Childs Nerv Syst (2004) 20:618–624DOI 10.1007/s00381-004-1001-8 S P E C I A L A N N U A L I S S U E

James Tait GoodrichDavid A. Staffenberg

Craniopagus twins:clinical and surgical management

Received: 15 April 2004Published online: 10 June 2004� Springer-Verlag 2004

J. T. Goodrich ())Division of Pediatric Neurological Surgery,Center for Craniofacial Disordersat the Children’s Hospital at Montefiore,Albert Einstein College of Medicine,Bronx, New York, USAe-mail: goodrich@aecom.yu.eduTel.: +1-718-9204187Fax: +1-718-5196876

D. A. StaffenbergDepartment of Plasticand Reconstructive Surgery,Center for Craniofacial Disordersat the Children’s Hospital at Montefiore,Albert Einstein College of Medicine,Bronx, New York, USA

Abstract Case report: We presentthe case of 2-year-old craniopagustwins who are in the process of un-dergoing a staged separation of theircraniums. In addition, they will un-dergo a full calvarial vault remodel-ing to reconstruct the lacking skulland scalp tissue in each child. Weelected to do a staged separation ra-ther than a single marathon operationfor various reasons, which we willdetail in this report. Discussion: Thephilosophy and surgical managementof each step will be discussed alongwith the pros and cons behind thereasoning that was used. At the timeof this report the children have un-dergone three of what are projected to

be four stages of separation. Each ofthe stages has had unique character-istics and surgical judgments in-volved, and those will be reviewed.Conclusion: Successful separation ofviable conjoined twins has been his-torically a great rarity. Successfulseparation of twins where both havecome out of surgery without anyneurological deficit remains a rareoccurrence and in our mind the ulti-mate goal.

Introduction

The craniopagus twins reported here were born on April21, 2002 via a cesarean section. A prenatal ultrasound hadbeen done and viable craniopagus twins were determinedat 30 weeks gestation. Their obstetrical deliveries wereuneventful and shortly after birth they were sent to a ma-jor university center where an extensive work-up wasdone to determine if any anomalies of the major organswere present; a multi-system work-up was unremarkable.A surgical separation was discussed by the local surgeonsbut the case was deemed too complex to be done at thelocal facilities. The Center for Craniofacial Disorders atthe Children’s Hospital at Montefiore was contacted andasked if we would be willing to undertake a separation ofthese craniopagus twins. After completion of an extensivemedical and radiological work-up in their country, wedecided to accept the children for a possible surgical sep-aration.

Our team initially asked the local physicians for a se-ries of medical and radiological studies to help us deter-mine what would need to be done for these children. Thereports that we were sent suggested that these were twohealthy 14-month-old children, with no medical issues; adifferent situation emerged when the children arrived inthe United States. Working with a local neurosurgeon wearranged a series of radiology examinations. Due to prob-lems with anesthesia and the inability to deliver it suc-cessfully, not all of these studies were completed. Wedeferred these studies until the children’s arrival at ourcenter. An early photograph of the twins sent suggested toour team that these children were underweight and thatboth had neurological problems. This proved to be thecase. We will discuss our management of both the pre-operative and operative phases of this separation. Wemade the decision to do a multiple-stage cranial separa-tion rather than a single-stage separation for variousreasons, which will be discussed (Fig. 1).

619

Fig. 1 Twin A on the left andtwin B on the right, at about1 year of age. Early neurologi-cal problems are clearly evidentin twin A with a partial left sideweakness secondary to posi-tional disuse of his left side

Fig. 2 a MR with twin A on topand twin B below, the interdig-itation of the respective brainscan be seen, a discrete planebetween the two brains can beappreciated. b The plane of fu-sion at the vertex of each childis clearly evident in this MR

Fig. 3 a CT angiograms show-ing the vertex vascular anoma-lies with twin A on the top andtwin B below. Twin A has alarge sagittal sinus draining thecircumferential venous plexus(seen clearly in B). Twin B hasminimal venous outflow in theposterior (occipital–parietal re-gion) portion of the vein. Thereis deep venous circulationclearly present in twin B and thegoal of a staged separation is tohave the deeper circulation takeover. b In this image twin B isnow at the top, the large midlinecircumferential venous plexuscan be clearly seen in the cen-ter, at the fusion of the twobrains, draining into a largeanomalous sagittal venousplexus

620

Case report

Six months elapsed between our first contact and the arrival of thechildren at the Children’s Hospital at Montefiore. During this in-terval, we attempted to obtain a series of radiological studies todetermine the anatomy of their conjoined brains. Due to aspirationpneumonia, chronic medical illness, and anesthesia problems,these studies were never completed. Due to their vertex joining,these children laid chronically on their backs. This position led torepeated aspiration pneumonia in both children with oral feeding.Both children arrived at about 60% of their normal body weight, aweight loss clearly due to the chronic aspirations and illness. Thechildren were initially placed at a local pediatric rehabilitationhospital for the purpose of health care, weight gain, and motor andintellectual development. In retrospect, this proved to be a mostimportant decision and one that became quite critical in the chil-dren’s care. Our initial reports on these children detailed no sig-nificant medical issues; such was not case on their arrival. Twin Awas severely hypertensive with a blood pressure of 210/125 mmHgand twin B was hypotensive at 60/40 mmHg. Both children hadaspiration pneumonia with fevers and chest X-rays showing diffuseinfiltrates. Neurologically, both children had mild but definitehemipareses from a chronic positioning on one side that did notallow them to use one side of their bodies. Both children weredevelopmentally delayed secondary to social isolation in theircountry and exhibited minimal smiles, no babbling and no form ofspeech development. Our pediatric team did an extensive work-upinitially to evaluate the blood pressure issues; a full systemic work-up, including an extensive renal evaluation was normal. Twin Ainitially required four cardiac medications to control the pressure,

bearing in mind that some degree of hypertension was essential forthe perfusion of both brains. Interestingly, medication given to onechild did not seem to affect the other. Twin A was routinely uri-nating but twin B rarely needed a diaper change. As the bloodpressures normalized twin B was no longer oliguric. Close col-

Fig. 4 A medical model reconstructed from the CT venous an-giogram showing the anomalous venous patterns. Twin A is on theright and twin B is on the left. Image courtesy of Medical Modeling

Fig. 5 A series of reconstruc-tions done by Medical Model-ing (Denver, CO) showing in 3-D different views of the venoussystem. Twin B is at the top andtwin A below. Twin A clearlyhas most of the venous outflowstructures. The anomalous cir-cumferential nature of the sag-ittal sinuses can be appreciatedhere

621

laboration with the rehabilitation team corrected the feeding issueswith the use of feeding tubes. Repositioning of the children duringfeeding led to an almost complete resolution of the aspirationpneumonia. Over a period of 6 weeks, adequate weight gain wasachieved in both children and a complete medical work-up wasfinished.

Our initial radiology work-up was directed towards determiningthe potential connection between the twins’ brains (Figs. 2, 3). Wewere particularly interested in any brain fusion, interdigitation, andpossible sharing of vascular supply. An MRI and CT study showedthat the brains abutted vertex to vertex with a 15% rotation, anO’Connell type I malformation. Twin A possessed the primaryvenous outflow pattern with the major sagittal sinuses. Twin Blacked any major venous outflow with the exception of somesmaller tributaries with his venous outflow directed to twin A. Afteran extensive discussion with our neuroradiology and pediatriccolleagues, it was determined that a staged separation would be theappropriate method, as we hoped to develop new deep venouscollaterals in twin B. Without these collaterals, the risk of venousstasis and stroke would be unacceptably high. We discussed butdecided not to perform any form of venous bypasses due to theanatomy and the known high failure rates of venous grafts.

Very helpful in our radiology work-up was the use of 3-D re-constructions of the arterial and venous circulations. The venouscirculation was anomalous but typical of what has been seen inprevious vertex-joined craniopagus twins. The sagittal sinus inthese children travels around the skull in a circumferential fashion.Rudimentary midline sagittal sinuses are seen in the forehead re-gion only in each child. As is typical in vertex joining there is amajor circumferential venous plexus at a fusion point of the twobrains. Key to making a separation plane is deciding which twin isto get this anomalous venous structure. Attempts to “split” thissinus will lead to major and unnecessary complications. We de-cided that the circumferential venous plexus was primarily pos-sessed by twin A. However, separating it from twin B would re-quire that twin B develop a deeper venous outflow pattern to avoidvenous stasis and congestion. Failure to develop this collateralpattern would lead to a serious stroke or even death secondary tosevere venous congestion, a point made by several previous sur-gical teams. With this anomalous anatomy in mind, our team de-cided to perform a staged separation in the hope of re-routing thesuperficial venous outflow present in twin A to a deeper cerebralcirculatory pattern in twin B. The model for this plan came from L.Marion Walker, M.D., whose team at Primary Children’s Hospitalin Salt Lake City, Utah, USA had performed three staged separa-tions of craniopagus twins with good results (discussed elsewherein this issue) (Figs. 4, 5).

Anesthesia issues

Several anesthesia issues became apparent early on. De-spite being conjoined twins, each twin has his own uniquecirculation and reaction to medications. Medication givento one twin does not necessarily affect the other twin.Giving intravenous fluids to one child can lead to the othertwin urinating. In our case one child was severely hyper-tensive and the other hypotensive. However, this began tonormalize as the separation went forward. Our pediatricteam recognized early on that normalizing the hypertensionin twin A might not necessarily be advantageous, as thehypertension was most likely needed to provide adequatecerebral perfusion for both twins. So a higher than normalblood pressure was considered acceptable (we used a range

of 140–180 mmHg systolic and 85–100 mmHg diastolic).These children needed separate anesthesia teams during thepre-operative work-ups. MRI machines are able to performonly one exam at a time so a separate and second monitorhad to be ordered to provide the necessary monitoringduring sedation for both twins in the MR machine. Eachtwin needed to be clearly documented in the MR or CTscanner so there would later be no confusion in identifyingeach child appropriately. In the operating room there wereseparate anesthesia teams for each child. In addition, athird senior-level anesthesiologist was present to act as abridge between the two teams. Slave monitors were madeavailable so that each team could see the other twin’sphysiologic responses.

Operating room issues

Due to the nature of these operations and the number ofnecessary personnel, the largest operating room was madeavailable. We positioned the children at the foot of twooperating tables with the heads suspended freely betweenthe two tables. Each child was draped separately givingfull exposure of each calvarium from the ears up. Pre-operatively, the plastic surgery team had marked out theflaps to be rotated for the final skin coverage. Using

Fig. 6 Intraoperative positioning of the twins, placed on separateoperating room tables with the heads suspended in between. Eachhead is draped separately allowing 360� access to each child. At thetime of separation each child goes onto the respective OR table forthe reconstruction

622

acrylic models provided by Medical Modeling (Denver,CO, USA), we were able to map out the skin flaps and thevarious craniotomies. We used only one OR scrub nursewho provided the instrumentation for both the plastic sur-geon and neurosurgeon. There were two separate circu-lating nurses, one for neurosurgery and one for plasticsurgery. The children were positioned so that two sur-geons could sit on opposite sides. By rotating the tableseach up to 30�, full circumferential exposure of the skullscould be obtained (Fig. 6).

At the initial surgery, tissue expanders were placed,but these later became infected with bowel flora and hadto be removed. Because of the subcutaneous seromas thatnormally developed and the risk of infections we latermade the decision to place the tissue expanders before the

final operation and not associate them with any craniot-omy

Staged separation issues

The principle behind the staged separation is to allowconjoined fused brains, which share arterial and venouscirculation, to develop adequate collateral circulation andavoid serious neurological sequelae. To accomplish thisgoal we decided that at least three, if not possibly fivestages would be needed to accomplish this. The initialcraniotomy was performed to separate a large frontalmidline sagittal sinus in twin B and this was achievedwithout sequelae. At this first operation, it became clearthat twin A embryologically owned the majority of the

Fig. 7a–d Imaging studiesdone just before the third sepa-ration phase. a 3-D CT recon-struction showing the first twostaged craniotomy flaps. b Op-erative positioning using acrylicmodels reconstructed from theCT studies (provided by Medi-cal Modeling. Denver, CO). c, dTwo CT angiograms showingthe collateral venous circulationthat has developed in twin B.These studies were done amonth after the second stagedseparation where approximately50–60% of separation had beencompleted

623

circumferential venous sinus and the natural plane was onthe side of twin B. In addition, twin A had a remnant duralmembrane present while twin B had none, although anarachnoid plane was present. At each stage of separation asilicone sheet was placed in the dissected plane to preventfusion and scarring of the plane. These sheets proved tobe extremely helpful as the staged separation went for-ward. To date, three separations have been done in acircumferential fashion and at this point, the twins are90% separated. At each stage approximately 3–4 largevenous vessels were gradually clamped and taken. Adural separation was done just lateral to the venous sinusof twin A, on twin B’s side. The brains were graduallyseparated following the arachnoid/dural plane. Metallicclips were used to seal the venous sinus allowing us toseparate the children with minimal bleeding. As each ofthe large veins was encountered, a temporary clip wasplaced and we carefully looked for brain swelling andvenous engorgement. By doing this in a staged fashion wehave not seen any significant venous stasis. As a result ofthis staging, twin B has developed a deep cerebral venousoutflow circulation that is clearly evident on MRV stud-ies. As each of the separation stages has been completed(three to date), the follow-up studies clearly reveal thisnecessary re-routing of venous outflow. Neither of thetwins has developed any neurological deficits, neithertemporary nor permanent, nor is there any evidence ofvenous or arterial infarction on MR. After the secondseparation stage, which involved taking a large posteriorsagittal vein, twin B was noted by the mother to be “notacting right.” Twin B seemed a bit slower and less affa-ble, but this personality change resolved over about 2weeks.

As the separations went forward, several interestingphysiological changes became evident. Twin A’s severehypertension began to normalize. Instead of four cardiacmedications for controlling the hypertension, twin A onlyrequired two and in lesser amounts. In addition, the wideswings in blood pressures began to normalize after thesecond and third operations. Twin B who was barelyurinating, now began to go from only one diaper a day tothree per day. So clearly some physiological dynamicswere evolving (Fig. 7).

Discussion

Vertex joined craniopagus twins clearly share some com-plex vascular anatomy. Most of the vascular sharing,however, appears to be on the venous side. Historically,these twins have been separated in a single-stage opera-tion, forcing both children to undergo rapid and some-times disastrous changes in blood flow patterns. Fol-lowing up on the initial studies of L. Marion Walker,M.D. and his team it seems intuitive that a staged sepa-ration is the ideal way to go in these complex cases.

Review of the previous cases in the literature on vertexcraniopagus twins typically reveal these anomalous ver-tex circumferential venous plexuses. One twin alwaysseems to have the major venous outflow while the othertwin has little and is shunting to its respective twin.Successful separation would dictate that the twin withoutthe collateral venous outflow needs to gradually developits own deep venous circulation before final separa-tion and a staged approach seems to us the ideal way toachieve that goal. An equally important goal is the avail-ability of adequate soft tissue coverage at the completionof the separation. Here, tissue expansion of the scalps arekey so that adequate scalp coverage can be obtained toprevent later CSF leakage and infection. Here, the co-operation between the plastic surgeon and neurosurgeonbecame essential and key to a good outcome. PersistentCSF leakage and meningitis can only lead to unaccept-able further neurological compromise. The anatomicalseparation of vertex-joined twins often leads to a dis-ruption of the CSF absorptive pathways over the vertex,so it is not uncommon for one or both twins to develop

Fig. 8 a Twin A to the left and twin B to the right, postoperativeday 2 after the third separation procedure. (Reprinted with per-mission by Montefiore Medical Center. Photography by AliceAddie.) b Twin A to the left and twin B to the right on postoper-ative day 3 after the third separation. At this point, about 90% of theseparation had been completed. Photograph by Alice Addie

624

hydrocephalus at some point. The neurosurgical team hasto be prepared to divert CSF in a satisfactory fashion atsome point. To date, neither of our twins has developedhydrocephalus, but we are prepared to deal with that itshould it occur.

For a separation to be successful an enormous amountof cooperative efforts are needed from the pediatric, an-esthesia, neurosurgical, and plastic surgery teams. In our

case, the contribution of the physical and occupationaltherapist was key in getting these children to the appro-priate body weight and in good health to tolerate thesecomplex surgical procedures. In our case, resolution ofsystemic high blood pressures and continual aspirationpneumonias had to be resolved before any staged cranialseparation could be accomplished (Fig. 8a, b).