One-year results of trabeculectomy with emphasis on the effect of patients’ age

In our examination of the relationship between age and trabeculectomy success through simple age-based comparisons and univariate regression, a notable trend emerged—older patients were more likely to experience trabeculectomy failure. These findings align with recent reports from South Korea, a region with a genetic background similar to that of Japan [5].

However, when we conducted analyses utilizing multiple regression analysis and propensity score matching to account for confounding factors, age ceased to be a primary risk factor for trabeculectomy failure. Instead, we observed that the prevalence of exfoliation glaucoma increased with aging, leading to poorer outcomes in the elderly. A previous report from South Korea [5], echoing our findings, categorizes exfoliation glaucoma as “Other secondary glaucoma” and highlights the risks associated with advanced age. However, the extent to which exfoliation glaucoma played a role in that report remains unclear. The authors did note differences in bleb height and vascularity when comparing elderly and young individuals, suggesting that elderly patients tend to have flatter bleb height and increased vascularity. We speculate that this characteristic may be related to exfoliation glaucoma. Indeed, a previous report from Sweden reveals that trabeculectomy outcomes for exfoliation glaucoma are significantly worse than for primary open-angle glaucoma (POAG), emphasizing smaller bleb height and extent along with increased vascularity in exfoliation glaucoma [14].

Exfoliation glaucoma, characterized by increased IOP due to extracellular deposits, becomes more prevalent with age [15]. Previous research proposes that the poor surgical outcomes in trabeculectomy for exfoliation glaucoma may be attributed to the extracellular material obstructing the pathway created during surgery. This pathway flows from the anterior chamber to the subconjunctiva through the peripheral iridectomy [14]. Additionally, studies report higher flare values in the anterior chamber after trabeculectomy in exfoliation glaucoma patients compared to POAG patients, and this increased postoperative inflammation has been identified as a potential cause of trabeculectomy failure [16]. Ayala et al. propose that the suboptimal performance of trabeculectomy in exfoliation glaucoma may be attributed to the advanced age of patients due to a lengthier exposure to glaucoma eye drops [14]. The prolonged use of these eye drops is posited to elevate the risk of inflammation, potentially exacerbated by additives like benzalkonium chloride present in the eye drops [17]. This observation aligns with our earlier findings, wherein an extended history of glaucoma eye drop utilization was associated with suboptimal postoperative outcomes in MIGS [18]. However, contrary to our previous study, the current investigation indicates that age no longer holds prominence as the primary risk factor for surgical outcomes in trabeculectomy. Thus, it is suggested that an extended history of eye drop use may not be a definitive determinant of poor outcomes in trabeculectomy, as demonstrated in this study.

Why do earlier reports, from before the early 2000s contend that the results of trabeculectomy are worse in younger patients? It is evident that the influence of hypotony, particularly its higher frequency in young individuals, played a significant role. However, this cannot explain why, in recent years young age is no longer considered a risk factor in trabeculectomy. We posit that this persistence may be attributed to the relatively high incidence of neovascular glaucoma (NVG) in young patients. In the era when tube shunt surgery was not prevalent, trabeculectomy stood as the primary option for NVG. However, with the increased effectiveness of tube shunts for NVG [19], tube shunt surgery is now often chosen as the primary choice. Consequently, recent reports are unlikely to report trabeculectomy failure in young individuals.

Our results reveal a surprising trend in which axial length tended to decrease in the elderly group. Myopia is a well-known risk factor for glaucoma patients [20]. We hypothesize that glaucoma patients affected by myopia might experience worsened glaucoma at a young age, leading to longer axial length in patients undergoing trabeculectomy at a younger age. Conversely, a survey of Japanese glaucoma patients indicates that the emmetropic or hyperopic group was more likely to experience worsening visual field defects than the myopic group [21]. Given that glaucoma deterioration is more common in the elderly, axial length tends to be shorter in the elderly group in the current study. However, this axial length was not associated with the trabeculectomy success rate. Previous reports indicate that eyes with a long axial length are more prone to developing hypotony maculopathy after trabeculectomy [22]. In the current study, with only 11 cases (3.6%) of hypotony among 305 eyes, the results may not have reached significance.

This study employed various analysis methods—multiple regression analysis and propensity score—in addition to simpler methods such as stratified analysis and univariate regression analysis to explore the impact of age on trabeculectomy outcomes. Simple analyses run the risk of overlooking confounding factors that may exist in the background. Association does not imply causation, as seen in the example of blood pressure and hair volume outcomes being correlated due to a hidden common cause, age. It is crucial to consider underlying confounds.

Two primary approaches exist to adjust for confounding factors: one measures the influence of confounding factors, and the other aligns all confounding factors. The former is multivariate regression analysis, while the latter is the propensity score method [23]. Multivariate regression analysis is useful when the exposure variable is continuous or a multigroup categorical variable. However, it has the disadvantage of less intuitive results’ interpretation and potential overfitting when numerous confounding factors need consideration [24]. On the other hand, the propensity score method allows consideration of an unlimited number of confounds and provides a clear understanding, as demonstrated in Table 5, but it is less popular than multivariate regression analysis and involves more analysis steps, making calculations cumbersome and time-consuming. The practice of verifying results’ robustness by using different statistical models is termed sensitivity analysis [25]. Through this comprehensive statistical approach, we determined that age is not a primary factor influencing surgery, but the rate of exfoliation glaucoma increases with aging, leading to poor outcomes in the elderly group.

To further strengthen the results of this study, a subanalysis was conducted excluding all cases of exfoliation glaucoma. As shown in Supplemental Table 1, the trend of increasing surgical failure with age disappeared in this dataset. In this study, bleb revision was not defined as a surgical failure. However, because some consider bleb revision to be a reoperation, we reanalyzed the data, considering bleb revision as a surgical failure. These results are shown in Supplemental Tables 2, 3, and 4. In this reanalysis, age was not a factor in the failure of trabeculectomy, but exfoliation glaucoma was.

There are several limitations to this study. First, the study was conducted at a single institution and included a biased population of only Japanese patients. In addition, the results are for short-term outcomes one year after surgery, and future studies will be required to determine long-term outcomes. Although the analysis focused on age, it is important to note that this dataset does not include a large number of patients under the age of 40.

In conclusion, age itself does not emerge as a primary factor indicating poor surgical outcomes for trabeculectomy. Instead, the prevalence of exfoliation glaucoma increases with age, resulting in inferior trabeculectomy results.

Comments (0)

No login
gif