An experimental in-vivo canine model for adult shunt infection

Roger Bayston¹ , Christine Brant² , Stephen M Dombrowski² , Geraldine Hall³ , Marion Tuohy³ , Gary Procop³ and Mark G Luciano²

¹BRIG, Division of Orthopaedic and Accident Surgery, C Floor West Block, Queen's Medical Centre, University of Nottingham, Nottingham, NG7 2UH, UK

²Department of Neurosurgery – S80, the Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH, 44195, USA

³Department of Clinical Pathology – L40 the Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH, 44195, USA

author email corresponding author email

Cerebrospinal Fluid Research 2008, 5:17doi:10.1186/1743-8454-5-17


Published:	24 October 2008

Abstract

Background

Detailed human studies of the mechanisms and development of shunt infection in real time are not possible, and we have developed a canine hydrocephalus model to overcome this. The intention of this pilot study was to show that the canine hydrocephalus model could be shunted using conventional "human" shunts, and that a shunt infection could be established so that further studies could then be planned.

Methods

Hydrocephalus was induced in seven dogs (Canis familiaris) by fourth ventricle obstruction. Four weeks later they were shunted using a Hakim Precision valve. Four of the dogs received shunts whose ventricular catheter had been inoculated with Staphylococcus epidermidis, and three were uninoculated controls. Four weeks after shunting the dogs were sacrificed and necropsy was performed. Removed shunts and tissue samples were examined microbiologically and isolates were subjected to detailed identification and genomic comparison.

Results

All the dogs remained well after shunting. Examination of removed shunt components revealed S. epidermidis in the brain and throughout the shunt system in the four inoculated animals, but in two of these Staphylococcus intermedius was also found. S. intermedius was also isolated from all three "negative" controls. There were slight differences between S. intermedius strains suggesting endogenous infection rather than cross- infection from a point source.

Conclusion

Shunt infection was established in the canine model, and had the experiment been extended beyond four weeks the typical microbiological, pathological and clinical features might have appeared. The occurrence of unplanned shunt infections in control animals due to canine normal skin flora reflects human clinical experience and underlines the usual source of bacteria causing shunt infection.