gms | German Medical Science

Objective: There is only little published on meningioma surgery in the elderly, with small numbers of patients per series. To improve selection for surgery, we investigated our collective, taking into consideration two proposed risk assessment systems: CRGS and SKALE score.

Methods: We assessed retrospectively morbidity and mortality in 166 patients aged 65 or older operated on an intracranial meningioma from 1995 to 2006 in our service. Data gathered included: age, gender, tumor location, symptoms/signs, perioperative complications, general health condition CRGS/SKALE and ASA scores, pre-/post-operative Karnofsky scores and Simpson's score. In order to investigate the ability of CRGS and SKALE to predict death up to 12 months postoperatively, we analyzed the relationship to all single factors of both grading systems using multivariate logistic regression modeling.

Results: 13 patients died (7.8%). Median Karnofsky score improved from 80 preoperatively to 90 postoperatively. Logistic regression for CRGS, SKALE, age and sex showed a significant relationship of the CRGS and SKALE with death at 12 months. Age and sex were no significant predictors when investigated alone. In the full logistic regression including all proposed factors, CRGS showed significant interactions between size, concomitant disease, eloquent location and edema. However, after stepwise reduction of the full multivariate regression model to its significant terms, only concomitant disease remained significant (OR: 0.133, 95% CI: 0.04–0.463, p<0.001). For SKALE, initially no interactions between factors showed significance. Stepwise deletion of the least important from the full model factor led to exclusion of all factors except the ASA score as only significant single predictor (OR: 0.435, 95% CI: 0.025–0.748, p=0.003).

Conclusions: Our results demonstrate that meningioma resection in the elderly is possible with few complications but with some mortality, which in our collective was similar to the published literature. However, using our database, we were unable to reproduce the utility of two proposed multi-item grading systems for prediction of death. Using multivariate logistic regression analysis of the proposed single factors, only concomitant disease and ASA score remained as significant predictors. We conclude that decision whether to operate or not should be taken individually. Patients with severe concomitant disease or high ASA score should be excluded from surgical therapy independently from other factors.