To define types of support failure after anterior compartment mesh placement and to investigate any association with predictors of recurrence.
To define types of support failure after anterior compartment mesh placement and to investigate any association with predictors of recurrence.
This was a retrospective study on patients ≥ 3 months after anterior mesh placement. All patients underwent a standardized interview, clinical examination using the International Continence Society Pelvic Organ Prolapse Quantification system (ICS POP-Q) and three-dimensional/four-dimensional (3D/4D) translabial ultrasound. Mesh failure was defined as recurrence of anterior/central compartment prolapse seen on ultrasound. Failures were classified as anterior, global and apical. Their association with hiatal area on Valsalva maneuver and levator avulsion was tested.
Three hundred and one patients were seen initially, of whom five were excluded because of missing data, leaving 296. Mean follow-up was 1.8 (range, 0.3–5.6) years. Mean age was 65 (range, 32–88) years. One hundred and thirty-nine (47%) women were fitted with a PerigeeTM mesh, 66 (22%) with an Anterior ProliftTM mesh and 91 (31%) with an Anterior ElevateTM mesh. Recurrent symptoms (lump/drag) were reported in 65 (22%), a recurrent cystocele was noted in 128 clinically (43%) and in 105 on ultrasound (35%). Avulsion was diagnosed in 117 patients (40%). Mean hiatal area on Valsalva was 33.3 (range, 14.1–60.0) cm2. Mesh failure was diagnosed in 112 patients (38%), comprising global failure in 81 (27%), apical failure in 23 (8%) and anterior failure in eight (3%). Apical and global failures were significantly associated with hiatal area, associations that remained after controlling for potential confounders.
Mesh failure, i.e. anterior or central compartment recurrent prolapse, was noted in 38% of patients on average 1.8 years after placement of anterior compartment mesh. Global and apical failures together constituted 93% of all mesh failures, both types of failure being significantly associated with hiatal area on Valsalva maneuver. Copyright © 2013 ISUOG. Published by John Wiley & Sons Ltd.
Cystocele is a challenging condition in pelvic reconstructive surgery, with a reported failure rate after anterior colporrhaphy of up to 40–63%[1-4]. To improve the surgical outcome of cystocele repair, there is a trend towards the use of synthetic implants to augment it. This has been shown to be effective in reducing recurrence[3, 5-8], even in women with levator avulsion, which appears to be a strong risk factor for prolapse recurrence[2, 9-11]. However, recurrence is still common after vaginal mesh interposition. The problem may lie partly in obtaining a strong attachment point for mesh anchorage. Failure of mesh anchoring structures can result in unusual, hitherto undocumented anatomical situations, leaving surgeons with unprecedented challenges in clinical management. Understanding the mechanisms of mesh-anchor failure may allow clinicians to treat recurrence more effectively and facilitate the optimization of implant design.
Modern polypropylene mesh implants in the anterior vaginal wall are highly visible on ultrasound, unlike on X-ray and magnetic resonance imaging, and can be reliably identified on three-dimensional/four-dimensional (3D/4D) ultrasound. Anterior compartment mesh is apparent on ultrasound as a highly echogenic linear structure, usually situated dorsal to the bladder neck, caudal and dorsal to the trigone and the posterior bladder wall. It is usually more clearly visible on Valsalva maneuver, rotating around the fulcrum of the symphysis pubis. It has previously been noted that several different types of mesh failure can be seen in women with recurrent anterior and central compartment prolapse, suggesting failure of distinct anchoring mechanisms, that is, dislodgement of anchoring structures such as transobturator tapes and mesh anchors.
The aims of this study were to define different types of mesh failure, to describe how commonly they occur and to determine any potential association with levator avulsion and hiatal area in patients after anterior mesh repair with/without concomitant prolapse repair in other compartments, as both avulsion and increased hiatal area have been shown to be significant risk factors for prolapse recurrence[10, 14].
This was a retrospective study using data obtained in four surgical audits (three external and one internal audit) of anterior compartment mesh approved by local Human Research Ethics Committees (SWAHS HREC 07-063, Greenslopes Private Hospital HREC 10-09; Townsville HREC 84/04; Victoria HREC No. 10310Q). Anterior compartment mesh was implanted by or under the direct supervision of five subspecialist surgeons at four tertiary urogynecological units. Patients with an anterior mesh repair undertaken between February 2004 and August 2011 were invited to a follow-up appointment at least 3 months after the procedure. The appointment included an in-house, non-validated questionnaire, clinical examination using the International Continence Society Pelvic Organ Prolapse Quantification system (ICS POP-Q) and 3D/4D translabial ultrasound examination using a GE Voluson 730, a GE Voluson 730 Expert or a Voluson I system with 8–4-MHz curved array volume transducer (GE Medical Systems, Zipf, Austria). Translabial ultrasound was performed in the supine position after voiding at rest, on Valsalva maneuver and on pelvic floor muscle contraction (PFMC), as previously described. At least three ultrasound volumes on Valsalva were acquired, and the volume on maximum Valsalva, i.e. the volume showing the most marked pelvic organ descent, was used for analysis. Offline analysis of ultrasound volumes was performed on a desktop personal computer using the proprietary software GE Kretz 4D View version 10.0 (GE Medical Systems), by an operator blinded to all clinical data. Cystocele recurrence on clinical examination was defined as ICS POP-Q ≥ Stage 2 and on ultrasound as a bladder that descended ≥ 10 mm below the pubis on Valsalva, which has been suggested as the most appropriate cut-off to define a significant cystocele on ultrasound, using receiver–operating characteristics curve statistics. Levator integrity was assessed on PFMC using tomographic ultrasound imaging or multislice ultrasound imaging. Avulsion was diagnosed if the muscle insertion was abnormal in the reference slice and within 5 mm cranial to this reference slice, as described previously. Hiatal area on Valsalva was determined in the axial plane, as described previously. In patients with prolapse recurrence on ultrasound, mesh failure fell into three distinct categories: (1) anterior failure: cystocele recurred ventral and caudal to a well-supported mesh (Figure 1); (2) apical failure: cystocele/anterior enterocele/uterine prolapse occurred dorsal to the mesh, with high mobility of cranial mesh aspects (Figure 2); and (3) global failure: prolapse recurred in association with high mobility of the entire mesh on Valsalva maneuver (Figure 3).
A test–retest series (n = 50) on mesh failure typing was performed by K.L.S. and V.W. to determine the interobserver repeatability of this visual classification. Statistical analysis was carried out using Minitab V13 (Minitab, State College, PA, USA), Cohen's kappa being used to determine interobserver agreement. Normality was assessed visually and checked using histograms and the Kolmogorov–Smirnov method. The t-test was used for continuous normally distributed data and the chi-square test for categorical variables. Binary regression analysis was performed to control for confounders of mesh failure. P < 0.05 was considered statistically significant.
Four hundred and sixty-eight patients had an anterior mesh repair during the inclusion periods of the audits. Three hundred and one patients (64%) attended an audit appointment. Five were excluded because of missing data, leaving 296 patients for the study. The mean follow-up interval was 1.8 (range, 0.3–5.6) years. One hundred and thirty-nine women (47%) had a PerigeeTM mesh fitted (American Medical Systems, Minetonka, MN, USA), 66 (22%) an Anterior ProliftTM (Ethicon, Somerville, NJ, USA) and 91 (31%) an Anterior ElevateTM (American Medical Systems). The mean age was 65 (range, 32–88) years and the mean body mass index (BMI) was 28 (range, 17–49) kg/m2. One hundred and forty-eight women (50%) had had a hysterectomy and 111 (37%) incontinence or prolapse surgery. Concomitant procedures were carried out in 217 patients (73%), comprising hysterectomy in 35, sacrospinous colpopexy in 27, sacrospinous hysteropexy in seven, Apogee in 24, posterior vaginal repair in 103, Posterior Elevate in seven, Posterior Prolift in 49 and a suburethral sling in 94. Two hundred and forty-two patients (82%) were satisfied with the outcome and 275 (93%) considered themselves cured/improved. Nineteen patients (6%) reported pain at interview, and current mesh exposure or a history of previously treated mesh exposure was noted in 28 cases (9%). Recurrent prolapse symptoms (lump/drag) were reported in 65 cases (22%). Median Ba was −2 (mean −1.6, range −3 to +2) and median C was −5 (mean −4.8, range −9 to +7). A recurrent cystocele was noted in 128 patients (43%) on clinical examination. On ultrasound, mean cystocele descent was 5.1 mm below the pubis (range, +21.8 to −44 mm; a negative value signifies descent below the pubic bone). Avulsion was diagnosed in 117 patients (40%), on the right side only in 39, the left side only in 16 and bilaterally in 62. Mean hiatal area on Valsalva was 33.3 (range 14.1–60.0) cm2. Mesh failure, that is, anterior or central compartment prolapse recurrence, was diagnosed in 112 women (38%). It was associated with recurrent cystocele in 107 (36%), uterine prolapse in four and an enterocele in one case. In a test–retest series (n = 50) on mesh failure typing, a kappa of 0.75 (95% CI, 0.60–0.91) was obtained between two observers, signifying substantial agreement. According to our classification of mesh-anchor failure, global failure was noted in 81 (27%), apical failure in 23 (8%) and anterior failure in eight patients (3%). Mesh failure was strongly associated with the recurrence of clinical prolapse in the anterior or central compartment (P < 0.001).
Clinical cystocele recurrence was associated with levator avulsion (51.3 vs 38.8%, P = 0.035) and hiatal area on Valsalva (34.34 vs 32.42 cm2), but the latter association did not reach significance (P = 0.06). Cystocele recurrence on ultrasound was significantly associated with both hiatal area on Valsalva (37.43 vs 30.8 cm2, P < 0.001) and levator avulsion (45 vs 31%, P = 0.014). The type of mesh failure was variably associated with these parameters (Table 1). On univariate analysis for predictors of either apical or global failure including age, BMI, vaginal parity, follow-up interval, levator avulsion, hiatal area on Valsalva, history of incontinence/vaginal repair, history of hysterectomy, concomitant surgery and type of anterior compartment mesh used, only follow-up interval, avulsion, hiatal area on Valsalva and mesh type were significant. On multivariate regression analysis, however, only the latter two parameters remained significant (Table 2), with each cm2 of hiatal area increasing the likelihood of recurrence by 13%, and with the Anterior Elevate most likely to result in anchoring failure.
|Type of mesh failure||Hiatal area on Valsalva (cm2)||Pa b||Levator avulsion (%)||Pa c|
|No mesh failure (n = 184)||30.81 ± 7.73 (n = 181)||34 (n = 184)|
|Any mesh failure (n = 112)||37.38 ± 8.46 (n = 109)||< 0.001||48 (n = 111)||0.017|
|Global failure (n = 81)||37.26 ± 8.31 (n = 79)||< 0.001||44 (n = 81)||0.114|
|Apical failure (n = 23)||41.01 ± 7.83 (n = 22)||< 0.001||65 (n = 22)||0.004|
|Anterior failure (n = 8)||28.59 ± 4.52 (n = 8)||0.226||38 (n = 8)||0.849|
|Variable||Odds ratio (95% CI)||P|
|Levator avulsion||1.18 (0.66–2.11)||0.6|
|Hiatal area on Valsalva (cm2)||1.13 (1.09–1.18)||< 0.0001|
|Follow-up interval (years)||1.23 (0.96–1.56)||0.1|
|Mesh type||< 0.0001|
|Perigee vs Anterior Prolift||2.60 (1.06–6.38)|
|Anterior Elevate vs Anterior Prolift||12.61 (4.54–35.01)|
The use of transvaginal mesh has increased substantially over the last decade, mainly owing to the development of widely marketed mesh kits employing transobturator or pelvic sidewall fixation with tapes or plastic anchors. In 2010 around 75 000 POP repairs were performed using transvaginal mesh in the USA alone. The increasing popularity of transvaginal mesh has led to growing concerns regarding mesh-related complications. In 2011 the US Food and Drug Administration issued a safety communication to alert clinicians to mesh complications. While mesh implants are not innocuous and complications such as chronic pain and vaginal mesh exposure do occur, mesh implants have been shown to be effective in reducing prolapse recurrence[3, 5, 6], especially in patients with levator avulsion. According to Rodrigo et al. the risk of recurrence in individuals has been shown to vary from 10 to 90%, depending on the state of the levator ani muscle and levator hiatal area. Clearly the balance between risks and benefits of mesh use varies from individual to individual and should be considered before mesh implantation is contemplated. However, even after mesh implantation, there remain a substantial number of patients in whom prolapse recurs. This implies that the implant is insufficiently load resistant and has failed in its primary purpose, i.e. to provide support when native tissue repair would be insufficiently load resistant. Understanding mechanisms of mesh support or anchoring failure may help to improve mesh design and performance and hence alter the risk–benefit balance of mesh use in patients at high risk of recurrence. In addition, patients with prolapse recurrence after transvaginal mesh use constitute a difficult group of patients, who confront clinicians with novel anatomical situations and equally new surgical challenges.
In this study on 296 women after anterior compartment mesh, 38% were diagnosed with mesh anchoring failure. Over two-thirds were ‘global failures’, consistent with dislodgement of all anchoring structures (that is, no remaining iatrogenic support), and about a fifth showed apical failure, implying dislodgement of the apical and/or superior lateral attachments. In the latter case, the inferior lateral attachments, usually close to the bladder neck, are still patent, resulting in appearances reminiscent of a Burch colposuspension with high cystocele. Apical failure can be dealt with by apical resuspension of the mesh to the sacrospinous ligament with a vaginal sacrocolpopexy or to the anterior longitudinal ligament of the spine in abdominal/laparoscopic colpopexy. Global failure constitutes a greater management problem, as even a highly effective apical suspension may not always result in sufficient Level II support owing to a lack of lateral attachment.
Both these commonest types of mesh failure were significantly associated with hiatal area on Valsalva maneuver. It is plausible that a larger hiatus subjects anchoring structures to greater loads, increasing the risk of dislodgement and failure. Future studies should focus on implant design to make anchoring structures more load resistant on the one hand, and on means to normalize levator morphobiometry on the other hand, in order to reduce the loads placed on anchoring structures during wound healing before the formation of permanent scar tissue. A recent pilot study on a surgical procedure designed to reduce hiatal dimensions has shown a mean reduction in hiatal area on Valsalva by at least 12 cm , suggesting that the latter approach is feasible. The risk profile and effectiveness of such procedures in reducing recurrence rates after prolapse surgery, however, remain to be determined.
Apart from global and apical failures, a small minority of mesh failure was caused by anterior support failure, in which a cystocele recurred anterior to a well-supported mesh implant. In contrast to global and apical failures, anterior failure was not associated with hiatal area on Valsalva maneuver or levator avulsion. It is probably due to suboptimal surgical technique, in that the caudal end of the implant was insufficiently secured to the bladder neck, leading to dislodgement of the mesh from the bladder base. Suturing the mesh implant back to the bladder neck seems to be the appropriate management in symptomatic patients. As in the case of other forms of recurrence, the ‘failed’ mesh can be used as an asset, provided that the surgeon is aware of the individual anatomical situation.
As far as we are aware this is the first attempt to define types of support failure after anterior compartment mesh placement and to define prevalence. The study was conducted on a large cohort of almost 300 patients with a mean follow-up interval of 1.8 years. Several weaknesses, however, should be acknowledged. The retrospective design is a limitation, and patients were seen in the context of four different (if methodologically identical) surgical audit projects. Preoperative POP-Q staging was not available for all patients. Furthermore, owing to limited equipment and staff availability, and the fact that many patients lived several hours from the location of the audits in Sydney, Brisbane and Townsville, we were able to see fewer than two-thirds of the original patients. This may have introduced a selection bias, and our findings may not truly reflect the prevalence of mesh failure in those populations.
The heterogeneity of the study populations may also be seen as a disadvantage. We saw patients after placement of three different anterior compartment meshes, which employ different means of anchoring, and the procedures had been performed by five senior surgeons and their trainees. Mesh type did influence the likelihood of recurrence, with the Prolift mesh the least likely to be associated with mesh anchoring failure and recurrence (Table 2). Our study was not large enough to differentiate type of failure according to type of mesh, and this issue may have to be examined using larger datasets. All three types of mesh anchoring failure observed by us occurred with all three meshes included in the study. Notwithstanding the comment above, the heterogeneity of our population may even be considered a strength rather than a weakness, as it might suggest that our results are more generally applicable.
Finally, some may regard the mean follow-up of 1.8 years as too short for a reliable assessment of long-term recurrence rates. However, length of follow-up was not a significant predictor of recurrence, an observation that parallels the findings of several other of our surgical projects. In general we have found that the prevalence of recurrent prolapse does not substantially rise after the first postoperative year.
In conclusion, in a cohort of 296 patients seen on average 1.8 years after anterior compartment mesh repair, we have documented anterior/central compartment recurrence in 38%. The majority were due to global and apical mesh anchoring failure, implying dislodgement of the apical and/or lateral attachments. A small minority were due to anterior failure, that is, dislodgement of the mesh from the bladder base, probably owing to suboptimal surgical technique. Global mesh failure and apical mesh failure were significantly associated with hiatal area.
There is a clear need for improvements in implant design, especially as regards the load-bearing capacity of mesh-anchoring structures, which seem to be insufficient in a substantial number of women receiving the three tested implants. In addition, it may be possible to develop means of normalizing hiatal dimensions in order to reduce the likelihood of fixation failure in women with abnormal levator morphology.
H.P. Dietz has received honoraria as a speaker for GE Medical, Astellas and AMS and has in the past acted as consultant for CCS, AMS and Materna Inc. He has received equipment loans from Toshiba and Bruel and Kjaer. H.P. Dietz and K.L. Shek have received unrestricted education grants from GE Medical. K.L. Shek has acted as consultant for Materna Inc. A. Rane is a consultant and preceptor for AMS. J. Lee is a preceptor for AMS and has received external research grant for a sling RCT, unrelated directly to this paper. A. Rosamilia and J. Lee are participating in investigator-led, company-supported studies with AMS and Boston Scientific and received honoriara from these and Astella. The other authors have no financial disclaimers or conflicts of interest to declare.