Most of the patients received a pulsatile device, and more than ADULT CARDIAC Table 1. Demographics Variable Group 1 CrCl 45 (n 38) Group 2 45 CrCl 65 (n 39) Group 3 Cr 65 (n 39) p Value Patient characteristics Age (y) Male (%) White (%) Ischemic CM (%) Diabetes (%) Prior sternotomy (%) VAD characteristics LVAD (%) VAD type (%) 0.13 HeartMate HeartMate II Novacor Thoratec BiVAD Thoratec IVAD Thoratec PVAD VentrAssist Bridge to transplant (%) Mean support (days) Transplant characteristics Donor age (y) Donor male (%) Donor white (%) CMV mismatch (D /R ) PRA 20 (%) Ischemic time (min) Induction (%) Tacrolimus (%) BiVAD biventricular assist device; CM cardiomyopathy; CMV cytomegalovirus; CrCl creatinine clearance; D donor; IVAD implantable ventricular assist device; LVAD left ventricular assist device; PRA panel of reactive antibodies; PVAD percutaneous ventricular assist device; R recipient; VAD ventricular assist device.

3 1350 SINGH ET AL Ann Thorac Surg MCS AND RENAL FUNCTION AFTER TRANSPLANT 2011;91: Table 2. Change in Creatinine Clearance With After Mechanical Circulatory Support Multiway Analysis of Variance F df Prob F Overall model Tertile tertile CrCl (ml/min) p Value a Overall Group 1 Group 2 Group 3 Overall 1 vs 2 2 vs 3 1 vs 3 Pre-MCS wk mo mo mo Pre-Xplant p value b a With Bonferroni correction for multiple comparisons. b Across time for each group. CrCl creatinine clearance; MCS mechanical circulatory support; Xplant transplantation. 90% of the patients had MCS implanted as a bridge to transplantation, although all of the patients subsequently underwent transplantation. There were no significant differences in donor characteristics among the three groups. There was a significantly higher use of induction therapy and a greater percentage of patients with an elevated panel of reactive antibodies in group 1. All patients had tacrolimus as their initial CNI (Table 1). The multiway ANOVA demonstrated significant differences in CrCl after MCS both as a function of time and across groups (Table 2). The interaction of time and tertile was also significant, demonstrating that the CrCl was changing by different rates with time across the groups. Overall there was a statistically significant improvement in CrCl from 58 ml/min to 69 ml/min during the first month after MCS (p ) as seen in Figure 1A. There were no further statistically significant improvements in CrCl beyond 1 month. As with the overall data, group 1 had a statistically significant improvement in CrCl during the first month after MCS, but no further improvement beyond 1 month (Fig 1B). Group 2 had insignificant improvements in CrCl with time when compared with baseline with the exception of the 3-month time (p 0.026). Similarly, group 3 did not show any statistically significant improvement in renal function as compared with baseline at any time after MCS. When comparing each period among the three groups, group 1 had significantly worse CrCl than group 2 until 1 month, after which there was no difference in CrCl between the two groups (Table 2). Group 1 and group 2 had statistically significantly worse (p 0.05) CrCl at all times after MCS compared with group 3. After CTX, the multiway ANOVA demonstrated significant differences in CrCL after transplantation both as a function of time and across groups (Table 3). The interaction of time and tertile was also significant, demonstrating different rates of change in CrCl with time by group. After CTX, there was an overall decline in CrCl from 70.4 ml/min before transplantation to 50.8 ml/min at 1 year after transplantation (Fig 2A). The CrCl at 2 weeks was significantly lower than the pretransplantation value, but beyond 2 weeks there was no further statistically significant decrease in the CrCl for the remainder of the first posttransplant year. Analysis of the cohort after transplantation with patients grouped according to their pre-mcs tertiles of CrCl reveals a similar pattern to the overall data (Fig 2B). As with the overall data, the CrCl at 2 weeks in both group 1 and group 3 was statistically significantly lower than the pretransplantation value, but thereafter the CrCl did not decline significantly. Group 2 had a trend toward a lower CrCl at 2 weeks as compared with baseline. There was no further statistical decline in CrCl as compared with baseline until month 6. When comparing each period among the three groups, groups 1 and 2 entered transplantation with statistically similar CrCl, 63 versus 66 ml/min (not significant), but both were significantly worse than pretransplantation value of 78 ml/min for group 3 (p and versus groups 1 and 2, respectively; Table 3). There were no statistically significant differences in CrCl between groups 1 and 2 or groups 2 and 3 at any time after transplantation. However, group 1 had statistically significantly worse CrCl for the first 6 months after transplantation as compared with group 3 (p 0.01). The multiway ANOVA demonstrated significant differences in tacrolimus levels after transplantation both with time and across groups (Table 4). However, the rate of change in tacrolimus levels with time was not different among groups. Mean tacrolimus levels were not statistically different across time after transplantation within each group with the exception of a higher tacrolimus

4 Ann Thorac Surg SINGH ET AL 2011;91: MCS AND RENAL FUNCTION AFTER TRANSPLANT ml/min. As with the overall data, the CrCl in group 1 remained statistically significantly lower at all times compared with group 3. Although the mean CrCl at the end of MCS remained lower for group 1 in comparison with group 3, some patients in group 1 had substantial improvements in their CrCl. Those patients who were able to achieve a CrCl of at least 60 ml/min by the end of support subsequently had equivalent posttransplantation renal outcomes across the three groups as seen in Figure 3. The percentage of patients able to achieve a CrCl at least 60 ml/min at the end of MCS was 53% in group 1, 74% in group 2, and 90% in group 3. When only the patients able to achieve a CrCl of at least 60 ml/min are examined, the only difference is a slightly higher pretransplantation CrCl in group 3 compared with group 2 (p 0.01). However, at every subsequent period there is no significant difference among the CrCl in any of the three groups. ADULT CARDIAC Fig 1. Change in creatinine clearance (CrCl) with time after mechanical circulatory support (MCS) shown for the overall population (A) and by preimplant tertile of creatinine clearance (B). Overall p ; Group 1, p ; Group 2, p 0.03; Group 3, p *p 0.05 compared with Pre-MCS (with correction for multiple comparisons); p 0.05 compared with 2 weeks (with correction for multiple comparisons). All other comparisons between time periods not significant (with correction for multiple comparisons). (Pre- MCS before mechanical circulatory support; PreXplant before transplantation.) level in group 1 at 1 month as compared with 2 weeks (Table 4). Comparisons of levels among the groups at each period also revealed no significant differences. After 2001 the goal tacrolimus levels at our institution were decreased, which is reflected in lower mean tacrolimus levels across all groups as seen in Table 5. When the mean tacrolimus levels from 2002 to 2008 are compared with those from 1995 to 2001, the mean trough levels were significantly lower for all periods for the 2002 to 2008 era. Given the lower mean tacrolimus levels starting in the year 2002, we reanalyzed the change in posttransplantation CrCl for each group. Each group s results mirrored the overall data with a decline in CrCl during the first 2 weeks, but no further significant decline thereafter. The baseline, 2-week, and 1-year values for CrCl for each group who underwent transplantation after 2002 were as follows: group 1, 58.4, 46.3, and 41.8 ml/min; group 2, 66.9, 60.6, and 51.5 ml/min; and group 3, 81.0, 66.7, and Comment In our study, we examined the renal function of patients with MCS both during support and for 1 year after undergoing transplantation. Overall there was an improvement in renal function during MCS. Those in the worst tertile of renal function had the greatest magnitude of improvement in renal function, but still did not achieve the level of renal function as those in the highest tertile of preimplant CrCl. After transplantation there was a decrement in renal function that was most marked in the first 2 weeks and less pronounced thereafter. Mean tacrolimus levels did not differ substantially among groups. Regardless of preimplant renal function, those patients who were able to achieve a CrCl of at least 60 ml/min had equivalent renal function throughout the first year after transplantation. Prior studies have also demonstrated improvement in renal function after the institution of MCS [3, 4], but not using CrCl for every assessment of renal function during the entire duration of support. For the three groups in our study, the greatest improvements in CrCl occurred during the first month of support, with little significant improvement thereafter. This is consistent with clinical experience as the greatest impact on the heart failure state is in the first several weeks of support, with normalization of cardiac output, a decrease of filling pressures, general resolution of the heart failure state, and the cessation of postoperative hemodynamic support. The lack of further improvement may represent the patient s underlying level of intrinsic renal dysfunction related to diabetes, hypertension, renovascular disease, or irreversible renal injury from the heart failure state. Our study is unique in the examination of renal recovery by the degree of preimplantation renal dysfunction. The group with the worst renal function enjoyed the greatest renal recovery, whereas those with the best renal function had little improvement in their renal function during support. This differential response can likely be explained by a more advanced state of preimplant heart

5 1352 SINGH ET AL Ann Thorac Surg MCS AND RENAL FUNCTION AFTER TRANSPLANT 2011;91: Table 3. Change in Creatinine Clearance With After Transplantation Multiway Analysis of Variance F df Prob F Overall model Tertile tertile CrCl (ml/min) p Value a Overall Group 1 Group 2 Group 3 Overall 1 vs 2 2 vs 3 1 vs 3 Pre-Xplant wk mo mo mo mo p value b a With Bonferroni correction for multiple comparisons. b Across time for each group. CrCl creatinine clearance; Xplant transplantation. failure in the former group as evidenced by the greater utilization of biventricular support. Despite this, group 1 saw the greatest relative improvement in renal function Fig 2. Change in creatinine clearance (CrCl) with time after transplantation show for overall population (A) and by preimplant tertile of creatinine clearance (B). Overall p ; Group 1, p ; Group 2, p ; Group 3, p p compared with pretransplant (with correction for multiple comparisons). All other comparisons between time periods not significant (with correction for multiple comparisons). (PreXplant before transplantation.) during support, such that by 1 month the renal function in group 1 was statistically indistinguishable from that of group 2. Group 3 had little change in renal function after MCS but began MCS with very modest renal dysfunction and not surprisingly had little change in their CrCl with time even with resolution of their heart failure state. The relationship between CNI use and post-ctx renal dysfunction is well established, but has not been examined in a large group after MCS. In our study, there was a 28% decrease in the renal function of the group as a whole at 1 year, with the largest decrement in the first 2 weeks. To assess the impact of preimplant renal dysfunction on posttransplantation outcomes, we continued to analyze patients after transplantation by their preimplant tertile of renal function. Interestingly, groups 1 and 2, which entered transplantation with statistically similar CrCl, also had similar renal function throughout the first posttransplant year. When assessing renal outcomes after transplantation, it is also important to account for the variation in trough CNI levels. All patients in our study population were receiving tacrolimus, and there were no significant differences in tacrolimus levels within each group across time or within each period among groups. Therefore the change in renal function after transplantation cannot be solely attributable to differential levels of calcineurin exposure. The mean levels for the group as a whole are much higher than currently accepted practice, mostly owing to the time frame of the study, in which our institution had higher goal tacrolimus levels before Even with significantly lower tacrolimus levels from 2002 to 2008, there was only a slight, but not clinically meaningful, improvement in the renal function after transplantation when compared with the group as a whole. Also of note, nearly half of the patients in group 1 received induction therapy, more than twice as many as in groups 2 and 3. Many of the patients received induction in an effort to delay the introduction of CNI

6 Ann Thorac Surg SINGH ET AL 2011;91: MCS AND RENAL FUNCTION AFTER TRANSPLANT Table 4. Tacrolimus Levels With After Transplant, by Preimplant Tertile of Creatinine Clearance Multiway Analysis of Variance F df Prob F Overall model Tertile tertile ADULT CARDIAC Tacrolimus (ng/ml) p Value a Overall Group 1 Group 2 Group 3 Overall 2 wk mo mo mo mo p value b a With Bonferroni correction for multiple comparisons. b Across time for each group. because of renal dysfunction. The tacrolimus levels were not significantly lower in the first 2 weeks in group 1 as compared with groups 2 and 3 for both the overall data (p 0.09) or for those who underwent transplantation after 2002 (p 0.08). Thus, despite induction therapy and an intention to limit CNI exposure, the early CNI levels were not statistically different and there was little impact on long-term renal outcomes with this induction strategy. Despite the fact that group 1 had worse renal outcomes after MCS, some patients in group 1 had substantial improvements in their renal function before undergoing transplantation. It remained unclear whether the worse posttransplant renal outcomes in group 1 were more dependent on the degree of renal dysfunction before MCS or the degree of renal dysfunction before transplantation. Therefore, we reassessed the renal outcomes based on whether or not patients were able to achieve a CrCl of at least 60 ml/min before transplantation. At this threshold the Table 5. Tacrolimus Levels With, Versus , by Preimplant Tertile of Creatinine Clearance Multiway Analysis of Variance F df Prob F Overall model Tertile tertile Tacrolimus Level (ng/ml) p Value 2 wk mo mo mo mo p value pretransplantation CrCl for all three groups was similar as was the posttransplantation CrCl throughout the first posttransplantation year. Thus the presence of renal dysfunction before MCS does not imply worse posttransplantation renal outcomes if there is reasonable renal recovery after MCS; however, the likelihood of achieving substantial improvements in CrCl is much less likely in those with worse CrCl before MCS. Many factors contribute to the timing of MCS, and data from INTERMACS suggest that MCS is still predominantly instituted in the sickest patients [11]. However, data from INTERMACS have also shown the sickest patients have the worst outcomes with MCS [11]. Given this and the improved outcomes demonstrated with the current generation of mechanical support [12, 13], the MCS community is moving toward a strategy of earlier implantation. Our data demonstrate that if MCS is implemented earlier in the course of heart failure and before development of significant renal dysfunction, those patients tend to have superior renal outcomes after both MCS and their subsequent transplantation. Patients who present acutely or have a sudden unexpected decompensation with concomitant renal dysfunction may still be supported with a reasonable expectation of renal recovery. However, our data would suggest that waiting to institute mechanical support until renal dysfunction occurs in patients with chronic heart failure may be jeopardizing their eventual renal outcomes, which, in turn, is a risk factor for worse overall posttransplantation outcomes [9, 10]. Limitations There are several limitations to this study that must be noted. This was a retrospective, single-center study of prospectively collected data for a long period with a relatively small number of patients. The numbers of patients in each group limit our ability to distinguish differences among groups and to investigate subgroup differences. Patients were only included if they survived to transplantation and were treated with tacroli-

7 1354 SINGH ET AL Ann Thorac Surg MCS AND RENAL FUNCTION AFTER TRANSPLANT 2011;91: Conclusions The use of MCS leads to improvements in renal function in patients with end-stage heart failure, but most markedly in those with the worst preimplant level of renal dysfunction. Most of the recovery in renal function tends to occur in the first month after MCS, with little change thereafter. However, the renal outcomes after transplantation seem to be more dependent on the level of renal function achieved during MCS than on the level of renal function before MCS. Fig 3. Change in creatinine clearance (CrCl) with time after transplantation by those who did and did not achieve a creatinine clearance of 60 ml/min before transplantation. *p is not significant for all comparisons among tertiles for each time period after transplantation. (PreXplant before transplantation.) mus. Inclusion of patients who died during support, did not undergo transplantation, or died within the first posttransplantation year may have changed the renal outcomes seen in this study. We recognize that renal dysfunction is a predictor of adverse events with both MCS and CTX; however, this study was designed to assess renal outcomes over time for patients who were successfully bridged to transplantation, rather than mortality. There are also many adverse events that occur both after MCS and transplantation that can substantially impact renal outcomes, but our database did not have a detailed collection of all of these adverse events that would have allowed us to examine the difference in event rates overall and among groups. The majority of assist devices used in this study were pulsatile. With the shift toward continuous-flow pumps being the dominant technology, it is uncertain whether these results will translate to a population of patients supported with the current generation of devices. However, organ recovery, including renal recovery, with continuous-flow devices is similar to that of pulsatile devices [4, 12]. We included both biventricular assist devices and left ventricular assist devices in the study, as this reflects the practical management of the full spectrum of heart failure patients eligible for MCS, especially given that those with biventricular failure are more likely to have worse renal function. References 1. Forman DE, Butler J, Wang Y, et al. Incidence, predictors at admission, and impact of worsening renal function among patients hospitalized with heart failure. J Am Coll Cardiol 2004;43: Hillege HL, Girbes AR, de Kam PJ, et al. Renal function, neurohormonal activation, and survival in patients with chronic heart failure. Circulation 2000;102: Butler J, Geisberg C, Howser R, et al. Relationship between renal function and left ventricular assist device use. Ann Thorac Surg 2006;81: Kamdar F, Boyle A, Liao K, et al. Effects of centrifugal, axial, and pulsatile left ventricular assist device support on endorgan function in heart failure patients. J Heart Lung Transplant 2009;28: Wilkinson AH, Cohen DJ. Renal failure in the recipients of nonrenal solid organ transplants. J Am Soc Nephrol 1999;10: Shiba N, Chan MCY, Kwok BWK, et al. Analysis of survivors more than 10 years after heart transplantation in the cyclosporine era: Stanford experience. J Heart Lung Transplant 2004;23: Ojo AO, Held PJ, Port FK, et al. Chronic renal failure after transplantation of a nonrenal organ. N Engl J Med 2003;349: Rubel JR, Milford EL, McKay DB, Jarcho JA. Renal insufficiency and end-stage renal disease in the heart transplant population. J Heart Lung Transplant 2004;23: Cantarovich M, Hirsh A, Alam A, et al. The clinical impact of an early decline in kidney function in patients following heart transplantation. Am J Transplant 2009;9: Odim J, Wheat J, Laks H, et al. Peri-operative renal function and outcome after orthotopic heart transplantation. J Heart Lung Transplant 2006;25: Kirklin JK, Naftel DC, Kormos RL, et al. Second INTERMACS annual report: more than 1,000 primary left ventricular assist device implants. J Heart Lung Transplant 2010;29: Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361: Pagani FD, Miller LW, Russell SD, et al. Extended mechanical circulatory support with a continuous-flow rotary left ventricular assist device. J Am Coll Cardiol 2009;54: INVITED COMMENTARY In this article, Singh and colleagues [1], from the University of Pittsburgh, give us a retrospective picture of the effect of mechanical heart support on renal function. They present 116 patients, collected over 13 years, who were supported to heart transplantation. Renal function improved. Does this matter? Well, yes! The International Society for Heart and Lung Transplantation (ISHLT) shows the effect of pretransplant recipient serum creatinine (Fig 1 [2]) on the relative risk of death at 1 year after heart transplantation (for SI units multiply by 88). This is taken as unity at a serum creatinine level of 1.2 mg/dl (106 moles/l) and doubles as creatinine rises to 2.4 mg/dl 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur