Ventriculoatrial and ventriculopleural shunts as second-line surgical treatment have equivalent revision, infection, and survival rates in paediatric hydrocephalus.

Ventriculoatrial (VA) and ventriculopleural (VPL) shunts are used as alternatives when CSF diversion to the peritoneal compartment with a ventriculoperitoneal (VP) shunt is not possible. The objective of this study is to compare directly the shunt survival and complications for both procedures in this setting in children. A retrospective analysis of 54 consecutive patients who underwent VA (36) or VPL (18) shunt insertion between January 2002 and December 2017 was conducted. The overall mean follow-up was 4.1 (SD 4.3) years, 2.8 (SD 4.1) for VPL and 4.7 (SD 4.4) for VA shunts, respectively (p = 0.11). Twenty-four (66.7%) patients in the VA group and 9 (50.0%) in the VPL group underwent shunt revision (p = 0.236); mean number of revisions was 2.2 (SD 3.0) and 0.94 (SD 1.4) in the VA and VPL groups (p = 0.079). Median time to failure was 8.5 (IQr 78, range 0-176) months for VA and 5.50 (IQr 36, range 0-60) for VPL shunts (log rank (Mantel-Cox) 0.832). Shunt survival at 3, 6, 12 and 30 months was 60.6, 51.5, 36.4 and 27.3%, respectively, for VA and 56.3, 43.8, 37.5 and 37.5% for VPL shunts (log rank (Mantel-Cox) test value 0.727). The infection rate was 13.8% for VA and 5.6% for VPL shunts (p = 0.358). Four patients with VPL shunts (22.2%) developed pleural effusions. Fourteen deaths (25.9%) were recorded during follow-up, 8 (22.2%) in the VA and 6 (33.3%) in the VPL group (p = 0.380); two of the deaths in the VA group were shunt-related. This study demonstrates that the outcomes of VA and VPL shunts, when used as second-line surgical treatment in paediatric hydrocephalus, were similar, as were the revision, infection and survival rates. The shorter longevity of these shunts compared with the general shunted population may reflect the complex nature of these children.