J Neurosurg 145:831–842, 2026
When symptomatic syringomyelia persists despite treatment of its underlying cause, choosing a drainage route requires more than selecting the shortest catheter trajectory. The condition of the subarachnoid space, the level of the cavity and the suitability of an extracranial receiving compartment all influence the operation and its subsequent failure modes. Khalafallah and colleagues compare three shunting strategies in a small adult cohort, illustrating the tension between early decompression and durable neurological benefit. Their findings support individualized planning and careful surveillance, rather than a universal hierarchy of shunt types.
Objective
To compare revision burden, time to failure and neurological outcomes after syringopleural, syringosubarachnoid and syringoperitoneal shunting for syringomyelia.
Methods
The authors retrospectively reviewed a prospectively maintained single-surgeon database covering 1997–2025. Thirty-one adults aged at least 18 years underwent 48 procedures: 20 syringopleural, 21 syringosubarachnoid and seven syringoperitoneal operations. Mean patient age was 47.2 years. Underlying conditions included trauma, tethering, meningitis and Chiari I malformation, as well as cases without an identified cause.
Shunt selection was determined by anatomy, presumed drainage physiology, previous operations and comorbidity, rather than random allocation. Four patients received different shunt types during their treatment and contributed to more than one group. Neurological status was classified as improved, unchanged or worse immediately after surgery, at three months and at the last available follow-up. Revision meant returning to the operating room for a shunt-related complication.
Main results
Ten of 31 patients required at least one revision; revision procedures accounted for 17 of the 48 operations. The reported revision proportions were 11/20 for syringopleural, 4/21 for syringosubarachnoid and 2/7 for syringoperitoneal procedures (55.0%, 19.0% and 28.6%; reported p = 0.003). These figures describe a dataset containing repeated operations and overlapping patient groups: they should not be interpreted as independent patient-level probabilities of failure after an initial shunt.
Catheter migration accounted for four of the 11 syringopleural revisions. Other revision indications included recurrent syrinx, obstruction, traumatic injury, subdural hygroma and infection. Neither time to revision nor the patient- or shunt-based revision-free survival comparisons showed a statistically significant difference between groups.
Immediate neurological improvement occurred after 11/20 syringopleural, 7/21 syringosubarachnoid and 1/7 syringoperitoneal procedures. Although numerically highest in the pleural group, the difference was not statistically significant (p = 0.117). At three months, improvement was recorded in 7/16, 6/13 and 2/6 evaluable cases, respectively, again without a significant difference.
Long-term observations were less reassuring but substantially incomplete. At the last assessment, improvement was recorded in one of ten evaluable pleural cases, two of ten subarachnoid cases and none of four peritoneal cases. Approximately half the clinical observations were missing in each group, and imaging completeness was even lower. These results raise concerns about durability but cannot establish the true frequency of late deterioration in the entire cohort.
Interpretation – operative relevance
The first decision is whether a shunt is appropriate at all. The authors place correction of the underlying pathology—such as craniocervical obstruction or tethering—before direct syrinx diversion when such treatment is feasible. Their comparison therefore informs drainage selection in selected refractory cases, not the replacement of cause-directed surgery by routine shunting.
For destination selection, their practice generally favored pleural drainage for cervical or upper thoracic cavities and peritoneal drainage for lower thoracic cavities. These preferences reflected anatomical access and presumed drainage characteristics, not proven superiority at particular spinal levels. Previous chest or abdominal surgery, pulmonary disease, peritonitis and earlier shunt failure also influenced the choice.
The subarachnoid option depended particularly on the receiving space. The authors increasingly restricted it to preserved subarachnoid anatomy or focal, single-level adhesions, avoiding reliance on a severely scarred or functionally obstructed compartment. Their practical message is to assess where fluid can drain effectively, rather than choosing a local shunt solely because it avoids a distal extracranial catheter.
The operative description includes laminectomy, intraoperative ultrasound localization, midline durotomy and myelotomy, and placement of a nonvalved silicone catheter. Pleural and peritoneal systems required subcutaneous tunneling to the selected destination; dural closure around the catheter was made watertight using sutures with sealant and/or graft material. These are the techniques used in the series, not independently tested determinants of outcome.
The observed pleural migrations make catheter stability an important planning concern. However, this study does not compare fixation techniques or establish a superior anchoring method. During revision, the authors considered changing the distal compartment when fibrosis, scarring or impaired absorption was suspected. Finally, early reduction in syrinx size should be interpreted alongside neurological examination: radiographic decompression did not consistently translate into sustained functional recovery.
Limitations
The retrospective, single-surgeon design, small and unequal groups, heterogeneous etiologies and anatomy-driven allocation prevent causal comparisons. Practice evolved over nearly three decades, introducing additional temporal confounding. Repeated procedures and crossover between shunt types complicate the interpretation of group statistics. Neurological outcomes were qualitative, follow-up duration varied and substantial missing clinical and imaging data create a serious risk of attrition bias. In particular, patients returning because of deterioration may be overrepresented among late observations. Seven peritoneal procedures are insufficient to establish that this destination is intrinsically inferior.
Clinical takeaway
When syrinx diversion is justified, match the destination to the spinal level, subarachnoid patency and suitability of the pleural or peritoneal compartment. Plan for mechanical failure and possible revision, and assess success through sustained neurological function rather than early cavity collapse alone. This series offers useful operative considerations, but does not identify a universally preferable shunt.