Clinical Application of IUGA/ICS Classification System for Mesh Erosion

Abstract

Aims

Our aim was to assess the usability of the IUGA/ICS classification system for mesh erosion in a tertiary clinical practice and to determine if assigned classification is associated with patient symptoms, treatment, and outcome.

Methods

We retrospectively identified women who had mesh erosion after prolapse or incontinence surgery. Each erosion was classified using the IUGA/ICS category time site (CTS) system. Associations between classification and presenting symptom (asymptomatic, pain, bleeding, voiding, or defecatory dysfunction, infection, prolapse), treatment type, and outcome were evaluated with chi-squared test, student’s t-test, and univariate logistic regression.

Results

We identified 74 subjects with mesh erosion; only 70% were classifiable. Asymptomatic patients (n = 19) (Category A) were more likely to be managed conservatively (P = 0.001). Symptomatic patients (n = 55) (Category B) were more likely to be managed surgically (P = 0.003). Other variables had no association with treatment. No variables were associated with outcome. Presenting symptom was associated with both treatment (P = 0.005) and outcome (P = 0.03). Asymptomatic subjects were more likely to have satisfactory outcome (P = 0.03). Urinary frequency and urgency were highly correlated with surgical management (P = 0.02).

Conclusions

One third of mesh erosions could not be retrospectively coded using the IUGA/ICS classification. The components of the system were not predictive of treatment nor outcome with exception of the Category A (asymptomatic) and Category B (symptomatic). Asymptomatic women with mesh erosion can be successfully managed with conservative measures. Use of a classification system may be enhanced if the system is simplified by limiting the number of variables to those associated with interventions and patient outcome.

INTRODUCTION

The prevalence of symptomatic pelvic floor disorders is 25% in the U.S.¹ The lifetime risk of undergoing surgery for prolapse or incontinence is 20%² and the risk of reoperation is 29%.³ The use of synthetic mesh was introduced to pelvic reconstructive surgery to decrease the rate of recurrence after surgical repair.⁴ However, there are unique complications associated with use of mesh, the most common being vaginal mesh erosion.⁵ The overall risk of erosion is 10% in patients who undergo prolapse repair with mesh^6–8 and 0.4–3.0% after anti-incontinence procedures with mesh.^9,10

A joint commission of the International Urogynecologic Association (IUGA) and the International Continence Society (ICS) published a classification system of mesh complications in January 2011.¹¹ The purpose of the classification system is to develop a common language and increase awareness of complications from procedures using prostheses and grafts in female pelvic surgery. A secondary aim of the classification system is to develop a registry of complications to inform and guide surgeons, patients, and industry. One of the intended applications of the system is retrospective use in analysis of medical records and surgical audits.¹² The system has been called the Category, Time, and Site (CTS) Classification. The CTS Classification seeks to encompass a vast range of clinical scenarios into a numerical and ordinal code so that no additional descriptors are necessary. The first domain—Category—describes the involved organ and size of the eroded mesh, the severity of the complication, and pain grade. The second domain—Time—refers to the interval between the time of the index surgery and the presentation of the complication. The third domain—Site—refers the location of the complication (see Fig. 1).

Fig. 1.

International Continence Society/International Urogynecological Association classification system.

The primary objective of this study was to assess the usability of the current classification system in an academic clinical practice. The secondary aim was to determine if the IUGA/ICS system for mesh-related complications is associated with treatment and outcome for mesh erosion. We hypothesized that the IUGA/ICS classification system would be cumbersome to use but would be associated with treatment and outcome.

MATERIALS AND METHODS

This IRB approved study is a retrospective cohort analysis of mesh erosion in patients who presented to a tertiary care referral center after surgery for pelvic organ prolapse or urinary incontinence with synthetic mesh (index surgery). There are numerous complications that can result from graft placement, however, we elected to look at mesh erosion rather than all possible complications given that there is a clear causal relationship between mesh placement and erosion.

Patients were included if they were evaluated during a 5-year period from 2007 to 2012. Subjects were identified using International Classification of Diseases, Ninth Revision (ICD-9) codes 996.30, 996.65, and 996.76 for mechanical complication of unspecified genitourinary device, implant, and graft; infection and inflammatory reaction due to other genitourinary device, implant, and graft; and other complications due to genitourinary device, implant, and graft, respectively. Subjects were included if they had one of the above ICD-9 codes associated with the following procedures: synthetic sling for stress urinary incontinence, vaginal mesh placed using a mesh-kit, hand-cut vaginal mesh, sacrocolpopexy, or sacrocervicopexy using synthetic mesh. Subjects were excluded if they had a complication presumed to be associated with mesh (e.g., isolated pain) in the absence of mesh erosion. Based on the newly proposed definitions from IUGA and ICS, we included the expanded definitions of mesh erosion including contraction, prominence, separation, exposure, extrusion, perforation, dehiscence, and sinus tract formation.¹¹

The medical records of eligible subjects were reviewed and demographic data, medical history, information about the index surgery, complications after mesh placement, presenting symptoms associated with mesh erosion, and treatment of the mesh erosion were abstracted.

The subjective complaints of patients with erosion were collected and grouped into the following categories: asymptomatic, pain, bleeding, voiding, or defecatory dysfunction, infection, and prolapse. Treatments were grouped into conservative and surgical management. Conservative management was defined as observation, use of vaginal estrogen, or office excision of exposed vaginal mesh. Outcome categories included (1) resolution of erosion and all or most symptoms (i.e., patient satisfied with outcome) and (2) persistence of erosion or symptoms or additional surgery (i.e., patient dissatisfied with outcome). The subjects were assigned a code describing the mesh erosion based on the CTS classification system (see Fig. 1 for description of classification system).

Study data were managed using REDCap (Research Electronic Data Capture)¹³ tools hosted at Massachusetts General Hospital. Categorical and continuous variables were analyzed with chi-squared and student’s t-test, respectively. Additionally, univariate logistic regression was used to assess correlation between the CTS category, treatment, and outcome. Differences that resulted in probability values of <0.05 were considered statistically significant. Statistical analyses were performed using SPSS software (version 20; SPSS Inc., Chicago, IL).

RESULTS

Using encompassing ICD-9 codes, 149 patients were initially identified. Based on eligibility criteria, 74 subjects with mesh erosion were included in the study. The majority of patients initially identified by ICD-9 code query but excluded in the study cohort were patients who did not meet criteria of a qualifying procedure (e.g., “complication of unspecified genitourinary device” that upon review of chart was complication of an intrauterine device rather than complication from prolapse or incontinence surgery) or did not have mesh erosion (e.g., isolated pain that resulted from mesh placement). Baseline characteristics and surgical information are summarized in Tables I and II. The mean age at the time of index surgery was 59 ± 12 years and mean body mass index was 27 ± 5 kg/m². For 80% (n = 49) of subjects, the index surgery using mesh was their first urogynecologic procedure. Sixty-one percent (n = 45) had prolapse surgery with mesh and 39% (n = 29) had incontinence surgery with mesh. Eighty-one percent (n = 60) of subjects had type I mesh (totally macroporous with pores >75 μm), no subjects had type II mesh (totally microporous with pores ≤10 μm), 7% (n = 5) had type III mesh (microporous with macroporous components), and no subjects had type IV mesh (non-porous). Fifty-eight percent (n = 43) of subjects had their index surgery at the study institution and 42% (n = 31) were referrals from outside hospitals. In 11% (n = 8) of subjects, the operative report from the index surgery was unavailable.

TABLE I.

Patient Demographics (n = 74)

Variable	N (%)
Age (years)	59 ± 12
Body mass index (kg/m2)	27 ± 5
Parity	2 (range 0–5)
Menopause
Before menopause	12 (16)
After menopause	61 (82)
Unknown	1 (1)
Tobacco use
Never	64 (86)
Former	7 (9)
Current	2 (3)
Unknown	1 (1)
Medical comorbities
None (healthy)	66 (89)
Immunocompromised	1 (1)
Diabetes	1 (1)
Cancer	6 (8)
Surgical history
No previous pelvic surgery	39 (53)
Previous hysterectomy for indication other than POP	28 (38)
Previous POP/UI surgery	15 (20)
ASA physical status classification
Class 1	2 (4)
Class 2	45 (61)
Class 3	4 (5)
Class 4	0
Class 5	0
Class 6	0
Unknown	23 (31)
Presenting symptom
Asymptomatic	19 (26)
Pain	21 (28)
Bleeding	14 (19)
Urinary frequency and urgency	16 (22)
Recurrent prolapse	3 (4)
Infection	1 (1)

TABLE II.

Characteristics of Index Surgery (n = 74)

Variable	N (%)
Index surgery
Mesh used for UI surgery	29 (39)
Mesh used for POP surgery	45 (61)
Type of POP surgery with mesh
Vaginal	39 (53)
Anterior compartment	11 (15)
Posterior compartment only	0
Apical compartment only	2 (3)
Apical and anterior compartments	10 (14)
Apical and posterior compartments	9 (12)
Apical, anterior and posterior compartments	7 (9)
Abdominal	6 (8)
Sacrocolpopexy	6 (8)
Sacrocervicopexy	0
No POP surgery with mesh	29 (39)
Type of UI surgery with mesh
Retropubic minimally invasive sling	24 (32)
Transobturator minimally invasive sling	2 (3)
Other type of mesh sling	3 (4)
No UI surgery with mesh	45 (61)
Type of mesh for all surgeries
Type I	60 (81)
Type II	0
Type III	5 (7)
Type IV	0
Other	1 (1)
Unknown	8 (11)
Surgeon
Female pelvic medicine and reconstructive surgeon	52 (70)
Other	22 (30)
Type of anesthesia
General	43 (58)
Regional or MAC	11 (15)
Unknown	20 (27)

Of the 74 subjects with mesh erosions, 30% (n = 22) were unable to be categorized due to missing information that was not included in the clinical documentation or operative report (Table III). Nine percent (n = 7) of the records were missing two or more variables needed to satisfy the CTS system. Twenty percent (n = 15) of records did not include Category information. The most common omitted Category variable was size (20%, n = 15), followed by anatomic location (8%, n = 6), and pain grade (4%, n = 3). Seven percent (n = 5) of records did not include Time from surgery to clinical diagnosis. Sixteen percent (n = 12) of records did not include Site of mesh erosion. Of the 22 subjects with mesh erosions that were unable to be categorized, two were due to sacrocolpopexy, five were due to urinary incontinence procedures (all retropubic slings), and 15 were due to vaginal prolapse repair. The erosions that occurred following vaginal prolapse repair were scattered among all vaginal compartments (i.e., apical, anterior, and posterior compartments). Of the unclassifiable mesh erosions, 64% (n = 14) had their index surgery performed at our institution and 36% (n = 8) were referrals from outside hospitals.

TABLE III.

Missing Variables in Patients Unable to be Fully Indexed Using the CTS System (n = 22)

Category location/size (1–7) n = 10	Category severity (A–D) n = 2	Category pain grade (a–e) n = 3	Time (T1–T4) n = 5	Site (S1–S5) n = 12
xAa-T3-x*	3xc-x-S3*	2Bx-T2-S1	2Aa-x-S1	xAa-T3-x*
xAa-T4-S1	xxa-T4-x*	3Bx-x-S3*	3Bd-x-S1*	xAa-T4-x*
xAa-T4-S1		xBx-T3-x*	3Bx-x-S3*	xBx-T3-x*
xAa-T4-x*			3xc-x-S3*	1Be-T4-x
xBa-T2-S2			xBa-x-x*	2Aa-T4-x
xBa-T3-S2				2Aa-T4-x
xBa-T4-S2*				2Bc-T4-x
xBx-T3-x*				3Aa-T3-x
xBa-x-x*				3Aa-T4-x
xxa-T4-x*				3Bc-T2-x
				xBa-x-x*
				xxa-T4-x*

Missing variable is denoted as x.

^*Denotes codes with two or more missing variables (n = 7).

Seventy percent (n = 52) of records were able to be fully categorized using the CTS system (Table IV). The vast majority of mesh erosions in our series were vaginal, falling into Category A or B with varying sizes (Category 1, 2, or 3). There was one vaginal abscess (2DaT4S3) and one vaginal infection (3CbT2S1) from mesh erosion. There were five erosions involving the urinary tract (Category 4).

TABLE IV.

CTS Codes of Subjects (n = 52)

CATEGORY	A ASYMPTOMATIC (n=20)			B SYMPTOMATIC (n=29)			C INFECTION (n=2)			D ABSCESS (n=1)			TIME
1 VAGINAL No epithelial separation (n=3)													Tl
	1AaT2S1												T2
													T3
				1BdT4S1									T4
				1BeT4S1
2 VAGINAL ≤1 cm exposure (n=26)													T1
	2AaT2S1												T2
	2AaT2S1
	2AaT2S1
	2AaT2S1
	2AaT3S1	2AaT3S2		2BcT3S1	2BaT3S2								T3
	2AaT3S1	2AaT3S2		2BeT3S1
		2AaT3S2
		2AaT3S2
		2AaT3S2
		2AaT4S2		2BcT4S1	2BcT4S2							2DaT4S3	T4
		2AaT4S2		2BcT4S1	2BcT4S2
		2AaT4S2		2BeT4S1	2BcT4S2
				2BeT4S1
				2BeT4S1
3 VAGINAL >1 cm exposure (n=18)													T1
				3BaT2S1			3CbT2S1						T2
				3BaT2S1
				3BeT2S1
	3AaT3S1		3AaT3S3	3BcT3S1									T3
	3AaT3S1			3BcT3S1
	3AaT4S1			3BaT4S1	3BeT4S2								T4
				3BcT4S1	3BeT4S2
				3BcT4S1	3BeT4S2
				3BeT4S1
				3BeT4S1
4 URINARY TRACT (n=5)			4AT1S5										T1
													T2
						4BT3S5							T3
						4BT4S3			4CT4S3				T4
						4BT4S5
No subjects in categories 5, 6, or 7 were found in this cohort

The numerical component of Category is on the right column and the ordinal component of Category is on the top row. Time is on the left column. Site is in sub-columns within the ordinal Category columns. No subjects in categories 5, 6, or 7 were found in this cohort.

In a post hoc analysis, we divided the subjects into two groups: those who were seen for evaluation of mesh erosion before the FDA notification regarding mesh safety was released in June 2011 (n = 50) and those who were evaluated after the FDA notification was released (n = 24). We found no difference in the number of subject able to be coded after the FDA notification (n = 19) compared to those who were seen before (n = 33) (P = 0.3).

We evaluated the association between CTS classification and treatment (conservative vs. surgical). Each domain and subdomain of the classification system was analyzed. Asymptomatic patients (Category A) were more likely to be managed conservatively than Categories B, C, and D (P = 0.001). Symptomatic patients (Category B) were more likely to be managed surgically than Categories A, C, and D (P = 0.003). There were insufficient patients in Category C and in Category D for independent evaluation. No association between treatment and Category 1–7 (P = 0.444), pain grade (P = 0.151), Time (P = 0.708), or Site (P = 0.985) was present. We also evaluated the association between CTS classification and the patient’s satisfaction with her final outcome (i.e., resolution of all or most symptoms or persistent complaints) independent of management choice and found no association with Category (P = 0.055), pain grade (P = 0.075), Time (P = 0.177), or Site (P = 0.098).

We evaluated the relationship between presenting symptom and treatment and found that presenting symptom was associated with conservative management or surgery (P = 0.005). Asymptomatic women were less likely to undergo surgical management (OR = 0.2 (95% CI 0.07–0.6), P = 0.03). For subjects who were symptomatic, each symptom was evaluated independently. Women presenting with urinary urgency and frequency were more likely to have surgical management (OR = 4.0 (95% CI 1.2–13.2), P = 0.02). Presenting symptom was also associated with outcome. Asymptomatic subjects were nearly five times more likely to have satisfactory outcome and be without persistent complaint (OR = 4.8 (95% CI 1.4–16.0), P = 0.03). When evaluated independently, the remaining groups of presenting symptoms (i.e., pain, bleeding, urinary frequency and urgency, recurrent prolapse, and infection) did not demonstrate significant association with outcome, that is, persistence or resolution of symptoms.

DISCUSSION

We found that 30% of mesh erosions could not be retrospectively coded using the IUGA/ICS classification of mesh-related complications in an academic tertiary care referral hospital with Female Pelvic Medicine and Reconstructive Surgery board certified surgeons. Additionally, the IUGA/ICS classification system was not associated with treatment or outcome of mesh erosion other than for patients who were asymptomatic. In contrast, the presenting symptom was associated with both treatment and outcome for those with urinary symptoms and those who were asymptomatic.

There has been a mixed reaction to the IUGA/ICS classification system in the literature. Some studies have also found difficulty assigning a code to patients¹⁴ while others have been able to classify all or most.^15,16 The interobserver reliability of the CTS system is poor^14,15 but may be improved with optimized instruction.¹⁷ Some studies have found good general applicability^16,18,19 but recommend altering the system to include voiding dysfunction and de novo overactive bladder as there is currently no way to account for these common complications.^18,20 We agree with criticism that the large number of variables makes comparing multiple subgroups challenging.²¹ The findings of our study supports the criticism of the classification system’s approach not being patient centered.²¹ Particularly when treating pelvic floor disorders, which are in general quality of life issues, we should strive to keep patient experience and satisfaction at the forefront of our actions and interventions. Complications schemas from other disciplines have been able to do this quite successfully.²² As we tailor and improve on our measurement tools in pelvic reconstructive surgery, we may do well as a specialty to take a patient centered approach in recording the complications of the surgeries we do to improve quality of life.

The principal strength of this study is its clinically relevant end-points providing the surgeon with information to manage complications of mesh erosion. Our study evaluated a wide variety of surgical approaches using mesh. This heterogeneous sample allows evaluation of the IUGA/ICS classification system in diverse contexts. A limitation of our study is the small number of subjects. Given the multiple components of this classification system, a meaningful power calculation is challenging. Including seven anatomic groups,^1–7 four symptom groups (A–D), five pain subcategories (a–e), four time groups (T1–4), and five site groups (S1–5) groups, there are over 560 possible codes for an individual mesh complication, making it difficult to draw conclusions about a single component. As we gather more information about mesh complications, a national or international mesh registry could provide data regarding a large number of patients; this could help to clarify the significance of the individual components that make up the classification system.

CONCLUSIONS

An ideal classification system for complications of mesh erosion would be patient centered, would adequately describe the complication, and be agile enough for use in clinical practice. The CTS system more specifically describes mesh erosion than other currently available surgical complication systems.²³ Based on our study, a more expansive description of associated symptoms (such as bowel and bladder symptoms) may enhance the ability to use the scale in counseling and management. Further research should be directed at determining which factors are integral and which are accessory to the IUGA/ICS’s stated aim of the classification system to “aid clinical practice and research.”

In our experience, 30% of mesh erosions were not able to be coded using the IUGA/ICS classification system for mesh-related complications in an academic tertiary care hospital with all surgeons board certified in Female Pelvic Medicine and Reconstructive Surgery. The classification system is expansive but symptoms frequently associated with mesh erosion such as voiding dysfunction are not included in the system. The components of the system were not predictive of treatment or outcome while patient symptoms were predictive of both. In busy practices, the use of a classification system in a mesh registry may be enhanced if the system has clinical significance and is patient centered.