TRABECULECTOMY
A RETROSPECTIVE FOLLOW-UP
OF 547 PATIENTS
M.A.KHALILI, M.DIESTELHORST 

Department of Ophthalmology, University of Cologne, Germany
Correspondence: M.A.Khalili, M.Diestelhorst, Joseph-Stelzmann-Str.9, 50931 Köln, Germany
Fax 49 - 221- 478 - 4609

M.Diestelhorst: E-Mail:  michael.distelhorst@medizin.uni-koeln.de

     Abstract: A major focus of our study was to determine the value of postoperative intraocular pressure (IOP) in predicting the outcome of trabeculectomy (TE). The medical charts of 547 patients undergoing glaucoma filtering surgery at the Department of Ophthalmology of the University of Cologne from 1987 to 1996 were reviewed. The status of the visual field, level of visual acuity, appearance of the bleb, cup/disc ratio and IOP were studied. Pre- and post- operative glaucoma medication was recorded. The eyes with congenital glaucoma and those treated with anti-metabolites were excluded.
     The results are presented with particular emphasis being placed not only on intraocular pressure (IOP) control but also on the progression of glaucomatous damage (deterioration of visual field or disc damage) and the decrease of visual acuity. The tonometric success rate of TE in controlling the IOP < 21mmHg was 61 %. Defining the rigid criteria for success of trabeculectomy as an IOP < 21 mmHg, no further visual field loss, no disc damage and no additionally required surgical intervention due to glaucoma, the success rate decreased to 44 %. The results indicate that other factors than normalization of IOP determine the success rate of TE. Should trabeculectomy be the therapy of first choice in the early stage of glaucoma?

Key words: Trabeculectomy - Retrospective long-term follow-up - Glaucomatous damage  

INTRODUCTION

    Trabeculectomy, introduced by Sugar [1] and Cairns [2], has become the favoured procedure for the surgical management of various types of glaucoma. Prior studies have described the short- and long-term effects of trabeculectomy on intraocular pressure (IOP) [2-7]. Few studies have addressed the outcomes with respect to optic nerve structure and function [8, 9]. This study presents the long-term outcomes of patients with various types of glaucoma after TE with respect to visual function.

  MATERIALS AND METHODS

    The medical charts of 547 patients undergoing glaucoma filtering surgery at the Department of Ophthalmology of the University of Cologne from 1987 to 1996 were reviewed. Patients with congenital glaucoma and trabeculectomies with anti-metabolites were excluded. The patients are referred to the University Eye Hospital of Cologne for treatment either because of deterioration of visual field or increase of IOP despite of maximum tolerable therapy. This is due to the fact that in Germany about 5500 Ophthalmologists are in private practice. Most of the trabeculectomies were performed by the consultant staff and senior registrars of the above hospital. One surgical unit used a fornix-based flap, the other a limbal-based technique. The status of the visual field, level of visual acuity, appearance of the bleb, description of the optic disc and lens transparency were documented at each visit. Visual field examinations were performed during the early years of the study on the Goldmann perimeter and later on the Humphrey automated perimeter. A diagnosis of progression of the disease was only made when comparing visual field from either manual or automated perimeter.

     The criteria for 'success' following glaucoma filtering surgery were defined as an intraocular pressure (IOP) < 21 mmHg throughout the entire study period, no evidence of further visual field loss, no decrease in visual acuity and no additionally required surgical intervention due to glaucoma. Criteria for tonometric success was an IOP < 21 mmHg throughout the entire post-operative period. The progression of visual field loss was determined by Goldmann perimeter and Humphrey perimeter. The visual field defects were classified in five stages according to Aulhorn [10]. Analysis of the visual fields by computerized perimetry were made by documentation of the mean deviation. A deterioration of the visual field was judged to be significant at a decrease of a mean deviation of 1.3 decibel (db) or more. Criteria for progression of the glaucomatous optic disc damage included an increase in the horizontal cup/disc ratio by 0.2 or greater (description). Eyes with a change of visual acuity of 0.2 were identified.

     Statistical analysis of the data included Whitney-Mann-U-Test, Wilcoxon-Test, Cross-stabs (Fisherďs Exact-Test) and construction of Kaplan-Meier curves. The comparison of survival curves was made by Log-rank-Tests. These statistics were performed using the statistical software package SPSS at the Department of Statistics and Biomathematics and Statistics, University of Cologne, Germany. P-values of 0.05 or less were considered statistically significant.

  RESULTS
PREOPREATIVE CHARACTERISTICS

Table 1 shows the preoperative characteristics of our patient population. Table 1: Preoperative characteristics of the patient population 

The mean duration of the follow-up period after trabeculectomy was 16.3 months (range: 0.1 - 116.6 months). The following results refer to the eyes on which trabeculectomy was performed at first (547 patients = 547 eyes).
IOP

     The IOP decreased from a mean preoperative value of 28.5 mmHg (( 9.8) to 15.8 mmHg (( 5.3) at the last post- operative visit. The average number of medications used decreased from 3.26 (( 1.59) preoperatively to 2.2 (( 1.2) at the last follow-up. 72.6 % (397 of 547 patients) required medication before trabeculectomy and 35.8 % (196 of 547 patients) after trabeculectomy.

     Simple statistics gave for the rate of IOP < 21 mmHg in the course of the entire post-operative period a figure about 61.2 % (335 Eyes) with or without anti-glaucomatous medication and a figure about 48.6 % without anti-glaucomatous medication. In terms of IOP surgical failure (IOP ( 21 mmHg) occurred in 38.8 % (212). In 78.8 % the tonometric failure took place within the first 3 months after trabeculectomy. Among the patients with tonometric failure (212) 60.4% (128) obtained anti-glaucomatous medication but only 68.75 % (88) of those showed an IOP < 21 mmHg taken at the latest clinical attendance. A second trabeculectomy on the same eye was performed on 55 patients (25.94 %) with tonometric failure, among those 85.45 % (47) had a final IOP below 21 mmHg.

The survival analysis by Kaplan-Meier (see Figure 1) shows cumulative one, two, three, five and six year tonometric success rates to be 61 %, 53.4 %, 50 %, 37.8 % and 37.6 % respectively for all type of glaucoma, with mean duration of intraocular pressure control being 1508 days.

Figure 1: 

     No significant difference in survival by sex (p = 0.88), by age (p = 0.53), surgical technique, limbal versus fornix - based flap (p = 0.28, see table 2) and by type of glaucoma was shown. The patient group with prior argon laser trabeculoplasty had a significant lower success rate in terms of normalization of IOP than the patient group without prior argon laser trabeculoplasty (p = 0.03, see Table 3). These two patient groups were similar in terms of a number of variables, including age (p = 0,6), sex (p=0,07), presenting preoperative IOP (p= 0,9) and visual fields (p= 0,12) (Whitney-Mann-test).
  Table 2: Influence of a fornix - or limbus-based conjunctival flap on the IOP after trabeculectomy 

  Table 3: Influence of prior LTP on the IOP after trabeculectomy 

Patients with a successful outcome following trabeculectomy in terms of lowering the IOP < 21 mmHg in the entire period after trabeculectomy had a significant lower (p < 0.01) mean IOP (IOP:10.3 ( 4.5 mmHg) at the date of leaving the hospital (median: 6 days) than those 212 patients with an IOP ( 21 mmHg in the course of the postoperative period (IOP:15.7 ( 8.8, mmHg). All in all bleb encapsulation was noted on 8 eyes. A second TE on the same eye had to be performed on 12.1 % (66). The second trabeculectomy was performed in the mean period of 173 days after trabeculectomy (median: 38 days).

VISUAL ACUITY

     The mean visual acuity decreased from 0.57 (( 0.32) preoperatively to 0.53 (( 0.31) at the last follow-up visit. 16.9 % (92) of the patients showed a decrease of visual acuity after TE. In 52.1% of those the decrease of visual acuity occurred within the postoperative period of 6 months (mean:449 d ( 63 d).
     The decrease of visual acuity could be explained in 22 patients (24 %) by progression of cataract (6 of those with miotic therapy) and in 35 patients (37 %) by development of cataract (17 of those with miotic therapy) which occurred in 16 patients (47.1%) within the post-operative period of 6 months. In 23 other eyes, visual acuity loss was attributed to worsening glaucoma, a second trabeculectomy was performed in these eyes.
     All in all 32 % (175) of the patients (n=547) had already cataract before trabeculectomy. In this patient group 14.3 % (25) showed a progression of cataract after trabeculectomy. A development of cataract was diagnosed in 10.7 % (59) of the patients (547) in a mean post-operative period of 475.2 ( 83.44 days after trabeculectomy. In 47.4 % the cataract was diagnosed within the first 188 days after trabeculectomy.
     In the patient group with decrease of visual acuity (92) 38.04 % received miotic therapy, in contrast to this in the patient group with stable visual acuity after trabeculectomy 23.5 % had miotic therapy. The difference between these two patient groups was highly significant (p = 0.001).

VISUAL FIELD

     The visual fields were classified in five stages according to Aulhorn (1979). Figure 2 shows the distribution of the visual fields of 465 patients before and after TE. 83.4 % (388) of the eyes remained stable, 6.5 % (30) showed an improvement, 10.1 % (47) a deterioration of the visual field. The deterioration of the visual field could be detected post-operatively after 602 ( 99 d (median: 326 d).

Figure 2: 

     The preoperative visual acuity in patients with deterioration of the visual field after trabeculectomy was 0.67 ( 0.15 and in patients with stable postoperative visual field 0.60 ( 0.15. The visual acuity at the latest clinical attendance was 0.51 ( 0.49 in patients with deterioration of the visual field and 0.56 ( 0.14 in patients with stable visual field. No significant difference in visual acuity before trabeculectomy (p = 0.11) and at the latest clinical attendance (p=0.29) after trabeculectomy was shown between these two groups.

     In Table 4 the mean preoperative and postoperative IOP and the mean percent reduction of IOP following surgery in patients with continued deterioration or with stable visual fields are shown. The postoperative IOP values are significantly different between these two groups of patients (p = 0.007).
Table 4:The mean pre- and postoperative IOP and the mean percent reduction of IOP at the latest clinical attendance in patients with or without continued deterioration of visual fields following trabeculectomy. 

     The influence of preoperative anti-glaucomatous drugs on the visual field was also investigated between these two groups. In the patientgroup with deterioration of visual field after trabeculectomy 87.2 % (41 of 47 patients) obtained anti-glaucomatous therapy and in the patientgroup with stable visual field after trabeculectomy 71.5 % (299 of 418 patients). These rates were significantly different between the two patientgroups (p=0.032).There was no significant difference in the number of anti-glaucomatous medication before and after trabeculectomy between the patientgroup with stable visual field and deterioration of the visual field after trabeculectomy.
Analysis of the visual fields by computerized perimetry revealed a decay of the visual field in 57 (16 %) of the patients, if we judge a deterioration of the visual field to be significant at a decrease of a mean deviation of 1.3 db or more.

OPTIC DISC

     Criteria for progression of the glaucomatous optic-disc damage was an increase in the cup/disc ratio by 0.2 or greater. Progressive glaucomatous damage could be detected in 4% (22) of the eyes. It occurred post-operatively after (414.18 ( 140.65) days (median: 215.5 days).
No correlation could be seen between the change of horizontal cup/disc ratio and the change of visual field after trabeculectomy ( Goldmann perimeter - classification of Aulhorn: r = 0.075, Humphrey perimeter - Mean deviation: r = 0.05).

'SUCCESS'

Defining the qualified criteria for success of trabeculectomy as an IOP < 21mmHg, no evidence of progressive disc damage or further visual field loss, no decrease in visual acuity and no additionally required surgical intervention due to glaucoma, there was a success rate of 44.4 % with or without therapy and a success rate of 35.3 % without additional therapy.
     Disease stability after trabeculectomy was defined as stable visual acuity, stable visual field (Goldmann Perimeter), stable optic disc appearance and no second intervention due to glaucoma. The rate of disease stability with or without medication was 58.1 %. According to this definition 41.9 % did not have a stable result after trabeculectomy. Among this patient group with post-operative  unstable result an IOP ( 21 mmHg in the post-operative period was only recorded in 48.9 %. In the patient group with a documented IOP ( 21 mmHg in the course of the post-operative time (212) 100 % had also an unstable result.
     To assess the long-term efficacy of surgical treatment of glaucoma the cumulative probability of disease stability was calculated. The cumulative probability of disease stability was 66 % after 1 year, 53 % after 2 years, 44, % after 3 years, 30 % after 5 years and 27.6 % after 6 years (figure 3). The disease stability showed a mean duration of 1262 ( 126 days.

  DISCUSSION

     Earlier studies [3-6, 9] have demonstrated trabeculectomy to be a relatively safe and effective procedure for short- and long-term control of IOP. But still important questions have remained unanswered: What are the long-term outcomes of trabeculectomy? How effective are our current surgical techniques in preserving visual function? There are only few studies which have addressed these issues [8, 11]. The majority of published series have dealt with short- and long-term tonometric results after trabeculectomy (Table 5). Table 5: Published long-term results of trabeculectomy 

    Long-term tonometric efficacy of trabeculectomy varies from 67.5 % - 98 % depending on the duration of follow-up and the criteria used for defining IOP. Since the aim of the TE is a maintenance of reduction of IOP in order to prevent further glaucomatous damage to the eye, criteria for success in this study was an IOP < 21mmHg at any single reading. This definition is in opposition to many other authors' definition whose criteria for success were less stringent [5,11]. Analysis of our data using a definition of failure similar to that of Bayer et.al. [11] and Mills [5] (IOP < 21 mmHg at the latest clinical attendance while receiving anti-glaucomatous medication) yielded a success rate of 86.8 % in our patient population.
     In the analysis by life-table method, the present study may be compared with the results of Inaba [3], who employed a similar method of analysis. Inaba using a definition of failure as an IOP > 20 mmHg for three consecutive months while receiving anti-glaucomatous medication found a 1-year cumulative probability of success (CPS) to be about 60 %, a 2-year CPS to be 55 % and a 5-year CPS to be 58 % among patients with various types of glaucoma. The CPS is higher than in the present one. This may be due to the less stringent criteria for failure. Our criteria for failure was more stringent (IOP > 21 mmHg at any single reading). A life-table analysis by Shirato [12] gave the 2-year CPS of about 57 % of an IOP < 21mmHg in POAG after TE. Analysis of our data using only those patients with POAG, showed a 2-year CSP of 55 %.

The fact that in this study patients with tonometric failure had a significant higher IOP in the first post-operative week than patients with tonometric success underlines that identification of patients at risk of failure in terms of IOP in the early post-operative period is possible and that closer follow-up and early medical or surgical intervention may be indicated.

     The significantly higher tonometric success rate of trabeculectomy in the primary trabeculectomy group as compared with that in the prior argon laser trabeculoplasty group agrees well with the findings of a prospective study of Midgal et al. [13] in 57 patients, which revealed a success rate of an IOP less than 22 mmHg of 98 % in the primary trabeculectomy group.

Although no significant difference in survival by surgical technique (limbal versus fornix based flap) could be calculated, it is remarkable that the median duration of IOP control was about three times longer in the patient group with limbal-based technique than in the patient group with fornix-based technique.  Our results are in agreement with other studies [4, 5, 8, 9] suggesting that reduction of visual acuity is a common event after TE. The most important causes are development or progression of cataract or progressive glaucomatous damage.

     The results in our patient population indicate that the patients with moderate preoperative field defects are not spared from the risk of further field loss. This finding is in good agreement with the study of Popovic (1991) [14].  Our study showed a highly significant difference (p = 0.007) in the postoperative readings of IOP between the 'no progression of field loss' group and the 'progression of field loss group'. In a study of Popovic et al.[14] no significant difference in post-operative IOP reduction was calculated.  Data from other studies suggested that the disease in patients with visual field loss remained stable with IOP of 18 mmHg or less [15, 16]. This is in good agreement with the finding of our study showing that the 'no progression field loss' group had a mean IOP of 15.71 ( 0.28 mmHg.

     The tonometric success rate after trabeculectomy was 61 %. But rigid criteria for success of trabeculectomy included an IOP < 21mmHg, no further visual field loss or disc damage and no decrease of visual acuity. There was an overall success rate of 44.4 %. The survival analysis by Kaplan-Meier (Figure 1) shows cumulative one, two, three, five and six year tonometric success rates to be 61 %, 53.4 %, 50 %, 37.8 % and 37.6 % respectively for all type of glaucoma, with mean duration of intraocular pressure control being 1511 days. But the cumulative probability of disease stability was 66 % after 1 year, 53 % after 2 years, 44, % after 3 years, 30 % after 5 years and 27,6 % after 6 years (figure 3).The disease stability showed a mean duration of 1262 ( 126 days. An obvious discrepancy between the cumulative tonometric success rates and the cumulative rates of disease stability places in the foreground from the second year onwards after trabeculectomy.

     The fact that only half of the patient population with unstable result had also an IOP 21 mmHg in the course of the post-operative period underlines that the clinically recorded measurements represent a very small sample which attempts to characterize the behaviour of pressure. A higher frequency of tonometric measurements after trabeculectomy might lead to an earlier identification and thus earlier care of those risk patients.

     The results of this study should be judged with the following caveats in mind. The pattern of referral to the clinic (which may not reflect true glaucoma in the population), criteria for patient selection, stage of glaucomatous damage and the definition of progression are factors likely to vary between studies and which strongly influence the overall finding. The results of this study underline that the success of trabeculectomy cannot be judged by the single criteria of IOP. Our criteria for success were more stringent  than in other studies. However, we feel that the aim of TE is a constant maintenance of reduction of IOP in order to prevent further damage to visual function with the main goal to improve or-at least- preserve the patients' quality of life. Further the results lead to the following question: Should trabeculectomy be the therapy of first choice in the early stage of glaucoma?

  REFERENCES  

1. Sugar AS: Experimental trabeculectomy in glaucoma. Am J Ophthalmol (1961) 51: 623-627.

2. Cairns JE (1969) Trabeculectomy: Preliminary report of a new method. Am J Ophthalmol 66: 673-679.

3. Inaba Z (1982): Long-term results of trabeculectomy in the Japanese: An analysis by life-table method. Jpn J Ophthalmol 26:361-373.

4. Jerndal T, Lundstrom M (1977): 330 consecutive trabeculectomies - A follow-up study through 0,5-3 years. Acta Ophthalmol 55:52-62.

5. Mills KB (1981) Trabeculectomy: A retrospective long-term follow-up of 444 cases. Br J Ophthalmol 65:790-795.

6. Watson PG, Grierson I (1981): The place of trabeculectomy in the treatment of glaucoma. Ophthalmol 88:175-196.

7. Wilson P (1977) Trabeculectomy: Long-term follow-up. Br J Ophthalmol 61:117-119

8. Caprioli J, Nouri-Mahdavi , Brigatti L, Weitzman M (1996): Outcomes of surgical treatment for Primary Open-Angle Glaucoma. In: Bucci MG. Glaucoma : Decision making in therapy. Springer Verlag Italia, Milano

9. D`Ermo F, Bonomi L, Doro D (1979): A critical analysis of the long-term results of trabeculectomy. Am J Ophthalmol 88: 829-835.

10. Aulhorn E (1978): Sensoric functional damage. In: Heilmann K, Richardson KT (ed). Glaucoma-conceptions of a disease. Georg Thieme Verlag Stuttgart: 157-180.

11. Bayer A, Erb C, Ferrari F, Knorr M, Thiel HJ (1995): The Tbingen glaucoma study. German J Ophthalmol 4: 289-293.

12. Shirato S, Kitazawa Y (1982): A critical analysis of the trabeculectomy results by a prospective follow-up design. Jpn J Ophthalmol 26: 468-480.

13. Migdal C, Gregory W, Hitchings R (1994): Long-term functional outcome after early surgery compared with laser and medicine in open-angle glaucoma. Ophthalmol 101:1651-1657.

14. Popovic V, Sj”strand J (1991): Long-term outcome following trabeculectomy II. Acta ophthalmologica 69: 305-309.

15. Kass MA, Kolker AE, Becker B (1994): Prognostic factors in glaucomatous visual field loss. Arch Ophthalmol: 1274-76.

16. Odberg T (1987): Visual field prognosis in advanced glaucoma. Acta Ophthalmol 65:27.

17. Akafo SK, Goulstine DB, Rosenthal R (1992): Long-term post trabeculectomy intraocular pressures. Acta Ophthalmol 70: 312-316.

18. Freedman J, Shen E, Ahrens M (1976) Trabeculectomy in a Black American glaucoma population. Br J Ophthalmol 60:573-574.

19. Roth SM, Spaeth GL et al. (1991): The effects of postoperative corticosteroids on trabeculectomy and the clinical course of glaucoma: A five-year follow-up study. Ophthalmic Surg: 724-729. 
   

Figures:

Figure 1:Kaplan-Meier curve of cumulative probability of tonometric success (IOP < 21 mmHg) against time.

 

Figure 2: Change of visual field [%] after trabeculectomy (classification by AULHORN)
in relation to the preoperative stage. Stable (white), loss of one stage or more (grey), increase of one stage or more (dark grey).

• n = 465