Cochlear implant and inner ear malformation. Proposal for an hyperosmolar therapy at surgery

The objective of this retrospective study is to evaluate the efficiency of hyperosmolar therapy for cerebrum spinal fluid (CSF) leakage in cochlear implant (CI) surgery in children with inner ear malformations. Between 1991 and 2006, 490 cochlear implantations were performed in Armand Trousseau Children's Hospital. Thirty-seven patients (7.5%) had inner ear malformation. They were classified as isolated enlargement of the vestibular aqueduct (EVA) (18 cases), incomplete partition (IP) (11 cases), common cavity (CC) (1 case) and variable canal and vestibular malformations (VSCC) (7 cases). A hyperosmolar protocol was applied during surgery to 13 patients after 2003 (Gp) to be compared to the 24 patients without treatment previously to this date (G0). Mean age at implant CI was 8.1 years (1-20 years), mean follow up was 3.9 years (1 month-15 years). Per operative observations were collected for all patients with an empiric method of evaluation of the leakage. A grading using five steps ranged from Grade 0 (no leak) to Grade 4 (gusher). Grading, complications and perceptive results in closed and open set word (Lafon lists) at respectively preoperatively, at 3 and 24 months were gathered and compared between the two groups. Important per operative leak was observed (Grade 4) in 24.3% cases (9/37) of Grade 4, 88.8% of them in G0 (8/9). In 66.6% cases there was a severe dysplasia (CC or IP) (6/9), to be compared to the 21.4% of cases of severe dysplasia with Grade<3 (6/28) (p=0.02). Grade 4 was seen in 45% cases of IP (5/11); it represented 33.3% of the IP in Gp (1/3), and 50% of the IP in G0 (4/8) (p>0.05). Grade 4 was seen in 16.6% cases of EVA (3/18); there were no Grade 4 observed in Gp (0/10), and 37.5% cases of EVA in G0 (3/8) (p=0.04). Grade 4 was observed in 100% case in CC in the G0 (1/1). Severe complications were misplacement of the electrode in one case (G0), persistent leakage in one case (G0) and meningitis in one case (Gp). Vertigo was observed in 29.7% of cases (11/37) in this population, 72.7% of them in G0 (8/11). Vertigo was associated to severe dysplasia in 75% cases in G0 (6/8), and to EVA in all cases in Gp. In G0, mean perceptive scores showed for G0, preoperatively and at 3 months, respectively, 1.3% and 50.6% in closed set word (CSW), and 65.9% in open set words at 2 years. In Gp mean perceptive scores showed preoperatively and at 3 months, respectively 6.1% and 69.8% in CSW, and 81% in open set words at 2 years. The differences between the two groups are not significant (p>0.05). Osmotherapy is known to be effective for cerebral oedema and regularly used in neurological surgery. In inner ear malformation, gusher at surgery is directly related to the intra-cerebral pressure (ICP). Corresponding to neurosurgical practice, the mainstay of our protocol rests on hyperosmolar treatment, to reduce the ICP the time of the surgery. Our results suggest that this treatment is effective for a better control of leakage at cochleostomy on EVA, and could be effective on more severe malformations. No severe complication related to surgery was seen in Gp. Its good tolerance could allow its use in most patients with inner ear malformation. Vertigo was a frequent complication. The possibility of vertigo depends on the initial vestibular status and on the course of the surgery. The protocol could protect the vestibular function, lowering the pressure and quantity of the liquid issue. The treatment does not seem to influence the perceptive results.