Secondary Logo

Journal Logo

Adult Circulatory Support

Left Lateral Thoracotomy for Centrifugal Continuous-Flow Left Ventricular Assist Device Placement: An Analysis from the Mechanical Circulatory Support Research Network

Maltais, Simon*; Anwer, Lucman A.*,†; Tchantchaleishvili, Vakhtang*; Haglund, Nicholas A.§,¶; Dunlay, Shannon M.*; Aaronson, Keith D.; Pagani, Francis D.; Cowger, Jennifer#; Salerno, Christopher T.#; Shah, Palak#; Khalpey, Zain**; Schmitto, Jan††; Stulak, John M.*

Author Information
doi: 10.1097/MAT.0000000000000714
  • Free


Recent advances in continuous-flow left ventricular assist device (CF-LVAD) technology have allowed significant improvements in outcomes for patients with end-stage heart failure (HF).1,2 As survival and quality of life on therapy continue to improve, adverse event (AE) rates have significantly decreased.3 The generalized acceptance of CF-LVAD therapy, coupled with a continued organ donor shortage, have led to increasing number of patients being supported for bridge-to-transplantation (BTT) and destination therapy (DT) indications.4

The HeartMate II (HMII) CF-LVAD (Abbott Laboratories, Lake Bluff, IL) is utilized for both BTT and DT indications,1 whereas the HeartWare CF-LVAD (HVAD) (Medtronic PLC, Dublin, Ireland) is currently approved for BTT patients only.5 Although large post-approval European registry studies have been published for the HVAD, no large multicenter experiences have been reported outside of landmark approval studies. Post-approval results and utilization of a durable device are generally considered important to introduce a “real world” dimension to reported clinical outcome; however, such results have not been reported yet. Furthermore, with a growing interest toward minimally invasive cardiac surgery, innovative surgical techniques for HVAD implantation have also been used in recent times,6–8 but no evidence-based consensus supporting these approaches exist. Understanding the challenges and impact on outcomes in a large cohort of patients implanted with the HVAD is of primary importance.

We sought to report contemporary outcomes in a large cohort of patients supported with a durable centrifugal continuous-flow LVAD (HVAD). We examined early and mid-term outcomes after implantation in a large collaborative effort of centers that comprise the Mechanical Circulatory Support Research Network (MCSRN). Finally, we assessed outcomes stratified by surgical approach for HVAD implantation: conventional sternotomy (CS) versus minimally invasive left lateral thoracotomy (MILT).

Materials and Methods

Study Population

Our study represents a retrospective analysis of prospectively collected data from May 2004 to August 2015. The data collection process and analysis were performed after informed patient consent and approval from the Institutional Review Board at each of the centers comprising the MCSRN: University of Michigan, Mayo Clinic College of Medicine, Vanderbilt Heart and Vascular Institute, St-Vincent’s Hospital, and Inova Fairfax Hospital.

Between May 2004 and August 2015, 1,150 patients underwent primary CF-LVAD implantation at our centers. Of these, 283 patients (25%) received a HeartWare Ventricular Assist System (HVAD); these patients represent the contemporary cohort for this analysis. Surgical approach included CS in 230 (81%) and MILT in 53 (19%) patients. Demographic and other patient-related data were obtained from patient medical records and prospectively-collected clinical databases of the MCSRN centers. Follow-up information was obtained from subsequent clinic visits and written correspondences from local physicians. The registry is regularly reviewed and audited by investigators for accuracy.

The primary outcomes of our study included: 1) all-cause mortality at 30 days, 90 days, and 1 year; 2) intraoperative blood product use, length of intensive care unit (ICU), and hospital stay; and 3) presence of AEs, including gastrointestinal bleeding (GIB), pump thrombus/hemolysis, or neurological events. All AEs were prospectively collected and are in accordance with the 2013 INTERMACS guidelines.2

Surgical Approach

CS versus MILT Approaches

For patients in the CS group, we used a standard implantation technique as described previously by Slaughter9 and Sileshi et al.10 For patients in the MILT group, we used a combination of previously described techniques by Schmitto et al.,7 Popov et al.,11 and Maltais et al.12 with some modifications, especially for off-pump patients. In summary, a left anterior thoracotomy was combined with either an upper hemi-sternotomy or a right anterior thoracotomy for outflow graft anastomosis. A small subset (8/53) of high-risk patients underwent a sole left anterior thoracotomy and the outflow graft was anastomosed to either the axillary artery or the descending aorta. Of 53 patients from the MILT cohort, 22 were implanted using an off-pump technique, whereas the remaining were done on cardiopulmonary bypass (CPB). All patients considered for MILT approach underwent a preoperative chest computed tomography to determine the position of the aorta. Patients with left atrial clots, significant patent foramen ovale, mitral stenosis, or severe tricuspid insufficiency were not considered for a MILT approach. Importantly, our experience with the MILT implantation technique spans only the last 3 years (2013–2015), as the technique has only recently started gaining acceptance within the field.

Statistical Analysis

Data were expressed as mean ± standard error of the mean for normally distributed data, and median with a range for non-normally distributed data. Data between two groups were compared using the χ2 test for continuous and dichotomous variables. A backward stepwise Cox regression analysis was used to identify those predefined perioperative variables that independently affected outcomes.

Kaplan–Meier survival analysis was used to evaluate time-related outcomes and produce plots, which were subsequently compared by the log-rank test. Statistical significance was considered at p < 0.05. Early operative mortality was defined as death occurring within 30 days of operation or at any time during the index hospitalization.


Patient Characteristics

Between May 2004 and August 2015, 283 patients underwent primary HVAD implantation. Of these, 230 patients (81%) were implanted using the CS approach, whereas 53 patients (19%) were approached through the MILT technique. Seventy-five percent (212/283) of our patients were males and the median age at implantation was 59 years (range, 18–76 years).

Detailed preoperative clinical characteristics are presented in Table 1. Heart failure etiology was ischemic in 42% of patients, and was comparable between both groups (p = 0.49). Significant differences were noted in INTERMACS profiles 1-2 (30% in CS cohort vs. 9% in the MILT cohort, p < 0.01), BTT indication (71% vs. 96%, p < 0.01), preoperative atrial fibrillation (35% vs. 60%, p < 0.01), and prevalence of hypertension (53% vs. 70%, p = 0.04).

Table 1.
Table 1.:
Preoperative Patient Characteristics: Comparison of Preoperative Patient Characteristics Between the Two Surgical Cohorts

Operative Data

A significantly higher number of patients in the CS cohort underwent concomitant tricuspid valve surgery (18% vs. 0%, p < 0.01). After accounting for MILT patients who underwent LVAD implantation off by-pass (22/53), the median CPB time for both cohorts was comparable (71 [6–255] min vs. 71 [11–188] min, p = 0.80). No significant differences were noted in intraoperative blood product use or concomitant aortic valve surgery between the two cohorts. Detailed operative data are presented in Table 2.

Table 2.
Table 2.:
Operative Data: A Comparative Analysis Between CS and MILT Cohorts

Perioperative Outcomes

A significant difference was noted in the length of hospital stay between the two cohorts (17 days vs. 12 days, p < 0.01); however, no significant difference was noted in the length of ICU stay. Right heart failure (RHF) and the need for right ventricular support for the overall cohort remained low; only five patients were implanted with an RVAD, and the results were comparable between both groups (p = 0.89). A total of 14 patients (4.9%) expired during the index hospitalization: 6% patients from the CS cohort and 3.9% patients from the MILT cohort (p = 0.15). Detailed early outcomes are presented in Table 3.

Table 3.
Table 3.:
Early Perioperative Outcomes: Univariate Outcome Data, Including Early Mortality

Late Survival and AEs

Follow-up was available in all early survivors (269 patients) for a median of 268 days (302 days of patient-years of support); 275 days for CS patients and 268 days for MILT patients (p = 0.26). Estimated survival for the overall cohort was 84% at 1 year, 75% at 2 years, and 66% at 3 years (Figure 1). When stratified by surgical approach, there was no difference in late survival between the two groups (Figure 2); 86% vs. 76% at 1 year and 78% vs. 71% at 2 years, p = 0.298. Over the course of the study period, 79/230 patients (34%) from the CS cohort and 13/53 (25%) from the MILT cohort were transplanted. These patients were excluded from all subsequent AE analyses based on a lost to follow-up principle to prevent confounding of outcomes for patients on device therapy.

Figure 1.
Figure 1.:
Overall survival: Kaplan–Meier survival curve showing overall survival for HVAD patients.
Figure 2.
Figure 2.:
Survival stratified by surgical approach: Kaplan–Meier comparison evaluating survival between CS vs. MILT patients. No differences in estimated survival were observed at 1 year or 2 years from implantation (p = 0.298). CS, full line; MILT, dashed line.

Detailed adverse events for the overall cohort are presented in Table 4. Patients in the MILT cohort had a significantly lower incidence of pump thrombus/hemolysis (18% vs. 4%, p = 0.02). However, the incidence of GIB was significantly higher in this cohort (13% vs. 26%, p = 0.02).

Table 4.
Table 4.:
Adverse Event Profiles: A Comparative Analysis Between CS and MILT Cohorts

Results of the Cox proportional multivariate model analysis for 30-day, 90-day, and 1-year survivals are presented in Table 5. Preoperative bilirubin level was identified as the only predictor of survival. Remaining preoperative patient characteristics and the surgical technique itself had no impact on survival; p = 0.067 for 30-day and p = 0.44 for 90-day survivals.

Table 5.
Table 5.:
Multivariate Model Analysis: Predictors of Survival in HVAD Patients


Main Findings of Our Study

Despite the variability of implant strategies and the multicenter post-approval nature of this trial, overall reported results for the entire cohort compares favorably to previously published trial outcomes.13,14 This large study further evaluates comparative outcomes between conventional sternotomy vs. minimally invasive thoracotomy approaches for HVAD implantation. Our comparative analysis for a total of 283 patients showed no significant differences in early survival between the two surgical cohorts. Estimated survival was comparable between the two groups, and multivariable analysis showed that surgical technique was not associated with increased mortality (all p > 0.05).

Expectedly, patients in the thoracotomy cohort appeared to be less ill by preoperative stratification, and the comparative analysis for perioperative outcomes revealed a significantly shorter median length of hospital stay for these patients (p < 0.01). However, no difference was observed in the median length of ICU stay between the two groups.

The cumulative AE incidence was comparable between the two cohorts. Interestingly, patients in the MILT cohort had a lower incidence of pump thrombus/hemolysis (p = 0.02), and the incidence of GIB was higher in these patients (p = 0.02). Multivariate analysis identified preoperative bilirubin level as the only independent predictor of survival.

AE Profile Differences

The current generation of CF-LVADs has superior AE profiles than the previously used pulsatile devices.1,15 Nevertheless, AEs such as GIB and pump thrombus continue to plague outcomes of these newer devices.16,17 This study reports in a large group of patients an overall comparable AEs profile for patients implanted with the HVAD device.10,18 Interestingly, our study also showed a significantly lower incidence of pump thrombus in the MILT cohort, whereas at the same time the incidence of GIB was significantly higher compared with the CS cohort. These results should be interpreted with caution as patients in the MILT group had a lower risk profile before surgery compared with patients with CS. Lower INTERMACS profile has been associated with increased risk of complications such as pump thrombosis.4 We have however found placement of the inflow cannula easier and reproducible in the MILT cohort due to the physiologic maintenance of the anatomic position of the heart within the pericardium. Although the impact of a suboptimal anatomic position of a continuous-flow centrifugal pump remains to be analyzed and interpreted, we believe a thoracotomy strategy allows for more intuitive assessment of pump positioning during implant, and this may play a potential role in the lower pump thrombus incidence in these patients. Finally, patients with descending aorta and axillary outflow graft anastomosis are typically maintained after implant with a higher INR goal (2.5–3.5 vs. 2–3 for all other patients) to avoid root thrombus formation.19 It is possible that this practice management difference has influenced our AE profile leading to a higher incidence of GIB in the MILT cohort.

Is a Thoracotomy a Suitable Implant Strategy for Selected Patients?

Minimally invasive LVAD surgery has been made possible by improvements in device performance and outcomes combined with miniaturization in device size.2,20 The newer techniques focus on two main areas: finding an alternate less invasive cardiac access apart from the conventional sternotomy approach, and moving from an on-pump to an off-pump technique, while maintaining outcomes for these challenging patients.21

The substitution of a full sternotomy with smaller incisions theoretically offers benefits like those achieved from other well-established minimal invasive cardiac surgery procedures.22–25 As these strategies evolve and become more popular, we must continue to hold the highest standards for outcomes and critically assess our results. Minimally invasive techniques can potentially reduce the potential trauma in patients with multiple previous sternotomies and preserve anatomy in BTT candidates.10,26 This study shows comparable early survival results for patients implanted using a CS or a MILT technique, but we were unable to confirm a blood product utilization or mechanical ventilation benefit for patients with the less invasive approach. Preserving the pericardium offers a potential protective role in maintaining right ventricular geometry and pressure-volume relationship, and this could be a major step towards reducing the incidence of RHF in LVAD patients at the time of implantation.27,28 The incidence of RHF was low in this cohort and future larger studies will be needed to investigate the benefits of this concept.

Activation of the systemic inflammatory response system by CPB, and its resultant effects on fibrinolysis, platelet sequestration, and degradation of coagulation factors, are well documented in the literature.26 Therefore, LVAD implantation without CPB could potentially reduce the burden of these complications and lead to an improved postoperative course.20,29–31 Despite limited data for off-pump LVAD implantation, the short-term benefits of using an off-pump technique to reduce blood product transfusion, reoperation for perioperative bleeding, acute kidney injury, and respiratory complications are well documented in the coronary artery bypass population.32 No significant advantage in blood product use was demonstrated for patients with an off-pump approach in this study (data not shown), but an off-pump strategy could reduce blood use and potentially decrease the risk of recipient sensitization, thus preserving donor pool availability.18

Everything Comes with a Price: Challenges of the MILT Technique

It is important to acknowledge that not all potential LVAD patients are suitable candidates for a minimally invasive or off-pump thoracotomy approach. Universal application of this strategy can lead to important complications and decrease the benefits of device implantation. Either patients who need a concomitant cardiac procedure or those with significant left-sided thrombus should be considered carefully for this strategy. A significant proportion of patients undergoing LVAD implantation have concomitant valve interventions33 or contraindications to an off-pump approach. With smaller thoracotomy and hemi-sternotomy incisions, access and exposure for both the left ventricular apex and ascending aorta also become technically more challenging and can result in malposition or unrecognized cardiac injuries.

The Importance of Liver Function Optimization

In accordance with several previous studies,34,35 our multivariate analysis continues to identify preoperative bilirubin level as an independent predictor of post-LVAD survival. Right heart failure is a common complication after LVAD implantation, and it has been repeatedly associated with increased morbidity and mortality in these patients.36,37 Although aggressive measures continue to be recommended for improvement of liver function before surgery, it might be beneficial to consider temporary mechanical support to assist hepatic recovery in the early pre- or postimplant phases.38


This study has limitations inherent in its study design and retrospective nature. The study includes data from five different institutions and is subjected to inherent limitations of data extraction from registries. Despite common practices, variability in clinical management during the postoperative period can also occur between institutions, and these can affect overall outcomes. Nevertheless, despite its limitation, reporting on a combined experience of several different institutions provides a real-world dimension to our results and accounts for institutional bias inherent to most single-center experiences. Second, our MILT cohort represents a combination of a few different approaches; hence, it is tough to draw conclusions for the entire spectrum and we base our outcomes on a specific “thoracotomy” approach. Last and very importantly, due to the highly selected nature of patients in the MILT cohort, patients in the CS cohort had significantly worse risk profiles preoperatively. Although this is a critical limiting factor, our multivariate analysis did not demonstrate a worse INTERMACS profile to be an independent predictor of postimplant survival in our study population.

Conclusions and Future Implications

This large multicenter experience confirmed excellent survival and a comparable AE profile for patients supported with the HVAD technology. Collectively, our results provide support to newer alternative implant techniques as outcomes in highly selected group of patients undergoing implantation via the MILT technique compare favorably to those of the conventional sternotomy strategy. Although survival is not affected by the thoracotomy implant strategy, it appears that patient’s AE profiles may be influenced by significant changes made to the clinical practice to adjust for alternative outflow graft placement. Careful assessment of outcomes when evaluating different implant approaches is necessary for widespread acceptance of the different techniques. Data from ongoing clinical trials will be essential to establish clinical guidelines and best practice for the HVAD device (HVAD Lateral Trial; NCT02268942).


The authors would like to acknowledge that this is a collaborative effort of all institutions of the Mechanical Circulatory Support Research Network.


1. Slaughter MS, Rogers JG, Milano CA, et al.; HeartMate II Investigators: Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009.361: 22412251.
2. Kirklin JK, Naftel DC, Kormos RL, et al. Fifth INTERMACS annual report: Risk factor analysis from more than 6,000 mechanical circulatory support patients. J Heart Lung Transplant 2013.32: 141156.
3. Birks EJ. Left ventricular assist devices. Heart 2010.96: 6371.
4. Kirklin JK, Naftel DC, Pagani FD, et al. Seventh INTERMACS annual report: 15,000 patients and counting. J Heart Lung Transplant 2015.34: 14951504.
5. Aaronson KD, Slaughter MS, Miller LW, et al.; HeartWare Ventricular Assist Device (HVAD) Bridge to Transplant ADVANCE Trial Investigators: Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation. Circulation 2012.125: 31913200.
6. Umakanthan R, Haglund NA, Stulak JM, et al. Left thoracotomy HeartWare implantation with outflow graft anastomosis to the descending aorta: A simplified bridge for patients with multiple previous sternotomies. ASAIO J 2013.59: 664667.
7. Schmitto JD, Molitoris U, Haverich A, Strueber M. Implantation of a centrifugal pump as a left ventricular assist device through a novel, minimized approach: Upper hemisternotomy combined with anterolateral thoracotomy. J Thorac Cardiovasc Surg 2012.143: 511513.
8. Sabashnikov A, Mohite PN, Simon AR, Popov AF. HeartWare miniaturized intrapericardial ventricular assist device: Advantages and adverse events in comparison to contemporary devices. Expert Rev Med Devices 2013.10: 441452.
9. Slaughter MS. Implantation of the HeartWare left ventricular assist device. Semin Thorac Cardiovasc Surg 2011.23: 245247.
10. Sileshi B, Haglund NA, Davis ME, et al. In-hospital outcomes of a minimally invasive off-pump left thoracotomy approach using a centrifugal continuous-flow left ventricular assist device. J Heart Lung Transplant 2015.34: 107112.
11. Popov AF, Hosseini MT, Zych B, Simon AR, Bahrami T. HeartWare left ventricular assist device implantation through bilateral anterior thoracotomy. Ann Thorac Surg 2012.93: 674676.
12. Maltais S, Danter MR, Haglund NA, et al. Nonsternotomy approaches for left ventricular assist device placement. Oper Tech Thorac Cardiovasc Surg 2014.19(3): 276291.
13. Pagani FD, Milano CA, Tatooles AJ, et al. HeartWare HVAD for the treatment of patients with advanced heart failure ineligible for cardiac transplantation: Results of the ENDURANCE destination therapy trial. J Heart Lung Transplant 2015.34: S9.
14. Slaughter MS, Pagani FD, McGee EC, et al.; HeartWare Bridge to Transplant ADVANCE Trial Investigators: HeartWare ventricular assist system for bridge to transplant: combined results of the bridge to transplant and continued access protocol trial. J Heart Lung Transplant 2013.32: 675683.
15. Molina EJ, Boyce SW. Current status of left ventricular assist device technology. Semin Thorac Cardiovasc Surg 2013.25: 5663.
16. Kirklin JK, Naftel DC, Kormos RL, et al. Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) analysis of pump thrombosis in the HeartMate II left ventricular assist device. J Heart Lung Transplant 2014.33: 1222.
17. Stulak JM, Deo S, Schirger J, et al. Preoperative atrial fibrillation increases risk of thromboembolic events after left ventricular assist device implantation. Ann Thorac Surg 2013.96: 21612167.
18. Sileshi B, O’Hara BK, Davis ME, et al. Outcomes of patients implanted using a left thoracotomy technique for a miniaturized centrifugal continuous-flow pump. ASAIO J 2016.62: 539544.
19. Kar B, Delgado RM 3rd, Frazier OH, et al. The effect of LVAD aortic outflow-graft placement on hemodynamics and flow: Implantation technique and computer flow modeling. Tex Heart Inst J 2005.32: 294298.
20. Maltais S, Davis ME, Haglund N. Minimally invasive and alternative approaches for long-term LVAD placement: The Vanderbilt strategy. Ann Cardiothorac Surg 2014.3: 563569.
21. George Makdisi, I-Wen Wang. Minimally invasive is the future of left ventricular assist device implantation. J Thorac Dis 2015.7(9): E283E288.
22. Mihaljevic T, Cohn LH, Unic D, Aranki SF, Couper GS, Byrne JG. One thousand minimally invasive valve operations: Early and late results. Ann Surg 2004.240: 52934; discussion 534.
23. Cohn LH, Adams DH, Couper GS, Bichell DP. Minimally invasive aortic valve replacement. Semin Thorac Cardiovasc Surg 1997.9: 331336.
24. Asher CR, DiMengo JM, Arheart KL, et al. Atrial fibrillation early postoperatively following minimally invasive cardiac valvular surgery. Am J Cardiol 1999.84: 7447, A9.
25. Murphy GJ, Reeves BC, Rogers CA, Rizvi SI, Culliford L, Angelini GD. Increased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery. Circulation 2007.116: 25442552.
26. Clive Landis R, Murkin JM, Stump DA, et al. Consensus statement: minimal criteria for reporting the systemic inflammatory response to cardiopulmonary bypass. Heart Surg Forum 2010.13: E116E123.
27. Dell’Italia LJ. Anatomy and physiology of the right ventricle. Cardiol Clin 2012.30: 167187.
28. Topilsky Y, Hasin T, Oh JK, et al. Echocardiographic variables after left ventricular assist device implantation associated with adverse outcome. Circ Cardiovasc Imaging 2011.4: 648661.
29. Cheung A, Lamarche Y, Kaan A, et al. Off-pump implantation of the HeartWare HVAD left ventricular assist device through minimally invasive incisions. Ann Thorac Surg 2011.91: 12941296.
30. Momeni M, Carlier C, Baele P, et al. Fibrinogen concentration significantly decreases after on-pump versus off-pump coronary artery bypass surgery: A systematic point-of-care ROTEM analysis. J Cardiothorac Vasc Anesth 2013.27: 511.
31. Van Meter CH Jr, Robbins RJ, Ochsner JL. Technique of right heart protection and deairing during heartmate vented electric LVAD implantation. Ann Thorac Surg 1997.63: 11911192.
32. Lamy A, Devereaux PJ, Prabhakaran D, et al.; CORONARY Investigators: Off-pump or on-pump coronary-artery bypass grafting at 30 days. N Engl J Med 2012.366: 14891497.
33. John R, Naka Y, Park SJ, Sai-Sudhakar C, et al. Impact of concurrent surgical valve procedures in patients receiving continuous-flow devices. J. Thorac. Cardiovasc. Surg. 2014.147:581589.
34. El-Banayosy A, Arusoglu L, Kizner L, et al. Predictors of survival in patients bridged to transplantation with the thoratec VAD device: A single-center retrospective study on more than 100 patients. J Heart Lung Transplant 2000.19: 964968.
35. Frazier OH, Rose EA, Oz MC, et al.; HeartMate LVAS Investigators. Left Ventricular Assist System: Multicenter clinical evaluation of the HeartMate vented electric left ventricular assist system in patients awaiting heart transplantation. J Thorac Cardiovasc Surg 2001.122: 11861195.
36. Kormos RL, Teuteberg JJ, Pagani FD, et al.; HeartMate II Clinical Investigators: Right ventricular failure in patients with the HeartMate II continuous-flow left ventricular assist device: Incidence, risk factors, and effect on outcomes. J Thorac Cardiovasc Surg 2010.139: 13161324.
37. Drakos SG, Janicki L, Horne BD, et al. Risk factors predictive of right ventricular failure after left ventricular assist device implantation. Am J Cardiol 2010.105: 10301035.
38. El Banayosy A, Kizner L, Schueler V, Bergmeier S, Cobaugh D, Koerfer R. First use of the Molecular Adsorbent Recirculating System technique on patients with hypoxic liver failure after cardiogenic shock. ASAIO J 2004.50: 332337.

continuous-flow left ventricular assist device; HeartWare centrifugal CF-LVAD (HVAD); surgical technique; left lateral thoracotomy

Copyright © 2017 by the ASAIO