Spine Surgery in Patients with Left Ventricular Assist Device

Heart failure is a complex clinical syndrome which affects around 26 million people worldwide with a significantly increased prevalcence among patients >64 years [1-4]. Demographic changes, the advancement in treatments and diagnostic technologies explain the increase in prevalence of heart failure, whereas the incidence is stable or slightly reduced [5,6]. A systematic review and metaanalysis including over 1.5 million heart failure patients estimated the 1, 2, 5 and 10-year survival to be 87%, 73%, 57% and 35%, respectively [7]. Despite significant advances in therapies and prevention, mortality and morbidity are still high, and the quality of life is poor without adequate tretament. Cardiac transplantation is the only causal therapy in the terminal phase of heart failure, but this therapy is limited due to the restricted donor availability and the required matching and compatibility [8]. Patients often wait for several month to years; 20% die on the waiting list each year without mechanical circulatory support. In these cases, assist devices are used to improve patient’s survival – Ventricular Assist Device (VAD). The Left Ventricular Assist Device (LVAD) is the most common implanted support system (97%, mostly continuous– flow), which demonstrates improved 1- and 2-year survival rates compared to optimal medical therapy [9]. LVAD can be used as bridge to transplant (patients awaiting a heart transplant) or as destination therapy (no candidate for heart transplant) or bridge to recovery (e.g., severe myocarditis), [10]. Since LVAD system provide complete support of the cardiac function with improved patient ‘s comfort and quality of life (discharge at home, ability to exercise), the number of implantations increased continuously [11]. As the number of LVAD grows, risks and complications in the long term follow up are recognized and noncardiac surgical procedures are necessary in some patients. The aim of this study is to evaluate risk factors and perioperative complications in LVAD patients who had to undergo spine surgery.

The authors retrospectively reviewed electronic medical records of patients, who underwent LVAD implantation from January 2011 to December 2020 and needed a spine surgery procedure. Patient related demographic parameters, indication for LVAD implantation, indication for spine surgery, preoperative anticoagulation, perioperative risk factors (evaluated by ASA classification), perioperative management and perioperative outcome (revision surgery) were analyzed.

Inclusion criterion

LVAD Patients that required NCS and general anasthetic from jan 2011-dec 2020.

Exclusion criteria

Endoscopic interventions, Surgery due to driveline-infection, ICD Implantation, Patients < 18years

The local ethic committee approved the collection of the data.

Anticoagulation protocol

All patients received long-term anticoagulation with vitamin K dependent anticoagulants in combination with platelet inhibitors (according to international normalaized ratio (INR) of prothrombin time 2-3 times the normal value): Heartmate3 ®: INR 2-2,5; HVAD®: INR 2,5-3. Preoperatively the anticoagulation therapy was bridged to continuous intravenous heparin therapy (target pTT 60-80 sec) when INR <2. The heparin therapy was stopped 4 hours prior to surgery and continued 6 hours postoperatively [12]. Bridging to peroral anticoagulation was performed postoperatively after secure wound healing.

A total of 626 LVAD implantations were performed from January 2011 to December 2020 in our institution. 83 LVAD patients (13,26 %) underwent general anesthesia for 100 noncardiac surgical procedures: Abdominal surgery: 58%; Orthopedic surgery: 29%; Others (endovascular procedures, soft tissue procedures): 13%; Revision surgery due to bleeding: 11%. Fig. 1 demnstrates the increasing number of NCS from 2011 to 2020 [Figure 1].

5 male LVAD patients (median 65 years, min 52 years, max 69 years) had to undergo 8 spine surgery procedures (8% of all NCS). 2 of the 5 patients received the LVAD system as bridge to transplant and 3 patients as destination therapy. The mean time from LVAD implantation to spine surgery was 223,5 days (min 112, max 3675). 8 spine procedures were performed in 5 patients: 4 surgeries were performed in prone position, 3 surgeries in the right lateral decubitus and 1 cervical spine surgery in supine position. Revision surgery was necessary in 1 patient with bleeding complication and hemothorax after thoracic vertebral body replacement. The mean duration of hospital stay was 92,3 days (min 26, max 164). Tab 1. demonstrates the patient related demographics and parameters [Table 1].

Table 1: LAVD patients with spine surgery (n=5).

Monitoring

All the patients were monitored with arterieal and central venous lines. A transesophageal echocardiogram was performed in 3 cases intraoperatively. LVAD monitoring ran continuously in all cases (flow, pulse index, speed and power) under supervision of a specialist.

Perioperative adverse events / complications

We did not recognize any right ventricular failure, no thrombembolic event, no pump dysfunction.

LVAD alarm

LVAD alert was registered in 4 cases due to a low-flow situation (hypovolamia). This problem was solved by LVAD pump speed reduction and kolloid infusion.

An intraoperative hypotension (MAP<70/20 min) was observed in 8 of 8 cases. This was mainly during the introduction of anaesthesia (n=3) and prone positioning (n=3). An intraoperative hypertension (MAP>90/20 min) was observed in 5 of 8 cases (introduction of anaesthesia n=3; during the operation n=2, lateral positioning n=1).

X-ray and hardware

In one case of percutaneous screw fixation the surgeon had visualization problems due to the hardware (driveline, pump).

Bleeding/ Thrombosis

No bleeding complication occured intraoperatively. Postoperative interventions due to continuous bleeding was necessary in two cases: One patient developed epistaxis and hematochezia. This was stopped bei tamponade and endoscopic cauterization in the colon sigmoideum. One patient developed an hemothorax due to bleeding out of an intercostal artery. This was stopped by endovascular embolization. A staged evacuation of the hematoma was necessary. No thrombotic complication has been observed during the perioperative period.

Blood product supply

The administration of red blood cells was necessary in 7 cases (n=55U), of PPSB-concentrate in 3 cases of fresh frozen plasma in 2 cases (n=11U) and of platelet concetrate in 1 case (n=1U) perioperatively.

A total of 626 patients received LVAD system from January 2011 to December 2020 as bridging or as destination therapy in an end stage heart failure. 100 non cardiac surgeries were necessary in 83 patients (13,26%). 8 spine surgeries were performed in 5 patients. More than the half of the NCS in our population were abdominal surgeries (58%), orthopedic surgeries occurred in 29%. This is consistent with most studies [13,14]. The mean age in our overall NCS population was 59,4 years (±10,34), 88% were male. The comorbidities in our NCS population did not differ from other reported LVAD patients with NCS: Most patients suffer from valvular heart disease, hypertension, coronary arterial disease. We regognized an overall mortality rate of 14,5% in the NCS group. The 30 days mortality after NCS is reported between 0 and 56% [15,8,16]. We observed a continuous increase of NCS from n=2 in 2011 to n=18 in 2020. This is consistent with other publications [14,12]. This is related to the demographic change, an increasing number of implantations, expansion of indication and better understanding in the perioperative management of NCS. The increasing number of LVAD implantation leads to an increasing number of noncardiac interventions and surgeries. The risk of readmission for NCS in LVAD patients is up to 10 % [17]. LVAD patients undergoing noncardiac procedures most frequently require upper or lower endoscopies (65%) and most of the procedures (95%) were elective or nonemergent procedures. Stone et al. reported of 138 LVAD patients who underwent 291 procedures in a study period of ten years (2003-2013). Upper and lower endoscopies were performed in 35,74% [18]. Degnan et al. reported of an average of 2.4 procedures per patient, 97% survived to discharge after their procedure [14]. A recent study of Berger et al. reported of a total of 105 surgical procedures in 63 LVAD patients. General anesthesia was required in 100 cases (abdominal surgery: 15,24%, neurosurgery: 8,57%, orthopedic surgery: 7,62%). It should be mentioned that they included LVAD related complications like late wound infections and driveline complications (24 cases; 22,86%) and pacemaker / ICD implantation and revisions (24 cases; 22,86%). The authors found a perioperative mortality of 3.8% with in the first 30 days and concluded that non-cardiac surgery can be performed on LVAD recipients with acceptable mortality [13]. To the best of our knowledge, this is the first study that investigates spine surgery procedures in LVAD patients. The mean time from LVAD implantation to surgery in our population was 223,5 days (min 112, max 3675). Berger et al. reported an interval between LVAD implantation and NCS of 645 days [13], Mathis et al. of 370 days [16] and Chen et al. of 329 days [19]. The median length of hospital stay due to NCS is 14.4 (±14.6) days and it is significantly higher compared with LVAD recipients admitted for other reasons except for NCS [17]. We recognized a much longer duration of hospital stay with at mean 92,3 days (min 26, max 164) in our spine surgery group. The reason for that is due to the necessity of preop and postop bridging (INR), the kind of surgery (staged anterior and posterior procedures) and revision surgery due to bleeding. In accordance with the literature our LVAD patients undergoing NCS are slightly older and have more cormorbidities compared to non NCS patients [17]. The mean age of LVAD patients undergoing noncardiac surgeries is 58.8 years [17]. Our spine patients were 65 years old and on average 5.6 years older than the most NCS patients. In our population two patients (40%) were classified as grade 3 and three patients (60%) as grade 4 according to the ASA classification. Mathis et al reported ASA 3 in 13,1%, ASA 4 in 85,8% and ASA 5 in 1,1% [16]. Ischemic cardiomyopathy (ICM) is the reason for LVAD implantation in most cases [16,20]. In our population, only one patient suffered from ICM, four patient (80%) demonstrated a dilated cardiomyopathy (DCM) as indication for LVAD. 2 of the 5 patients (40%) received the LVAD system as bridge to transplant and 3 patients (60%) as destination therapy. Other authors reported a more balanced distribution between destination therapy and bridg to transplant therapy [16]. Spine surgery was performed in prone position in 4 cases, in right lateral decubitus in 3 cases and in supine position in one case. Especially the prone position needs attention in LVAD patients: Pulling and kinking of the driveline should be avoid. As X-ray visualization is mandatory in percutaneous spine surgery, the pump and the driveline should be out of the operation field. Further the contact of disinfectants can lead to LVAD pump failure or damage. Hemodynamic alterations during prone positioning can lead to major problems in LVAD patients. When moving a patient into the prone position an average decrease in cardiac index (CI) of 24% occurs (decrease in stroke volume) [21]. Positioning in a convex frame further can cause a decrease in CI and stroke volume index with no significant increase in inferior vena cava (IVC) pressure. The obstruction of the inferior vena cava by abdominal compression leads to decreased cardiac output, increased bleeding, venous stasis and consequent thrombotic complications. Careful positioning with free hanging abdomen is decisive in LVAD patients especially because these patients have reduced compensatory mechanisms [22,21]. Therefore, MAP of <85-90 mmHg is recommended perioperatively [22]. Hypertension during the induction of anaesthesia should be avoided to minimize pump failure [8]. Arterial and central venous lines are recommended for continuous hemodynamic monitoring. An intraoperative TEE should be performed in case of suspected severe volume alteration [24,18]. The international Society for Heart and Lung Transplantation recommends a continuously LVAD monitoring under supervision of a specialist [25]. We observed hemodynamic alteration outside the proposed MAP target area without instability or pump dysfunction in 8 of 8 cases, mostly during the introduction of anaesthesia and prone positioning. Intraoperative Hypotension is reported by other authors in 27% [16]. MAP> 90 mmHg is associated with a 34% increased risk of perioperative stroke events [26]. Therefor a strictly anticoagulation regime is crucial in LVAD patients to avoid thrombotic events, especially pump thrombosis. In case of major noncardiac surgical procedures perioperative bridging is necessary to avoid bleeding problems. In non-emergency surgeries the switch of the longterm warfarin therapy to an intravenous heparin therapy with target partial thromboplastin time between 60 and 80 seconds is recommended [8]. In our population the mean INR was 1,5 and the mean PTT 44,5 sec 4 hours prior surgery. This corresponds to the general recommendations and reports of other authors [16,27]. Bleeding is one of the major complications in LVAD patients with NCS [19,12]. Abnormal bleeding occurs in almost 50% of patients. The yearly risk for developing a GI bleed is almost 9% [28].

The bleeding risk is based on various factors:

1. An acquired von Willebrand deficiency due to shear forces and enzymatic degradation of the factor by the pump [29-31].

2. The development of gastrointestinal arteriovenous malformations [32].

3. The necessitty of longterm antikoagulation with warfarin (INR: 2 - 2.5) [33].

We recognized a lower GI bleed in one case, which was controlled by endoscopic cauterization. Other authors demonstrated clinical benefits of octreotide in the treatment of LVAD-associated GI bleeds [34,35]. In one case we had to deal with an intercostal artery bleed, which was stopped by selective embolization. Regarding the literature NCS is associated with the transfusion of perioperative blood products in 62% [14,28,36]. Therefore, a sufficient blood product allocation is recommended preoperatively. In our small group of 5 patients and 8 spine procedures the administration of red blood cells was necessary in 7 cases (n=55U) perioperatively. The application of cell saver systems is recommended in case of extended bleeding and procedures with high bleeding risk. 3 LVAD patients in our study needed surgery due to spinal infection and proofen bacteriemia, which represents a contraindication for the use of cell savers. One patient underwent percutaneous stabilization due to a pathologic fracture. Tumor condition represents a contraindication as well. Antidote or blood supplements are recommended in anticoagulated patients who need emergency surgery to avoid bleed and circulation problems. But the risk of LVAD pump thrombosis and pump failure must be considered [13]. Our patients received PPSB-concentrate in 3 cases, fresh frozen plasma in 2 cases (n=11U) and platelet concetrate in 1 case (n=1U) perioperatively. We did not recognize thrombotic events or pump failure. In one case of transthoracic vertebral body replacement, we had access difficulties due to an extensive scarring after LVAD implantation. These problems might be reduced in the future by optimizing device design and minimally invasive LVAD implantation [37].

Due to the retrospective nature of our study, data were limited to that which was recorded for clinical care purposes. LVAD patients who were admitted to external institutions for NCS might be missing. The number of spine procedures in this NCS group is very small, major complication and adverse events might be undetected.

To the best of our knowledge this is the first report of spine surgery procedures in LVAD patients. We did not find considerably differences in morbidity and complication rates compared to other reported NCS. The risk of perioperative bleeding is known and should be controlled by strictly bridging management and blood product supplementation. Prone positioning might be a problem for hemodynamics, thats why continuously monitoring is mandatory. No patient died perioperatively. As LVAD implantation will increase, the number of NCS and also spine surgeries will increase in the future. Spine procedures are feasible in LVAD patients. We recommend surgery in LVAD patients in specialised centers to reduce morbidity rates in these complex patients.

None.

No conflict of interest.

1. Sherrington C (1952) The integrative action of the nervous system CUP Archive.
2. Todd ME (1937) The Thinking Body A Study of the Balancing Forces of Dynamic Man, PB Hoeber, Incorporated.
3. Gibson JJ (1966) The senses considered as perceptual systems. Boston Houghton.
4. Kim W (2020) The Knee Proprioception as Patient-Dependent Outcome Measures within Surgical and Nonsurgical Interventions. Proprioception IntechOpen.
5. Han J, G Waddington, R Adams, J Anson, Y Liu (2016) Assessing proprioception: What do you really want to know? Response to Krewer et al. Journal of sport and health science 5(1): 93 -94.
6. Kim W (2022) Study about Rules for the Haptic Control of Locomotion. Research Developments in Science and Technology Vol. 1: 1-5.
7. Kim W, D Araujo, SS Kohles, SG Kim HH Alvarez Sanchez (2021) Affordance-Based Surgical Design Methods Considering Biomechanical Artifacts. Ecological Psychology 33(1): 57-71.
8. Han J, G Waddington, R Adams, J Anson, Y Liu (2016) Assessing proprioception: a critical review of methods. Journal of Sport and Health Science 5(1): 80-90.
9. Bushnell M, G Duncan, R Hofbauer, B Ha, JI Chen, et al. (1999) Pain perception: is there a role for primary somatosensory cortex? Proceedings of the National Academy of Sciences 96(14): 7705-7709.
10. Kim W, A Veloso, J Tan, C Andrade (2010) A Reciprocal Connection at Knee Joint. ASME 2010 Summer Bioengineering Conference, Naples FL.
11. Hunt K H (1990) Kinematic geometry of mechanism. Oxford Clarendon Press.
12. Kim W, D Araujo, SS Kohles, SG Kim HH Alvarez Sanchez (2020) Affordance-Based Surgical Design Methods Considering Biomechanical Artifacts Ecological Psychology: 1-15.
13. Conconi M, N Sancisi, V Parenti-Castelli (2019) The geometrical arrangement of knee constraints that makes natural motion possible Theoretical and experimental analysis. Journal of biomechanical engineering 141(5): 051001.
14. Wojtys EM, JA Ashton-Miller, D Fry-Welch, LJ Huston (2001) Can proprioception really be improved by exercises?
15. Lanczos C (2012) The variational principles of mechanics Courier Corporation.
16. Merlet J P (1989) Singular configurations of parallel manipulators and Grassmann geometry The international journal of robotics research 8(5): 45-56.
17. Fregly BJ, TF Besier, DG Lloyd, SL Delp, SA Banks, et al. (2012). Grand challenge competition to predict in vivo knee loads. Journal of Orthopaedic Research 30(4): 503-513.
18. Kim W, A P Veloso, D Araújo, V Vleck, F João (2013) An informational framework to predict reaction of constraints using a reciprocally connected knee model Computer Methods in Biomechanics and Biomedical Engineering 18(1): 78-89.
19. Kim W, E A Vela (2022) Freedom in Osteoarthritis of the Knee. Applied Sciences 12(2): 839.
20. Kim W, EA Vela (2021) A Tensional Network in the Knee. Biomedical Journal of Scientific &amp; Technical Research 40(2): 32073-32078.
21. Witchalls J, P Blanch, G Waddington, R Adams (2012) Intrinsic functional deficits associated with increased risk of ankle injuries: a systematic review with meta-analysis. British journal of sports medicine 46(7): 515-523.
22. Gibson JJ (1979) The ecological approach to visual perception Houghton Mifflin.
23. Sharma L, DE Hurwitz, EJ Thonar, JA Sum, ME Lenz, et al. (1998) Knee adduction moment, serum hyaluronan level, and disease severity in medial tibiofemoral osteoarthritis. Arthritis Rheum 41(7): 1233-1240.
24. Andriacchi TP, TM Griffin, RF Loeser, CR Chu, EM Roos, et al. (2020) Bridging Disciplines as a pathway to Finding New Solutions for Osteoarthritis a collaborative program presented at the 2019 Orthopaedic Research Society and the Osteoarthritis Research Society International. Osteoarthritis and Cartilage Open 2(1): 100026.
25. Gibson JJ (1986) The Ecological Approach to Visual Perception Lawrence Erlbaum Associates.
26. Kim W (2021) A Critical Study on the Duality of Functional Knee Axes and Foot Contact. Research Trends and Challenges in Medical Science Vol. 8: 62-71.
27. Kim W, Y Choi, HG Lee (2016) The Duality of Knee Functional Axes and Foot Contact. Journal of Functional Morphology and Kinesiology 1(4): 387-395.
28. Semple JG, GT Kneebone (1960) Algebraic Projective Geometry Clarendon Press.