Low-dose kaolin-induced feline hydrocephalus and feline ventriculostomy: an updated model

Abstract

Object

Intracisternal injection of kaolin is a well-described model of feline hydrocephalus. Its principal disadvantage is a high rate of procedure-related morbidity and mortality. The authors describe a series of modifications to a commonly used protocol, intended to ameliorate animal welfare concerns without compromising the degree of ventricular enlargement.

Methods

In 11 adult cats, hydrocephalus was induced by injection of kaolin into the cisterna magna. Kaolin doses were reduced to 10 mg, compared with historical doses of ~ 200 mg, and high-dose dexamethasone was used to reduce the severity of meningeal irritation. A control cohort of 6 additional animals received injections of isotonic saline into the cisterna magna.

Results

The mean ventricular volume increased from a baseline of 0.183 ± 0.068 ml to 1.43 ± 0.184 ml. Two animals were killed prior to completion of the study. Of the remaining animals, all were ambulatory by postinjection Day 1, and all had resumed normal oral intake by postinjection Day 3. Two animals required subcutaneous fluid supplementation. Ventriculostomy using anatomical landmarks was performed to ascertain intraventricular pressure. The mean intraventricular pressure after hydrocephalus was 15 cm H₂O above the ear (range 11–20 cm H₂O).

Conclusions

Reduction in kaolin dosage and the postoperative administration of high-dose corticosteroid therapy appear to reduce morbidity and mortality rates compared with historical experiences. Hydrocephalus is radiographically evident as soon as 3 days after injection, but it does not substantially interfere with feeding and basic self-care. To the extent that animal welfare concerns may have limited the use of this model in recent years, the procedures described in the present study may offer some guidance for its future use.

Of the various animal models of obstructive hydrocephalus, kaolin injection into the cisterna magna is among the oldest and most comprehensively studied. It was originally described by Dixon and Heller³ in 1932 and has since been used extensively in the feline population.^{1–5,8–10,12–15} As described in the literature, however, it suffers from substantial morbidity and mortality. Using kaolin doses of 175–225 mg, various authors have described mortality rates as high as 20–50%^10,14 and neurological disability or prolonged anorexia in 25–50% of subjects.^1,2,8–10 Indeed, the bulk of the literature surrounding this method dates back to the 1970s or earlier, perhaps reflecting evolving societal attitudes about what constitutes acceptable animal experimentation. We hypothesized that a substantial reduction in the rates of subject morbidity and mortality could be obtained through simple modification of existing protocols.

Methods

Twelve purpose-bred, adult female domestic felines weighing 2.0–3.5 kg were obtained. Animals were given free access to food and water and kept in individual cages. Normal light and dark cycles were maintained. Enrichment in the form of toys was provided to each animal. Animals were quarantined for 3 days prior to the beginning of the experiment. On the day of cisternal injection, general anesthesia was induced using 100 mg of subcutaneous ketamine HCl (100 mg/ml, IVX Animal Health, Inc.) followed by 2.0 ± 0.5% of inhaled isoflurane in 100% oxygen. Endotracheal intubation was performed, and a peripheral intravenous catheter was placed. For the first 7 animals, 1 mg of acepromazine maleate (10 mg/ml, IVX Animal Health, Inc.) was added to the initial ketamine dose; however, this was found to prolong emergence from anesthesia, rendering postoperative neurological examination difficult, and was abandoned. All animals also received 8-mg atropine sulfate (0.54 mg/ml, Phoenix Scientific, Inc.) subcutaneously to reduce airway secretions.

Once general endotracheal anesthesia had been accomplished, all animals underwent baseline MR imaging of the brain. One cat was excluded from further study after lobar holoprosencephaly with massive ventriculomegaly was demonstrated on the initial MR images, leaving 11 animals for the present study.

Each animal then received a cisternal injection of kaolin and was placed prone on a heated operating table with the neck partially flexed. The head was stabilized with soft rolls. Adequate separation of the posterior occiput–C1 interspace was confirmed with C-arm fluoroscopy, and care was taken to ensure that the cisterna magna remained dependent relative to this interspace. The midline was marked, and the occiput and upper cervical spine were prepared and draped in the usual sterile fashion. Fifty milligrams of cefazolin (100 mg/ml, Watson Laboratories, Inc.) was given intravenously. A 22-gauge, 1.5-inch pediatric spinal needle with a Quincke tip (Becton, Dickinson and Company) was then inserted in the midline under fluoroscopic guidance into the occiput–C1 interspace until clear CSF was returned (Fig. 1). The needle was secured using a mobile clamp affixed to the table (Fig. 2). Five drops of CSF were allowed to escape before a 1-ml slip tip syringe (Becton, Dickinson and Company) containing a sterile colloidal suspension of kaolin was attached to the needle. Initially, the kaolin dose was 50 mg in 0.5 ml of sterile saline, but because of complications at the higher dose, it was reduced to 10 mg in 0.25 ml of sterile saline for the remaining 7 animals. Slow injection of kaolin was completed over 10 minutes. Each animal was maintained in the neck-flexed position under general anesthesia for 15 minutes before it was allowed to awaken.

Fig. 1.

Lateral fluoroscopic view of injection via occiput–C1 interspace.

Fig. 2.

Operative photograph of 1-ml syringe affixed to stabilizing clamp.

Postoperative Care

Postoperatively, the animals were maintained in individual cages where oral intake and urine output could be monitored. Subcutaneous injections of 50-mg cefazolin were given every 8 hours for 3 postoperative doses. Buprenorphine 0.0125 mg was given subcutaneously twice daily for 2 days. In an effort to minimize pain secondary to kaolin-induced aseptic meningitis, 1-mg dexamethasone was given subcutaneously twice daily for 3 days. If oral intake was inadequate on postoperative Day 1, the animals were offered more palatable soft food or NutriCal. No animal was force-fed. Postoperative subcutaneous fluids were administered to the first 2 cats only.

Follow-up MR imaging was performed between postoperative Days 3 and 6 with the animals in a state of general anesthesia. Once ventriculomegaly was confirmed, a ventricular catheter connected to a subcutaneous reservoir was placed in the right frontal region to enable subsequent measurement of ICP. Each animal was prepared and draped, and a preoperative antibiotic was given. A curvilinear incision was made in the right frontal region, and a bur hole drilled using standard surface landmarks: 14–15 mm anterior to the external auditory meatus, and 6–7 mm from the midline.^6,7 Because of the mobility of the feline scalp, exclusive use of preincision markings carries the risk of misplacement of the bur hole. To prevent this, we used a clear plastic drape so that anatomical landmarks such as the nose and eyes could be continually referenced. We also identified the sagittal suture after skin incision and verified the correct lateral distance from this midline structure before drilling. Correct bur hole placement typically corresponded to a position just medial to the insertion of the temporalis muscle; placement of the reservoir beneath the temporalis muscle was unnecessary and led to an excessively lateral position of the bur hole in 1 animal. A small ventricular catheter (In-tegra LifeSciences) was cut according to a distance measured on MR images (15–19 mm total length) and was attached to a mini reservoir (Integra LifeSciences). The catheter–reservoir construct was inserted into the right lateral ventricle using a central stylet. Postoperatively, the animals received the same schedule of cefazolin and buprenorphine as after cisternal injection; dexamethasone was not included in this postoperative regimen.

A third and final MR imaging session was scheduled between 7 and 17 days after the initial injection, depending on MR imager availability. Following this imaging session, the animals were humanely killed with 3–4 ml of pentobarbital sodium (Euthasol, 390 mg/ml; Virbac Animal Health, Inc.) administered intravenously.

All MR images were reviewed for signs of hydrocephalus. Various prospectively determined parameters were measured, including the volume of both lateral ventricles, the maximal thickness of the right temporal horn, and the maximal distance between the lateral margin of the frontal horns on a coronal slice. Ventricular volumes were computed using DICOM viewer software (OsiriX DICOM viewer) after segmentation of the lateral ventricles on MR imaging.

Pathological Analysis

After they were killed, the cats’ brains were explanted, fixed in formalin, and embedded in paraffin. Sections were taken of the cerebrum in the coronal plane at the level of the foramen of Monro and the cervicomedullary junction in the axial plane; staining was performed with H & E and with a stain specific for glial fibrillary acidic protein. Blinded microscopic analysis was performed by a veterinary pathologist. Cerebral sections were inspected for signs of hydrocephalus, with particular attention to white matter interstitial edema, ependymal disruption/hyperplasia, and subependymal astrocytosis. Comparison was made to a control cohort of 6 additional animals who underwent injection of isotonic saline into the cisterna magna instead of kaolin. Cervicomedullary sections were inspected for leukocyte infiltration that would suggest an inflammatory response at the fourth ventricular outflow. For each animal, each variable was quantified using the following scheme: 0 for absent, 1 for mild, 2 for moderate, and 3 for severe. The chi-square test was used to compare the frequency of absent and mild classifications with moderate and severe classifications in the hydrocephalic and control groups.

Results

Two animals were killed prior to completion of all imaging sessions. The first of these experienced significant clinical worsening on Day 9 and was found somnolent and recumbent in its cage. The subcutaneous reservoir was tapped using a 21-gauge butterfly needle and an elevated opening pressure of 20 cm H₂O above the ear was obtained. The animal failed to improve after CSF drainage. Necropsy revealed unilateral pulmonary congestion consistent with early pneumonia. The second animal was found to have a nonfunctioning reservoir on Day 5, the day after reservoir placement. Because posthydrocephalus imaging had already been performed, and a nonfunctioning reservoir precluded assessment of ICP, a humane death was chosen for the animal. The remaining 9 animals completed all imaging sessions as planned.

The duration of postoperative recovery from kaolin injection was variable. All animals regained consciousness after termination of anesthesia. A majority of animals were ambulatory on the day of injection, although 4 animals did not walk independently until postoperative Day 1. All animals demonstrated some degree of anorexia after injection, ranging from 1 to 3 days in duration, and 2 animals required postoperative subcutaneous fluid supplementation. There were no wound complications or infections.

After the first 4 injections, which included the 2 early deaths described above, a number of modifications to the protocol were made. These modifications did not affect the development of hydrocephalus (Table 1), but the clinical course of the remaining 7 animals was generally better than that of their predecessors. First, the dose of kaolin was decreased, as described above. Second, the time frame for postinjection imaging was reduced, so that subsequent animals completed the experiment in 8 days rather than the planned 14 days. After these modifications, there were no procedure-related deaths and no need for supplemental subcutaneous fluid administration. All subsequent animals were ambulatory on the day of the injection and eating by postoperative Day 2.

TABLE 1.

Summary of operative and postoperative data^*

Cat No.	Kaolin Dose (mg)	Vol of Injected Solution (ml)	POD of Awakening	POD of Walking	POD of Eating	Days of Subcut Fluid Supplement	Baseline Ventricular Vol (ml)	Wk 1 Hydrocephalic Ventricular Vol (ml)	Wks 2–3 Hydrocephalic Ventricular Vol (ml)	ICP After Hydroceph (cm H2O)
1	50	0.25	0	1	2	4	0.128	1.040	killed	20
2	20	0.50	0	1	1	1	0.0972	0.652	1.19	15
3	20	0.50	0	1	2	0	0.145	0.979	killed	NA
4	20	0.50	0	1	3	0	0.193	1.28	1.25	18
5	10	0.25	0	0	1	0	0.200	1.20	1.46	15
6	10	0.25	0	0	1	0	0.194	1.00	1.37	NA
7	10	0.25	0	0	2	0	0.088	0.868	1.48	21
8	10	0.25	0	0	2	0	0.229	1.25	1.50	11
9	10	0.25	0	0	2	0	0.171	1.66	1.66	NA
10	10	0.25	0	0	2	0	0.250	0.918	1.70	12
11	10	0.25	0	0	2	0	0.320	1.66	1.23	NA

^*NA signifies inability to measure ICP either because the animal had been killed early or because the catheter was not functional.

Abbreviations: Hydroceph = hydrocephalus; POD = postoperative day; Subcut = subcutaneous; Supplement = supplementation.

All animals demonstrated substantial ventricular enlargement over baseline at both the second and third imaging sessions (Fig. 3). The mean baseline ventricular volume was 0.183 ± 0.068 ml, which increased to 1.14 ± 0.315 ml at the second imaging session, and 1.43 ± 0.184 ml at the third imaging session. The mean maximal distance between the lateral aspect of the frontal horns was 10 ± 0.93 mm at baseline, increasing to 13 ± 1.1 mm at the second imaging session, and 14 ± 1.3 mm at the third imaging session. The mean thickness of the right temporal horn as measured on coronal images was 2.6 ± 1.0 mm at baseline, 6.3 ± 0.87 mm at the second imaging session, and 6.4 ± 1.0 mm at the third. It is difficult to draw conclusions about the relationship between ventricular size and duration of anorexia and between ventricular size and kaolin dose, because of the small numbers of animals involved. Among the 3 animals that experienced < 24 hours of anorexia, the mean final ventricular volume was 1.34 ± 0.135 ml compared with 1.47 ± 0.198 ml among animals with > 24 hours of anorexia (p = 0.35). Only 2 animals who received > 10 mg of kaolin completed all the imaging sessions; the mean final ventricular volume in this group was 1.22 ± 0.040 ml, compared with 1.49 ± 0.16 ml among the animals that received 10 mg of kaolin (p = 0.06).

Fig. 3.

A–C: Coronal T2-weighted MR images through the frontal and temporal horns at sessions 1, 2, and 3, respectively, demonstrating progressive ventricular enlargement. Placement of right frontal ventriculostomy is also demonstrated in panel C.

Intraventricular pressures after induction of hydrocephalus ranged from 11 to 20 cm H₂O above the ear. Mean intraventricular pressure for all animals was 14.7 ± 3.35 cm H₂O. Mean ICP was generally higher in the higher dose kaolin group than in the lower dose group, 17.6 ± 2.51 cm H₂O versus 12.5 ± 1.73 cm H₂O, respectively.

Ventricular cannulation was confirmed on MR imaging in 9 of the 11 animals studied. In 1 animal, the ventricular catheter occupied a position in the brain parenchyma just lateral to the ventricle. Another animal was killed before postventriculostomy imaging could be performed. In 2 other animals, choroid plexus and brain parenchyma occluded correctly placed catheters.

Postmortem sectioning of the brain was performed in 16 animals with blinded histopathological analysis performed by a veterinary pathologist. Formalin-fixed specimens could not be located for 2 animals in the kaolin injection group, and these animals were excluded from the analysis. Inflammation in the cervicomedullary region was markedly greater in the kaolin injection group than in the saline injection group, with a severity score of 2.7 ± 0.48 versus 0.5 ± 0.71, respectively (Figs. 4 and 5). Analysis of the histopathological correlates of hydrocephalus was confounded by additional exclusions in the saline control group. In 4 of the 6 animals in this group, a robust neutrophil infiltrate consistent with inflammatory response was found in the choroid plexus and ependyma of the lateral ventricles; in each case, this was more severe on the side of ventriculostomy. The presence of this infiltrate made analysis of white matter edema and ependymal and subependymal architecture impossible, and so these animals were excluded from the quantitative analysis for these variables. No clinical correlation for this finding was present in any of the affected cats; each subject completed the trial without neurological deficits. For the remaining cats, the severity score for each hydrocephalus variable was higher in the kaolin injection group: for white matter edema, 1.5 ± 0.37 versus 1.3 ± 0.35; for ependymal hyperplasia, 1.6 ± 0.46 versus 0.75 ± 0.35; and for periventricular astrocytosis, 1.5 ± 0.37 versus 1 ± 0. The hydrocephalic cats were significantly more likely to demonstrate moderate to severe cervicomedullary inflammation (p = 0.0005) and moderate to severe white matter edema (p = 0.03) than the normal controls. Differences in the remaining variables did not reach statistical significance.

Fig. 4.

Photomicrograph showing region of the fourth ventricular outflow, demonstrating robust inflammatory cell infiltration of the leptomeninges in this region. Cerebellum is located at the bottom of the image, brainstem at the top. H & E, original magnification × 10.

Fig. 5.

Photomicrograph showing the region of the fourth ventricular outflow, demonstrating normal anatomy without evidence of significant inflammation. Cerebellum is located at the top of the image, brainstem at the bottom. H & E, original magnification × 10.

Discussion

In this paper we describe modification of a well-established model of obstructive hydrocephalus, intended to limit subject morbidity and procedure-related death. Our principal modifications involved reducing the dose of kaolin more than 10-fold and instituting an aggressive regimen of postoperative corticosteroid administration. Among animals that received the final kaolin dose of 10 mg/0.25 ml, the mortality rate was 0%, all were ambulatory on postoperative Day 1, and 10 of 11 were eating independently on postoperative Day 2. This represents a substantial improvement over experiences documented in the literature.^{1,2,4,8–10,12,14,15} There are multiple possible explanations for this observation. One possibility is that there is a reduction in the severity and duration of the chemical meningitis that accompanies kaolin injection. It is known that kaolin injection creates a robust inflammation at the outflow of the fourth ventricle, which is followed by fibrosis and constriction.¹¹ Reducing the dose of the caustic agent is likely to cause a reduction in the degree of inflammation-related symptoms. Corticosteroids also have well-known antiinflammatory properties. A second possibility is that there is a reduction in the degree of hydrocephalus itself. Although a similar ventricular size was seen in the higher- and lower-dose injections, both doses are an order of magnitude less than that described in the literature and may be less likely to cause complete occlusion of fourth ventricular outflow. The trend toward lower intraventricular pressure in the lower-dose group lends some support to this.

Shortening the period of observation may also have reduced the incidence of unacceptable hydrocephalus-related morbidity. It is not possible to know how many cats would have developed deficits if followed over the course of many weeks. Of the cats who completed all imaging sessions, the clinical course was one of stability or gradual improvement. Of note, all cats in this study demonstrated substantial ventriculomegaly as early as 3 days after injection. Thus, for investigations of acute hydrocephalus, a prolonged waiting period may not be necessary.

It is important to note that the historical experience with feline kaolin-induced hydrocephalus differs in the adult and neonatal populations. While the adult model has traditionally been associated with substantial morbidity, infantile models of kaolin-induced hydrocephalus appear to be more durable. A review of publications utilizing a neonatal model reveals substantially less morbidity than that described in the adult population.^1,2,5,12,13 Although this may be partially attributable to variations in the dose of kaolin or other experimental modifications, a more likely explanation involves the increased compliance of the skull prior to suture fusion. Our modifications are therefore principally relevant to the adult model, in which humane survival of the subjects after kaolin injection has traditionally been a challenge.

We have described an 81% success rate in cannulating the lateral ventricle of hydrocephalic cats, using a free-hand technique guided by anatomical landmarks. The catheter patency rate was 64%, reflecting intervening obstruction of the catheter by tissue. Because ventriculostomy success rates have not been widely reported in the literature, it is difficult to place this result in context. Specifically, it is uncertain whether the cannulation success rate might have been improved with the use of a stereotactic frame. However, because a stereotactic frame requires a considerable capital investment and may, by virtue of its size, interfere with certain surgical procedures, we note that in a substantial majority of cases, a nonstereotactic technique is adequate.

Conclusions

Reduction in kaolin dose and the postoperative administration of high-dose corticosteroids appear to reduce subject morbidity and mortality in an adult feline model of kaolin-induced hydrocephalus, compared with historical experiences. Hydrocephalus is radiographically evident as soon as 3 days after injection, but it does not substantially interfere with feeding and basic self-care. To the extent that animal welfare concerns may have limited the use of this model in recent years, the procedures described herein may offer some guidance for its future use.

Abbreviation used in this paper

ICP

intracranial pressure