Advertisement

del Nido versus Buckberg cardioplegia in adult isolated valve surgery

Open ArchivePublished:October 22, 2014DOI:https://doi.org/10.1016/j.jtcvs.2014.10.085

      Background

      del Nido solution is a non–glucose-based, single-dose cardioplegic solution with few data supporting its safety in adults. We hypothesized that it and Buckberg solution offer myocardial protection of equivalent safety for isolated adult valve surgery.

      Methods

      Adult patients undergoing primary isolated aortic or mitral valve surgery with del Nido or Buckberg solution from January 2010 to September 2013 were 1:1 propensity matched (85 aortic valve, 110 mitral valve), and outcomes were compared.

      Results

      After aortic valve operations, no hospital deaths occurred, and troponin T levels (median 0.19 ng · mL−1 for del Nido vs 0.21 ng · mL−1 for Buckberg) were similar, with no statistically significant change in left ventricular ejection fraction (P = .4). Aortic clamp, bypass, and operating room times were shorter with del Nido solution (44 ± 14 vs 56 ± 19; 56 ± 18 vs 70 ± 24; and 285 ± 44 vs 308 ± 61 minutes, respectively; P < .0001). Peak intraoperative glucose levels (170 ± 31 vs 240 ± 41 mg · dL−1; P < .0001) and postoperative insulin-drip requirements (46% vs 82%; P < .0001) were lower. After mitral operations, there were no hospital deaths and no statistically significant cardioplegia-specific changes in troponin T levels (median 0.37 ng · mL−1 for del Nido vs 0.4 ng · mL−1 for Buckberg) or postoperative left ventricular ejection fraction (P = .13). We found no clear time differences with del Nido solution in mitral cases, but intraoperative glucose levels and postoperative insulin-drip requirements (184 ± 37 vs 250 ± 60 mg · dL−1 and 50% vs 67% mg · dL−1, respectively; P = .009) were lower.

      Conclusions

      del Nido solution can be used safely and effectively as an alternative to Buckberg solution in adult isolated valve surgery and is associated with lower insulin requirements and potential time and cost savings.

      CTSNet classification

      Abbreviations and Acronyms:

      CPB (cardiopulmonary bypass), LA (left atrial), LV (left ventricular), LVEF (left ventricular ejection fraction), SD (standard deviation)
      See related commentary on pages 637-8.
      At our institution, Buckberg cardioplegic solution
      • Follette D.M.
      • Steed D.L.
      • Foglia R.
      • Fey K.
      • Buckberg G.D.
      Advantages of intermittent blood cardioplegia over intermittent ischemia during prolonged hypothermic aortic clamping.
      • Follette D.M.
      • Mulder D.G.
      • Maloney J.V.
      • Buckberg G.D.
      Advantages of blood cardioplegia over continuous coronary perfusion or intermittent ischemia. Experimental and clinical study.
      • Follette D.M.
      • Fey K.
      • Buckberg G.D.
      • Helly Jr., J.J.
      • Steed D.L.
      • Foglia R.P.
      • et al.
      Reducing postischemic damage by temporary modification of reperfusate calcium, potassium, pH, and osmolarity.
      has been the standard used in adult cardiac surgery for several decades. An induction dose is given (mixed with patient blood in a 1:4 ratio; Table 1) to arrest the heart; a maintenance solution is administered every 15 to 20 minutes thereafter. During a typical procedure, multiple doses are given, which may interrupt the flow of the operation. A reperfusion solution is administered just before release of the aortic clamp.
      Table 1Cardioplegia compositions
      Buckberg solutiondel Nido solution
      Cold inductionComponents
       D5 ¼ NS392 mL Plasma-Lyte A1200 mL
       Tromethamine 0.3 M60 mL Mannitol 20% (3.2 g · L−1)16 mL
       C-P-2-D30 mL MgSO4 50% (2 g · L−1)4 mL · L−1
       KCl (2 mEq · mL−1)36 mEq NaHCO3 (1 mEq · mL−1)13 mL · L−1
      Total volume500 mL KCl (2 mEq · mL−1)13 mL · L−1
      Maintenance Lidocaine 2%7.8 mL
       D5 ¼ NS798 mLVolume given20 mL · kg−1 up to maximum dose of 1000 mL
       Tromethamine 0.3 M123 mL
       C-P-2-D61 mL
       KCl (2 mEq · mL−1)36 mEq
      Total volume1000 mL
      Reperfusion
       70% dextrose26 mL
       Tromethamine 0.3 M56 mL
       C-P-2-D113 mL
       KCl (2 mEq · mL−1)15 mEq
       Glutamate/aspartate62.5 mL
      Total volume500 mL
      D5, 5% Dextrose solution; NS, normal saline; C-P-2-D, citrate-phosphate-2-dextrose.
      An alternative, del Nido solution,
      • Matte G.S.
      • del Nido P.J.
      History and use of del Nido cardioplegia solution at Boston's Children's Hospital.
      has also been used for decades in other institutions, primarily for pediatric heart surgery. It is administered as a single dose that is said to last up to 180 minutes. It is a calcium-free, potassium-rich, non–glucose-based solution (in contrast to Buckberg solution, which is dextrose-based) and has an electrolyte composition similar to extracellular fluid (Table 1). It is delivered along with fully oxygenated patient blood in a 4:1 ratio, providing potassium-based myocyte depolarization with concurrent lidocaine sodium channel blockade. Just as Buckberg solution has additives for substrate enhancement, calcium regulation, and buffering, del Nido solution contains additives to effect free-radical scavenging, calcium channel blockade, and buffering (Table 1).
      • Matte G.S.
      • del Nido P.J.
      History and use of del Nido cardioplegia solution at Boston's Children's Hospital.
      Given the ease of administration, and the potential for reducing surgical interruption and cost, interest in del Nido solution has been increasing in the adult cardiac community.
      • Matte G.S.
      • del Nido P.J.
      History and use of del Nido cardioplegia solution at Boston's Children's Hospital.
      Although there is a large body of unpublished experience regarding its use in adults at some centers, and a small amount of animal data,
      • Govindapillai A.
      • Hua R.
      • Rose R.
      • Friesen C.H.
      • O'Blenes S.B.
      Protecting the aged heart during cardiac surgery: use of del Nido cardioplegia provides superior functional recovery in isolated hearts.
      • O'Blenes S.B.
      • Friesen C.H.
      • Ali A.
      • Howlett S.
      Protecting the aged heart during cardiac surgery: the potential benefits of del Nido cardioplegia.
      published clinical data are scant. No randomized trial has been launched to investigate the comparative safety of these 2 cardioplegia methods. Therefore, to estimate sample size and other details to facilitate such a trial, the present propensity-matched study addresses whether del Nido solution provides myocardial protection equivalent to that of Buckberg solution during primary isolated cardiac valve surgery.

      Patients and Methods

       Patients

      From August 29, 2012, to September 1, 2013, a total of 394 patients underwent surgery in which del Nido solution was used, including isolated valve procedures, septal myectomies, maze procedures, limited aortic procedures, and multiple valve procedures. Of these, 205 received del Nido solution, 89 for aortic valve procedures and 116 for mitral valve procedures. These isolated aortic and mitral valve operations are the focus of our study.
      To find a sufficient number of patients for propensity matching, we expanded the inclusion criteria for Buckberg cardioplegia to 2010. From January 1, 2010, to September 1, 2013, a total of 2429 patients underwent primary isolated valve surgery. Of these, 2224 received Buckberg solution, 1074 for aortic valve procedures (Table E1) and 1150 for mitral valve procedures (Table E2).

       Data

      Data were retrieved from the prospective Cleveland Clinic Cardiovascular Information Registry and the Anesthesia Record Keeping System, and were intraoperatively collected from quality improvement records. All data were approved for use in research by the institutional review board, with patient consent waived. Patient characteristic variables and in-hospital outcomes were as defined by The Society of Thoracic Surgeons National Adult Cardiac Surgery Database.

       Surgical Technique

      Conventional general anesthesia was used in all patients regardless of surgical approach. The surgical approach (full sternotomy or upper hemisternotomy for aortic valve operations and full sternotomy, partial sternotomy, right thoracotomy, or robotic for mitral valve operations) was chosen by the surgeon. Although the details of cardioplegia administration varied, in general, antegrade and retrograde induction doses of Buckberg solution were administered (approximately 500 to 600 mL each), with maintenance doses (approximately 400 mL) given every 15 to 20 minutes thereafter. Controlled warm reperfusion in the form of reanimation or “hotshot” cardioplegia
      • Teoh K.H.
      • Christakis G.T.
      • Weisel R.D.
      • Fremes S.E.
      • Mickle D.A.
      • Romaschin A.D.
      • et al.
      Accelerated myocardial metabolic recovery with terminal warm blood cardioplegia.
      (approximately 300-400 mL, with both retrograde and antegrade delivery) was generally performed before removal of the aortic clamp. In general, del Nido solution was administered in a single antegrade dose (generally 1000 mL). Redosing (Appendix E1) of del Nido solution and systemic cooling were performed at the discretion of the surgeon; however, by and large, systemic hypothermia was not used in either group.
      Intraoperative transfusion, fluid administration, and use of inotropes and pressors were carried out at the discretion of the anesthesiologists. Intraoperative insulin-drip administration was driven by protocol, which aimed at maintaining blood glucose between 70 and 150 mg · dL−1, with insulin drip started for levels >120 mg · dL−1 before cardiopulmonary bypass (CPB), or for any single level >150 mg · dL−1 on or after CPB.

       Endpoints

      The primary endpoint assessed was highest postoperative in-hospital troponin T level. A sample was routinely obtained 6 to 12 hours postoperatively. Multiple samples were obtained only if clinically indicated. Secondary endpoints included in-hospital death and additional measures of myocardial injury (postoperative left ventricular ejection fraction [LVEF] assessed by transthoracic echocardiography, use of post-CPB chemical inotropic/pressor therapy, and new-onset atrial fibrillation or flutter). Tertiary endpoints included intraoperative hematocrit levels, volumes of resuscitative fluid and red blood cell transfusions, trends in intraoperative glucose levels and immediate post-CPB insulin-drip requirements, and aortic clamp time, CPB time, and total operating room time.

       Data Analysis

      All analyses were performed using SAS statistical software (SAS version 9.2; SAS Institute, Cary, NC).

       Propensity-score development and use

      To simulate a randomized trial, propensity-score methodology was used to construct comparable del Nido and Buckberg cardioplegia groups, separately for aortic valve replacement and mitral valve surgery.
      • Rubin D.B.
      The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials.
      Multivariable logistic regression analysis was performed to identify patient characteristics associated with the del Nido group. Variables considered in the analyses are listed in Appendix E2. Initially, a parsimonious model was developed using automated analysis of 500 resampled data sets, followed by tabulating the frequency of occurrence at P ≤ .05, of both single factors and closely related clusters of factors.
      • Breiman L.
      Bagging predictors.
      • Sauerbrei W.
      • Schumacher M.
      A bootstrap resampling procedure for model building: application to the Cox regression model.
      Factors with ≥50% occurrence were retained in the parsimonious models. The only reliable factor associated with use of del Nido solution in aortic valve patients was a less invasive approach (P < .0001); several variables, including approach, were associated with its use in the mitral valve group (Table E3).Thereafter, we augmented the parsimonious model with other variables representing patient demographics, symptoms, and cardiac and noncardiac comorbidities to form saturated models (Appendix E2).
      A propensity score was then calculated for each patient by solving the saturated model for the probability of being in the del Nido group. Next, using only the propensity score, del Nido patients were matched to Buckberg patients using a greedy matching strategy
      • Bergstralh E.J.
      • Konsanke J.L.
      Computerized Matching of Cases to Controls. Technical Rep. No. 56.
      without replacement. Patients whose propensity scores deviated by more than 0.15 were considered unmatched. This process yielded 85 well-matched pairs from the 89 del Nido cases (96% matched) for aortic valve replacement (Table 2), and 110 well-matched pairs from the 116 del Nido cases (95% matched) for mitral valve surgery (Table 3). Propensity-matched patients in both valve groups were drawn from across the entire spectrum of propensity scores (Table 2, Table 3; Figure E1, A and B), and patient pairs were well-matched (Figure E2, A and B).
      Table 2Characteristics and operative details of matched patients undergoing aortic valve surgery
      VariableBuckberg group (n = 85)del Nido group (n = 85)P value
      n
      Patients with data available.
      No. (%) or mean ± SDn
      Patients with data available.
      No. (%) or mean ± SD
      Demographics
       Women8524 (28)8526 (31).7
       Age (y)8568 ± 158569 ± 14.8
       Body mass index (kg · m−2)8529 ± 8.18529 ± 6.6.8
      NYHA functional class7285.2
       I24 (33)22 (26)
       II24 (33)42 (49)
       III21 (29)19 (22)
       IV3 (4.2)2 (2.4)
      Aortic valve pathophysiology
       Aortic valve area (cm2)740.76 ± 0.41790.78 ± 0.40.8
       Mean aortic valve gradient (mm Hg)7948 (36, 63)
      Median (15th, 85th) percentiles.
      8251 (36, 63)
      Median (15th, 85th) percentiles.
      .2
       Peak aortic valve gradient (mm Hg)8080 ± 268284 ± 27.17
      Cardiovascular comorbidity
       Calculated LV relative wall thickness (cm)800.52 ± 0.16830.54 ± 0.14.3
       LV ejection fraction (%)8357 ± 8.58458 ± 9.8.4
       Calculated LV mass (g)80223 ± 8683224 ± 71.6
       Atrial fibrillation or flutter847 (8.3)8510 (12).5
       Peripheral arterial disease8514 (16)8513 (15).8
       Hypertension8576 (89)8580 (94).3
      Noncardiac comorbidity
       Diabetes8585
      Pharmacologically treated18 (21)17 (20).8
      Insulin treated5 (5.9)6 (7.1).8
      Non–insulin treated16 (19)13 (15).5
       Potassium (mmol · L−1)844.2 ± 0.60854.1 ± 0.57.4
       Glucose (mg · dL−1)85106 ± 2785103 ± 26.5
       Creatinine (mg · dL−1)851.01 ± 0.54850.98 ± 0.26>.9
       Cockcroft-Gault creatinine clearance8588 ± 358587 ± 34.9
       Hematocrit (%)8540 ± 4.58539 ± 4.9.9
      Operative details
       Aortic valve surgery
      Repair856 (7.1)856 (7.1)>.9
      Replacement8579 (93)8519 (93)>.9
       Incision/approach
      Full sternotomy858 (9.4)859 (11).8
      Hemisternotomy8577 (91)8576 (89).8
       Year of operation8585<.0001
      201012 (14)0 (0)
      201123 (27)0 (0)
      201238 (45)8 (9.4)
      201312 (14)77 (91)
      SD, Standard deviation; NYHA, New York Heart Association; LV, left ventricular.
      Patients with data available.
      Median (15th, 85th) percentiles.
      Table 3Characteristics and operative details of matched patients undergoing mitral valve surgery
      VariableBuckberg group (n = 110)del Nido group (n = 110)P value
      n
      Patients with data available.
      No. (%) or mean ± SDn
      Patients with data available.
      No. (%) or mean ± SD
      Demographics
       Women11040 (36)11032 (29).3
       Age (y)11056 ± 1211056 ± 12.8
       Body mass index (kg · m−2)11025 ± 3.911026 ± 3.5.11
      NYHA functional class101110.2
       I59 (58)50 (45)
       II37 (37)56 (51)
       III4 (44)3 (2.7)
       IV1 (0.99)1 (0.91)
      Mitral valve pathophysiology
       Mitral regurgitation grade
      1+1102 (1.8)1101 (0.91).9
      2+1102 (1.8)1102 (1.8).9
      3+11019 (17)11023 (21).9
      4+11087 (79)11084 (76).8
       Mitral stenosis1101 (0.91)1101 (1.8).6
      Cardiovascular comorbidity
       LV ejection fraction (%)10959 ± 5.910860 ± 5.4.3
       Calculated LV mass (g)96244 ± 75102227 ± 70.14
       Atrial fibrillation or flutter1105 (4.5)1100 (0).04
       Peripheral arterial disease1105 (4.5)1104 (3.6).7
       Hypertension11060 (55)11059 (54).9
      Noncardiac comorbidity
       Diabetes110
      Pharmacologically treated1102 (1.8)1102 (1.8)>.9
      Insulin treated1100 (0)1091 (0.92).3
      Non–insulin treated1102 (1.8)1091 (0.92).6
       Potassium (mmol · L−1)1094.1 ± 0.521104.2 ± 0.53.7
       Glucose (mg · dL−1)11092 ± 1411092 ± 18.5
       Creatinine (mg · dL−1)1101.03 ± 1.031101.02 ± 0.79.9
       Cockcroft-Gault creatinine clearance11092 ± 2511097 ± 26.19
       Hematocrit (%)11040 ± 5.011041 ± 4.9.19
      Operative details
       Mitral valve surgery
      Repair110108 (98)110106 (96).4
      Replacement1102 (1.8)1104 (3.6).4
       Incision/approach
      Full sternotomy1109 (8.2)11011 (10).6
      Minimally invasive (hemisternotomy, thoracotomy, or robotic)110101 (92)11099 (90).6
      Robotic11038 (35)11044 (40).4
       Year of operation110110<.0001
      201030 (27)0 (0)
      201123 (21)0 (0)
      201236 (33)16 (15)
      201321 (19)94 (85)
      SD, Standard deviation; NYHA, New York Heart Association; LV, left ventricular.
      Patients with data available.

       Factors associated with myocardial injury

      Cumulative distributions for all troponin values were constructed for unmatched and matched patient pairs. Linear regression models for postoperative troponin T were created that included cardioplegia group, myocardial ischemic time, and the interaction of cardioplegia group and myocardial ischemic time for each valve cohort of matched patients. Troponin values were log-transformed for analysis because of their skewed distribution. Data transformations of the ischemic times were investigated (logarithmic, inverse, exponentials) to find the best model to fit the data.
      Nested linear regression models were used to test group differences in postoperative LVEF (SAS PROC GLM). Within each valve cohort of matched patients, 2 models were constructed. The first model included preoperative LVEF, cardioplegia group, their interaction, and myocardial ischemic time. The second included only preoperative LVEF and myocardial ischemic time. The F test was used to test the hypothesis that the smaller model fit the data adequately and that the larger model (inclusion of cardioplegia group) is unnecessary.

       Presentation

      Continuous variables are summarized as mean ± standard deviation (SD) or as equivalent 15th, 50th (median), and 85th percentiles when data were skewed; comparisons were made using the Wilcoxon rank sum test. Categoric data are summarized using frequencies and percentages; comparisons were made using the χ2 test or Fisher exact test when the frequency was <5.

      Results

       Myocardial Injury

      No patient died in either group. The highest measured postoperative troponin T level was higher after mitral valve surgery than after aortic valve surgery (Figure 1), but in matched patients within each valve group, the distribution of values was similar for del Nido and Buckberg patients (aortic valve, 0.19 vs 0.21 ng · mL−1, P = .05; mitral valve, 0.37 vs 0.40 ng · mL−1, P = .2), although slightly lower with del Nido cardioplegia in both valve surgery groups (Table 4, Table 5). Longer ischemic times were associated with higher troponin T levels in both del Nido and Buckberg patients after mitral or aortic valve operations (P < .0001; Figure 2 and Table E4). Postoperative LVEF, adjusted for preoperative LVEF, was unaffected in both the del Nido and Buckberg groups after aortic valve (P = .4) and mitral valve surgery (P = .13) (Table E5).
      Figure thumbnail gr1
      Figure 1Highest measured troponin T levels after isolated AV and MV surgery: (A) after AV surgery, unmatched patients; (B) after AV surgery, propensity-matched patients; (C) after MV surgery, unmatched patients; (D) after MV surgery, propensity-matched patients. Solid lines represent patients receiving del Nido solution; dashed lines represent those receiving Buckberg solution. AV, Aortic valve; MV, mitral valve.
      Table 4Management variables of matched patients undergoing aortic valve surgery
      VariableBuckberg group (n = 85)del Nido group (n = 85)P value
      n
      Patients with data available.
      No. (%) or mean ± SDn
      Patients with data available.
      No. (%) or mean ± SD
      Myocardial injury
       Highest measured troponin T (ng · mL−1)840.11/0.21/0.48
      15th/50th/85th percentiles.
      850.08/0.19/0.41
      15th/50th/85th percentiles.
      .05
      Inotropic and vasoactive agents
       Epinephrine8515 (18)8512 (14).5
       Norepinephrine8521 (25)8526 (31).4
       Milrinone855 (5.9)852 (2.4).3
       Vasopressin853 (3.5)850 (0).08
      Blood products
       Albumin (mL)54500 ± 24027420 ± 200.15
       Crystalloid (mL)842900 ± 1000852900 ± 850.8
       Any blood product
      Intraoperative or postoperative.
      8521 (25)8521 (25)>.9
      Glucose and insulin
       Peak glucose (mg · dL−1)84240 ± 4185170 ± 31<.0001
       Insulin8570 (82)8539 (46)<.0001
      Operative times (min)
       Aortic clamp time, all approaches8556 ± 198544 ± 14<.0001
      Full sternotomy869 ± 27953 ± 22.16
      Upper hemisternotomy7754 ± 177643 ± 13<.0001
       CPB time, all approaches8570 ± 248556 ± 18<.0001
      Full sternotomy887 ± 37970 ± 34.3
      Upper hemisternotomy7768 ± 217654 ± 14<.0001
       Total OR time, all approaches85310 ± 6185280 ± 44.009
      Full sternotomy8340 ± 719300 ± 46.3
      Upper hemisternotomy77300 ± 5976280 ± 44.01
      Intraoperative laboratory values
       Peak potassium (mmol · L−1)844.8 ± 0.51854.6 ± 0.53.09
       Nadir hematocrit (%)8428 ± 4.38528 ± 3.9.4
      SD, Standard deviation; CPB, cardiopulmonary bypass; OR, operating room.
      Patients with data available.
      15th/50th/85th percentiles.
      Intraoperative or postoperative.
      Table 5Management variables of matched patients undergoing mitral valve surgery
      VariableBuckberg group (n = 110)del Nido group (n = 110)P value
      n
      Patients with data available.
      No. (%) or mean ± SDn
      Patients with data available.
      No. (%) or mean ± SD
      Myocardial injury
       Highest measured troponin T (ng · mL−1)1100.21/0.40/1.03
      15th/50th/85th percentiles.
      1070.19/0.37/0.76
      15th/50th/85th percentiles.
      .2
      Inotropic and vasoactive agents
       Epinephrine11014 (13)11021 (19).2
       Norepinephrine11015 (14)11011 (10).4
       Milrinone1101 (0.91)1101 (0.91)>.9
       Vasopressin1102 (1.8)1101 (0.91).6
      Blood products
       Albumin (mL)57440 ± 17923400 ± 146.5
       Crystalloid (mL)1092400 ± 8201102500 ± 736.18
       Any blood product
      Intraoperative or postoperative.
      11018 (16)11016 (15).7
      Glucose and insulin
       Peak glucose (mg · dL−1)108250 ± 60110180 ± 37<.0001
       Insulin11074 (67)11055 (50).009
      Operative times (min)
       Aortic clamp time, all approaches11071 ± 2311067 ± 22.3
      Full sternotomy971 ± 281148 ± 14.03
      Partial sternotomy3074 ± 292662 ± 18.3
      Right thoracotomy3361 ± 142959 ± 20.12
      Robotic3876 ± 224480 ± 21.13
       CPB time, all approaches110100 ± 3411098 ± 33.3
      Full sternotomy996 ± 331167 ± 15.02
      Partial sternotomy3098 ± 352685 ± 19.3
      Right thoracotomy3391 ± 202979 ± 23.008
      Robotic38112 ± 4044125 ± 29.03
       Total OR time, all approaches110342 ± 80110335 ± 60.8
      Full sternotomy9318 ± 6911280 ± 32.09
      Partial sternotomy30351 ± 8526341 ± 66.8
      Right thoracotomy33345 ± 10429316 ± 48.3
      Robotic38337 ± 4944359 ± 56.08
      Intraoperative laboratory values
       Peak potassium (mmol · L−1)1085.01 ± 0.651104.9 ± 0.54.16
       Nadir hematocrit (%)10829 ± 4.311029 ± 4.4.9
      SD, Standard deviation; CPB, cardiopulmonary bypass; OR, operating room.
      Patients with data available.
      15th/50th/85th percentiles.
      Intraoperative or postoperative.
      Figure thumbnail gr2
      Figure 2Relationship of troponin T level to myocardial ischemic time. Filled circles represent values for del Nido patients; open squares represent values for Buckberg patients. Solid lines are regression estimates (). A, Patients undergoing primary isolated aortic valve operations. B, Patients undergoing primary isolated mitral valve operations.

       Inotropic Support and Other Secondary Outcomes

      Following aortic or mitral valve surgery, use of inotropic and vasopressive agents in the del Nido and Buckberg groups was similar (Table 4, Table 5). Prevalence of atrial fibrillation, volume of intraoperative resuscitative fluids, and transfusion requirements were also similar. Occurrence of permanent stroke, reoperation for bleeding/tamponade, renal failure, and prolonged (>24 hours) ventilation after either aortic or mitral valve surgery was similar in both cardioplegia groups (Table 6).
      Table 6Postoperative complications in matched patients undergoing aortic or mitral valve surgery
      ComplicationsBuckberg group (n = 85)del Nido group (n = 85)P value
      No. (%)No. (%)
      Aortic valve surgery (n = 85)
       Hospital death0 (0)0 (0)
       Permanent stroke0 (0)1 (1.2).3
       Renal failure1 (1.2)2 (2.4).6
       Reoperation for bleeding/tamponade3 (3.5)0 (0).08
       Atrial fibrillation27 (32)30 (35).6
      Mitral valve surgery (n = 110)
       Hospital death0 (0)0 (0)
       Permanent stroke1 (0.91)1 (0.91)>.9
       Renal failure0 (0)0 (0)
       Prolonged ventilation (>24 h)0 (0)1 (0.91).3
       Reoperation for bleeding/tamponade0 (0)1 (0.91).3
       Atrial fibrillation32 (29)35 (32).7

       Intraoperative Glucose Levels and Insulin-Drip Requirements

      Peak intraoperative glucose level was lower (172 ± 31 vs 236 ± 41 mg · dL−1, P < .0001), as was the need for insulin drips (46% vs 82%, P < .0001), after aortic valve operations performed with del Nido solution (Table 4). After mitral valve operations using del Nido solution, peak intraoperative glucose levels and insulin requirements were similarly lower (184 ± 37 vs 247 ± 60 mg · dL−1 and 50% vs 67%, respectively, P = .009; Table 5).

       Operating Room Times

      For aortic valve operations overall, aortic clamp, CPB, and total operating room times were shorter with del Nido than with Buckberg solution (44 ± 14 vs 56 ± 19 minutes, 56 ± 18 vs 70 ± 24 minutes, and 285 ± 44 vs 308 ± 61 minutes, respectively, P < .0001; Table 4). However, when evaluated by approach, only upper hemisternotomy cases were associated with shorter times when del Nido solution was used. In contrast, aortic clamp, CPB, and operating room times were similar for mitral valve operations overall (Table 5). However, for those undergoing full sternotomy, they were shorter (48 ± 14 vs 71 ± 28 minutes, P = .03; 67 ± 15 vs 96 ± 33 minutes, P = .02; and 280 ± 32 vs 320 ± 69 minutes, P = .09, respectively; Table 5) when del Nido solution was used, although total operating room time was similar. Only right thoracotomy CPB times were statistically significantly lower in the del Nido group (79 ± 23 vs 91 ± 20 minutes, P = .008).

      Discussion

       Principal Findings

      This study demonstrates no disadvantages to using del Nido solution in primary isolated adult aortic or mitral valve surgery. Measures of myocardial injury, such as troponin T levels, postoperative LVEF, and postoperative inotropic/pressor support were similar to those of patients for whom Buckberg solution was used.
      Use of del Nido solution has several advantages. The crystalloid component of del Nido solution is not glucose-based (unlike Buckberg solution), and patients receiving it had fewer blood glucose perturbations, and less need for postoperative insulin drips. The groups were well matched with regard to presence or absence of diabetes, as well as preoperative glucose levels. However, more patients who had Buckberg solution required insulin drips afterward than did those with a history of diabetes. Specifically, only 46% of aortic valve patients receiving Buckberg cardioplegia had a history of diabetes, but 82% were placed on insulin drips.
      Operative times were shorter with certain operative approaches. The reduction in time may be due to fewer interruptions of the surgical repair and perhaps to a decreased need for coronary sinus cannulation (in patients without important aortic regurgitation, only antegrade cardioplegia was used in del Nido cases). We speculate, however, that some of these differences between operative approaches (eg, full sternotomy or right thoracotomy) were surgeon-specific.
      An additional advantage of using del Nido solution is its lower cost. Although we did not formally collect cost data in this study, multiple areas of potential cost savings are associated with this form of cardioplegia. At our institution, del Nido solution costs approximately $29 per dose; Buckberg solution costs approximately $75 per dose. Assuming 1 dose of del Nido solution per case, and at least 1 maintenance dose in a Buckberg solution case, the cardioplegia costs associated with del Nido solution are about $29, versus $225 for Buckberg solution (a savings of approximately $196).
      When retrograde cardioplegia is not used, additional savings are possible. For instance, the percutaneous retrograde cannula we use in robotic cases costs roughly $1000, whereas in open surgery, such equipment costs about $30, resulting in an overall savings of at least $1190 per robotic case and $220 per open case. However, equipment cost savings may vary among institutions, depending on the retrograde cannula used. Additional areas for savings include decreased operating room time and costs associated with insulin drips. In an era when healthcare costs are increasingly scrutinized, and maximal efficiency has become a vital concern, these differences have important implications.

       Clinical Implications

      Our adoption of del Nido solution has been intentionally cautious. Two critical keys to the success of “one-shot” cardioplegia are adequate delivery and excellent venous return. Thus, we have elected not to use del Nido solution in patients with important coronary artery disease, in an effort to avoid inhomogeneous or incomplete delivery, and this study cannot speak to its proper use in such patients. Impaired venous return tends to rewarm the heart (increasing myocardial metabolic demand), and the attendant increased venous back pressure encourages cardioplegia washout.
      Patients in this study were in good overall health, with preserved left ventricular systolic function, and they underwent procedures with aortic clamp times that rarely exceeded 90 minutes. This investigation does not allow us to comment reliably on optimal redosing strategies or the safety of del Nido solution in sicker patients (eg, those with depressed LVEF), or on those undergoing operations with longer aortic clamp times.

       Limitations

      This is a single-institution study, which limits its generalizability. We did not capture surgeon-specific details, such as adjuncts to myocardial protection (eg, systemic/topical cooling or use of a myocardial temperature probe to assess adequacy of delivery). Also not captured were perfusionist-specific maneuvers, such as intraoperative hemoconcentration, that may have played a role in avoiding hemodilution in cases using del Nido solution (which has a 4:1 crystalloid:blood ratio compared with the 1:4 ratio of Buckberg solution). Similarly, the potential effects on procedural costs associated with hemoconcentration (at a rough cost of $70, if performed) were not captured. Finally, patients were not matched for operative year; this was necessary to provide an adequate pool for propensity matching, because use of del Nido solution in isolated valve surgery quickly took hold after its introduction at our institution.

      Conclusions

      del Nido solution seems to be safe for use in adult primary isolated aortic or mitral valve operations. It has the advantages of decreased potential for surgical interruption, lower perturbations in intraoperative blood glucose levels, less need for postoperative insulin drips, decreased surgical times in certain surgical approaches, and lower costs. Caution is still warranted in adopting this form of cardioplegia, and further study is required to establish its indications and optimal use.
      S.M. thanks Dr Pedro del Nido for initiating her interest in del Nido cardioplegia and for his and Greg Matte's (chief perfusionist at Children's Hospital Boston) insights and guidance during the institution of its use at Cleveland Clinic. The authors are grateful to the surgeons (particularly Drs Michael Argenziano and Craig Smith) and perfusionists (particularly Linda Mongero) of New York Presbyterian Hospital—Columbia for allowing us to observe their use of del Nido solution and for graciously sharing experiences and insights with us. S.M. thanks Patrick Grady and the entire perfusion team at Cleveland Clinic for their support in instituting this novel type of cardioplegia. Special thanks also go to Catherine Torma, CCP, and Annmarie Fatula, CCP, for their intraoperative data collection and management thereof.

      Appendix E1. Redosing

      We reviewed every case in which the total dose of del Nido solution was larger than 1 dose per aortic clamp period. Among the 195 matched del Nido cases, 21 were redosing cases. Two were aortic valve cases in which a small initial antegrade dose was not sufficient for arrest, so an aortotomy was made, and an additional dose was administered directly down the coronary ostia. One case involved a patient who fibrillated after aortic clamp removal and was rearrested briefly after electrical defibrillation failed. Intraoperative redosing was highly surgeon-dependent and generally involved a partial (200-500 cc) additional dose of del Nido solution after 45 to 90 minutes. This occurred in 3 aortic valve cases (3.5%) and 15 mitral valve cases (14%).

      Appendix E2. Variables Considered in Propensity Analyses

       Demography

      Age
      In final propensity models.
      (years); sex
      In final propensity models.
      ; race
      In final propensity models.
      (black, white, other); height
      In final propensity models.
      (cm); weight (kg); body surface area (m2); body mass index
      In final propensity models.
      (kg · m−2).

       Presentation

      New York Heart Association functional class (I-IV)
      In final propensity models.
      ; prior stroke; prior myocardial infarction.
      In final propensity models.

       Cardiac Comorbidity

      Preoperative atrial fibrillation
      In final propensity models.
      ; heart failure
      In final propensity models.
      ; coronary artery disease
      In final propensity models.
      (number of systems with ≥50% stenosis).

       Noncardiac Comorbidity

      Preoperative glucose
      In final propensity models.
      ; pharmacologically treated diabetes
      In final propensity models.
      ; hypertension
      In final propensity models.
      ; peripheral arterial disease
      In final propensity models.
      ; history of smoking
      In final propensity models.
      ; chronic obstructive pulmonary disease; renal failure requiring dialysis; blood urea nitrogen
      In final propensity models.
      (mg · dL−1); creatinine
      In final propensity models.
      (mg · dL−1); bilirubin
      In final propensity models.
      (mg · dL−1); cholesterol (total, high-density lipoprotein
      In final propensity models.
      [mg · dL−1]; low-density lipoprotein [mg · dL−1]); triglycerides (mg · dL−1)
      In final propensity models.
      ; hematocrit (%).

       Echocardiography Data (Preoperative)

      Aortic valve regurgitation grade
      In final propensity models.
      ; mitral valve regurgitation grade
      In final propensity models.
      ; tricuspid valve regurgitation grade
      In final propensity models.
      ; aortic valve stenosis
      In final propensity models.
      ; mitral valve stenosis
      In final propensity models.
      ; left ventricular (LV) ejection fraction
      In final propensity models.
      (%); LV internal diameter in diastole (cm); LV internal end-systolic diameter (cm); LV end-diastolic volume (mL); LV end-systolic volume; LV mass (g); LA (left atrial) diameter (cm); LA volume
      In final propensity models.
      (mL); posterior wall thickness
      In final propensity models.
      (cm); aortic valve area
      In final propensity models.
      (cm2).

       Valve Etiology

      Rheumatic
      In final propensity models.
      ; degeneration
      In mitral valve propensity model only.
      ; endocarditis.
      In mitral valve propensity model only.

       Procedural

      Figure thumbnail fx1
      Figure E1Mirrored histogram of distribution of propensity scores. Darkened areas represent matched patient pairs: (A) isolated aortic valve surgery; (B) isolated mitral valve surgery.
      Figure thumbnail fx2
      Figure E2Propensity-score matching for del Nido versus Buckberg solution in adult isolated valve surgery. A, Covariable balance for selected variables before (triangles) and after (squares) propensity matching, expressed as standardized difference: aortic valve surgery. B, Covariable balance for selected variables before (triangles) and after (squares) propensity matching, expressed as standardized difference: mitral valve surgery. AV, Aortic valve; COPD, chronic obstructive pulmonary disease; AVR, aortic valve replacement; NYHA, New York Heart Association; Clr, clearance; LVEF, left ventricular ejection fraction; PAD, peripheral arterial disease; BC, Buckberg solution; DNS, del Nido solution; TV regurg, tricuspid valve regurgitation; MVR, mitral valve replacement; LA, left atrial.
      Table E1Characteristics and operative details of patients undergoing aortic valve surgery: Before matching
      VariableBuckberg group (n = 1074)del Nido group (n = 89)P value
      n
      Patients with data available.
      No. (%) or mean ± SDn
      Patients with data available.
      No. (%) or mean ± SD
      Demographics
       Women1074440 (41)8928 (31).08
       Age (y)107466 ± 158969 ± 14.17
       Body mass index (kg · m−2)107330 ± 7.48929 ± 6.6.3
      NYHA functional class87589.6
       I279 (32)23 (26)
       II419 (48)44 (49)
       III158 (18)20 (22)
       IV19 (2.2)2 (2.2)
      Aortic valve pathophysiology
       Aortic valve area (cm2)8120.74 ± 0.29830.81 ± 0.50.6
       Mean aortic valve gradient (mm Hg)92931/46/65
      Median (15th, 85th) percentiles.
      8634/51/63
      Median (15th, 85th) percentiles.
      .4
       Peak aortic valve gradient (mm Hg)94180 ± 308683 ± 28.2
      Cardiovascular comorbidity
       Calculated LV relative wall thickness (cm)9810.54 ± 0.17860.54 ± 0.14.5
       LV ejection fraction (%)106057 ± 9.08860 ± 10.04
       Calculated LV mass (g)980250 ± 9586230 ± 77.2
       Atrial fibrillation or flutter106659 (5.5)8910 (11).03
       Peripheral arterial disease107473 (6.8)8915 (17).0006
       Hypertension1074850 (79)8984 (94).0005
      Noncardiac comorbidity
       Diabetes
      Pharmacologically treated1073234 (22)8920 (22).9
      Insulin treated107172 (6.7)898 (9.0).4
      Non–insulin treated1071177 (17)8914 (16).8
       Potassium (mmol · L−1)10734.2 ± 0.62894.1 ± 0.56.13
       Glucose (mg · dL−1)1074102 ± 2789108 ± 39.2
       Creatinine (mg · dL−1)10741.04 ± 0.67890.98 ± 0.26.6
       Cockcroft-Gault creatinine clearance107391 ± 398987 ± 34.4
       Hematocrit (%)107438 ± 5.38939 ± 4.9.16
      Operative details
       Aortic valve surgery
      Repair107463 (5.9)897 (7.9).5
      Replacement10741011 (94)8982 (92).5
       Incision/approach
      Full sternotomy1054355 (34)899 (10)<.0001
      Hemisternotomy1054699 (66)8980 (90)<.0001
       Year of operation107489<.0001
      2010331 (31)0 (0)
      2011308 (29)0 (0)
      2012312 (29)8 (9.0)
      2013123 (11)81 (91)
      SD, Standard deviation; NYHA, New York Heart Association; LV, left ventricular.
      Patients with data available.
      Median (15th, 85th) percentiles.
      Table E2Characteristics and operative details of patients undergoing mitral valve surgery: Before matching
      VariableBuckberg group (n = 1150)del Nido group (n = 116)P value
      n
      Patients with data available.
      No. (%) or mean ± SDNo. (%) or mean ± SD
      Demographics
       Women1150398 (35)11636 (31).4
       Age (y)115057 ± 1211656 ± 11.2
       Body mass index (kg · m−2)114926 ± 4.911625 ± 3.4.3
      NYHA functional class904116.01
       I401 (44)51 (44)
       II385 (43)61 (53)
       III109 (12)3 (2.6)
       IV9 (1.0)1 (0.86)
      Mitral valve pathophysiology
       Mitral regurgitation grade1150116.7
      1+21 (1.7)1 (0.86)
      2+41 (3.6)2 (1.7)
      3+249 (22)23 (20)
      4+836 (73)90 (78)
       Mitral stenosis115031 (2.7)1162 (1.7).5
      Cardiovascular comorbidity
       LV ejection fraction (%)113159 ± 6.111460 ± 5.4.11
       Calculated LV mass (g)1044240 ± 80108220 ± 70.02
       Atrial fibrillation or flutter114339 (3.4)1160 (0).04
       Peripheral arterial disease115035 (3.0)1164 (3.4).8
       Hypertension1150619 (54)11662 (53).9
      Noncardiac comorbidity
       Diabetes
      Pharmacologically treated115058 (5.0)1162 (1.7).11
      Insulin treated115023 (2.0)1161 (0.87).4
      Non–insulin treated115043 (3.7)1161 (0.87).11
       Potassium (mmol · L−1)11474.1 ± 0.561164.1 ± 0.53.5
       Glucose (mg · dL−1)115094 ± 2011692 ± 18.04
       Creatinine (mg · dL−1)11491.01 ± 0.661161.01 ± 0.78.5
       Cockcroft-Gault creatinine clearance114995 ± 3011697 ± 25.2
       Hematocrit (%)115040 ± 5.011641 ± 4.8.2
      Operative details
       Mitral valve surgery
      Repair11501041 (91)116112 (97).03
      Replacement1150109 (9.5)1164 (3.4).03
       Incision/approach
      Full sternotomy1016279 (27)11614 (12).0003
      Minimally invasive (hemisternotomy, thoracotomy, or robotic)1016737 (73)116102 (88).0003
      Robotic1150389 (34)11645 (39).3
      Year of operation1150116<.0001
      2010365 (32)0 (0)
      2011360 (31)0 (0)
      2012299 (26)16 (14)
      2013126 (11)100 (86)
      SD, Standard deviation; NYHA, New York Heart Association; LV, left ventricular.
      Patients with data available.
      Table E3Factors associated with use of del Nido solution in mitral valve operations
      FactorEstimate ± SEP valueReliability (%)
      Percentage of times factor appeared in 500 bootstrap models.
      Greater height (taller)
      (180/height)2, inverse squared transformation.
      −2.3 ± 0.92.01455
      Lower TR grade−0.93 ± 0.28.001193
      Heart failure1.4 ± 0.24<.000199
      Smoking−0.97 ± 0.40.0251
      Thinner posterior wall thickness
      Logarithmic transformation.
      −2.3 ± 0.58<.000194
      Larger left atrial volume (index)
      Logarithmic transformation.
      0.87 ± 0.29.00290
      Approach
       Less-invasive1.5 ± 0.34<.0001100
       Robotic−1.16 ± 0.39.003100
      Intercept−3.8 ± 1.4.009
      C-statistic = .80. SE, Standard error; TR, tricuspid regurgitation.
      Percentage of times factor appeared in 500 bootstrap models.
      (180/height)2, inverse squared transformation.
      Logarithmic transformation.
      Table E4Postoperative troponin T levels after aortic valve and mitral valve surgery according to type of cardioplegia
      ParameterAortic valve surgeryMitral valve surgery
      Estimate ± SEPEstimate ± SEP
      Ischemic time
      (Hours)2, squared transformation.
      0.39 ± 0.103.00020.31 ± 0.055<.0001
      Ischemic time (h) in del Nido group (interaction)
      Versus Buckberg cardioplegia.
      −0.34 ± 0.36.3−0.25 ± 0.21.3
      de Nido group
      Versus Buckberg cardioplegia.
      0.19 ± 0.30.50.22 ± 0.26.4
      Intercept−1.8 ± 0.122<.0001−1.3 ± 0.105<.0001
      Troponin T modeled on natural logarithmic scale, log(troponin T). SE, Standard error.
      (Hours)2, squared transformation.
      Versus Buckberg cardioplegia.
      Table E5Postoperative left ventricular ejection fraction after aortic valve and mitral valve surgery according to type of cardioplegia
      ParameterAortic valve surgeryMitral valve surgery
      Estimate ± SEP valueEstimate ± SEP value
      Preoperative LV ejection fraction in Buckberg group0.63 ± 0.103<.00010.15 ± 0.108.16
      Preoperative LV ejection fraction in del Nido group0.78 ± 0.076<.00010.25 ± 0.125.05
      del Nido group
      Test of group effect: aortic valve surgery P = .4; mitral valve surgery P = .13.
      −9.5 ± 7.54.2−7.9 ± 10.2.4
      Ischemic time (min)0.062 ± 0.047.190.014 ± 0.0208.5
      Intercept17 ± 7.12.0244 ± 6.81<.0001
      SE, Standard error; LV, left ventricular.
      Test of group effect: aortic valve surgery P = .4; mitral valve surgery P = .13.

      References

        • Follette D.M.
        • Steed D.L.
        • Foglia R.
        • Fey K.
        • Buckberg G.D.
        Advantages of intermittent blood cardioplegia over intermittent ischemia during prolonged hypothermic aortic clamping.
        Circulation. 1978; 58: I200-I209
        • Follette D.M.
        • Mulder D.G.
        • Maloney J.V.
        • Buckberg G.D.
        Advantages of blood cardioplegia over continuous coronary perfusion or intermittent ischemia. Experimental and clinical study.
        J Thorac Cardiovasc Surg. 1978; 76: 604-619
        • Follette D.M.
        • Fey K.
        • Buckberg G.D.
        • Helly Jr., J.J.
        • Steed D.L.
        • Foglia R.P.
        • et al.
        Reducing postischemic damage by temporary modification of reperfusate calcium, potassium, pH, and osmolarity.
        J Thorac Cardiovasc Surg. 1981; 82: 221-238
        • Matte G.S.
        • del Nido P.J.
        History and use of del Nido cardioplegia solution at Boston's Children's Hospital.
        J Extra Corpor Technol. 2012; 44: 98-103
        • Govindapillai A.
        • Hua R.
        • Rose R.
        • Friesen C.H.
        • O'Blenes S.B.
        Protecting the aged heart during cardiac surgery: use of del Nido cardioplegia provides superior functional recovery in isolated hearts.
        J Thorac Cardiovasc Surg. 2013; 146: 940-948
        • O'Blenes S.B.
        • Friesen C.H.
        • Ali A.
        • Howlett S.
        Protecting the aged heart during cardiac surgery: the potential benefits of del Nido cardioplegia.
        J Thorac Cardiovasc Surg. 2011; 141: 762-770
        • Teoh K.H.
        • Christakis G.T.
        • Weisel R.D.
        • Fremes S.E.
        • Mickle D.A.
        • Romaschin A.D.
        • et al.
        Accelerated myocardial metabolic recovery with terminal warm blood cardioplegia.
        J Thorac Cardiovasc Surg. 1986; 91: 888-895
        • Rubin D.B.
        The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials.
        Stat Med. 2007; 26: 20-36
        • Breiman L.
        Bagging predictors.
        Machine Learn. 1996; 24: 123-140
        • Sauerbrei W.
        • Schumacher M.
        A bootstrap resampling procedure for model building: application to the Cox regression model.
        Stat Med. 1992; 11: 2093-2109
        • Bergstralh E.J.
        • Konsanke J.L.
        Computerized Matching of Cases to Controls. Technical Rep. No. 56.
        Dept Health Science Research, Mayo Clinic, Rochester, Minn1995