SEARCH

SEARCH BY CITATION

Keywords:

  • preputial/penile skin flap;
  • dorsal;
  • onlay

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Study Type – Therapy (case series)

Level of Evidence 4

What's known on the subject? and What does the study add?

For long complex anterior urethral stricture augmentation urethroplasty is considered the standard procedure but the best substitute material is still to be ascertained.

Preputial/penile skin is a very good substitute especially when used as a dorsal onlay. It demonstrates exceptional functional and cosmetic results even in patients with unsuitable oral mucosa.

OBJECTIVE

  • • 
    To present our experience of single-stage reconstruction of urethral stricture with preputial/penile skin flap, as a dorsal onlay flap (DOF) where there is an adequate urethral plate and as a tubularized flap (TF) where there is a compromised urethral plate, in cases of complex anterior urethral strictures.

MATERIALS AND METHODS

  • • 
    We retrospectively reviewed 144 patients, who underwent single-stage repair of pendular /bulbar urethral strictures with preputial/penile flap as either a DOF or a TF, between January 2001 and December 2008.
  • • 
    Patients were divided into three groups: Group 1 consisted of patients who underwent transverse preputial DOF; Group 2 consisted of those who underwent tube urethroplasty; and Group 3 consisted of those patients who were circumcised and for whom the penile skin was used as a DOF (circumpenile flap).
  • • 
    Patients were followed up by physical examination, retrograde urethrography, uroflowmetry and post-void residual urine measurement.

RESULTS

  • • 
    The mean follow-up was 40.1 months (range 36–84 months).
  • • 
    The primary success rates at 1 year follow-up were 90, 85 and 93.3% for Groups1, 2 and 3, respectively, and at 3-years follow-up they were 85, 75 and 86.7%, respectively.
  • • 
    Half of the recurrences were successfully managed with a single visual internal urethrotomy or dilatation.
  • • 
    The secondary success rate was defined as recurrent stricture managed by a single endoscopic procedure and was 5, 10 and 6.8% in Groups 1, 2 and 3, respectively. The overall success rate was 90.85 and 93.3%, respectively.
  • • 
    A total of 75% of the patients in the study completed 60 months of follow-up with no additional recurrence.

CONCLUSIONS

  • • 
    A preputial/penile flap for complex anterior urethral stricture is a good treatment option, with results similar to other techniques, has acceptable donor site morbidity and is effective even in circumcised patients and for those patients with unsuitable oral mucosa.
  • • 
    A DOF is less likely to lead to diverticula formation and post-void dribbling. TFs have a higher failure rate than DOFs but, when combined judiciously with secondary endoscopic procedures, can provide good results.

Abbreviations
DOF

dorsal onlay flap

TF

tubularized flap

VIU

visual internal urethrotomy

BMG

buccal mucosa graft

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Anterior urethral reconstruction has undergone significant changes over the last 60 years. Excision with end-to-end anastomosis remains the ‘gold standard’ for short-segment post-traumatic bulbar strictures; however, for long-segment strictures not amenable for end-to-end anastomosis, numerous options are available and the best is yet to be ascertained. Augmentation urethroplasty is needed if the stricture is lengthy or involves the penile urethra. Augmentation can be performed in either one or two staged procedures. Use of a graft or flap for augmentation is a source of much controversy, with re-stricture rates of ∼15% [1,2]. Full-thickness skin graft urethroplasty was popular in the 1950s. This was later replaced by the penile skin fasciocutaneous flaps as first described by Orandi [1] and modified and popularized by Quartey [2], McAninch [3] and Jordan and Stack [4] during the early 1990s. Most authors have reported similar success rates of 83–93% for both the free grafts and pedicled skin flaps, and generally poorer results with the formation of a tube [5–9].

In the current era of buccal mucosa, now considered the donor substitute of choice for augmentation, the success of the fasciocutaneous flaps in the management of anterior urethral strictures should be reassessed.

The transverse island flap, and especially the circumpenile flap, first described by McAninch [3] for complex anterior urethral stricture, provides excellent cosmetic and functional results. It reliably provides up to 15 cm of non-hirsute, well-vascularized tissue that can be used in all areas of the urethra, from the meatus to the membranous area, as a tubularized flap (TF) or dorsal onlay flap (DOF). With this technique, anterior urethral reconstruction can be performed as a single-stage procedure in cases of severely compromised urethral plate, offering potentially lower costs and social acceptability as compared with staged repair [10,11]. We use a transverse island preputial/penile island flap, including the circumpenile flap as a DOF for single-stage reconstruction of such complex anterior urethral strictures where there is an adequate urethral plate, and as a TF where there is a compromised urethral plate.

The aim of the present study was to present the technique and outcomes of preputial/penile skin flap urethroplasty in 144 patients at our institution.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Between January 2001 and December 2008 a total of 144 patients with complex anterior urethral strictures underwent single-stage anterior urethral reconstruction with preputial/penile flap (augmentation) or tube. Four patients were lost to follow-up and were excluded from the study. Complex strictures were defined as the presence of long-segment strictures not suitable for end-to-end urethroplasty, those with extensive spongiofibrosis, or deficient urethral plate or previous surgery for stricture repair.

Preoperative assessment consisted of history, examination, uroflowmetry, and retrograde and voiding cysto-urethrography. Preoperative variables are shown in Table 1. To assess the urethral mucosa and urethral stricture calibre, a preliminary urethroscopy was carried out using a 6-F paediatric cystoscope in all patients. A transverse island preputial flap or circumpenile flap was used as a DOF only if stricture calibre was >6 F with an acceptable urethral plate (Fig. 1A). If there was severe spongiofibrosis and the urethral lumen was <6 F in calibre, or if there were multiple, focally dense stricture segments and the urethral plate was inadequate (Fig. 1B), urethral substitution was performed using a TF, with proximal and distal anastomosis to the normal urethra.

Table 1. Preoperative variables
VariableGroup 1: transverse preputial DOFGroup 2: TFGroup 3: circumpenile DOF
  1. IIEF, International Index of Erectile Function.

No. of patients404060
Median age (range)38.0 (17–56)35.4 (11–66)39.5 (19–60)
Mean stricture length, cm (range)07.0 (6–12)08.2(5.5–10)8.9 (7–10.5)
Mean no. of procedures performed previously1.821.8
Stricture location, n (%)
Penile4 (10)04 (10)04 (6.7)
Bulbar16 (33.3)12 (30.0)10 (16.7)
Peno-bulbar18 (45)24 (60.0)46 (75.9)
Mean (sd) IIEF erectile function domain score23.9 (5.80)24.4 (3.96)23.62 (5.53)
Mean (sd) IIEF orgasmic function domain score6.9 (2.31)7.12 (5.18)6.5 (4.29)
image

Figure 1. A, RGU/MCU of 48-year-old man with long penobulbar stricture with adequate urethral plate. B, retrograde urethrogram and micturating cystourethrogram of 52-year-old man with a history of prolonged urethral catherization and dense stricture with poor urethral plate in penobulbar area.

Download figure to PowerPoint

The primary success rate was defined as the number of patients without any urethral intervention at 1 year and then at 3 years. At 3 years, the secondary success rate was defined as the total number of patients managed successfully by a single endoscopic intervention with at least 2 recurrence-free years from the last endoscopic intervention. The overall success rate was defined as a combination of the primary success rate and the secondary success rate at 3 years.

SURGICAL TECHNIQUE

Stricture exposure

For penile strictures the urethral reconstruction was carried out using a circum-coronal incision. For more proximal strictures (penobulbar, bulbar urethra) a midline perineal incision was used. In the standard lithotomy position, the strictured part of the urethra was completely mobilized from the corpora cavernosa. If the stricture extended proximally high in the bulbomembranous urethra, an inferior pubectomy was performed to facilitate exposure of the membranous urethra (as was seen in four of the present patients). Urethral stricture with a preserved urethral plate was opened longitudinally on its dorsal aspect (Figs 2, 3 and 4) for DOF repair, otherwise the scarred urethra was completely excised for TF replacement. The length of stricture was assessed intra-operatively by measurement.

image

Figure 2. RGU/MCU of the patient in Fig. 1B after preputial flap urethroplasty.

Download figure to PowerPoint

image

Figure 3. Raising the prepatial flap.

Download figure to PowerPoint

image

Figure 4. Complete proximal mobilization of the preputial flap to allow tension free transposition to the perineum.

Download figure to PowerPoint

For a preputial flap ∼ 0.5 cm proximal to the corona, a transverse or circular island flap (∼1.5 cm width for a DOF and 2.5 cm width for a TF) from the inner layer of the prepuce was developed over the dartos pedicle (Fig. 3). Circumpenile skin was used in patients who were circumcised. After splitting the flap ventrally (Fig. 5), the pedicle was mobilized proximally to an extent that allows ventral transposition of the flap without tension (Fig. 4). Utmost care was taken to preserve the vascularity of the flap. From a technical point of view we tried to keep the dartos pedicle as thick as possible. Three patients had skip strictured areas with normal lumen in between. In these patients, the flap was bivalved for substitution.

image

Figure 5. Transposition of the flap through the scrotal tunnel.

Download figure to PowerPoint

For a proximal (bulbar) repair, the flap was passed through a scrotal tunnel to the bulb without torsion and without placing excessive tension on the pedicle. For DOF repair, the fasciocutaneous flap was then brought on to the exposed dorsally opened portion of the urethra and sutured to adjoining edges of the urethra (Fig. 6) using continuous sutures over a 16-F Foley catheter, as we have previously described [12].

image

Figure 6. Suturing of the flap to the adjoining edges of the urethra.

Download figure to PowerPoint

For tube substitution, the flap was tubularized around a 16F Foley catheter. The distal and proximal ends of the native urethra were adequately spatulated ventrally and the TF was thus anastomosed to the native urethra. The suture line was then covered with the second layer, usually the dartos fascia or the tunica vaginalis harvested from the ipsilateral testis [13]. Foley catheter and suprapubic catheter, if placed preoperatively, were left indwelling for 3 weeks.

After 3 weeks a per urethral catheter dye study was carried out. If there was no significant extravasation, the perurethral catheter and suprapubic catheter were removed on the same day. Any transient extravasation was managed by extending the period of catheterization. Follow-up consisted of uroflowmetry and urethral calibration with a 16-F Foley catheter at 1 month, and at 3-monthly intervals for the first year and 6-monthly intervals thereafter. Contrast studies were performed as and when required, depending on the uroflowmetry and calibration findings. The incidence of fistula, penile skin complications, post-void dribbling and recurrence were recorded. Failure was defined as the need for any subsequent urethral procedure (internal urethrotomy, urethral dilatation or urethroplasty).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

The mean age, stricture length and proportion of pendulous and bulbar strictures were similar in the three groups (Table 1). The incidence of penobulbar stricture was higher in Group 3 than in the other two groups. The cause was unknown in 20 (50%), traumatic in 10 (25%) and iatrogenic in 10 (25%) patients in Group 1, unknown in 20 (50%), traumatic in eight (20%) and iatrogenic in 12 (30%) patients in Group 2. In Group 3 the cause was iatrogenic in 16 (26.67%), traumatic in 16 (26.67%) and unknown in 28 (46.67%) patients. All patients had a history of urethral intervention in the form of dilatations, visual internal urethrotomy (VIU) or anastomotic urethroplasty (Table 1). The median follow-up was 42.3, 41.8 and 46.4 months in Groups 1, 2 and 3, respectively.

The mean operating time was 3 h for Groups 1 and 3 and 4 h for the TF repair. This was expected because flap tubularization has a long suture line and is more technically demanding than DOF repair. There was no significant intra-operative blood loss in any group. Postoperative complications are shown in Table 2. Immediate postoperative complications included donor site penile haematoma in four patients, and two each in Groups 1 and 2; all required surgical evacuation and compressive dressing and healed uneventfully.

Table 2. Outcomes for all groups.
 Group 1: transverse preputial DOFGroup 2: TFGroup 3: circumpenile DOF
  1. IIEF, International Index of Erectile Function.

Median follow-up, months42.341.846.4
Patients completing follow-up at 5 years, n (%)30 (75)28 (70)40 (76.67)
Recurrence at 1 year, n (%)4 (10)6 (15)4 (6.67)
Primary success rate at 1 year (%)908593.3
Recurrence at 3 years2 (5)4 (10)4 (6.67)
Primary success rate at 3 years, %857586.66
Recurrence at 5 years, n000
Recurrent stricture needing only a single VIU/dilatation4 (10)4 (10.0)4 (6.67)
Secondary success rate, %5106.77
Overall success rate, %908593.3
Mean (sd) IIEF erectile function domain score24 (5.87)24.1 (4.96)23.78 (6.59)
Mean (sd) IIEF orgasmic function domain score7.1 (2.36)7.64 (4.08)6.9 (4.94)

Eight (20%) patients in Group 1, eight (20%) in Group 2 and 10 (16.67%) in Group 3 had donor site superficial penile skin necrosis which healed within 6 weeks of surgery without scarring, deformity or any need of subsequent procedure. Two (5%) patients in Group 1 and two (3.33%) in Group 3 had extensive skin loss and required split-skin grafting. Two (5%) patients developed a urethrocutaneous fistula in Group 1 at the site of distal anastomosis and underwent successful fistula repair 6 months after initial surgery. Four (10%) patients in Group 1, two (5%) in Group 2 and two (3.3%) in Group 3 had penile torsion of <60°. Six (15%) patients in Group 1, four (10%) in Group 2 and eight (13.3%) in Group 3 had penile hypoaesthesia, which did not interfere with sexual function. Three weeks after surgery, four (10%) patients in Group 2 had mild extravasation of dye at the proximal anastomotic site on per urethral study, and required an additional week of urinary diversion. Two neourethral diverticula were seen on urethrogram in Group 2 (5%). The diverticula were asymptomatic, relatively small and required no surgical intervention. Eight (20%) patients in Group 1, 10 (25%) patients in Group 2 and 12 (20%) in Group 3 had some degree of post-void dribbling. This symptom was considered bothersome if patients needed manual urethral compression after voiding to empty residual urine (Table 2).

In follow-up, recurrent strictures were found in four (10%) patients in Group 1, six (15%) patients in Group 2 and four (6.67%) in Group 3. All of them presented with worsening of stream. At 3-year follow-up two more patients in Group 1, four in Group 2 and four in Group 3 had recurrences. Fourteen of the 24 recurrences were within the first 6 months of surgery (mean 4.5 months). Five more recurred at a mean of 1 year. The initial success rates of 90, 85 and 93.3% were achieved in Groups 1, 2 and 3, respectively; these dropped to 85, 75 and 86.7%, respectively at 3 years. All the failures happened in the first 18 months with no new recurrences noted in the subsequent years. Of the patients in Groups 1, 2 and 3, respectively, 75, 70 and 76.7% completed > 5 years of follow-up with predicted success rates almost equivalent to those at 3 years, as there were no new recurrences after 1.5 years. In Group 2, strictures usually occurred at the proximal anastomostic site. Twelve out the 24 (50%) recurrences stabilized after the first attempt of either VIU or dilatation, followed by weekly calibration. The secondary success rate for Groups 1, 2 and 3 were 5%, 10% and 6.77%, respectively; therefore the overall success rates were 90, 85 and 93.3% (Table 3). Twenty recurrences were at the anastomotic sites, either proximal, distal or both. Three had re-stenosis of the entire graft needing layopen with perineal urethroplasty. Five patients needed end-to-end urethroplasty for stricture at the proximal anastomotic site. In these cases, an advancement of the flap was carried out, requiring only the removal of the strictured site and not the whole of the flap. All five patients were doing well with a mean follow-up of 2 years after the second urethroplasty.

Table 3. Management of stricture recurrence
 Single VIUSingle dilatationMore than one VIU/dilatationEnd-to-end urethroplastyPerineal urethroplasty
Group 131011
Group 222321
Group 322121

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Most authors have reported identical success rates for both the free grafts and fasciocutaneous flaps when used as an onlay, and generally poorer results with the TFs [5–9,14–17]. In a retrospective analysis of 63 patients, Barbagli et al. [16] reported a lower success rate with a pedicle skin penile flap (67%) as compared with free grafts (80%), oral mucosal grafts (82%) and skin grafts (78%). In these studies the most frequent complications requiring intervention were graft contracture and urethral stricture. Long follow-up intervals are critical for determining the continued reliability of non-genital skin for stricture repairs, despite encouraging short- and mid-term results. According to Metro et al.[18], late urethral strictures are a significant long-term risk for substitution urethroplasty using non-genital skin. The advent of genital/extragenital grafts, especially the buccal mucosa graft (BMG), as the ‘gold standard’ treatment for long anterior stricture urethra places into question the use of a preputial flap. Interim and long-term results of penile skin graft urethroplasty show a declining success rate (73% and 66%, respectively) [19,20]. BMGs, which are now being considered as the graft of choice, provide good interim results in various series but many of these have been reviews rather than original articles and, in many, the main site of stricture has been bulbar. In addition, the mean stricture length in these studies is 4.8 cm with short-to-interim follow-up periods suggesting selection bias. The combined mean success rates for various series on ventral and dorsal buccal mucosa onlay grafts for bulbar and penile stricture range from 77.5 to 98% after more than 3 years of follow-up [5,21–34]. The success rate for one-stage buccal mucosa urethroplasty for penile strictures is 85%. In the present study, we obtained the best results for bulbar onlay stricture and slightly poorer results for the penobulbar stricture. Many of the penobulbar strictures were relatively longer with poorer urethral plates. Longer stricture requires a longer flap, which places the vascularity of the flap at risk and is also suggestive of more aggressive disease. The use of dorsal BMGs in two stages for long complex anterior strictures has not much better results, with a success rate of 87% at 3 years [17] and with further costs for an intervening secondary procedure. Many authors believe that a two-stage procedure becomes more like a three-stage procedure. With judicious use of a single VIU, Whitson et al. [17] achieved a success rate of 79% at a mean follow-up of 10 years, with their fasciocutaneous circumpenile flap, although their study did suggest that patients who had a previous history of urethroplasty or hypospadias repair were at more risk of failure.

The reported incidence of stricture after preputial flap urethroplasty in different series is 5–33% [6,12,15–17,35–37] for DOFs and 30–60% for TFs [14,35]. The high recurrence of stricture when TFs are used can be attributed to a poor quality urethral plate as well as the two anastomotic lines proximally and distally, coupled with an extensive suture line for creation of the tube. We believe that a single judicious use of VIU or dilatation should not be considered as a failure as patients in our study had a long strictures with history of some procedure performed beforehand. In the present study, 50% of recurrences stabilized after the first attempt of either VIU or dilatation. The overall success rate therefore for Groups 1, 2 and 3 are 90, 85 and 93.33%, respectively. All the failures happened in the first 18 months with stabilization in the subsequent years. Totals of 75, 70 and 76.7% of the patients in Groups 1, 2 and 3, respectively, completed more than 5 years of follow-up with predicted success rates almost equivalent to those at 3 years. In the present series stricture recurrences were also more common with TF than with DOF repairs, but were fewer than in other studies. This higher success rate with TFs might be attributable to good technique, meticulous tissue handling, good haemostasis, use of bipolar cautery and the use of tunica vaginalis or dartos fascia as a second layer closure over the TF.

In the literature, the rate of occurrence of urethrocutaneous fistula ranges from 0 to 13% [5,12,15,35,36], and is higher for TFs. By contrast, in the present study, only two patients (5%) developed a urethrocutaneous fistula in Group 1 and there were none in Groups 2 or 3. The non-occurrence of urethrocutaneous fistula with TFs may again be attributed to second line closure with tunica vaginalis or dartos fascia.

The incidence of pseudodiverticular formation in flap repair ranges from 0 to 5% [5,12,17,35,36] in various series. In the present series, diverticula did not develop in Groups 1 and 3 but they were noted in 5% cases in the TF group (Group 2). In addition, there was a lower incidence of bothersome post-void dribbling in both the DOF repair groups. Some degree of post-void dribbling is inherent with any form of substitution urethroplasty. The flap in dorsal onlay requires limited spongiosal opening and is smaller and tacked down to the urethral plate on both sides, as opposed to TFs and ventral onlay, which are not supported by the urethral plate and therefore more prone to sacculation. In this respect, DOFs are more advantageous than ventral placement.

Necrosis of the penile skin proximal to the flap results when the vascular supply of the sub-dermal plexus is compromised and its incidence varies between 2 and 27% [5,12,15,17,19,35,36]. Although minor in most cases, it can result in delayed healing and prolonged hospitalization. This is an inherent disadvantage of any pedicle penile skin flap, although in experienced hands its incidence is lower [35].. In the present study, the incidence of superficial penile skin necrosis in all groups was between 16.6 and 20%.From a technical point of view we tried to keep the dartos pedicle thick and if we had to err we did it on the skin side.

The reported incidence of penile hypoaesthesia after harvesting of a preputial flap in various studies is ∼11% [15,17,35], and for penile torsion it is ∼13% [5,15,36]. In the present patient cohort the incidence was similar for each of these complications. The torsion was <60° and did not interfere with erection and intercourse. There was no significant difference in the pre- and postoperative erectile and orgasmic function scores. As suggested by McAninch [3], to prevent anaesthesia in the glans and distal foreskin, injury to the dorsal neurovascular bundles must be avoided when dissecting the dorsal lamella of Buck's fascia off of these structures, whereas dissecting the flap up to the root of the penis eliminates the risk of penile torsion.

In the present study, 30 patients had considerably damaged oral mucosa owing to the use of tobacco and other substances that irritate the oral mucosa, and so making it unsuitable for substitution. There was a significant amount of sub-mucosal fibrosis and, as is already known, the alkaloid and tannin content of Areca nuts (betel nuts) are responsible for fibrosis [38]. In South-East Asia, as tobacco chewing habits are very prevalent, so buccal mucosa is often not a suitable candidate for substitution. The prepuce, therefore, represents a very good alternative to BMG with good and durable mid-term and long-term results.

The present results show the mid-term reliability of the preputial flap, even for patients with unsuitable oral mucosa, and in our practice each procedure is implemented in different clinical scenarios as described. Cosmetically, the circumpenile flaps leave only a circumcision scar. Also penile torsion is rarely seen if the flap has been mobilized adequately and, even if it occurs, it rarely hinders sexual intercourse. The superficial necrosis of the skin is rarely severe and usually heals with conservative management alone. The complications of a preputial flap are similar to those of a BMG, which include pain, numbness and difficulty in opening the mouth and, rarely, stenson duct injury. Futhermore, in the present study, the results of urethroplasty using preputial skin or penile skin in circumcised patients were similar to each other.

Any substitution urethroplasty deteriorates over time [39]. As the mean follow-up in the present series is a little over 6 months, long-term follow-up is necessary to provide more precise analysis of the longevity of the procedure; however, in the present series ∼75% of patients completed 5 years of follow-up with no additional recurrence after initial 18 months. Dorsal placement of the flap could be the reason for the lower number of failures over extended periods of follow-up.

In conclusion, unlike substitution urethroplasty using non-genital skin, vascularized flaps of genital and preputial skin appear to be unique, in that they have relatively lower long-term stricture rates and are easy to harvest. Preputial/penile flap repair provides excellent long-term cosmetic and functional results. TF repair appears to result in a higher incidence of stricture recurrence, but is nevertheless a good alternative when combined judiciously with secondary procedures, enabling single-stage reconstruction in the majority of cases where there is a compromised urethra. In our experience, skin flaps based on preputial/penile skin clearly demonstrate exceptional versatility with good mid-term results for patients with complex anterior urethral stricture, including those who had unsuitable oral mucosa.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES