As was the case for our patient, thoracic migration can lead to somnolence and respiratory failure, yet it can also be asymptomatic in the absence of marked hydrothorax [23]. Aside from dyspnea being the most common presenting symptom, tension hydrothorax, shock symptoms and pleuritic chest pain have also been reported [20]. Differential diagnoses for somnolence and respiratory failure in children with thoracic migration of VPS include valproate toxicity [8, 11, 20], pneumonia [6], or other causes for restrictive lung defects causing dyspnea. According to institutional preference or availability, diagnostic tests for CSF hydrothorax include cell count, biochemical tests, and bacteriological culture from the thoracic fluid, radionuclide shuntogram, beta-2 transferrin, as well as beta-trace protein [8, 11, 20]. However, in our opinion, these should only be considered in case there is no visible migration of the catheter within the thoracic cavity on imaging studies, as to exclude any other probable cause.
Supradiaphragmatic migration of a VPS has been noted to appear most commonly after an erroneous subcutaneous passage while tunneling [21]. Nevertheless, in this setting, the symptoms may arise within days after the intervention and not in a delayed fashion. Blunt subcutaneous passers, as the ones more recently utilized, are much less likely to result in an incorrect tunneling through the intercostal spaces [10]. Yokoya Shigeomi reported a case wherein an incorrect tunneling led to the penetration of the thoracic wall through the intercostal space, arguing that the position of the passer should be confirmed by touch [31]. While this is a reasonable recommendation, especially in adults, in newborns and infants any alteration in the subcutaneous passage may be readily visible. Moreover, the postoperative control thoracic CT can demonstrate an improper subcutaneous trajectory and signs of pneumothorax pertaining to surgical penetration of the pleural cavity [1]. Therefore, the improper placement of the catheter in the initial surgery has been effectively ruled out in our patient. Breaches within the intercostal muscles and pleura can subject the catheter to negative thoracic pressure, ultimately leading to migration [2]. The intercostal space is also more prone to this kind of incident in lean individuals, as is the case for small children or undernourished patients [15]. Such a migration was also reported after VPS placement in a patient who had previously undergone sternotomy [24]. The evidence on hand points to the likelihood that supradiaphragmatic migration can occur under the condition of thinned or weakened intercostal muscles.
However, intrathoracic displacement may also occur through small defects in the diaphragm, a phenomenon known as transdiaphragmatic migration [20, 6]. In this scenario, pleural effusion can ensue even if the catheter is still situated within the peritoneal cavity, particularly if the resorptive capacity of the peritoneum is reduced [8]. This may be accompanied by peritoneal ascites, however it is not a universal condition [11]. CSF malabsorption is nonetheless a key element for this event, expedited by the reduced surface of the peritoneum in small children. On the other hand, one case of a VPS catheter passing through the liver to finally reach into the lung has also been described [17]. Among the risk factors, a personal history of abdominal infection, previous abdominal surgical interventions, and the development of peritoneal pseudocysts have also been recounted [7, 11]. While it had not been the case for our patient, leakage from the VPS valve may also be a possibility, especially if the valve is not properly secured to the segments of the catheter.
Under general anesthesia, hydrothorax or pleural effusion may lead to an increase in airway pressure coupled with a reduction of dynamic lung compliance and abdominal distension [26]. Pulmonary compliance mirrors the distensibility of the respiratory system, being described as the difference in pressure necessary to expand the lung by a given volume. A safe method of avoiding intraoperative complications related to anesthesia is by closely supervising respiratory system parameters like dynamic lung compliance. It was generally recommended that significant pleural effusions should be drained before the induction of general anesthesia, as these may impede lung expansion, whereas drainage will improve ventilation [14, 19]. However, in the report by Lee et al., general anesthesia could be induced in patients with large pleural effusions before performing thoracostomy and without any deleterious effect on ventilation or cardiovascular functions [14]. However, none of the patients in their report were of pediatric age. Moreover, pulmonary edema arising from post-induction pleural drainage has also been reported [4]. As such, the recommendations and issues concerning the induction of anesthesia in infants and young children with hydrothorax remain elusive.
Regardless of the theories offered, the exact mechanism of catheter thoracic migration is unknown. We hypothesize that, in the case presented, an erosion between the muscle fibers of the intercostal muscles and thin costal cartilages was formed as a reaction to the direct contact with the foreign body represented by the catheter itself. This breach then could have increased in diameter in time as a combined result of compression, local reaction, and muscle contraction while breathing. Once this breach was wide enough, it allowed the slippage of a segment of the catheter, which later formed a loop, slowly and steadily increasing with each respiratory cycle due to inspiratory negative pressure. As the distal end of the shunt was pulled out of the peritoneal cavity and moved further away from the abdomen, the CSF coursed through the tunnel created, yet towards the thoracic cavity, again because of the negative pressure. This could have happened only when the catheter was close enough to the thorax for the alternating intrathoracic pressures to generate the pump effect on the fluid, and sufficiently distant from the abdomen for the CSF not to seep into the peritoneal cavity. However, it cannot be accurately estimated when this phenomenon started occurring in relationship with the onset of the symptoms. Considering that she was symptom-free for two weeks following the initial thoracostomy, it is however likely that the phenomenon began in earnest within a month prior to that intervention. Adding to the fact that she did not present any respiratory symptoms until then and that the expansion of the lungs was normal after VPS revision, it can be realistically excluded that the migration had a iatrogenic cause.
Depending on the experience of the treating surgeon, treatment for this complication consists of a VPS repositioning, catheter externalization and ulterior reinternalization [11], switching to an alternative CSF diverting procedure (such as a ventriculoatrial shunt) [8, 9, 18], or the complete removal of the shunt if there is evidence of hydrocephalus resolution [1]. Thoracocentesis or thoracostomy are also recommended, as they result in swift amelioration of the complaints and dyspnea, while the former also provides evidence for the source of hydrothorax [12, 16, 20]. For older patients, especially those with recurrent pleural effusion, positive pressure ventilation through nasal continuous positive airway pressure (nCPAP) may prove a valuable therapeutic tool [27]. We argue that for our patient, since she suffered from congenital hydrocephalus, which implies a lifelong dependency on the shunt, plus the fact that there were no signs of peritoneal malabsorption or bacteriological contamination, our decision to reinsert the distal end into the opposite side of the abdomen was the most suitable.
Highlights of this case report include the young age of the patient, the extremely rare and perplexing complication she sustained, and the timeframe between the surgical placement of the VPS and the occurrence of the supradiaphragmatic migration. Also of note are her fast recovery and complete amelioration of symptoms.