396

Prospective Comparison BetweenTotal Sternotomy and Ministernotomyfor Aortic Valve ReplacementFadi Farhat, M.D., Zhiqian Lu, M.D., Mathilde Lefevre, M.D., Pietro Montagna,

M.D., Philippe Mikaeloff, M.D., and Olivier Jegaden, M.D.

Department of Cardiac Surgery, Louis Pradel Hospital, Claude Bernard University,Lyon, France

ABSTRACT Ministernotomy (MS) is an alternative for total sternotomy (TS) in aortic valve replace-

ment. We compared these two approaches for results and adverse effects in a prospective study.

From January to December 2000, 100 patients who underwent aortic valve replacement were

included in two groups of 50 according to the surgical approach that used MS or TS; one senior

surgeon performed all cases in each group. Valvular pathologies were either stenosis or insuffi-

ciency. Mean age was 63 ± 14 years in MS, 67 ± 12 in TS (p = ns). NYHA class was 2.7 ± 0.5 in

MS, 2.8 ± 0.6 in TS (p = ns). Left ventricular ejection fraction was 58 ± 12% in MS, 57 ± 12% in

TS (p = ns). There was a significant difference between MS and TS in aortic cross-clamping (66 ±14 min vs 48 ± 9 min) and cardiopulmonary bypass (88 ± 18 min vs 69 ± 10 min, p < 0.01),

but not in intervention times (2.8 ± 0.4 hours vs 2.7 ± 0.4 hours). Mean intensive care stay was

reduced in MS (1.7 ± 1.6 days vs 2.6 ± 6 days, p < 0.05). Intubation times (12 ± 7 hours vs 14 ±9 hours), 24 hours bleeding (394 ± 219 mL vs 465 ± 318), reintervention for hemostasis (4% vs

2%), rhythmic complications (14% vs 14%), and mortality at 1 month (2% vs 2%) were comparable

in MS and TS. In aortic valve surgery, ministernotomy is technically more demanding and needs

more time. It is as safe and as effective as conventional sternotomy but its eventual benefits,

excepting upon cosmesis, are still to be defined. (J Card Surg 2003;18:396-401)

Minimally invasive cardiac surgery has evolvedin these past four years to the point that it can of-fer benefits over conventional procedures. Differ-ent strategies were adopted with variable resultsto determine the effects in terms of lesseningthe invasiveness. Minimally invasive approachesare based upon two principles, which are to avoidcardiopulmonary bypass and its objective prob-lems,1or to reduce the incision and to preserveas far as possible the bony structures.2,3 Minister-notomy (MS) is a well-known alternative to totalsternotomy (TS) in aortic valve surgery.4,5 Many

Address for correspondence: Prof. Olivier Jegaden, M.D.,Hopital Louis Pradel, BP Lyon-Monchat, 69394 Lyon cedex03 France. Fax: +33 4 75357532; e-mail: ojegaden@compuserve.com

different techniques via partial sternotomy havebeen described.6-8 Some authors have also pro-moted transthoracic approaches, generally usinga right parasternal procedure.9,10 However, de-spite worldwide enthusiasm for minimally inva-sive approaches, one should take into account thefact that valvular surgery via total sternotomy hashad long-term success with decreasing morbidityand mortality and does not suffer from a learningcurve.2 Thus, a minimally invasive approach mustbe at least as safe as the gold standard in aorticvalve surgery.

We present the results of a prospective studyby comparing both total sternotomy (TS) and min-isternotomy (MS) through reverse T incision ap-proaches in aortic valve surgery.

J CARD SURG FARHAT, ET AL. 3972003;18:396-401 MINISTERNOTOMY FOR AVR

MATERIALS AND METHODS

From January to December 2000, all the pa-tients undergoing first-time, elective, isolatedaortic valve replacement (AVR) were analyzedprospectively. Fifty patients underwent AVR viaTS while 50 others underwent the procedure withMS. Two senior surgeons performed one elec-tive technique each. Age, valvular diseases, LVEF,mean pulmonary arterial pressure, cardiac index,and NYHA class were comparable between boththe groups (Table 1); none of the patients had mi-tral or coronary disease. Anesthetic protocol wascomparable in all patients. All cardiopulmonarybypass (CPB) procedures were led under mild(35◦C) hypothermia. Postoperative pain manage-ment consisted of intravenous propacetamol andmorphine (patient-controlled analgesic or system-atic). Patients received continuous unfractionedheparin postoperatively until epicardial electrodesand drains removal, followed by oral warfarin. Allpatients had 1-month follow-up after surgery.

OPERATIVE TECHNIQUE

In the MS group, the patients were placedin a supine position. External defibrillating padswere placed behind right scapula and in apicalposition. A 5-cm midline skin incision was madestarting 3 cm from the sternal notch. The ster-num was cut vertically from the manubrium tothe second or third intercostal space using an os-cillating saw. Another sternal incision was madehorizontally to form a reversed T, and the inter-

TABLE 1

Preoperative Patient Characteristics

Characteristics MS Group TS Group

Age (range) 63 ± 14 (30–87) 67 ± 12 (24–89)Gender 30M/20F 34M/16FAI 10 11AS 37 33AI+AS 3 6LVEF (%) 58 ± 12 57 ± 12MPAP (mmHg) 25 ± 12 22 ± 8Cardiac index (l/min/m2) 2.5 ± 0.8 2.7 ± 0.7NYHA class 2.7 ± 0.5 2.8 ± 0.6CSS 0.9 ± 0.7 1.2 ± 0.5

MS = ministernotomy; TS = total sternotomy; M = male;F = female; AI = aortic insufficiency; AS = aortic stenosis;LVEF = left ventricular ejection fraction; MPAP = meanpulmonary arterial pressure; NYHA = New York HeartAssociation; CCS = Canadian cardiac score.

nal thoracic arteries were preserved. The level ofthe used ntercostal space was decided preoper-atively from the view of the profile chest-X ray.A children’s Finochietto retractor was inserted toexpose the upper and anterior mediastinal space.Thymic remnant and fat were excised to improveexposure. The pericardium was opened in its up-per portion. At this point, the ascending aorta andthe right atrial appendage were in sight. Pericar-dial edges were sutured to the skin and the re-tractor repositioned. Thus, pericardium protectedthe bony structures and elevated the heart intothe operative field. In the first 34 patients, centralvenous cannulation using a double-stage cannulawas made through the right atrial appendage. Forthe following cases, a 25-Fr double-stage venouscannula (Heartport®, Inc., Pinewood, CA) was in-serted through the right groin using the Seldingertechnique under transesophageal echocardiogra-phy (TEE) control into the right atrium. It was con-nected to a centrifugal pump to improve venousdrainage (Biomedicus®, Medtronic, Grand Rapid,IL). Arterial cannulation was performed into theascending aorta using a disposable 21-Fr DLPTM

cannula and an aortic needle was placed for car-dioplegia and venting. No left ventricular ventingwas used. Antegrade cold crystalloid cardioplegiawas given directly into the ascending aorta, or incase of aortic insufficiency it was directly deliv-ered through the coronary ostia. Proper de-airingwas done and monitored by TEE via the aorticroot by means of an aortic venting needle, withthe patient placed in Trendelenburg position. Aftersurgical hemostasis and before protamin admin-istration, the venous cannula was removed and aU-stitch secured the groin hemostasis.

In the TS group, a two-stage venous cannulawas placed into the right atrium through theatrial appendage. A left ventricular venting wasinserted through the ventricle’s apex. Aortic can-nulation, cardioplegia delivery, and aortic ventingwere done as in the MS group. TEE was not sys-tematically used. Prothesis implantation, tempo-rary stimulation, drainage, and sternum closurewere done in the same manner in both groups.

STATISTICAL ANALYSIS

The two groups were compared with Stu-dent’s t-test for unpaired data, chi-square test, orFisher’s Exact Test when appropriate. A p valueless than 0.05 was considered significant.

398 FARHAT, ET AL. J CARD SURGMINISTERNOTOMY FOR AVR 2003;18:396-401

RESULTS

No stentless valves or homografts were usedfor any patient. Prosthetic diameters were com-parable in both groups (22.3 ± 2.3 vs 22.9 ±2.3, p = ns). Mechanical versus bioprosthesesrepartition was comparable in both groups. Cross-clamping and CPB times were significantly higherin the MS group (p < 0.01, Table 2). Operativetimes were comparable between MS and TS (p= ns). No patient scheduled for MS was con-verted to TS. No preoperative death occurred.There was one postoperative death (1 month) ineach group. In the MS group, one 87-year-old mandied on postoperative day (POD) 10 from suddendeath. This patient had AVR for AS. Preoperativedata showed NYHA class 3, a low cardiac index(1.5 l/min/m2), and a mean pulmonary arterialpressure of 40 mmHg. The postoperative coursewas marked by atrial fibrillation and a long intu-bation (24 hours), and intensive care unit (ICU)stay (5 days). Postoperative echography showeda perfect prosthetic function, no atrial thrombus,and no pericardial effusion. In the TS group, one89-year-old man died on POD 26 from severepneumopathy. Preoperative data revealed NYHAclass 3 with good LVEF. The patient was intubatedfor 24 hours postoperatively. Intubation times andpostoperative 24 hours bleeding were compara-ble in both groups. Mean ICU stay was lower inMS group (p < 0.05) but discharge from hospi-tal was approximately done on POD 10, whichwas comparable in both groups (Table 3). In theMS group, 22% of the patients went home atdischarge instead of going to a rehabilitation cen-ter, against 12% in TS group (p < 0.05), whichmeans a better postoperative recovery after min-isternotomy approach.

TABLE 2

Operative Data in Ministernotomy (MS) and

Total Sternotomy (TS) Groups

MS (n = 50) TS (n = 50) P

Cross clamp time (min) 66 ± 15 49 ± 10 <0.01CPB time (min) 89 ± 18 70 ± 11 <0.01Intervention time 2.8 ± 0.4 2.7 ± 0.4 NsMechanical valve 39 37 –Bioprostheses 11 13 –Diameter (mm) 22.3 ± 2.3 22.9 ± 2.3 Ns

(19–29) (19–29)

CPB = cardiopulmonary bypass.

TABLE 3

Postoperative Outcome in Ministernotomy

(MS) and Total Sternotomy (TS) Groups

MS TS(n = 50) (n = 50) P

Intubation time (hours) 12 ± 7 14 ± 9 NsPostoperative bleeding 394 ± 219 465 ± 318 Ns

24hrs (ml)TransfusionICU stay (days) 1.7 ± 1.6 2.6 ± 6 <0.05Hospital stay (days) 9.4 ± 3.8 10.0 ± 3.1 NsPostoperative destination

Other medical department 16 (32%) 18 (36%)Postoperative reeducation 22 (44%) 25 (50%)Home 11 (22%) 6 (12%)

ICU = intensive care unit.

Major postoperative complications are repre-sented in Table 4. There was no statistical dif-ference in postoperative rhythmic complicationsbetween the two groups (8 in MS, 7 in TS).One patient in each group suffered mediastini-tis treated by debridement and irrigation drainagein association with 6 weeks of intravenous an-tibiotherapy with a good outcome. Except thesetwo cases, there was no sternal instability in bothgroups.

In the MS group, one patient had a minorcerebrovascular accident without sequelar deficit

TABLE 4

Postoperative Mortality and Major Morbidity

in Ministernotomy (MS) and Total

Sternotomy (TS) Groups

MS TS(n = 50) (n = 50)

Mortality at one month 1 1Morbidity

Atrial fibrillation 2 2Other rhythmic complications 4 4Atrioventricular block (PM) 2 1Infarct 0 1Mediastinitis 1 1Cerebrovascular accident 1 0Pulmonary infection 0 1Low cardiac output 2 0Reintervention 0 1Reexploration for bleeding 2 1Postoperative percutaneous drainage 1 0

CRMM 15 13% of total 30 26

PM = pacemaker; CRMM = cumulated rate of mortality andmorbidity.

J CARD SURG FARHAT, ET AL. 3992003;18:396-401 MINISTERNOTOMY FOR AVR

(mental confusion that cleared before discharge).Two others had low cardiac output in postop-erative needing inotropic support. Both were inNYHA class 3 and had good LVEF (>65%). Twopatients had revision for bleeding. Mediastinal ex-ploration revealed important clotting without iden-tifiable surgical bleeding. One patient requiredreadmission on POD 15 for pericardial effusionwith right atrial compression due to an overdoseof oral anticoagulation. He required percutaneousneedle drainage without further complication.

In the TS group, one patient with chronic ob-structive pneumopathy had pulmonary infectiontreated with antibiotics for 15 days. One 75-year-old man with important annular calcifications andlow LVEF had re-operation on POD 4 for paravalvu-lar leakage that was not revealed by preoperativeechocardiography. One patient underwent revi-sion for bleeding, without any identifiable surgicalcause. Another patient had a postoperative infarctin the posterior territory (postoperative troponin =25 mg/mL), with favorable evolution.

Cumulated rate of mortality and morbidity wasnot reduced in the MS group in comparison withthe TS group. The majority of the complications inthe MS group occurred in the first 30 patients (14vs 2). In TS patients, complications were encoun-tered all over the study. The complications werenot related to the preoperative condition (LVEF orNYHA class). No specific study was released forpain after comparing MS and TS.

DISCUSSION

In the past five years, minimally invasive cardiacsurgery has evolved to the point that it could offerbenefits over conventional procedures. Two ma-jor strategies were adopted to determine the less-ening of invasiveness, which are avoiding CPB orreducing the incision. The second option was mo-tivated by the fact that the progress in surgicaltechniques in all fields, and especially in video as-sistance, rendered the cardiac patient more de-manding of reduction in the “physical aggres-sion” he had to face.2,11 Different incisions weredescribed to approach the aortic valve: Transversesternotomy or “Z” sternotomy, superior hemis-terotomy, inverted T, reversed L or J, and paraster-nal. After an initial experience, the parasternal ap-proach popularized by Cosgrove et al.9 was aban-doned because of cartilage complications (pseu-darthrosis and lung hernia), esthetic disagree-

ments (paramedian skin incision), and the poten-tial complications from femoral arterial cannula-tion.9,12 Transverse sternotomy or “Z” were notideal because they required division of both inter-nal thoracic arteries.7

This study presents the results of a prospec-tive nonrandomized comparison between minis-ternotomy (MS) and total sternotomy (TS) in aorticvalve replacement. All the procedures were per-formed by two surgeons, but each one alwaysrealized the same procedure (MS or TS). MS wasperformed via inverted T incision. This was pre-ferred to reversed L because it seemed to al-low a better exposition when passing throughthe second intercostal space, with lower risk tothe internal thoracic arteries. There still are somelimitations in this study, even when it was doneprospectively: First, patients were not random-ized, which could introduce the question of sys-tematic bias; second, a standard technique and anew one with a subsequent learning curve werecompared. Finally, although anesthetic protocolsand postoperative care are standardized, the dura-tion of endotracheal intubation, ICU stay, or trans-fusion criteria are ultimately the decision of theanesthesiologist.

Aortic valve replacement through a partial up-per sternotomy is a safe procedure and it seemsas efficacious as total sternotomy. We did not en-counter any difficulty for aortic cusps resection,aortic suture, and temporary pacemaker wire im-plantation on the anterior wall of the right ven-tricle because of the small access. A word ofcaution should be added for patients with rightventricular dysfunction: these patients could ben-efit from topical cooling in association with coldcrystalloid cardioplegia and partial sternotomy lim-its the amount of topical ice application. Smallersternotomy does not limit exposure to the aor-tic root, but it is inconvenient because it hidesthe pulmonary veins and a large part of the leftventricle, rendering a left venting even risky orimpossible. Some authors reported conversionsto total sternotomy for left atrial lesions espe-cially after pulmonary vein venting. For this rea-son, we decided from the beginning of the se-ries not to vent the left cavities and we think thata simple aortic needle was sufficient for properde-airing. Left cavities control was always madewith the help of transesophageal echocardiogra-phy (TEE).13 However, three postoperative eventscould be related to air embolism: One patient

400 FARHAT, ET AL. J CARD SURGMINISTERNOTOMY FOR AVR 2003;18:396-401

suffered postoperative mental confusion, brainCT scan showed no abnormality, and clinical re-covery was obtained before ward discharge; andtwo patients had postoperative low cardiac outputneeding postoperative inotropic support. Both pa-tients had good preoperative LVEF, and presentedpostoperatively with temporary right ventriculardysfunction. Air embolism into the right coronaryartery could be incriminated even if none of thesetwo patients had significant ascension of postop-erative troponin level. Because left ventricle ac-cess and shaking are limited, secondary mobiliza-tion of the air trapped in LV cavities is possibledespite favorable TEE images, and this remains aconcern. One postoperative infarct was encoun-tered in the TS group, but no angiographic controlwas demanded.

In our series, cross-clamping and CPB timeswere longer in MS compared to TS. By contrast,operative times were comparable in both groups,due to a shorter sternum opening and closure,and surgical hemostasis. Liu et al. have reportedcomparable cross-clamping, CPB, and operativetimes in both groups,14 but the partial sternotomythat was performed passed through the fourthintercostal space or was a reversed “C.” We thinkthat sternum stability is not secured when onlyone intercostal space is left untouched. In thiscondition, a mini-incision only brings the limitationof exposure and does not seem to fit the idea ofreducing trauma. None of our MS patients hadsternal instability postoperatively.

In the last 16 patients we used a long femoralcannula that was placed into the right atriumunder TEE control and was inserted with theSeldinger technique. That was supposed to clearthe operative field and allow a better view ofthe aorta. Groin venous cannulation allowed asmaller incision and later on the use of thesecond intercostal space was more systematic.We did not a encounter venous problem dueto peripheral cannulation, such as phlebitis orlymphocoele.

Although, the smaller incision seems to be ap-pealing cosmetically, this study did not providegood evidence of the superiority of MS for thepostoperative course by reducing morbidity ormortality. We did not find any statistical differ-ence in postoperative arrhythmia between bothgroups. Asher et al.15 reported 26.3% versus38.0% of atrial fibrillation in MS and TS in a ret-rospective analysis. Postoperative rhythmic prob-

lems are often related to the preoperative sta-tus and to the pericardial irritation more than thelength of the sternal incision. Risk of revision forbleeding was comparable in both groups. Bleed-ing in the first 24 hours was not reduced in the MSgroup as previously reported. Transfusion statis-tics were not analyzed separately because the cri-teria of transfusion depend upon the anesthesiol-ogists in our institution and are not totally stan-dardized.

Most of the authors agree that the minimally in-vasive approach to aortic valve replacement doesnot reduce the risk of infection,16 especially inthe case of the division of both internal thoracicarteries. Nevertheless, sternal stability after re-intervention for mediastinitis is probably easier toobtain when part of the sternum has not beendivided as in a J incision.

Concern for cost was another major reason forpromoting minimally invasive aortic valve surgery,based first upon the reduction of postoperativemorbidity and the subsequent length of hospi-tal stay. Cohn et al. have reported that minimallyinvasive aortic valve replacement could be lessexpensive than standard surgery when done fre-quently.17 We did not study this factor in our se-ries. Both groups had comparable postoperativemorbidity and hospital stay, so we can hypoth-esize that the general cost was approximatelycomparable. In our institution, hospital dischargeis made after intravenous-oral anticoagulant treat-ment relay. Gundry et al. have performed minister-notomy with a mean postoperative length of stayof 2.6 days without early re-admission.18 Manyof their patients were children and had fast re-habilitation. Fast tract in minimally invasive aorticvalve surgery is possible since patient selectionand treatment observance are severe and wellplanned before intervention.

CONCLUSION

In aortic valve surgery, ministernotomy is tech-nically more demanding, needs more cross-clamping and CPB times, and does not seem toreduce postoperative bleeding or infection risk.It is as safe and effective as conventional ster-notomy but its eventual benefits, excepting forcosmesis, are still to be defined. Nevertheless,it is the technique of reference in our institutionfor standard aortic valve replacement. Thus, totalsternotomy is to be reserved for high risk patients,

J CARD SURG FARHAT, ET AL. 4012003;18:396-401 MINISTERNOTOMY FOR AVR

such as left ventricular dysfunction or NYHA class4 when esthetics is not the concern of the surgicalprocedure.

REFERENCES

1. Wouters R, Wellens F, Vanermen H, et al: Stern-itis and mediastinitis after coronary artery bypassgrafting. Analysis of risk factors. Tex Heart Inst J1994;21(3):183-188.

2. Lytle B: Minimally invasive cardiac surgery. J Tho-rac Cardiovasc Surg 1996;111:554-555.

3. Mack M, Landreneau R: Minimally invasive cardiacsurgery. Semin Laparosc Surg 1996;3(4):259-267.

4. Benetti FJ, Mariani MA, Rizzardi JL, et al: Minimallyinvasive aortic valve replacement. J Thorac Cardio-vasc Surg 1997;113(4):806-807.

5. Tam RK, Garlick RB, Almeida AA: Minimally inva-sive redo aortic valve replacement. J Thorac Car-diovasc Surg 1997;114(4):682-683.

6. Gundry SR, Shattuck OH, Razzouk AJ, et al: Facileminimally invasive cardiac surgery via minister-notomy. Ann Thorac Surg 1998;65(4):1100-1104.

7. Nair RU, Sharpe DA: Minimally invasive reversed Zsternotomy for aortic valve replacement. Ann Tho-rac Surg 1998;65(4):1165-1166.

8. Izzat MB, Yim AP, El-Zufari, et al: Upper T mini-sternotomy for aortic valve operations. Chest1998;114(1):291-294.

9. Cosgrove DM III, Sabik JF: Minimally invasive ap-proach for aortic valve operations. Ann Thorac Surg1996;62(2):596-597.

10. Minale C, Reifschneider HJ, Schmitz E, et al: Min-imally invasive aortic valve replacement without

sternotomy. Experience with the first 50 cases.Eur J Cardiothorac Surg 1998;14(Suppl I):S126-S129.

11. Ehrlich W, Skwara W, Klovekorn W, et al: Do pa-tients want minimally invasive aortic valve replace-ment? Eur J Cardiothorac Surg 2000;17(6):714-717.

12. Cuenca J, Rodriguez-Delgadillo MA, et al: Is thefemoral cannulation for minimally invasive aorticvalve replacement necessary? Eur J CardiothoracSurg 1998;14(Suppl I):S111-S114.

13. Secknus MA, Asher CR, Scalia GM, et al: Intraop-erative transesophageal echocardiography in min-imally invasive cardiac valve surgery. J Am SocEchocardiogr 1999;12(4):231-236.

14. Liu J, Sidiropoulos A, Konertz W: Minimally inva-sive aortic valve replacement (AVR) compared tostandard AVR. Eur J Cardiothorac Surg 1999;(16Suppl II): S80-S83.

15. Asher CR, Dimengo JM, Arheart KJ, et al: Atrial fib-rillation early postoperatively following minimallyinvasive cardiac valvular surgery. Am J Cardiol1999;84:744-747.

16. Massetti M, Babatasi G, Neri E, et al: Minimallyinvasive, but too many infections. Ann Thorac Surg2000;69(3):977-978.

17. Cohn LH, Adams DH, Couper GS, et al: Minimallyinvasive cardiac valve surgery improves patient sat-isfaction while reducing costs of cardiac valve re-placement and repair. Ann Surg 1997;226(4):421-6;discussion 427-428.

18. Gundry SR: Aortic valve surgery via limited inci-sions. In: Oz MC, Goldstein DJ, eds. ContemporaryCardiology: Minimally Invasive Cardiac Surgery.Totawa, NJ: Humana Press, 1999, pp. 205-214.