Craniopagus Twins

Published on 26/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2990 times

CHAPTER 180 Craniopagus Twins

Samuel R. Browd, James Tait Goodrich, Marion L. Walker

Craniopagus twins represent one of the rarest and most complex congenital anomalies seen in pediatric neurosurgery. It was not until 1952 that either twin even survived an attempt at surgical separation. Since then, modern neurosurgical techniques have advanced the opportunities for separation, but the surgery to separate craniopagus twins remains one of the most complex and treacherous procedures in all of neurosurgery. Until very recently, outcomes have been mixed, with one twin usually suffering severe neurological consequences after separation. The potential for disastrous outcomes makes the decision to separate craniopagus twins not only a complex surgical dilemma but also an ethical and philosophical challenge. As with many innovative and complex procedures, two opposing strategies have emerged for separating craniopagus twins. Both single-stage and multistage separations have been performed, with staged separations leading to the best neurological outcomes, especially in recent cases. In the following sections, we describe a brief history of craniopagus, review the epidemiology and classification, illustrate strategies for presurgical evaluation, discuss the rationale and technique for staged separation, and share examples of cases deemed nonseparable with current techniques and technology.

Historical Perspective

Because of their rarity and the profound curiosity they elicit, conjoined twins were often represented in ancient literature and art. Historical narratives suggest that although conjoined twins were sometimes revered, persecution and exploitation were more common among those who survived infancy.

The first documented illustration of craniopagus twins dates from the 15th century, describing a set born in Bavaria, Germany, in 1491 as a “warning from God” (Fig. 180-1). A famous case of craniopagus parasiticus was reported by Sir Everard Home¹ (1756-1832) in his collected works as the “Boys of Bengal,” a set of twins born in India in the 1770s (Fig. 180-2). In the early 19th century it became quite popular to exhibit conjoined twins in traveling road shows and circuses as “freaks” of nature and even to distribute postcards and autographed pictures of them (Fig. 180-3).

FIGURE 180-1 Early image of conjoined twins from Cosmographia universalis by Sebastian Münster (1488-1552), a Basel edition printed by Sebastian Heinrich-Petric in 1550.

(Courtesy of James T. Goodrich.)

FIGURE 180-2 Preserved skull of the Boys of Bengal.

(Courtesy of the Hunterian Museum Collection.)

FIGURE 180-3 The McCarther twins. These photographs were given out at events. The bottom photo contains the contact information for their press agent.

(Courtesy of James T. Goodrich.)

Several sets of craniopagus twins have lived into adulthood, although more than 90% of craniopagus twins die by age 10 years, regardless of attempts at surgical separation. The first successful craniopagus surgery in which one twin lived involved Roger and Rodney Brodies, separated in stages by Oscar Sugar in 1952-1953. Although Roger died a month after surgery, Rodney lived to the age of 11 and died of complications of hydrocephalus.

Classification and Demographics

In his fundamental work on teratology, Forster ² introduced the term craniopagus to describe twins joined at the head. Craniopagus twins are defined by a union of the calvaria only; unions involving the foramen magnum, skull base, vertebrae, or face are classified separately (e.g., cephalopagus, parapagus, rachipagus, diprosopus) (Fig. 180-4).³ In craniopagus twinning, the union is rarely symmetrical and may involve any portion of the head. An infinite number of configurations can occur, based on the attachment location and the degree of individual rotation of one head on the other. Variations in the degree of conjoining or sharing of underlying structures such as the meninges, venous sinuses, and cortex add to the individual nature of each case.

FIGURE 180-4 Variations of the union in craniopagus twins.

(Redrawn from Spencer R. Conjoined Twins: Developmental Malformations and Clinical Implications. Baltimore: Johns Hopkins University Press; 2003).

Several authors, beginning with O’Connell,⁴ have attempted to classify the numerous phenotypic variations seen in craniopagus twins. According to O’Connell’s classification schema, partial craniopagus twins have limited surface area affected, with intact crania or minimal cranial defects. Total craniopagus twins share an extensive surface area with widely connected cranial cavities. O’Connell⁴ subclassified “vertical” craniopagus twins—”parietal craniopagus,” according to the classification of Bucholz and associates⁵ (Fig. 180-5)—into types I through III, based on the degree of rotation of one on the other. Stone and Goodrich,⁶ in an effort to evaluate outcomes among different types of twins, used the partial and total divisions described by O’Connell and also defined two main subtypes: angular and vertical. Angular craniopagus twins have an intertwin longitudinal angle less than 140 degrees, regardless of axial rotation. Vertical craniopagus twins display a continuous longitudinal cranium and are further subdivided based on the degree of intertwin axial facial rotation: type I, rotation less than 40 degrees (formerly O’Connell type I); type II, rotation between 140 and 180 degrees (O’Connell type II); and type III (“intermediate”), rotation between 40 and less than 140 degrees (O’Connell type III).⁶

FIGURE 180-5 Examples of the four general subclassifications of craniopagus twins (top to bottom): frontal, parietal, temporoparietal, and occipital.

(Redrawn from Bucholz RD, Yoon KW, Shively RE. Temporoparietal craniopagus. Case report and review of the literature. J Neurosurg. 1987;66:72-79.)

Craniopagus twins are extremely uncommon, occurring once in every 0.6 to 2.5 million live births (Table 180-1).⁶^,⁷ The cause of the conjoining of twins is unclear; both “incomplete fission” and “fusion” hypotheses have been proposed by embryologists.³^,⁸ There are no known environmental or genetic factors. Interestingly, the female-to-male ratio is nearly 4 : 1, although this has not been correlated with any cause of the condition.

TABLE 180-1 Demographic and Embryologic Characteristics of Craniopagus Twins

INCIDENCE

Conjoined twins: 10 to 20 per million births in the United States

Craniopagus twins: 0.6 per million births in the United States

GENETICS

Identical (monozygotic)

Female-to-male ratio 4 : 1

CAUSE

No known genetic or environmental causative factors

Fusion or incomplete fission

MORTALITY Conjoined twins: 40% stillborn; 33% die in perinatal period

An extensive review of the literature from 1919 through 2006 notes 64 well-documented cases of craniopagus with 41 separation attempts—29 performed as single-stage separations and 12 as multistaged separations.⁹

Surgical Separation

Risk Stratification

The decision to separate craniopagus twins is obviously fraught with serious considerations. Stone and Goodrich ⁶ proposed their classification system in part to better understand the outcomes after surgical separation. It is now recognized that shared dural venous sinuses present significant difficulties when proceeding with separation, but numerous other risk factors must be assessed when evaluating craniopagus twins for separation. Critical to the surgical decision-making process is an understanding of the degree of shared scalp, calvaria, and dura (independent dural envelopes or incomplete dural separation); the amount of separation, interdigitation, or fusion of cortex or deeper structures; the extent of shared arterial connections and cross-flow; the extent of shared or common venous sinuses and drainage; the presence of paired or separate venous outflow; the presence or absence of independent deep venous drainage; and the presence of common or separate ventricular systems and hydrocephalus.¹⁰ Because each of these factors has serious implications for the surgical strategy, we proposed a scheme by which individual cases can be evaluated to determine the surgical risk (Table 180-2).¹⁰ Although each case involves many unique considerations, this scheme, which assigns point values based on the aforementioned criteria, can be used to gain a preliminary understanding of the degree of risk associated with a separation attempt; a higher score is indicative of a more difficult separation.

TABLE 180-2 Surgical Risk Stratification

CHARACTERISTIC	SCORE*
Scalp
Minor surface area shared (<10 cm²)	1
Major surface area shared (>10 cm²)	2
Calvaria
Minor surface area shared (<10 cm²)	1
Major surface area shared (>10 cm²)	2
Dura Mater
Independent dural envelope surrounds cortex	1
Dura shared along one or more planes	2
Neural Tissue
Completely separate	1
Interdigitated but not fused	2
Minor areas of fusion (<5 cm² total surface area)	3
Major areas of fusion or involvement of eloquent cortex (>5 cm² total surface area)	4
Arterial Connections
None	1
Minor feeding branches (M4, A4, P4)	2
Distal branches (M3, A3, P3)	3
Proximal branches (M2, A2, P2)	4
Major vascular trunks (M1, A1, P1, ICA)	5
Venous Connections
None	1
Separation of major sinuses with minor shared draining veins	2
Shared along anterior of superior sagittal sinus and/or shared distal (transverse/sigmoid)	3
Shared along posterior of superior sagittal sinus without involvement of the torcula	4
Shared along posterior of superior sagittal sinus with involvement of the torcula	5
Deep Venous Drainage
Present	1
Absent	2
Cerebrospinal Fluid (Ventricular Anatomy)
Separate ventricular systems	1
Shared ventricular system	2
Venous Outflow
Ipsilateral	1
Contralateral (crossed)	2
Arterial Flow
Ipsilateral	1
Contralateral (crossed)	2

A, anterior cerebral arteryica; ICA, internal cerebral artery; M, middle cerebral artery; P, posterior cerebral artery.

* Higher score equates with more difficult separation. Minimum score = 10; maximum score = 28.

From Browd SR, Goodrich JT, Walker ML. Craniopagus twins. J Neurosurg Pediatr. 2008;1:1-20.

Preoperative Assessment

The preoperative evaluation of patients before a separation procedure is critical, and planning generally takes months of clinical monitoring, imaging, and modeling. In addition to cranial considerations, cardiac, pulmonary, and renal function determine the feasibility of undertaking the prolonged surgical procedures involved in separation. Separation cannot be undertaken until the patients have reached an age and body weight that allow them to tolerate the potentially significant blood loss that can occur during surgery.

New techniques for image reconstruction and flow modeling have provided surgeons with valuable tools to examine critical neural and vascular structures in patients and develop detailed preoperative plans.¹¹^–¹⁵ Because of the unique anatomy with regard to the sharing of critical and noncritical structures, the goal of imaging is to define the patients’ anatomy—especially the vascular anatomy, with particular attention paid to the venous sinuses. Early computed tomography (CT) and magnetic resonance imaging (MRI) data allow a preliminary understanding of the extent of fusion and the complexity of the shared arterial and venous architecture, whereas later imaging can provide more extensive modeling of the critical structures for surgical planning.

Recent cases have benefited from the ability to create detailed acrylic three-dimensional models based on imaging studies (Fig. 180-6). High-resolution digital reconstructions of CT images of the head can be used to develop three-dimensional models to assess the degree of bony fusion. Information from the bony architecture is used to develop the surgical corridor, site of separation, and plan for postseparation reconstruction of the cranial vault. Newer helical CT scanners can also provide high-resolution angiography that complements traditional digital subtraction angiography.

FIGURE 180-6 Three-dimensional model based on computed tomography and magnetic resonance imaging data from craniopagus twins separated at Montefiore.

(Courtesy of Medical Modeling Inc., Golden, CO.)

MRI provides highly detailed evidence of shared cortex, ventricular anatomy, hydrocephalus and transependymal flow, dural folds, and septations. Because of their significance in successful outcomes, these and other issues that are directly related to surgical risk determine the strategy employed during separation and reconstruction. Advanced MRI techniques such as magnetic resonance angiography and venography detail the vascular architecture—specifically, evidence of shared arterial communications, a shared or common venous sinus (most commonly a circumferential sagittal sinus, transverse sinus, or torcula), anomalous venous structures, and collateral or deep venous drainage. Arterial flow characteristics (vector and velocity) can also be assessed with magnetic resonance angiography and venography.

Digital subtraction angiography remains the principal technique for assessing the vascular architecture. These angiographic studies provide a blueprint for the appropriate steps during separation. Knowledge of the venous anatomy is particularly crucial in staged separation because cortical veins are selectively pruned from the shared sinus, forcing collateralization of deep venous drainage. Detailed evaluations of arterial development and architecture, arterial collateralization or cross-flow, direction and degree of any cross-flow, shared or anomalous venous anatomy, and direction and timing of venous outflow are developed from four-vessel cerebral angiography of the arterial and venous phases. These assessments should be augmented by angiographic imaging of the cortical veins and their insertion locations along the sagittal or shared sinus, evidence of collateral or alternative venous drainage and deep venous drainage patterns, and outflow dominance in the transverse sinuses. The external carotid circulation should be studied in detail for evidence of collateralization or retained fetal anastomosis.

Stereotactic imaging sequences for both CT and MRI are recommended and can be used for both preoperative planning and intraoperative guidance, although rigid registration is less effective because of the need for frequent repositioning. Newer magnetic tracking systems allow patient repositioning without loss of registration data. Functional MRI and positron emission tomography have been used in some cases to assess cortical function. The practicality of functional imaging has to be assessed on an individual basis; however, any additional imaging studies that might provide useful clinical or research data should always be considered.

Single-Stage Separation

There has been limited success with single-stage separation as reported by other groups.¹⁶ Generally, the surgery is undertaken as a surgical tour de force in which the length of cases often surpasses 20 to 30 hours. Single-stage separation attempts are predicated on the twins’ ability to handle independent venous drainage immediately at the time of separation. Cases in which reconstruction of the sagittal sinus was attempted have been described,⁹ but these efforts are rarely successful because of the complexity of this procedure and the instant change in hemodynamic forces and stresses that occur when trying to bypass the entire cerebral venous outflow. Given the complexity of these cases and the difficulty of predicting the suitability of individual venous outflow, which must function immediately on separation, we favor a staged surgical approach.

Theory of Staged Separation

Shared venous anatomy is the cause of the majority of surgical morbidity encountered in modern separation attempts.⁴ Until recently, death or insurmountable neurological deficits were expected in one twin because of the devastating bleeding that can occur during separation of the shared dural venous sinuses. In these cases, the surviving twin was the one who received the entire superior sagittal sinus during separation, because there was no way during single-stage surgery to prepare the other twin for the loss of this drainage.

In a staged separation, however, the twins can be prepared for the separation by the gradual occlusion of venous outflow, which allows for remodeling and collateralization. In contrast to acute occlusion, which results in venous hypertension proximal to the occlusion and leads to severe brain edema or hemorrhage, gradual occlusion of major venous structures is well tolerated by patients. The staged approach allows the donating twin to develop a robust collateral venous drainage system using the deep veins that join the petrosal sinus and other deep venous outflow channels.

Because of the slow process of collateralization, a staged separation generally takes many months from the initial procedure to final separation. Each case is unique, so the exact timing and length of the process vary. Six to eight procedures involving circumferential craniotomies and arterial-venous ligations are generally undertaken, with 1 to 2 months between surgeries, before the final separation surgery.¹⁷

Surgical Techniques

The successful separation of twins who share a venous sinus depends on the surgeon’s ability to disconnect one twin from that structure. On preoperative vascular imaging, one of the twins can be identified as having the majority of the outflow to the common sinus, while the other generally displays evidence of a more robust deep venous collateral system. The former is selected as the twin who will receive the shared venous sinus, and the other will be forced to further develop an alternative deep venous drainage system.

During the preliminary surgeries, several bridging veins are selected for pruning in the donor twin (Fig. 180-7). Initially, temporary clips are placed across a bridging vein, and the venous drainage field is monitored for evidence of hypertension. Alternative bridging veins are selected for pruning, or the procedure is aborted, if brain edema occurs or marked cortical erythema develops. Only a couple of bridging veins are ligated at each surgery to allow the development of alternative deep venous collaterals. After each of the surgeries, repeat imaging is performed to look for evidence of alterations in the venous structures as the donor twin is disconnected from the shared sinus. These pruning surgeries are continued until the donor twin is completely disconnected from the shared sinus and the final separation surgery can be undertaken.

FIGURE 180-7 A, Intraoperative photograph shows pruning of the draining vein to the circumferential sinus. B, Artist’s representation of the first surgical stage: opening a bifrontal flap and identifying the cleft between the two brain hemispheres. C and D, Photographs show intraoperative positioning and the initial bifrontal flap and exploration. E, Photograph of twins separated using a staged approach.

(A, C, D, and E, From Browd SR, Goodrich JT, Walker ML. Craniopagus twins. J Neurosurg Pediatr . 2008;1:1-20. B, From Walker M, Browd SR. Craniopagus twins: embryology, classification, surgical anatomy, and separation. Childs Nerv Syst 2004;20:554-566, with the permission of Springer Science and Business Media.)

Risks and Complications

The principal risk among the myriad possible complications is venous hypertension and subsequent intraparenchymal hemorrhage while separating one of the twins from the shared venous structures. The process of staged separation should allow the development of more extensive deep venous drainage in the donor twin. To minimize the risk associated with this dramatic venous remodeling, intraoperative temporary clipping and inspection of cortical areas for evidence of venous hypertension are critical. The decision to prune a few bridging veins selectively and then terminate a surgical session requires forethought and consideration of the overall goal of a multistaged separation. Careful preoperative planning can assist with decisions about which veins to select for pruning and how many veins to prune at each operation.

Another issue associated with staged procedures is the need for the surgeon to know at all times the overall anatomic location of the circumferential sinus and to have a conception of where the final ligation will occur.¹⁷ Placing Silastic sheets between adjacent cortical areas can help preserve natural tissue planes and provide a means for reorientation at the time of repeated procedures.¹⁸

The dural venous sinus is not the only critical structure shared by craniopagus twins. Shared or fused cortex is often reported as well. Our strategy is to coagulate and divide shared cortex; this has been well tolerated, without obvious cognitive consequences in the twins. Cross-filling of arterial blood between twins is also a possibility, and the surgeon must know the direction of flow and the contribution of the shared arterial branches to decide when and where to coagulate the arteries.

Large defects created by the separation often need to be surgically repaired for a good functional and cosmetic outcome. Reconstruction involves multiple anatomic structures, including the dura, calvaria, and scalp. Dura is reconstructed with a variety of allograft products, including onlay allograft dural substitutes. Large calvarial defects may be left open after the initial or subsequent separation surgeries; however, the use of split-thickness autograft or the later fabrication of custom methyl methacrylate cranioplasties can lead to good cosmetic reconstructions of the cranial vault. The use of tissue expanders during multiple surgical procedures allows the growth of additional scalp and reduces the need to incorporate rotational or transpositional flaps.¹⁸ Although some have reported the long-term (3- to 6-month) use of tissue expanders before a single separation procedure,¹⁹ we advise delaying the placement of tissue expanders during staged separations until the penultimate procedure to avoid the common complication of infection.

Despite the absence of major portions of the dura mater, falx, and superior sinuses, craniopagus twins do not present with hydrocephalus, indicating that cerebrospinal fluid is adequately reabsorbed.²⁰ As surgical separation progresses, however, the absorptive capacity may be altered in one or both twins, requiring cerebrospinal fluid diversion. Routine postoperative scanning with CT and MRI can provide evidence of hydrocephalus, and shunt placement can be performed as the need arises.

As we indicated earlier, the potential for infection in the months between the first and last surgical procedures is high, and postoperative infections frequently occur. Indwelling catheters, shunts, tissue expanders, and drains provide a direct route for infection. Antibiotics are routinely administered both intraoperatively and in the postoperative setting, and the use of potential vehicles for infection should be minimized to reduce the risk.

Nonoperable Craniopagus

Although great strides have been made in our ability to separate craniopagus twins, some cases are still nonoperable with current techniques and technology. The major reason for inoperability continues to be shared venous anatomy and, less frequently, shared eloquent structures. Patients with complex venous anatomy and fusion anomalies that equate with risk stratification scores in the upper 20s are still thought to be inoperable. Two cases are illustrated that are currently thought to be nonseparable.

Illustrative Case
Craniopagus Twins with Conjoined Brain Bridge

In this case, a set of angular craniopagus twins joined side to side, with conjoined parietal-occipital regions, was diagnosed by prenatal ultrasonography. Extensive vascular and neurological imaging studies revealed that their brains were conjoined by a central diencephalic bridge (Fig. 180-8). Because of the severity of the conjoining at the diencephalon, the morbidity and mortality risks were considered unacceptably high, and separation was not performed.

FIGURE 180-8 A, Prenatal magnetic resonance imaging (MRI) of conjoined brains. B, Photograph of the children shortly after birth reveals an “angular” type of conjoining. (Photo courtesy of BC Children’s Hospital.) C, Coronal MRI shows the “complex” conjoined diencephalic bridge between the two brains. As a result of this anatomic bridge, the morbidity and mortality of separating these twins were unacceptably high. D, Three-dimensional reconstruction using computed tomography data to reconstruct the conjoined venous system in this set of twins.

(D-Courtesy of Medical Modeling Inc., Golden, CO).

Illustrative Case
Effect of Cultural Views on Craniopagus Management

This case illustrates the complex social and regional philosophical views that can complicate craniopagus management. These nonoperable craniopagus twins had a partial angular frontal configuration (Fig. 180-9). They were diagnosed prenatally by ultrasonography and delivered via elective cesarean section. The children were conjoined at the forehead with a shared anterior fontanelle, and both were neurologically intact. Because of financial limitations and cultural prejudice against conjoined twins in the family’s community, imaging studies were not done until the 24th day of life. CT ultimately showed that the cerebrum and cerebellum of twin B were better developed than those of twin A and that the frontal lobes were fused. By day 25, twin B developed respiratory distress, which led to cardiac arrest. The premature death of twin B necessitated an emergency separation attempt, which was unsuccessful. Although it is unlikely that the outcome would have been different, the complex cultural and social issues that complicated the neurosurgical management of this set of craniopagus twins cannot be overstated.

FIGURE 180-9 Photograph of conjoined twins shortly after birth with a frontal angular conjoining. The extensive shared structures prevented the safe separation of these twins.

It is our hope that as technology evolves and our understanding of the physiology of these unique children advances, improved staged surgical techniques will enable cases once thought to be inoperable to proceed to surgery with good functional and cosmetic outcomes.

Acknowledgment

This chapter is based on articles by Walker and Browd ¹⁸ and Browd and colleagues.¹⁰ Readers are referred to these sources for further information on the history and management of craniopagus twins.

We thank D. Douglas Cochrane, MD, and his team at the Children’s and Women’s Health Center in Vancouver, British Columbia, for sharing the details of one of their cases. We thank our colleagues in Lagos, Nigeria—in particular Okezie O. Kanu, MD, consultant attending neurosurgeon at Lagos University Teaching Hospital, and his neurosurgical team—for allowing us to present their case as part of this chapter. We thank Kristin Kraus for her editorial assistance in preparing and organizing this chapter.