Clinical Challenges in Otolaryngology

Clinical Challenges in Otolaryngology

Rejuvenation of the Midface

Edwin F. Williams III, MD, FACS

Edward D. Buckingham, MD

Dawn C. Buckingham, MD

Hypothesis: Midfacial rejuvenation can best be safely and effectively accomplished through endoscopic browlift incisions.

Background

Surgeons have for many years recognized the youthful contour of the neck and jawline. Additionally, forehead rejuvenation has been well described and undertaken for many years through trichial, coronal, and more recently endoscopic techniques. As the aesthetic face is divided in to horizontal thirds, forehead and lower face techniques accomplished adequate reversal of facial decent in the inferior and upper facial regions while leaving the middle facial region essentially unchanged. In the last decade, several authors have recognized and brought needed attention to surgical management of the middle third of the face. (1-4) Within the midfacial region two primary areas are addressed: 1) the inferior orbital area and 2) the nasolabial fold. However, more recently attention has been drawn to the benefit that midfacial elevation has on the jawline through superior vector repositioning of the midface and attached jowl. (5)

The undesirable changes that occur surrounding the inferior orbit have been described as the “nasal-jugal” or “tear trough” deformity. The creation of this appearance is multi-factorial. In youth there is a gentle convexity of the pre-tarsal and pre-septal eyelid that gently interfaces with the inferiorly progressing cheek convexity. As aging occurs the gentle pretarsal/preseptal curvature is overwhelmed by the forward migration of orbital fat as a result of a weakened orbital septum. This harsh orbital convexity, because of fibrotic attachments of the orbital septum to the orbital rim at the arcus marginalis, is then abruptly transitioned to a skeletonized inferior orbital rim. This inferior rim skeletonization is worsened because the malar fat pad laterally and sub-orbicularis oculi (SOOF) fad pad centrally and medially have descended from their youthful roles of camouflaging the orbital rim. Decent of these fat pads also leads to deficiency of the malar eminence and loss of the full heart-faced shape of youth exchanging instead for the narrow pear shaped face of late adulthood. Much discussion has occurred regarding the excision, repositioning, or mobilization of orbital fat for treatment of this lower eyelid deformity in conjunction with management of the midfacial fat pads. The proposed hypothesis is intended to discuss treatment of tissue descent outside of the orbital septum, the SOOF and malar fat pad. As the orbital fat debate is a separate topic and any of these orbital fat techniques could be combined with procedures for midfacial elevation, surgical management of orbital fat will not be discussed further. Correction of the nasojugal deformity centers on resuspension of the SOOF and malar fat pad as well as tightening and elevation of the orbicularis muscle sling. Resuspension of this complex will not only correct the nasojugal line, but will also optimally restore the malar fat pad to its position over the bony malar eminence, therefore replenishing the youthful facial contour. This goal can be accomplished theoretically through an abundance of techniques that will be the focus of later discussion.

The second common area discussed in regard to midfacial rejuvenation is the nasolabial fold. Ramirez provides an excellent discussion of the definitions and descriptions of the anatomy relevant to the nasolabial fold. (2) In order for clear discussion a distinction must be drawn between the nasolabial fold and the nasolabial sulcus or crease. The nasolabial fold consists of the cheek fat (malar) as it descends and lies over the nasolabial sulcus. Prominence of the nasolabial complex with age occurs because of greater descent of the malar fat pad relative to the more fixed sulcus. The fixed sulcus is created by the medial dermal attachments of the SMAS and elevators of the corner of the mouth. For ease of discussion this entire complex will further be referred to as simply the nasolabial fold. A great deal of debate exists in the literature regarding improvement in this region or lack thereof, especially in the context of SMAS facelift techniques as compared to deep plane techniques. The major difficulty in improving this area is that elevation of the malar fat pad must occur without an equivalent elevation of the sulcus, otherwise the relationship of the fold to the sulcus remains the same and the overall complex appears unchanged except its position is then cephalic and lateral.

Many different techniques have been recently published addressing midfacial rejuvenation. Some only address one or the other of the regions previously discussed, but all have in common an attempt at elevation of either the SOOF orbicularis complex, malar fat pad or both. The major difference between the publications lies in surgical approach and dissection plane. The options are enumerable, but taken as a whole a natural division in philosophy occurs between those techniques using an approach through the lower eyelid and those using an approach either open or endoscopic from a temporal hairline incision. A key anatomical point to make is that those approaches from the temporal region must pass deep to the temporal (frontal) branch of the facial nerve, and this structure is therefore at risk, while those approaches performed through a lower eyelid incision avoid dissection around this important motor nerve. An author familiar with the endoscopic midface lift would argue its ease and safety; however to other surgeons this fact may not be so acceptable. Thus the purposeful wording of the hypothesis that the midface is “best” approached through this technique. This obviates a discussion comparing the advantages and disadvantages of the endoscopic as well as other techniques with thoughtful comparisons of success, risks and limitations. Several different authors’ approaches will be discussed with an emphasis again placed on the major division in technique being an approach either superior or inferior to the zygomatic arch.

Pro

As mentioned previously multiple techniques have been described in regard to elevation of the midface. Finger describes an approach through a temporal extension of a facelift incision utilizing a small stab over the malar process to gain access to the subperiosteal plane in the midface which is then suspended to the deep temporal fascia. (6) Psillakis performs a coronal browlift in which the dissection is extended in a subperiosteal plane, originally described by Tessier, over the midface. Once again the periosteum underlying the malar fat pad and origins of the zygomatic major and minor muscles is suspended to the deep temporal fascia. (7) Byrd approaches the midface utilizing endoscopic brow incisions and extends the dissection subperiosteal over the arch then again supraperiosteal over the body of the zygoma freeing the malar fat pad and orbicularis muscle from the underlying zygomatic muscles. The dissection is not carried further to “avoid injury to nerve branches to the orbicularis oculi muscle”. (8) Several other variations described by other authors include retrograde dissection through a subciliary or gingival buccal incision that connects to an open or endoscopic temporal incision. (2,3,9-11) The procedure with which our hypothesis is concerned is a purely “endoscopic” surgical approach to the midfacial region and is described by Rounds and Williams in two publications. (5,12) While this endoscopic dissection can be aided by a gingival-buccal incision and retrograde dissection, this incision is unnecessary and may lead to complications. Lam and Williams reported on 325 patients undergoing “endoscopic” midfacial elevation. The first 56 patients had a gingival-buccal incision as part of the procedure, however two of these patients required incision and drainage for a subperiosteal abscess. The remainder of 269 patients approached without this incision did not experience infectious complications. (5)

While the procedure as described by Williams is termed “endoscopic” midfacial elevation no endoscopes are actually used. The temporal and midfacial dissection is performed with the aide of a converse retractor and headlight for direct vision of the temporal and zygomatic arch dissection where the temporal branch of the facial nerve is at risk. (13) The remainder of the midface dissection is performed by external hand palpated dissection taking care to avoid the infraorbital neurovascular foramen. The often accompanied forehead procedure is performed through a central sagitally oriented incision and two parasagital incisions centered radially from the junction of the middle two thirds and lateral one third of the hair bearing eyebrow. The midface dissection is performed through an incision in the temporal area approximately one centimeter behind and parallel to the hairline. The subperiosteal forehead dissection is connected to the temporal dissection which is performed in a plane just superficial to the deep layer of the deep temporal fascia. At the level of the superior helix the superficial temporal fat pad is encountered and entered. Staying on the superficial aspect of the fat pad the dissection is carried inferiorly releasing the periosteum off the body of the zygoma and anterior portion of the zygomatic arch staying deep to the frontal branch of the facial nerve which is protected by the superficial layer of the deep temporal fascia and lies within the superficial temporal fascia which is arguably the superior extension of the SMAS. Once the zygoma is reached hand guided dissection can then occur over the entire midface taking care to avoid the infraorbital nerve. After the release is complete, a retractor and headlight are used to visualize the origin of the zygomaticus major muscle, malar fat pad, and underlying periosteum. A stitch is then placed into the periosteum and muscle and secured in a directly vertical vector to the deep temporalis fascia. Two more stitches are then placed from the deep aspect of the anterior portion of the temporal incision through the temporal parietal fascia and secured to the deep temporal fascia. (5) During placement of the malar and temporal parietal suspension stitches the entire lower face including the jowl area, the malar fat pad, obicularis complex and lateral brow can be visualized to undergo vertical suspension.

Lam and Williams retrospectively reviewed 325 patients, selecting 100 randomly with 6 to 50 months of follow-up and rated improvement in the periorbital region, nasolabial fold, and jowl. The improvements were graded as marked, mild, or none. In the periorbital area, 70% of patient showed marked, 30% mild, and 0% no improvement. 57% of patients had marked or mild improvement in the jowl region, and the nasolabial region showed 4% with marked improvement, 60% mild, and 36% no improvement. (5) This study indicates that the endoscopic midfacial lift is very effective in rejuvenating the periorbital region but provides only mild improvement in the majority of patients in the difficult nasolabial area. Additionally, even though lower facelift techniques would be touted to improve the jowl, this procedure also appears to improve this region in a majority of patients. Ramirez, cites several advantages of the subperiosteal midface technique: 1) If endoscopes are chosen the subperiosteal plane provides a better optical cavity, 2) The bony landmarks of the facial skeleton aide the dissection, 3) The subperiosteal dissection plane is quicker and safer, 4) The forehead and midface are dissected in the same plane, 5) The zygomatic arch “barrier” to vertical lifting of the face is overcome, 6) Allows vertical lifting of the upper and midface in continuity, 7) Increased vascularity flaps allowing use in smokers or simultaneous resurfacing, and 8) Rigidity of the periosteum allows a better more long lasting structure to support the face. (3) Williams experience supports that the endoscopic approach to the midface is safe and effective with mild to marked objective improvement in the three facial zones analyzed. Additionally, the use of “endoscopic” incisions and lack of retrograde dissection from the lower eyelid or gingival buccal area maximize the minimally invasive nature of the procedure and scar camouflage. Ramirez and others advocate the subperiosteal plane and while using other surgical approaches to achieve the same goal, advocate that subperiosteal technique provides a superior dissection plane and results compared to other techniques.

In order to evaluate the hypothesis however, the disadvantages of this technique must be discussed as well as the advantages of other perhaps less invasive methods of correcting these aging concerns. While the disadvantages of the other techniques would support the endoscopic technique and could therefore be placed in the “pro” section, discussion of these disadvantages will be more useful following a description of the technique and so will be included in the following section. Contrarily, since the endoscopic technique has already been presented the limitations and disadvantages of the technique will be discussed in the “Con” section where they are rightly suited.

Con

To begin this section the risks and limitations, those items disfavoring the endoscopic technique, will be presented. Following that, alternative techniques will be discussed focusing on both the advantages and disadvantages of each.

The proposed disadvantages of the endoscopic subperiosteal technique can be divided into potential risks, side effects, and limitations. The most obvious and severe risk is of facial nerve paralysis. The quoted rates of facial nerve paralysis, most often the frontal branch, is between 0% and 7% for various techniques with all being transient. (6-11) Williams’ experience with the endoscopic technique yielded three patients with frontal branch and one with buccal branch paresis in 325 cases and all resolved within 6 months. (5) Sensory abnormalities occur around the incision sites regularly and resolve. Williams reported one patient with maxillary division nerve distribution loss that did not resolve. Additionally, scarring, alopecia, and temporal wasting, all potential complications, have not been encountered with the endoscopic midface technique. (5) According to Rounds, significant facial swelling and lateral canthal angle narrowing occur and are transient side effects of the operation. (12) While lateral canthal angle repositioning is a major criticism of the endoscopic procedure, Williams found no consistent significant change on horizontal or vertical lateral canthal position. (5) There is however consistently a change in the orientation of the lateral “crows feet” rhytids that imparts a slight oriental appearance to the eye for a short period. This universally resolves with time and as further repetitive contraction of the orbicularis sling reorients these rhytids to a more natural configuration. Prolonged swelling has not been a concern perhaps because the entire operation including suspension of the forehead is routinely performed in less than 1.5 hours.

The major limitation of the endoscopic procedure is the limited improvement in the nasolabial fold complex. While a majority of patients had mild improvement very few had marked improvement. (5) The other limitation is that the procedure does require intimate knowledge of the tissue planes of the temporal region, zygoma, lateral orbit, and maxilla. While the dissection is relatively easy once understood and performed, the initial experience can be quite anxiety provoking. Additionally, placement of the midfacial suspension suture from the temporal incision can be challenging even for the most experienced surgeon. Another criticism cited is the expense of instrumentation; however Williams’ modifications do not require use of an endoscope and requires only two to three endoscopic periosteal elevators which can be purchased for around $225 per dissector.

The arguments for and against the hypothesized procedure have now been presented, but in order to fully evaluate the usefulness other alternatives must be compared. As mentioned this other group consists of those procedures elevating the malar complex through periorbital incisions. Many authors have published different techniques for elevation of the SOOF and/or malar fat pad though inferior eyelid incisions. The following articles are a representation of the differing techniques and discussion of proposed benefits and limitations. Items to consider when comparing the different techniques are the type of incision, plane of dissection, use of lateral canthal tightening procedure, lateral crows feet extension of the incision, and purported lower facial benefits. Moelleken (13) and a similar procedure by Turk (14) perform the cheek lift through a subciliary incision with a 4-5 mm lateral extension with an “optional” canthoplasty. The dissection creates a skin muscle flap with dissection that continues inferiorly between the orbicularis oculi and zytomaticus major. The SOOF and orbicularis is then suspended to the intermediate temporalis fascia and in Moelleken’s technique liposuction is performed to flatten a “lateral bulge” that is created. The results for 71 patients state a reduction in infraorbital hollowness, cheek pad ptosis, and upper nasolabial fold depth. The study reports three patients with temporary (2 week) hypotonicity of the lower eyelid and three patients with mild permanent increased scleral show. Additionally, minor scar revision had to be performed in eight patients due to inexact skin excision, palpable deep suture, or fullness at the lateral incision margin. The problems in technique creating these complications have been resolved according to the article however. (13) Scalafani reports using a transconjunctival incision with a one centimeter cutaneous lateral extension with canthotomy and cantholysis. A preseptal dissection is performed 8-10 mm inferior to orbital rim where the maxillary periosteum is incised and a subperiosteal dissection is performed to the inferior orbital foramen. A maxillary gingival buccal incision is then performed and the remainder of the maxillary and zygomatic periosteum is released up to the anterior zygomatic arch. This subperiosteal cheek flap is then advanced and secured to the anterior cuff of periosteum at the orbital rim. A lateral canthoplasty is then performed in all cases and a skin pinch is performed. No specific patient results are presented however the article states generalized success. Transient maxillary division hypoesthesia, lateral ectropion, and 2-3 weeks of facial edema are reported as possible complications and side effects. A major stated limitation of the procedure is inability to affect the nasolabial crease. (15) McCord’s variation describes using a subciliary incision and after preserving the pre-tarsal orbicularis a skin muscle flap is elevated to the orbital rim where a subperiosteal dissection is performed. The skin and subcutaneous tissue is then separated from the orbicularis to allow separate skin and muscle flap elevation. The muscle and skin are then resuspended superiorly and a lateral canthoplasty performed. The article does not discuss complications or specific success, but reports generalized aesthetic success in a large group of patients. (16) Fayman also performs a subciliary incision and utilizes a subperiosteal dissection. (17) Freeman reports using a transconjunctival incision and once the arcus marginalis is reached the dissection over the maxilla is performed in a supraperiosteal plane. The SOOF is then lifted and suspended to the periosteum. Excellent to above average results are reported for 90% of patients with regard to correcting the nasojugal deformity. (18,19) Zeph reports utilizing either a subciliary incision for patient with taunt lower lids or a transconjuctival incision with lateral canthotomy and canthopexy for patients with lower lid laxity. A supraperiosteal dissection is then performed to the nasolabial fold and the SOOF is lifted and resuspended. Performance of the procedure in 64 patients reported “excellent” subjective results. The study commented that the recovery was longer than for standard blepharoplasty with 2-4 weeks of malar edema. Also 40% of patients experienced transient diplopia which resolved within one week due to orbital fat repositioning affecting the inferior oblique muscle. (20)

Many different techniques to rejuvenate the nasojugal deformity through a lower lid blepharoplasty incision have been presented, and additional methods are certainly described. Many if not all patients with nasojugal deformity undergo some type of lower eyelid surgery regardless of whether the midface is to be addressed.. Addressing the midface through the lower blepharoplasty incision is therefore practical; as no new remote incisions need to be made. Additionally, the midfacial fat pads, SOOF and malar, are in close proximity to the lower eyelid and are easily accessed through lower eyelid incisions with a variety of dissection planes that are easily defined and completed. Suture placement and fixation of the fat pads is then easily accomplished. Since the entire dissection is inferior to the zygomatic arch no risk is present for damage to the frontal nerve. Several authors report success with the procedure especially for correction of the nasojugal deformity and ptotic malar fat pads and some report success in elevation of the nasolabial fold.

Disadvantages of the lower eyelid approach to the midface are as follows. Many of the techniques involve a lateral cutaneous extension of the blepharoplasty incision whether it be subciliary or transconjunctival. Additionally, several of the techniques involve rotation and excision of inferior cheek skin. The disadvantage of this incision is two fold: 1) it is highly conspicuous and requires an extended (approximately 4 -6 week) time period to become concealed and 2) cheek skin is of different thickness and so excising lower eyelid skin and lifting cheek skin creates a thickness mismatch which can be difficult to conceal. Also the more the tissues are elevated and lifted, the more skin that needs to be excised laterally to avoid a standing cutaneous deformity and thus the longer the lateral extension needs to be. Additionally, many of the procedures recommend lateral canthal tightening. The fact that this is required suggests that undesirable tension on the lower lid is created; this places the lower lid at risk not only at risk for retraction, scleral show or ectropion, but also lateral canthal repositioning risks changing the shape of the lateral eye, which most patients will not tolerate well. A small but mentionable risk of maxillary nerve division hypoesthesia is present as well. The major limitation of the lower eyelid procedures is that the affect on the nasolabial fold is probably minimal, similar to the endoscopic midfacial technique.

Bottom Line

Over the last decade awareness of midfacial decent on the overall appearance of the aging face has fostered enthusiasm toward procedures to elevate the ptotic midfacial tissue. Two primary groups of procedures have been described, those through a temporal incision and those through a lower eyelid incision. Of the procedures described from the temporal region, the endoscopic technique described by Rounds and Williams requires the smallest incisions and avoids the complications associated with retrograde dissection through a gingival buccal incision. This procedure is the focus of the hypothesis.

The primary areas of concern addressed by the various midfacial procedures are the periorbital area as a result of SOOF displacement, the malar eminence with inferior displacement of the malar fat pad, and the nasolabial fold which is a complex result of malar fat pad decent and muscle sling support from dermal attachments of the oral commissure elevators medial to the nasolabial sulcus. The most controversial of the three areas with regards to possible improvement is certainly the nasolabial fold. Without regard to the ongoing debate over lower facelifting procedures to improve this area, the best information of the articles presented was by Williams in regard to the endoscopic technique. In that study, only 4% of patients showed marked and 60% mild improvement in the nasolabial fold appearance. Of the lower eyelid procedures only Moelleken described improvement of this region and the results were not apparently scientifically evaluated. It is doubtful that any of the described procedures significantly improve the nasolabial fold without additional adjuvant intervention. The endoscopic midface technique as well as the lower eyelid techniques all report good success for improvement of the nasojugal deformity and many of the studies also tout improvement in the appearance of the malar region. The only study that comments on improvement of the jowl area is the endoscopic midfacial technique by Williams.

Both the endoscopic and lower eyelid procedures have the potential for hypoesthesia; however this complication is uncommon and almost universally temporary. Additionally, both the endoscopic and lower eyelid procedures have the potential for complications associated with the lateral canthus. It appears that with careful dissection around the lateral canthal area taking care not to detach the tendonous insertions to Whitnall’s tubercle, that the endoscopic technique does not significantly alter this area. The lower eyelid techniques do not state outwardly problems with lateral canthal angle changes, but many describe the necessity of either a canthopexy or canthotomy and “formal” lateral canthal tendon repositioning and/or shortening. Any time manipulation of the lateral canthal complex is undertaken the risk exists for altering the shape of the lateral eye, which is undesirable and highly noticed by most patients. The major risk of which the temporal approach claims sole possession is to the frontal branch of the facial nerve. While it appears that this nerve is rarely if ever permanently disabled with the endoscopic technique, the lower eyelid techniques avoid this dissection altogether and so therefore must be safer in this regard. The major unilateral disadvantage to the lower eyelid procedure in many instances is the creation of a lateral standing cutaneous deformity which may require excision of lower eyelid skin, increased lateral extension into the crow’s feet which requires prolonged healing time, and creation of a tissue thickness mismatch between the thin lower eyelid and thicker cheek skin.

Lower eyelid approaches to the midface, because of the ease of dissection and direct access to the midface, may be considered useful procedures for correction of the nasojugal deformity and ptotic malar fat pad. Additionally, these procedures are useful in the correction of isolated malar festunes. However for overall improvement of the aged appearance of the midface the endoscopic subperiosteal approach via a temporal brow lift incision is superior for the following reasons. The endoscopic temporal approach addresses the entire midfacial region as a composite flap containing not only periosteum, but also temporal-parietal fascia and SMAS, allows excellent improvement of the malar and inferior orbital area, mild improvement of the nasolabial area, and improvement in the jowl. The dissection allows a complete vertical vector of suspension of the entire midface through small incisions concealed in the hairline. The risks to the facial nerve as well as other risks while possible are apparently infrequently encountered in experienced hands. Therefore, midfacial elevation achieved with subperiosteal dissection through endoscopic incisions is probably superior, once mastered, when compared to other variations of procedures through a temporal incision or approaches utilizing lower eyelid incisions.

References

1) Hamra ST. The Zygorbicular dissection in composite rhytidectomy: an ideal midface plane. Plast Reconst Surg. 1998;102:1646-1657.

2) Ramirez OM. The subperiosteal approach for the correction of the deep nasolabial fold and the central third of the face. Clin Plast Surg. 1995;22(2):341-356.

3) Ramirez OM. Endoscopic subperiosteal browlift and facelift. Clin Plast Surg. 1995;22(4):639-660.

4) Hester TR, Codner MA, McCord CD. The “centrofacial” approach for the correction of facial aging using the transblepharoplasty subperiosteal cheek lift. Aesthet Surg Q. 1996;16:51-58.

5) Williams EF, Lam SM. Midfacial rejuvenation via an endoscopic-browlift approach: a critical evaluation of a five-year experience. Arch Facial Plast Surg. In Print.

6) Finger ER. A 5-year study of the transmalar subperiosteal midface lift with minimal skin and superficial musculoaponeurotic system dissection: a durable, natural-appearing lift with less surgery and recovery time. Plast Reconstr Surg. 2001;107:1273-1283.

7) Psillakis JM, Rumley TO, Camargos A. Subperiosteal approach as an improved concept for correction of the aging face. Plast Reconstr Surg. 1988;82:383-392.

8) Byrd HS. The extended browlift. Clin Plast Surg. 1997;24:233-246.

9) Burnett CD, Rabinowitz S, Rauscher GE. Endoscopic-assisted midface lift utilizing retrograde dissection. Ann Plast Surg. 1996;36:449-452.

10) Hobar PC, Flood J. Subperiosteal rejuvenation of the midface and periorbital area: a simplified approach. Plast Reconstr Surg. 1999;104:842-851.

11) Anderson RD, Lo MW. Endoscopic malar/midface suspension procedure. Plast Reconstr Surg. 1998;102:2196-2207.

12) Rounds MR, Cheney ML, Quatela VC. Endoscopic facial surgery. Facial Plast Surg. 1998;14:217-226.

13) Moelleken B. The superficial subciliary cheek lift, a technique for rejuvenating the infraorbital region and nasojugal groove: a clinical series of 71 patients. Plast Reconstr Surg. 1999;104:1863-1873.

14) Turk JB, Goldman A. SOOF lift and lateral retinacular canthoplasty. Facial Plast Surg. 2001;17:38-48.

15) Sclafani AP. The multivectorial subperiosteal midface lift. Facial Plast Surg. 2001;17:29-36.

16) McCord CD, Codner MA, Hester TR. Redraping the inferior orbicularis arc. Plast Reconstr Surg. 1998;102:2471-79.

17) Fayman MS, Potgieter E. Zygomaticus major advancement as an adjunct to lower blepharoplasty. Aesth Plast Surg. 2002;26:26-30.

18) Freeman MS. Transconjunctival SOOF lift blepharoplasty. Facial Plast Surg Clin. 2000;8:291-302.

19) Freeman MS. Transconjunctival sub-orbicularis oculi fat (SOOF) pad lift blepharoplasty. Arch Facial Plast Surg. 2000;2:16-22.

20) Zeph RD. Contemporary management of the aging lower eyelid. Facial Plast Surg Clin. 2000;8:259-276.