Changes in CT cerebral blood flow and volume associated with rapid maxillary expansion in a rabbit model

Dynamic perfusion computed tomography (p-CT) has recently gained widespread application in the study of brain hemodynamics. With its ability to provide quantitative measurements of cerebral blood flow (CBF) and cerebral blood volume (CBV), p-CT has been proven to be an efficient and reliable imaging tool in the assessment of cerebral perfusion.1 Most clinical applications of p-CT are in the diagnosis and treatment of acute ischemic stroke, chronic ischemic cerebral disease, brain tumors, etc.2–4 

Rapid maxillary expansion (RME), invented in 1860 by Angell,5 is used for the orthopedic treatment of transverse maxillary deficiency. RME exerts high forces that can easily split the midpalatal suture in young patients, forcing the maxillary halves apart.6 Because the maxilla is connected to neighboring bones by sutures, RME therapy appears to involve an ample portion of the craniofacial complex. With the help of multislice computed tomography, Leonardi7 found that early treatment with RME produced a significant bony displacement by opening the circummaxillary suture. Opening of the spheno-occipital synchondrosis was found in both animal models and in growing patients.8–10 Several finite element analyses of the effect of RME found signs of high stress around the circummaxillary sutures.11–16 

According to Holberg,13 the superior orbital fissure, the lacerated foramen, and the carotid sulcus are particularly affected by RME. Close to these sutures are some important and vulnerable vascular structures that play essential roles in the blood supplying the brain. There are several publications reporting on vascular complications in adults after RME. One concern is whether the high stress distribution in the cranial base (foramina) caused by RME will influence brain hemodynamics. With the introduction of p-CT to evaluate RME, this speculation can be verified.

The aim of this study was to evaluate both treatment and posttreatment changes produced by RME in brain hemodynamics on a rabbit model.

Twenty-four male New Zealand white rabbits (12 weeks old; 2.8–3.2 kg each, mean weight, 3.01 ± 0.22 kg) were bred at Nanjing Medical University Animal Center. The rabbits were all fed with commercial pellet food and water ad libitum, and they were housed in cages under standardized laboratory conditions of a regular light and dark schedule and relative humidity. The rabbits were randomized into two groups: an experimental group (12 rabbits with RME treatment) and a control group (12 rabbits, no treatment). Approval for the study was obtained from the Animal Ethics Committee of Nanjing Medical University.

The acrylic resin bonded appliance used in this study has been described by Memikoğlu and Iseri19 as a device in which a Hyrax-type screw (Dentaurum, ref 600-300, Ispringen, Germany) was embedded in acrylic resin between the first premolars as close as possible to the palate, with the resin covering the occlusal and labial surfaces of the teeth (Figure 1). The resin was trimmed thin enough to preserve the interocclusal space while allowing maximum occlusal contact bilaterally. The activation protocol included two turns per day and lasted for 10 days. Then the acrylic splint remained in situ for retention for another 10 days. In the control group the screw was not embedded in acrylic.

P-CT examination was performed on a 16-slice CT scanner (SOMATOM Sensation 64, Siemens, Munich, Germany). Baseline p-CT imaging was performed immediately after the RME appliance was bonded. In all, four serial p-CT scans were performed on each rabbit: on day 0 (T0, which was selected as baseline), on day 5 of the activation period (T1), on day 10 (T2, at the end of the activation period), and on day 20 (T3, day 10 of the retention period).

Anesthesia was induced with intravenous pentobarbital (3%, 30 mg/kg) through the auricular vein. Animals were placed in a prone position on a custom device with their heads secured stationary and parallel to the CT couch. A low-radiation, noncontrast CT head scan was performed before perfusion to determine the scope of the cerebrum (80 kV, 100 mA, matrix 512 × 512, slice thickness 2 mm). For perfusion imaging, a fixed 5-mm-thick slice was selected to cover the whole brain. The contrast material (Omnipaque, 300 mg/mL; 1.5 mL/kg) was injected at a rate of 0.3 mL/s through the auricular vein with an automatic injector (Meorao). After 5 seconds, a cine (continuous) scan was initiated with the following parameters: 130 kV and 160 mA, 1 second per rotation for a duration of 50 seconds, and a 512 × 512 image matrix. Four 5-mm-thick axial sections were acquired with the 16-section CT scanner, and approximately 70 images were collected with a section thickness of 5 mm.

All the images were transferred to an imaging workstation (Syngo MultiModality Workplace, Siemens), where functional analyses were conducted on the images with p-CT software (Syngo Neuro Perfusion CT, Siemens). To obtain time density curves, the larger caliber superior sagittal sinus is traditionally used as the reference vessel. A region of interest (ROI) was manually drawn on each functional map of CBF, including the entire cerebral cortex area bilaterally, while ROIs were automatically positioned onto the corresponding CBV and mean transit time (MTT) maps. The ROIs were drawn on a representative transverse plane.

CBF, CBV, and MTT maps were generated by Syngo Neuro Perfusion software (Siemens) for each time point. Each color perfusion map was measured by two doctors. Parametric values of the ROIs representing the cerebral cortex were obtained from the CBF, CBV, and MTT maps. Bilateral brain values were averaged.

All analyses were performed with SPSS for Windows (version 13, SPSS Inc, Chicago, Ill), and significant differences were defined as P < .05. Data were represented by means ± standard deviations. The Kolmogorov-Smirnov test was used at baseline to determine whether the values were normally distributed. Two-way analysis of variance (ANOVA) was used as an omnibus test to identify main significant differences between the experimental and control group. Differences within groups were assessed by one-way ANOVA, with post hoc analysis between different time points.

The measurement results from the p-CT studies in the two groups are given in Table 1. The serial pixel-based color-coded parametric p-CT maps showing the CBV, CBF, and MTT data are shown in Figure 2.

Normal distribution was confirmed with the Kolmogorov-Smirnov test (P < .05) at baseline.

The mean CBF within the RME group increased from 37.10 to 46.87 mL/100 g/min from T0 to T2 and then decreased to 43.68 mL/100 g/min at T3; however, these changes were not statistically significantly different. From T0 to T2, a similar tendency was found for CBV and MTT: mean CBV and MTT increased significantly, from 18.48 to 24.46 mL/100 g and from 74.21 to 87.75 seconds, respectively (Figure 3).

From T0 to T3, all perfusion parameters showed increases in the RME group. CBF grew from 37.10 mL/100 g/min at T0 to 43.68 mL/100 g/min at T3, CBV increased from 18.48 mL/100 g to 23.61 mL/100 g, and MTT increased from 74.21 to 78.20 seconds; however, these differences were not statistically significantly different.

In the control group, no parameters changed significantly at any time points. Two-way analysis of variance indicated significant differences with respect to time and treatment on CBF, CBV, and MTT (P  =  .20).

P-CT has always played an important role in detecting and treating many types of cerebrovascular diseases. However, the orthodontic application of p-CT is rarely seen. As an efficient and noninvasive application, p-CT may be suitable for monitoring the follow-up treatment responses caused by some orthopedic therapies on the hemodynamic changes of the brain. Through the use of quantitative parameters, p-CT can analyze the hemodynamic changes caused by RME.

Although dynamic contrast-enhanced imaging can be obtained by both magnetic resonance imaging and CT, p-CT is more available and efficient and less expensive than magnetic resonance.20 Unlike the technique proposed by Feng et al.,21 which requires the removal of a section of the skull to place an ultrasonic flow probe, p-CT is a noninvasive method that can be measured repeatedly in vivo, and no surgery is needed.

RME has been used routinely to correct maxillary transverse deficiency. Although the RME force is concentrated on widening the maxilla, the transverse forces are transmitted via the pterygomaxillary connection to the unpaired sphenoid bone. Therefore, high stresses are observed in the sphenoid bone, particularly the greater wing and the superior orbital fissure, as confirmed by a number of finite element analyses11–16 and two photoelastic analyses.22,23 Through the use of fresh pig heads, Sun et al.24 recently proved that RME can induce significant physiologic-level strains at the circummaxillary sutures. However, the extent to which the RME forces affect brain hemodynamics has not yet been studied extensively.

In this study, although an increase in CBF was seen from T0 to T2 during the activation period, this difference was not significant between the two time points. As we know, a decrease in CBF always indicates an insufficient blood supply, sometimes to the point of ischemia. The absence of a decrease in CBF indicates the safety of the RME therapy as an orthopedic method in accordance with the long-term stability of clinical application. The relatively stable CBF values seen here might be a result of the cerebral circulation reserve and the autoregulatory ability of the brain.

In the RME group, the CBV showed a statistically significant increase during the activation period; the CBV also increased after the retention period, although insignificantly. The increases in both CBV and CBF may help to explain why RME might alleviate some symptoms of enuretic children. Children ages 4–6 years who wet their beds more than two nights per month are diagnosed with nocturnal enuresis (NE). Several studies have investigated the effects of RME on NE.25–28 These studies showed that NE was greatly reduced or had ceased completely within a few months of maxillary expansion. A 10-year follow-up study showed that nearly 50% of the positive effects were still stable 10 years after the end of orthodontic treatment.29 One of the main features of the children with NE is high arousal thresholds during sleep.30 The high stress distribution focused on several foramina in the cranial base caused by RME may cause the nearby blood vessels to dilate. As a result of the vasodilator effect, manifested as increased CBV and CBF, the blood and oxygen supply to the brain is also increased, which might make children with NE sleep less “deeply.” According to Foltán et al.,31 the respiratory disturbance index and obstructive apneas of patients during sleep improved significantly when treated with surgically assisted RME. In a recent study, Schütz et al.32 analyzed, through polysomnography, the nocturnal breathing of 16 adolescents who underwent maxillary expansion and mandibular advancement with a Herbst appliance. The number of respiratory effort–related arousals and the respiratory disturbance index decreased after treatment. Foltán et al.31 and Schütz et al.32 attributed these changes to an increased nasal volume. But according to Deeb et al.,33 Langer et al.,34 and Giuca et al.,35 there was no significant correlation between the increase in nasal volume and transverse expansion. Therefore, there may be some connection between the increased CBV and the examined sleep parameters.

In his work, Nevéus36,37 pointed out that RME as a presumptive sleep-oriented therapy still needs further research. As we know, with an increase in age, NE children self-heal at a rate of about 15% per year. Herein lies the hypothesis that as children grow, their craniocerebral size and CBV also increase, leading to the 15% self-healing rate. Carotenuto et al.38 also found a higher prevalence of a dolicofacial pattern in primary NE children, which correlated craniocerebral size with NE.

Also, as a result of the vasodilator effect, MTT increased from baseline to T2 during the activation period and then decreased slightly after retention, in accordance with the change in the expansion force. Because the activation lasted for 10 days, the cumulative strength of maxillary arch expansion achieved its peak at T2.

• Our results suggest that for the treatment of transverse maxillary deficiency, RME is a safe method despite its high orthopedic expansion force.

• In addition to its common use in cerebrovascular diseases, the application of p-CT in orthodontics also proved effective in evaluating the follow-up hemodynamic changes of the brain caused by therapy with RME.

• The fact that there is such an increase in both CBV and CBF may help to explain why RME could alleviate the symptoms of enuretic children.

We gratefully appreciate the support of the subjects who participated in this study. This study was supported in part by the National Natural Science Foundation of China (30973361).