"Shunt pumping test": detecting its efficacy through an experimental model.

Purpose: Shunt pumping test has often been used clinically to detect functional status of ventriculoperitoneal (VP) shunt. Its ability to correctly predict the status is not reliably known. Ethical dilemmas make it difficult to perform any studies in patients with blocked shunts, and hence, a requirement of devising an experimental model was felt.

Method: An experimental model was devised using a Chhabra Slit N Spring shunt. The pressures were maintained in the proximal and distal chamber by real-time monitoring and maintained similar to intra-ventricular and intra-abdominal pressures. Three such models with scenarios of proximal block (PB), distal block (DB), and a functional shunt (BO) were created. Twenty-five participants were tested using these models to assess the efficacy of shunt pumping test.

Results: The experimental model could be used successfully to perform the test. The sensitivity of the test to detect a shunt with block on any side (AB) was found to be 0.79 (95% confidence interval 0.72-0.85) and specificity to be 0.69 (95% confidence interval 0.59-0.80). Its ability to detect the side of block was also evaluated. Absolute correctness value, odds ratio, and interpersonal heterogeneity were also evaluated. Pressure changes in proximal and distal catheter on compressing the chamber in various scenarios were recorded.

Conclusion: The shunt pumping test has moderate ability to predict a blocked shunt and can aid clinical assessment of shunt block. It has only limited ability to detect the side of block. Pressure changes in the proximal and distal catheters on chamber compression are commensurate with the rationale of "shunt pumping test."