Introduction: The Design and Creation of a Grading System for Surgery of Intracranial Aneurysms.
By conducting a review of clinical outcomes for patients with aneurysms treated using current microneurosurgical techniques and intensive care unit management, we determined that grading systems based only on the clinical condition of the patient failed to produce a significant stratification of outcome between individual grades of patients. We hypothesized that outcome prediction for patients surgically treated for intracranial aneurysms could be improved by including factors other than clinical condition that were also strongly associated with outcome.
Because of the limitations associated with grading systems that are based purely on the clinical condition of the patient, we analyzed other factors that have been hypothesized to be important for outcome in a large series of aneurysm operations. Mulitivariate logistic regression analysis revealed five factors associated with outcome, including patient age, aneurysm size, aneurysm location, density of SAH, and whether the patient was comatose. These factors were then incorporated into a simple grading scale to create a composite system capable of predicting risk in both unruptured and ruptured lesions. The system was designed to be simple to use, being based on data readily available very early in the course of treatment of most patients with aneurysms, including data from the physical examination of the patient and the computed tomographic (CT), magnetic resonance imaging, and angiographic studies that define the amount of SAH. and the size and location of the lesion. In this report, we describe the development and use of the grading system, including its application in a large series of surgically treated patients with cerebral aneurysms.
We initially studied several factors known in the literature and our experience to influence outcome in order to gauge the magnitude of effect of these specific factors. They included:
Hunt-Hess Grade: Univariate logistic regression analysis confirmed the increased risk of a poor outcome or death with increasing Hunt and Hess grade. Of note, the risk increase was greatest for patients who presented in coma (Hunt and Hess Grades IV and V). Whereas the relative risk estimates ranged from 1.33 to 2.52 for Hunt and Hess Grades I through III, a much higher estimate of 13.66 was obtained for Hunt and Hess Grade IV patients, and all six Hunt and Hess Grade V patients had poor outcomes or died. Furthermore, in the lower Hunt and Hess grades, only the risk estimate for the Hunt and Hess Grade III patients was significantly different from the baseline group of Hunt and Hess Grade 0. Thus, stratification of outcome for patients in Grades 0 through II was not possible based on using Hunt and Hess grading alone.
Fisher Scale Univariate logistic regression analysis of the Fisher Scale subcategories revealed that compared to patients without any SAH, only patients with a Fisher Scale score of 3 (relative risk [RR] = 3.2; 95% confidence interval, 1.41–7.24) and patients with a Fisher Scale score of 4 (RR = 14.8; 95% confidence interval, 6.6–33.3) were at significantly elevated risk. The patients with small amounts of hemorrhage, Fisher Scale scores 1 and 2, did not have outcomes different from those of the patients with no SAH in this univariate analysis.
Patient age Univariate logistic regression analysis showed a stepwise increase in risk (Table 7), with an increase from the second quartile of age (43–52 yr, RR = 2.94) and third quartile of age (53–63 yr, RR = 5.83) to the fourth quartile of age (>63 yr, RR = 7.7). Each of the quartile risk estimates was significantly greater than the risk for the youngest patients in the first quartile (<43 yr).
Aneurysm size The size distribution of the aneurysms that were surgically treated in this series is shown in Figure 1. In analyzing the association between aneurysm size and outcome, the population was divided into quartiles based on the number of patients in each group. One hundred seven patients had aneurysms that were 6 mm in diameter or smaller, 108 patients had aneurysms that were 7 to 8 mm, 107 patients had aneurysms that were 9 to 11 mm, and 112 patients had aneurysms that were 12 mm or greater. For patients with aneurysms in the third and fourth size quartiles (>9 mm), the risk of poor outcome was significantly increased over the baseline category of an aneurysm of less than 6 mm.
Giant Posterior Circulation Lesions:
We did not note a significant location effect in the data series with a similar percentage of patients experiencing good or excellent outcomes when anterior versus posterior location was compared (85 versus 88%). However, when giant aneurysms were considered specifically, there was a trend for an increased risk when located in the posterior circulation as compared to the anterior circulation.
Multivariate logistic regression of surgical outcomes
To confirm whether each of the previously identified factors was independently associated with outcome, all four variables were combined into a multivariate logistic regression equation. For the purpose of this analysis, each variable was dichotomized as in the final grading scale. As noted, each variable was very significantly associated with final outcome.
A new grading system: outcome and mortality
Based on the above analysis, a grading system was devised. The grading system incorporated each factor that seemed to be independent in terms of its relation to outcome. The proposed grading system is based on the indivdual factors analyzed above. With this new grading system, 98 patients were assigned Grade 0, 160 patients were assigned Grade 1, 118 patients were assigned Grade 2, 47 patients were assigned Grade 3, 11 patients were assigned Grade 4, and no Grade 5 patients were operated on. When average outcome or percent good or excellent outcome was then analyzed using the new grading system, a strong correlation developed between grade and outcome. There was no statistical difference between Grade 0 and Grade 1 patients, yet every patient with a grade greater than Grade 1 (Grades 2–4) had statistically worse outcomes compared to the next lower grade. It was determined that mortality correlated with preoperative Hunt and Hess grade, as it has in other studies Using the new grading system, mortality was also observed to tightly correlate with the new grade assigned.
SUMMARY The MGH grading system we describe is designed to help more accurately predict outcome for patients with ruptured and unruptured aneurysms managed surgically. Advantages of the system include its ease of use, its derivation from clinical data available from the initial examinations (including the standard Hunt and Hess grade) and radiological studies of most patients with aneurysms, and its ability to be comprehensive in risk stratification for both unruptured and ruptured aneurysms. Future work to validate the system will require the prospective application of the scale to larger numbers of patients from different institutions. In addition, the grading system may provide a unique tool for application to patients treated via endovascular techniques as well. It is only through accurate assessment of potential risks and efficacy of surgery and endovascular treatment that it will be possible to achieve a detailed comparison of various treatment modalities. By using a common grading system that incorporates not only patient-specific factors but also aneurysm-specific factors, a more meaningful analysis of outcome using surgical or endovascular techniques could be obtained.