VP shunts remain the preferred CSF diversion procedure to manage hydrocephalus, with the use of non-VP shunts reserved for instances in which the peritoneum is contra-indicated or has failed previously, except where ETV is appropriate. [15, 16]
Efficacy of VA and VPL Shunts
The Shunt Design Trial and Hydrocephalus Research Network’s studies demonstrate a shunt failure rate of up to 47% in the first 2 years. [17, 18] By comparison, the VA shunt failure rate of 45% (median follow up 1,9 years) in our current series, is not dissimilar. However, the VPL shunt group had a 62% failure rate (median follow up 4,2 years). Although the duration of follow-up for the latter group is longer, there appears to be a difference at 12 months (Figure 1). Although differences between the groups due to the non-randomized selection are unavoidable, it is notable that the VA shunted patients were younger and therefore a higher rate of shunt dysfunction may have been expected.
The lower shunt survival rate for VA and VPL shunts is expected as these procedures are typically performed as second-line measures within a very heterogenous subset of patients who already have had several shunt complications. The patient demographic (low- and middle-income) and spectrum of disease may contribute to this. Warf et al., reported that up to 60% of hydrocephalus in Africa is due to infection. [19] In our study 9% of hydrocephalus was related to myelomeningocele, 11,6% to post-infectious cases, and 9,3% to TBM-related hydrocephalus. It is likely that many of the unknown aetiology group were post-infectious cases. Additionally, the insertion of VA and VPL shunts are undertaken in complex patients as evidenced by the high rate (48,8%) of medical co-morbidities and the high number of previous shunt-related procedures in our sample. One child had 28 prior shunt procedures.
Direct comparison between VPL and VA shunts as a second-line option is sparse, and is compromised by small numbers, relatively short follow-up, and compounded by the greater complexity of these patients in recent series. [20] Much of the available literature derives from historical series where these systems were inserted as a primary procedure and may not be directly applicable because surgical and perioperative techniques have changed over the intervening years. These earlier studies, however, highlight the efficacy of these systems in the absence of other compounding comorbidities. Keucher and Mealey in 1979 demonstrated similar mortality and infection rates for VA and VP shunts in 228 patients with infantile non-neoplastic hydrocephalus, although VA shunts had more revisions, and late complications were more frequent and severe. [12]
Vernet et al., from 1970 to 1991, reviewed 120 cases of infantile hydrocephalus who underwent VA shunting as a primary procedure. [21] With an average follow up of 11 years, they demonstrated no operative mortality and only 1 shunt-related death which was secondary to shunt nephritis. Their infection rate was 4.2% with an average revision rate of 2.2 per patient. Of note, 66% of revisions were for elective lengthening of the atrial catheter. [21] Due to this disadvantage they supported VP shunts over VA shunts as a primary procedure. [21] A Norwegian study of 128 children followed up children who received a VA shunt as a primary procedure between 1967 to 1970 over a 45-year period: 30% of shunts were revised within a year, and 73% within the first decade, with 26,3% of revisions done for elective lengthening of the catheter.[22] Rymarczuk et al. showed no difference in the survival of VA vs VP shunts, excluding elective lengthening procedures in the VA group. [23]
Of interest in the adult population with VA shunts, where fewer revisions due to growth are needed, Lam and Villemure (49 patients), and Al-Schameri et al. (255 patients) demonstrated similar infection and complication rates between VP and VA shunts. [24, 25] Both favoured VP shunts as a primary procedure due to ease of placement and less potential for a severe complication. [24]
Yavuz et al. studied VA shunts as a second-line option in 10 patients aged 5 to 13 years. They reported 3 revisions due to thrombosis, endocarditis and pulmonary embolus.[26] Clark et al., also studied 94 VA shunt insertions in 38 patients as a second-line intervention. They reported higher revision rates to ours, with shunt survival rates of 53%, 43% and 27% at 6, 12 and 24 months respectively, and an overall infection rate of 11%.[27] They concluded that the percutaneous ultrasound-guided technique was safe with a serious adverse event rate of only 2%.
We prefer to reserve the use of VPL shunts for children over the age of 4 years, due to concerns about pleural effusions where lung capacity and compliance may be reduced. Hoffman et al., had a similar approach; 12 (20%) of their 59 patients developed pleural effusions, 6 of which were under 11 months of age. Twenty-three of their patients required no revision. [10] Jones et al., in 52 VPL shunt patients (mean age of 8 years) reported 3 shunt infections, 4 catheter obstructions, 1 symptomatic pleural effusion, and 1 death from shunt malfunction. [13] Martiınez-Lage et al., in 6 patients (5 to 13 years) noted no revisions after a mean follow-up of 2,5 years. [28] In an adult population, where pleural effusion may be less concerning, Craven et al. in 2017 studied 22 VPL shunts and reported a median shunt survival of 14 months. [29]
In a recent review Forte et al., found similar results to ours in their VPL and VA shunt comparison.[20] In a series of 36 VA shunt and 18 VPL shunt insertions over 15 years, VA shunt survival was 60.6%, 51.5% and 36.4% at 3, 6 and 12 months respectively, while VPL shunt survival was 56.3%, 43.8% and 37.5% respectively.[20] Median time to shunt revision was 8,5 and 5,5 months for VA and VPL shunts respectively. We concur with their conclusions about the role of VA or VPL shunts as a second-line procedure. They advised consideration of VA over VPL shunt insertion in those under 5 years. [20] Rymarczuk et al. in their review of 85 VA shunt patients over a 13 year period, further agrees with the second-line role these shunt systems serve and demonstrated similar outcomes to those above. [23]
VA and VPL shunt use may afford time for the peritoneum to heal, allowing later re-introduction of a VP shunt. In our series VP shunting was undertaken for 50,0% of VA (n=5) and 57,9% (n=11) of VPL shunt revisions. Once the original insult contraindicating the peritoneum has resolved, a VP shunt can be reconsidered in the settings of non-VPS failure.
Current Technique of VA shunts
With our shifting focus from VPL to VA shunts, one of the objectives was to evaluate the VA shunt placement technique. The Seldinger technique (percutaneous guidewire assisted placement), first described in 1981, has become preferred.[30] Clark et al. described the assistance of ultrasound guidance and intra-operative fluoroscopy to confirm distal tip position. [27] More recently, Della Pepa et al., reported venous catheter insertion under ultrasound guidance with ECG-guided distal tip positioning. [31] This technique utilises an electrode-integrated venous catheter and relies on predictable changes in the ECG p-wave trace as the atrium is approached. [31] This technique was similarly described by Muhammad et al. and by McCracken et al. [32, 33]
In our study 23 (92,0%) VA shunts were inserted with the Seldinger technique under ultrasound guidance, with intra-operative fluoroscopy performed in 7 (28,0%). We noted that a technique using patient measurements together with chest lead ECG monitoring, to aid in correct catheter placement within the lower third of the superior vena cava, is safe and effective. Deep IVC placement was seen in only 2 cases (neither of these cases were done with fluoroscopy nor measurement to the angle of Louis techniques, with rather an estimate of 10cm used instead by the surgeon); both these patients remained well and have not required revision as of 2022.
Complications
Short and long-term complications were more common in the VPL group, most of which (n=7/21,9%) were related to pleural effusions with 1 case of pleural empyema (3,1%). This compared to the cohort by Forte et al., which reported a rate of pleural effusions at 22,2%.[20] In our study the shunt sepsis rate for the VA and VPL shunt group was 4% (n=1) and 15,6% (n=5) respectively, compared to infection rates reported by Forte et al., of 13.8% and 5.6% for VA and VPL shunts. [20]
Reported complications for VPL shunts include pneumothorax, lung injury, ventilatory difficulties, pneumocephalus, tension hydrothorax and fibrothorax.[4] Small asymptomatic pleural effusions are also commonly described.[34] Reported VA shunt complications include catheter thrombosis, thrombo-embolism (including pulmonary emboli), vessel perforation, bacterial endocarditis, arrythmia, nephritis, pulmonary hypertension and cor pulmonale. [4, 11, 12, 35] Interestingly, Vandersteene et al., demonstrated a pro-coagulant effect of CSF which is attributable to coagulation proteins and tissue factor.[36] Generally CSF concentrations in the venous system are well below the critical threshold required; however in certain circumstances they may increase the risk of clot formation. [36] Shunt-nephritis was first described in 1965 and is typically thought to arise from infection with low virulence organisms, triggering an immune complex deposition at the glomerular basement membrane. [11, 35, 37] In our one case of shunt-nephritis, a VP shunt was inserted after antibiotic treatment. At the time of this study this patient is still doing well, with no signs of recurrence and no long-term sequelae.
Limitations of this study include its retrospective nature, descriptive methodology, limited long-term follow-up and the small cohort, the size of which precluded a reliable comparative statistical analysis. Due to the more recent insertion of VA shunts a lower median follow up time was encountered, which may also contribute to the apparent higher survival rate of these shunt systems. Comparison to conventional VP shunt survival and complications is limited by the selection criteria. Larger studies with longer term follow-up are recommended in order to establish robust clinical guidelines. However, due to small numbers, a multi-centre approach is necessary.