Vanneste J, Augustijn P, Tan W F, Dirven C
Department of Neurology, St Lucasziekenhuis, Amsterdam, The Netherlands.
J Neurol Neurosurg Psychiatry. 1993 Mar;56(3):251-6. doi: 10.1136/jnnp.56.3.251.
The value of an ordinal global scale derived from combined clinical and CT data (clin/CT scale) to predict the clinical outcome in 112 patients shunted for presumed normal pressure hydrocephalus (NPH) was analysed. The clinical data were retrospectively collected, all CT scans were re-evaluated, and the clin/CT scale was determined blind to the results of further ancillary tests and to the post-surgical outcome. The scale ranked three classes of prediction: on the basis of clinical and CT characteristics, improvement after shunting was probable, possible, or improbable. The predictive value of the clin/CT scale for the subgroup of communicating NPH was established for two different strategies, depending on the strictness of selection criteria for shunting. In the subgroup of patients with presumed communicating NPH, the prevalence of shunt responsiveness was 29%; the best strategy was to shunt only patients with probable shunt-responsive NPH: the sensitivity was 0.54, the specificity 0.84, and the predictive accuracy 0.75, with a limited number of ineffective shunts (11%) and missed improvements (13%). The study illustrates its need to assess the pre-test probability of NPH based on combined clinical and CT data, before establishing the clinical usefulness of an ancillary test.
分析了从临床和CT数据综合得出的序贯性整体量表（临床/CT量表）对112例因疑似正常压力脑积水（NPH）而接受分流术患者临床结局的预测价值。回顾性收集临床数据，重新评估所有CT扫描结果，并在不了解进一步辅助检查结果和术后结局的情况下确定临床/CT量表。该量表将预测分为三类：根据临床和CT特征，分流术后改善可能、有可能或不太可能。根据分流选择标准的严格程度，针对两种不同策略确定了临床/CT量表对交通性NPH亚组的预测价值。在疑似交通性NPH患者亚组中，分流反应性的患病率为29%；最佳策略是仅对可能有分流反应性的NPH患者进行分流：敏感性为0.54，特异性为0.84，预测准确性为0.75，无效分流数量有限（11%），改善遗漏率为13%。该研究表明，在确定辅助检查的临床实用性之前，需要基于临床和CT数据综合评估NPH的预测试概率。
