The Incidence ofComplications in Endoscopic Anterior

Reprinted from SPINE, Vol. 20,No. 14, July 15, 1995Copyright © 1995, SPINE J, B.Lippincott Company

Printed in the U.S.A.

The Incidence of Complications inEndoscopic Anterior ThoracolumbarSpinal Reconstructive SurgeryAProspective Multicenter Study Comprising theFirst 100 Consecutive Cases

Paul C. McAfee, MD. John R. Regan, MD, Thomas Zdeblick, MD,James Zuckerman, MD, George D. Picetti, III, MD, Stephen Heim, MDW. Peter Geis, MD, and Ira L. Redder, MD

SPINE Volume20, Number 14, pp 1624-163201995, Lippincott-Raven Publishers

The Incidence of Complications inEndoscopic Anterior ThoracolumbarSpinal Reconstructive SurgeryA Prospective Multicenter Study Comprising theFirst 100 Consecutive Cases

Paul C. McAfee, MD,* John R. Regan, MD,t Thomas Zdeblick, MD,tJames Zuckerman, MD,S George D. Picetti, III, MD,|| Stephen Helm, MD,1IW. Peter Gels, MD,* and Ira L Redder, MD*

Study Design. A prospective multicenter study on100 consecutive surgical procedures.

Objectives. A prospective multicenter study was performed to evaluate the early perioperative complications in 100 endoscopic spinal procedures—78 video-assisted thoracic surgical procedures and 22laparoscopic lumbar instrumentation and fusion procedures.

Summary of Background Data. Endoscopic procedures have been widely applied in general surgery forappendectomy, cholecystectomy, liver resection, Nissenfundoplication, colon resection, and hernia repairs. Video-assisted thoracic surgery is widely used for pleuralbiopsy, lung resection, and sympathectomy. This is thefirst large series to date investigating the safety andpotential complications using endoscopic surgery foranterior decompression or fusion of the thoracolumbarspine.

Methods. Video-assisted thoracic surgical procedures included multilevel anterior thoracic releases fordeformity, 27 patients: anterior thoracic discectomieswith spinal canal decompression, 41 patients; pyogenicvertebral osteomyelitis decompression, 2 patients; andvertebral corpectomy for neurologic decompression, 8patients. Mean operative time was 2 hours, 34 minutes(range, 45 minutes to 6 hours), and mean length of staywas 4.97 days (range, 2-21 days).

Anterior laparoscopic interbody stabilization and fusion at L4-5 or L5-S1 was performed in 22 patients.The mean operative time was 4 hours, 17 minutes(range, 2 hours, 40 minutes to 9 hours), and the meanlength of stay was 5.6 days (range, 1-23 days).

From *1116 Scoliosis and Spine Center, Baltimore, Maryland, fTheTexas Back Institute, Dallas, Texas, ^University of Wisconsin Hospital, Division of Orthopaedic Surgery, Madison, Wisconsin, ||The Per-manente Med. Group, Inc., Sacramento, California, ^OrthopaedicAssociates of Dupage, Ltd., Carol Stream, Illinois, $St. Mary's Hospital, San Francisco, California.Presented at the 29th Annual Meeting of the Scoliosis Research Society, Portland, Oregon, September 21—24, 1995.Accepted for publication February 1, 1995.Device status category: 5.

1624

Results. The most common video-assisted thoracicsurgical complications were transient intercostal neuralgia (six patients) and atelectasis (five patients). Themost common laparoscopic complication was bonegraft donor site infection (two patients). There were twoendoscopic cases that were converted to open procedures, one for extensive pleural adhesions and one fora common iliac vein laceration.

Conclusions. The endoscopic spinai approachesproved to be safe operative procedures in 100 consecutive cases. There were no permanent iatrogenic neurologic injuries and no deep spinal infections. [Keywords: endoscopy, laparoscopy, thoracoscopy, VATS]Spine 199S;20:1624-1632

The goal of this preliminary study is to analyze theperioperative early complicationsof endoscopicsurgeryto evaluate the feasibility of VATS '̂̂ '̂̂ ^ and laparoscopic surgical approaches.'®'̂ '̂̂ ^ It is not the intent orgoal of this study to evaluate the incidence of obtainingan endoscopic spinal fusionor to evaluatepostoperativeinstability because that requires 2 years or more follow-up time. The present study is an article concerningcomplications,not an outcome study of an intervention.

• Materials and Methods

This prospective multicenter study performed from 1990 to1994 evaluates the early perioperative complications in thefirst 100 consecutive patients undergoing endoscopic spinalreconstructive procedures—78 video-assisted thoracic surgical(VATS) procedures from T3 to T12 and 22 laparoscopiclumbar instrumentationand fusion procedures.TTie VA*I3 procedures were performed at one of three institutions—^The Scoliosis and Spine Center of Baltimore, Maryland, 20 patients;The Department of Orthopaedic Surgery, Kaiser Permanente,Sacramento, Califomia, 20 patients; and The Texas Back Institute, Dallas, Texas, 38 patients. There were 22 laparo-scopic-assisted lumbar fusion instrumentation proceduresperformed at one of four institutions—Scoliosis and SpineCenter,

One Hundred Endoscopic Anterior Spinal Procedures • McAfee et al 1625

Table 1. Endoscopic Spinal Reconstructive Procedures

Procedure

Thoracic discectomiesMultilevel anterior discectomy for correction of scoliosisAnterior release for kyphosisExcision of hemivertebraeVertebral corpectomy for neurologic decompressionPyrogenic vertebral osteomyelitis decompressionsVATS total

Anterior laparoscopic Interbody stabilization and fusionprocedures—L4-5 or L5-S1

Total

VATS = video-assisted thoracic surgery.

Number of

Patients

Baltimore, Maryland; Spine Center, SanFrancisco, California;Departmentof Onhopaedic Surgery, University of Wisconsin,Madison,Wisconsin; and Mona Kea Medical Park, Chicago,Illinois.

The distribution of endoscopic procedures is listed in Table1. The most common surgical procedure was a VATS forthoracic discectomy and spinal canal decompression, 41 patients. Therewere threemajortypes of anteriorprocedures forrelease of deformities: anterior multiple level discectomies(usually six levels), 20 patients; anterior release for kyphosis,four patients; and anterior resection of a thoracic hemivertebrae, three patients. There were eight patients with fracturesor tumors causing incomplete neurologic deficit who underwent anterior thoracic corpectomy and spinal cord decompression (Figure 1). Lastly, two patients underwent anteriordecompression of pyogenic vertebral osteomyelitis.

The 22 laparoscopic procedures reported in thisstudycomprised the first cases enlisted in the laparoscopic Bagby andKuslich (BAK) protocol for theInvestigational Device Exemp

Blunt Nerve Hook

Rlght T7Nerve Root

m Remnant of^^T7Vert. Body

Base of Right17 Pedicle

B

tion (IDE Study) filed with the United States Food and DrugAdministration.^'

To date, the BAK device is an investigational device and isapproved for IDE in several medical centers for anterior laparoscopic insertion into the L4—5 and L5-S1 intervertebraldisc spaces. There is a very strict protocol with specific indications, namely degenerative disc disease with neuroforaminalnarrowing and radicuiopathy. Radiographic criteria are concordant pain reproduction on discography and an abnormalmagnetic resonance imaging. For inclusion into the BAK laparoscopic study, the patient needs to fit a particular psychologic profile, mustnot havehad a previously attemptedlumbarfusion, must not be a cigarette smoker, and needs to haveinternal disc disruption and collapse at either L4—5 or L5-S1.

The range in ages for the major groups of thoracoscopicand laparoscopic procedures followed the same trends as fortheir corresponding open procedures. The patients undergoinganterior release for scoliosis or kyphosis were the youngest,mean = 19.0 years (range, 2—44 years). The thoracoscopicdiscectomy patients were the next to youngest age group,mean = 41.4 years (range, 22-84 years). The patients undergoing corpectomy were older because six were performed formetastatic lesions causing incomplete neurologic deficits,mean = 51.9 years (range, 28-84 years). The patients undergoinglaparoscopic lumbar procedures had a meanageof 40.6years (range, 28-84 years).

In general, the indications for anterior thoracoscopic orlaparoscopic surgery were identical to the analogous openthoracotomy or iaparotomy spinal procedure. The only relativecontraindications for anterior endoscopic surgery are multiple previous surgical procedures or empyema causing suchextensive adhesions that portal placement cannot be accomplished with good visualization. Additionally, anterior instrumentation for thoracic stabilization hasnot been developed asyet for endoscopic applications. Cases with fractures or tu-

r. -

Figure 1. Aright-sided VATS T7 corpectomy is shown. (A) The right 17 nerve root and the anterior aspect of the thecal sac are visualizedafter 77 VATS corpectomy. Cephalad is to the right and caudad is to the left. (B) The corresponding intraoperative view through a 30®thoracoscope is pictured. The anterior aspect of the spinal canal is well visualized as is an epidural plexus of veins anterior to the dura!sac.

1626 Spine • Volume 20 • Number 14 • 1995

Table 2. Perioperative Data

Procedure Type (number of patients)

Thoracoscopic (78)Mean operative time; 2 hours, 34 minutes (range, 45 minutes-6 hours)Average chest tube duration: 1.44 days (range, 0-3 days)Length ofstay: 4.97 days (range, 3-21 days)ICU: 41 patients overnight in a combined RR/ICU

Laparoscopic (22)Length of stay: 5.6 days normalized to4 days (range, 1-23 days)Intraoperative blood loss 194 cc (range, 50 cc to 800 cc): no transfu

sions neededMean operative time: 4 hours, 17 minutes (range, 2 hours 40 min-

utes-9 hours)

ICU = intensive care unit. RR/ICU = recovery room/intensive care unit.

mors with 3-column spinal instability were notdecompressedendoscopically. In our institutions, we have virtually supplanted thoracotomies and laparotomies with endoscopic procedures aside from cases requiring anterior instrumentationcases, such as anterior Texas Scottish Rite instrumentation forthoracolumbar scoliosis.

• Results

Video-Assisted Thorecic SurgeryTheplanned procedure wasaccomplished in all but onepatient, who required conversion to an open procedurebecause of scarring from a previous costotransversec-tomy (Table 2). The mean operative time was 2 hours,34 minutes, (range, 45 minutes to 6 hours). Theaveragechest tube duration was 1.44 days (range, 0-3 days).Forty-oneof 78 (52.5%) patients were monitored in anintensive care unit (ICU) setting for the first night aftersurgery. There was an obvious institutional bias in thisfactor because in Baltimore the recovery room and theintensive care unitarea combined postoperative facility.When this factor is discounted, the normalized incidenceofpostoperative ICU use was 19of58 (32.8%) patientsundergoing VATS. The mean postoperative length ofstay was 4.97 days (range, 2-21 days).

Laparescopic Lumbar Interbody Fusionand Stabiiaation

The mean operative time was 4 hours, 17 minutes(range, 2 hours 40 minutes to 9 hours). The mean lengthof staywas 5.6 days (range, 1-23 days).

intraoperative Estimated Blood LossThe operative estimated blood loss was specific to thetype of spinal procedure listed in Table 3. The anteriorthoracic releases for deformity had a mean intraoperative blood loss of 94 cc (range, negligible to 300 cc).Thoracic discectomy and spinal canal decompressionhad a mean estimated blood loss of 400 cc (range,25—2500 cc). The most extensive procedures were anterior thoracic corpectomies with bleeding from rawcancellous bone surfaces and epidural vessels, and themean estimated blood loss was 1175 cc (range, 250-2800 cc; Figure 2).

The lumbar interbody BAK and fusion procedurenever required a transfusion, even in a patient with anintraoperative common iliacveininjury. Aside from thispatient, the majorityof the operativeblood lossresultedfrom the harvesting of bone graft from the iliac crestdonor site rather than the spinal procedure. The meanlaparoscopic estimated blood loss was 194 cc (range,50—800 cc).

• Complications

Vidoo-Assisted ThoracicSurgeryThe major complicationsare listedin Table 4. The mostcommon complication was six cases of postoperativeintercostal neuralgia. The etiology was thought to be acombination of factors—electrocauterization of thehead of rib before excision, the use of rigid 10-mmthoracoports rather than flexible intercostal portals, orcompression of a spinal nerve with a Kerrison rongeur(Rock Surgical, Baltimore, MD). In six of the patients,thepainandparesthesias were transient andresolved by6 weeks after surgery.

There were five patients with postoperative ateleaa-sis significant enough to prolong the patient's hospital-ization, and one patient had a loculated pleural effusionthat resolved by 1 month after surgery.

One patient had a thoracoport penetrate the elevatedright hemidiaphragm at the beginning of the operativeprocedure (Figure 3). This was an 84-year-old womanwith a previous empyema, causing extensive pleural adhesions. The 10-mmhole in the diaphragm wassuccessfully repaired thoracoscopically with endoscopic staples, and there was no postoperative sequelae.

There were no permanent iatrogenic spinal neurologicinjuries. One patient undergoing a thoracicreleasefor scoliosis from T5 to TIG developed transient legweakness from occult spinal stenosis at T12-L1. Thisfully resolved by 6 weeks after surgery. To prevent thiscomplication, we no longer "jackknife" the patient foroperative positioning.

Table 3. Intraoperative Estimated Bleed Loss

Endoscopic ProcedureNumber of

PatientsMean

(cc)Range(cc)

Multiple anterior discectomies. 27 94 (negligible to 300)scoliosis or kyphosis release

Thoracic discectomy and spi 41 400 (25-2500)nal canal decompression

Thoracic corpectomy and spi 8 1175 (250-2800)nal cord decompression

Lumbar interbody BAK fusion 22 194 (50-800)procedure

Total 98*

* Note, the group with two patients with pyogenic vertebral osteomyelitiswas too small to analyze.BAK = Bagby and Kuslich.


Figure 2. This 38-vear-oid woman presented to us with a denseparaparesis secondary to a 112burst fracture instrumented elsewhere. The anterior posterior (A)and lateral (B) myelogram showhigh grade block of contrast. Thepatient underwent endoscopicT12 anterior corpectomy performed through four thoracicportals without need of COj insufflation. The mean estimatedblood loss of 1175 cc (range,250-2800 cc) in the eight corpectomy patients was the highest ofany of the four groups of anteriorendoscopic surgery. (C) The pre-operative axial computed tomography image at the T12 fractureis shown. The correspondingpostoperative computed tomography image (D) shows that thespinal canal has been well decompressed using endoscopictechniques.

Laparoscopic ComplicationsThere were no complications from pneumoperitoneumor CO2 insufflation used exclusively in laparoscopiccases (not used in thoracoscopy; Figure 4).

In 22 anterior lumbar laparoscopic BAK procedures,there was one with a vascular injury that occurred during the operative exposure of the L5-S1 intervertebraldisc space below the "crotch" of the aortic bifurcation.It was only the second BAK implantation done at thisparticular institution, and the complication occurredduring the use of a retractor. Pins were being embeddedinto the body of the L5 vertebrae to perform retractionof the posterior peritoneum and iliac vessels.

Unfortunately, during implantation of the retractor,a flange on the pin on the patient's left side impinged onthe left iliac vein, causing a small tear. Hemorrhage wascontrolled using direct pressure using a laparoscopicinstrument, and open conversion to a mini-laparotomywas performed. Simple repair of the tearwas performed

t "X

|VB

Table 4. Complications

Procedure Type

ThoracoscopicIntercostal neuralgia (all transient)AtelectasisExcessive epidural blood loss, more than 2500 ccConversion to open thoracotomy caused by previous cos-

totransversectomyPenetration of right hemidiaphragm from thoracoport in

patient with previous empyemaTransient paraparesis related to spinal stenosis at a dif

ferent vertebral level and operative positioningLaparoscopic

Conversion to open laparotomyfor repair of left commoniliac vein

Bone graft donor site infectionPostoperative upper gastrointestinal bleed in patient anti-

coagulated with Coumadin up to 2 weeks before sur-gery

No cases of permanent iatrogenic neurologic deficit.No cases of spinal wound infection.

Number ofPatients


ft.Collapsed |S^Righl Lungl-J;

-.Ij ,•. U DepressedInflated , Diaphraam

_ B

Figure 3. From a thoracoscopic portal In the sixth intercostal space, a hole In the right hemidiaphragm is visualized (A). The underlyingliver was not injured but is visible. The diaphragm was inadvertently entered by a blunt 10-mm thoracoport placed too interiorly inresection of a T11 tumor (B). The diaphragm was repaired thoracoscopically using a hernia stapler without long-term sequelae. To preventthis complication, the first portal should be placed in the midaxillary line In the sixth or seventh intercostal space. All subsequent portalsare placed under thoracoscopic visualization from within the chest cavity.

with minimal blood loss {800cc estimated blood loss forthe entire procedure). The procedure was completedthrough a 4-inch Pfannenstiel incision. Thepatient wasdischarged 6 days after surgery with no long-term sequelae.

^^Janetai^Peritoneum

MM

Small

intestine

Figure 4. An illustration is shown of a loop ofsmall bowel beingforced up a laparoscopic trochar. This is not a complication,provided the surgeon visualizes each trochar endoscopically before an Instrument is placed into the abdomen. Another importanttechnical rule is to always visualize the portals at the end of thelaparoscopic case after the CO2 pressure has been reduced.Venous hemorrhage can be tamponaded off by the intraoperativeCO2 pneumoperitoneum. Inspecting each portal site at the end ofthe laparoscopic procedure reduces thechances of postoperativeIntra-abdominal hemorrhage and hypotension in the recoveryroom.

• Discussion

Perhaps the most important factor to minimize complications in endoscopic spinal surgeryis to work with anexperienced general surgical laparoscopic or thoracoscopic specialist. '̂̂ '̂ '̂'® The majority of cases in thisstudy were performed at institutions with regionalteaching facilities for minimally invasive surgery. Thereare four factors that are different for endoscopic procedures compared with conventional "open" procedures,which help diminish complications.

Minimally invasive surgery or VATS procedureswithin the torso require special considerations to successfully and safely perform operative procedures ineither the thoracic or the abdominal cavity and in adjacent tissues. The first key element is the ability to visualize the operative site and the areas surrounding theoperative location. Visualization primarily requires aport of entry (either though the thoracic or abdominalwall), which requires a camera, a videoscope, and appropriate light source. The second important factor iscreation of a workspace to perform the operative procedure. The technique to produce a workspacewillvarydepending on whether the operative procedure is to beperformed in the thoracic cavity, the mediastinum, theabdominal cavity, or the retroperitoneum. The thirdfactor important to minimally invasiveprocedures is theuse of instrumentation that will allow the surgeon toaccomplishthe same technicalgoalsas are accomplishedin open surgery. The most important set of equipment inthe endoscopic room is the "emergency set" of equipment and retractors necessary if the procedure has to beconverted to an open case. A fourth, and final, issue isthe provision of anesthetic support to help to maintain a


•

. LUNG

Camera

Endoshears

Figure 5. Adhesions involving the chest wall (pleura! adhesions) (A) or the abdominal wall (peritoneal adhesions) (B) should be takendown using endoscopic techniques before placing trochars orportals in their vicinity. Endoshears are used to lyse adhesions tomobilizethe underlying lung parenchyma (video-assisted thoracic surgery) orbowel (laparoscopy) toallow correct placement oftrochars withoutdanger to underlying viscera (C). Triangulation is used to lyse adhesions overlying the T9-T10 levels.

steady statefor the surgeon during the operative procedure.

Methods to Avoid Compiications inSpinal TTioracoscopy

Complications resulting from anesthesia or one-lungventilation are covered in full-length endoscopic textbooks.Failure to collapse the lung on the operative side usually results from improper placement of the

i i

Figure 6. A schematic diagram illustrates the correct insertion ofthe Bagbyand Kuslich fusion cage at L5-S1. Noticethat the fusioncage can be inserted laparoscopically. The titanium cage distractsthe L5-S1 intervertebraldisc space, and the cancellous iliac bonegraft facilitates L5-S1 fusion.

double lumen or univent tube. The position of the tubeneeds to be rechecked with a bronchoscope after positioning the patient in the lateral decubitus position. It isnot uncommon for the tube to shift when the patient isbeing repositioned, so the ability to collapse the lungneeds to be shown before trochar incisions are made.

The sixth or seventh intercostal space in the midax-illary line is the safest place for the first thoracoscopicportal. This is the only portal that is made withoutthoracoscopic visualization inside the chest, and, insome respects, it is the most dangerous. With the lungcollapsed, the central tendon of the diaphragm can bemigrate as high as the eighth intercostal space. Althoughthe disposable thoracoscopic 12-mm portals are blunttipped, we had a case where the initial portal penetratedthe right hemidiaphragm. The underlying liver was exposed and visible thoracoscopically. Fortunately, therewas no visceral injury, and the diaphragm was successfully repaired using a laparoscopic hernia stapler. Thisunderscores the importance of making the first thoracoscopic portal in the sixth or seventh midaxillary line.The correct technique is to enter the chest using thesame gentle technique as with chest tube insertion—a


blunt Kelly rides over the cephalad portion of the rib,avoiding the neurovascular bundle. The most commoncomplication was six cases of intercostal neuralgia. Itwas probably attributable to rigid thoracic trochars.The largest diameter trochar used in the chest should be12 mm. There is only a fixed amount of intercostalspace. As the surgeon levers and torques large spineinstruments, it is possible to exertpressure on the intercostal nerves. Fortunately, in all of our cases, the intercostal neuralgia was transient and resolved spontaneously within 6 weeks. In thepastyear, wehave switchedexclusively to theEthicon flexible thoracic portals (Ethi-con Endosurgery, Cincinnati, OH) to minimize the incidence ofintercostal neuralgia. Another useful step is toavoid monopolar cautery at the inferior margin of therib when skeletonizing the head of the rib before removal. Electrocautery injuryto the intercostal nerve hasbeen observed by this mechanism.

Every timethe scope is placed down a portal, it mustbe donegently and slowly—sometimes this is difficult ina lengthy procedure when thescope is being repeatedlycleaned. TTie lung can be inadvertently reinflated, anddirect injury to the lung parenchyma has been observedin laboratory training sessions.

Once thefirst portal is made, the 10-mm, 30®-angledscope is inserted, and subsequent portals are made under thoracoscopic visualization inside and direct visionoutside the chest. It is important to take down all adhesions between the chest wall and the visceral pleurabefore inserting a portal in the area. The lung can betented andplastered to thechest wall (Figure 5).Usuallyendoshears can easily take downthe adhesions, but withextensive pleural scaring, the adhesions can be sweptaway by digital sweeping movements from the portalsites.

To avoid direct lung injury, all instrumentsneed to bevisualized, but in particular fan retractors need to bevisualized when they are opened and closed. A fan retractor should never be extracted from the chest in asemi-opened position because the lung can be pinchedwithin the fingers of the fan.

We do not use positive pressure CO2 insufflation tocollapse the lung in thoracoscopy. Krasna et aF foundthat pressure greater than 12 mm Hgis associated withmediastinal shift and rapid changes in cardiac output.Subcutaneous emphysema is more likely to occur whentrochars are dislodged or incompletely pulled out of thechest cavity. Gas embolism and subcutaneous or evenmediastinal emphysema have been reported—thesecomplications can be avoided simply by doing all thoracoscopic cases as a gas-less procedure.

Significant injuries to the aorta, superior vena cava,or pulmonary vessels will require immediate conversionto open thoracotomy for control. For this reason, allpatients areprepared for thispossibility. All experiencedendoscopic surgeons state that the most important surgical instruments in the thoracoscopic operating room

are those instruments that are required for open thoracotomy.

Intrathoracic breakage of instruments, such as graspers and pituitary rongeurs,can occur.This ismore commonthan withopenthoracotomies because theworkingend of the instruments are identical, but theendoscopicinstruments usually have 250—300-mm shaft lengths.Thisincreases the frictional forces along theshafts as thesliding surfaces become stuck with bone fragments.Manyof the disposable and semi-disposable endoscopicinstruments are simply not durable enough to resectbone or disc tissue. One of our investigators had threepituitary rongeurs break within the thoracic cavity. Fortunately, the retrieval of the broken hardware was ableto be accomplished endoscopically, and an open conversion was not required.

In a prospective study of 78 consecutive thoracoscopic cases, there were no infections, no permanentiatrogenic neurologic injuries, and no major vascularinjuries.There was one patient who required conversion to an open thoracotomy because she had denseadhesions resulting from a previous costotransversec-tomy.

Techniques to Avoid CompUcations inSpinal Laparoscopy

One advantage of the anesthetic considerations of laparoscopy compared with thoracoscopy is that routineendotracheal intubation without one lung ventilation ismore easily accomplished. There are still ventilatoryriskswith laparoscopic spineprocedures because of theCO2 pneumoperitoneum and the Trendelenburg position used in approaches to the lumbosacral junction.Any closed technique (as opposed to the Hasson^ oropen technique) of establishing pneumoperitoneum involves percutaneous placement of a needle within theperitoneal cavity for CO2 insufflation beforethe trocarsor cannulas are placed. To avoid complications, it isimportant to elevate the umbilicus with towel clips before insertion of the Veress needle. Analogous to thethoracoscopic sequence of portal placement, the first10-mm portal placed in the umbilicus is used for thelaparoscopic camera. All subsequent trocharplacementsare done under endoscopic visualization. It is recommended to aim the sharp introducing trochars in acaudaddirection toward the pelvis to reduce thechancesof hitting the underlying bowel or great vessels.

The Hasson or open cannula systemis usedfor directopen insertion of a blunt cannula. It is used in patientswith unsuccessful Veress needle insertion or multipleprevious laparotomies with abdominal adhesions. Analogous to thoracoscopic techniques, it is often useful totake down abdominal wall adhesions before insertingtrochars in the lower quadrants. We frequently use endoshears to mobilize the uterus if endometriosis is extensive. It is important to have an unrestricted access tothe L5—SI intervertebral disc space through a suprapu-

One HundredEndoscopic Anterior Spinal Procedures • McAfee et al 1631

bic portal, and to permit this, the uterine-abdominalwall adhesions should be released.

Toprevent hemorrhage after surgery, we recommendsupplementing endoscopic vascular clips with loop ligatures (Endoloops—Ethicon or Surgities—US Surgical,Norwalk, CT). It is helpful to decrease the insufflationpressure from 20 mm Hg to 10 mm Hg before the endof the procedure to ensure all venous bleeding vesselshave been coagulated.

As a general rule, it is important to ensure compatibility between the shafts ofthe endoscopic spinal instrumentsand the laparoscopictrocars and cannulas. Unlikethoracoscopic procedures, there needs to be an airtightseal throughout the procedure or pneumoperitoneumwill be lost, and visualization will be impaired. Theinsertion of osteotomes and sharp curettes needs to bedone carefully, as we have had an osteotomecut the sealon the laparoscopic cannula. Curved curettes, inparticular, need to be carefully withdrawn through the tro-chars because they can get stuck in the flapper valvemechanism. It is important before beginning the procedure to make sure all orthopedic spinal instruments arecompatible with the laparoscopic trochars because theyare usually manufactured by different companies.®'̂

There are several complications specific to the anterior laparoscopic approaches to the L5-S1 interverte-bral discs that are worth emphasizing (Figure 6). In malepatients, it is important to use bipolar electrocauterybelow the sacral promontory to avoid injury to thesympathetic plexus and retrograde ejaculation. Identifythe peristalsis in the ureter bilaterally and confirm theirlocation. We have seen anomalous ureters come verycloseto the junction of the common iliacvessels and theL5—SI intervertebral disc. The most treacherous complication is laceration of the common iliac vein whileattempting to expose the L5—SI intervertebral discspace. We have observed this in the teaching laboratoryand in one clinical case. Hemostasis was successfullyachieved clinically but required conversion to an openlaparotomy. There is more success inmobilizing the iliacvessels by ligating thesmall tethering branches andcompletely avoiding the tendency to stretch any veins regardless of their size. Orthopedic spinal implants areusually made for patients of average size. If a patientsimply has too little room for two fusion cages betweenthe iliac vessels, it is better to use just one fusion cage orjust one large bone graft dowel rather than to undulymanipulate the iliac veins or the inferior vena cava.

Comparison to Published LiteraturoThe landmark comparison of laparoscopic techniquesversus open laparotomy was the publicationof Reddickand Olsen's series of 100 patients undergoing cholecys-tectomy in 1989 (Table 5).'̂ * A huge array of generalsurgical procedures have been described laparoscopi-cally—appendeaomy,^^ liver resection, '̂ Nissen fun-doplication, splenectomy, selective vagotomy, nephrec-

Table 5. From Eddie Joe Reddick's Original Series 1989*

Cholecystectomy

Patient Characteristics Laparoscopic Minilaparotomy

Mean age (yr) 40 40Postoperative length of stay (days) 1.96 2.80Return to work (days) 6.5 34Operative time (min) SO 65

* From Reddick and Olsen [14|.

tomy, colon resection, and hernia repairs. Theexplosive enthusiasm and rapid technical advances withthe commercial outpouring of resources is attributableto one procedure—laparoscopic cholecystectomy.^ '̂̂ '*'16,18,19 -pjjg chief advantage oflaparoscopic cholecystec-tomycompared with the open procedure is a morerapidreturn to work of patients (Table 5).

A similar comparison of open versus endoscopic thoracic discectomies is not possible at the current time.The four largest published series of open anterior thoracic discectomies or costotransversectomies are—Bohlman and Zdeblick,^ 22 patients (1988); Otani etal,*^ 23 patients (1988); Albrand and Corkill,^ sevenpatients (1979); andSimpson et al,^^ 23patients (1993).In these combined series of 75 patients, there were twopatientswith transient iatrogenic paraparesis after openanterior thoracic discectomy and no reports of spinalinfections after surgery. None of these four studies ofopen thoracic discectomies reported the length of stay,the intraoperative blood loss, the mean operative time,or chest tube duration.

• Conclusions

The endoscopic spinal approaches proved to be safeoperative procedures in 100 consecutive cases; therewere no permanent iatrogenic neurologic injuries andno deep spinal infections. One VATS procedure wasconverted to an open thoracotomy becauseof extensivepleural adhesions from a previous costotransversec-tomy. One laparoscopic procedure was converted to amini-laparotomy because of a left common iliac veininjury. There were no long-term sequelae in these twoprocedures.

Acknowledgment

The authors thank Michael Mack, MD, Chief of Video-Assisted Thoracic Surgery in Dallas, Texas, for assistance. His abilities allowed for the favorable clinicaloutcomes of many patients illustrated in this study.

References

1. Alband OW, Corkill G. Thoracic disk hemiation. Treatment and prognosis. Spine 1979;4:41-6.2. Bagby G. Arthrodesis by the distraction-compression

methods using a stainless steelimplant. Orthopedics 1988;11:931-4.


3. Bohlman HH, Zdebiick TA. Anterior excision of hemi-ated thoracic discs.J BoneJoint Surg [Am] 1988;70:1038-47.

4. Brodke DJ, Zdebiick T, Kunz D, McCabe R. Biomechan-ical comparison of posterior lumbar interbody fusion including a new threaded titanium cage. Presented at the AnnualMeetingof The International Society for Studyof the LumbarSpine, Marseilles, France, June, 1993.5. Hasson HM. Modified instrument and method for lap-

aroscopy. Am J Obstet Gynecol 1971;110:886-7.6. Jacobs M, Verdeja JC, Goldstein HS.Minimally invasive

colon resection (laparoscopic colectomy). Surg Laparosc En-dosc 1991;1:144-50.7. Krasna MJ, Mack MJ. Adas of Thoracoscopic Surgery.

St. Louis: Quality Medical Publishing, 1994.8. McAfee PC, Bohlman HH, Yuan HA. Anterior decom

pression of traumatic thoracolumbar fractures with incomplete neurological deficit using a retroperitoneal approach. JBoneJoint Surg [Am] 1985;67:89-104.

9. McAfee PC, Zdebiick TA. Tumors of the thoracic andlumbar spine: Surgical treatment via the anterior approach. JSpinal Disord 1989;2:145-54.10. Obenchain TG,Cloyd D.Outpatient laparoscopic lumbardiskectomy: Description of technique and review of first twenty-one cases. Surgical Technology International 1994;2:415-8.

11. Obenchain TG. Laparoscopic lumbar diskectomy: Casereport. J Laparoendosc Surg 1991;1:145-9.12. Otani K, Yoshida M, Fujii E, Nakai S, Shibasaki K. Thoracic Diskhemiation. Surgical treatment in 23 patients. Spine1988;13:1262-7.

13. Pier A, Gotz F, BacherC. Laparoscopicappendectomy in625 cases: From innovation to routine. SurgicalLaparoscopyand Endoscopy 1991;1:8-13.14. Reddick EJ, Olsen DE. Laparoscopic laser cholecystec-tomy, A comparison with mini-lap cholecystectomy. Surg En-dosc 1989;3:131-3.15. Regan JJ, McAfee PC, Mack MJ. Atlas of EndoscopicSpine Surgery. St. Louis: Quality Medical Publishing, Inc.,1995.

16. Scott TR, Graham SM, Flowers JL, Bailey RW, ZuckerKA. An analysis of 12,397 laparoscopic cholecysteaomies.Surgical Laparoscopy and Endoscopy 1992;2:191-8.17. Simpson JM, Silveri CP, Simeone FA, Balderson RA, AnHS. Thoracic disk hemiation. Re-evaluation of the posteriorapproach using a modified costotransversectomy. Spine 1993;18:1872-7.

18. Southern Surgeon's Club.A prospective analysis of 1518laparoscopic cholecystectomies. New EngiJ Med 1991;325:1073-8.

19. Zucker KA, ed. Surgical Laparoscopy Update. St. Louis:Quality MedicalPublishing, 1993.

Address reprint requests to

Paul C. McAfee, MDThe Scoliosisand Spine Center

7SOS Osier Drive, Suite 104Baltimore, MD 21204

Documents

The Incidence ofComplications in Endoscopic Anterior