Change in average peak cerebrospinal fluid velocity at the cerebral aqueduct, before and after lumbar CSF tapping by the use of phase contrast MRI, and its effect on gait improvement in patients with normal pressure hydrocephalus
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.241406
Source of Support: None, Conflict of Interest: None
Keywords: Cerebrospinal fluid, flow, gait, improvement, normal pressure hydrocephalus, tapping, velocity
Normal pressure hydrocephalus (NPH) is a disease of the elderly consisting of a triad of gait disturbance, dementia, and urinary incontinence. The diagnosis of NPH is supported by imaging findings such as ventricular dilatation out of proportion to sulcal enlargement, upward bowing of the corpus callosum, flattened gyri against the calvarium, and increased flow voids. The importance of diagnosing this condition depends on the fact that there are many other conditions having similar clinical presentation. It is one of the causes of dementia which is treatable, and radiologists have an important role in confirming the diagnosis of NPH leading to an accurate treatment.
Lumbar puncture (LP) is a well-documented invasive test for draining cerebrospinal fluid (CSF). Studies have proven the significance of LP in clinical improvement of patients with NPH., The first and predominant symptom to develop in NPH is gait disturbance.,, It is also the first symptom showing improvement after shunt surgery. Therefore, studying the improvement in gait is more feasible in NPH. To study the improvement in gait before and after lumbar CSF tapping, the gait scale is used.
The treatment option for NPH patients is shunt surgery, and 80% of the patients show a clinical improvement following a CSF diversion procedure.,,,,, It is desirable to consider a non-invasive method for diagnosing NPH. Phase contrast magnetic resonance imaging (PC-MRI) is an important noninvasive tool for measuring various CSF flow parameters. Various studies showed a decrease in CSF flow parameters at the cerebral aqueduct after shunt placement. Flow parameters available in our study were peak positive velocity, peak negative velocity, average peak velocity, average flow, positive pixel flow, negative pixel flow, and average pixel flow. Studies have shown less interobserver variations for peak CSF flow velocities among the different flow parameters.
Upon reviewing literature, a comparison of peak flow velocity by PC-MRI and the gait disturbance in the patients sounded feasible. There are no prior prospective studies to the best of our knowledge on this subject. Hence, in patients suffering from NPH, we tried to study the different CSF flow parameters across the aqueduct utilizing PC-MRI before and after lumbar CSF tapping and tried comparing it with gait improvement.
Suspected patients with NPH, who presented with an unsteady gait, urinary incontinence, or memory impairment, along with CT/MRI evidence of ventriculomegaly out of proportion to sulcal enlargement, were selected as cases in our study. These patients were admitted in the neurology department. The clinical evaluation was done by a mini-mental status evaluation (MMSE) and by examining the gait of patients, which was assessed by the gait scale. Thirty milliliter of CSF was tapped on 3 consecutive days in them by performing a lumbar puncture. CSF was tapped over a period of 1–2 hours daily, until 30ml was tapped. The opening pressure was measured and was in the normal range of <20 cm/s. The patients were evaluated before and after tapping. All participants gave a written consent for the study either themselves; or, if they were unable to provide consent, then their close relative provided the consent. The study period was from 1st November 2014 to 30th April 2016. Patients having contraindications for performing an MRI, such as the presence of a cardiac pacemaker or a dedicated metallic implant, and/or the occurrence of severe claustrophobia; patients who refused to participate in the study; and, patients having NPH due to secondary causes such as subarachnoid hemorrhage, head trauma, tumor, central nervous system (CNS) infections or following a surgery were excluded from our study.
PC-MRI was used for measuring various CSF flow parameters. These parameters were compared before and after the tapping. Cine PC – MRI through plane flow encoding, and two dimensional (2D) fast cine –PC MRI through plane flow encoding, are the two acquisition methods supported in the flow analysis.
All patients were scanned on a 1.5-T OPTIMA 450W MRI scanner (General Electric Milwaukee, WI, USA). This scanner used a radiofrequency (RF) coil with polarized head array placed centrally. MRI examination was initialized using a localizing sequence. The MRI scan parameters are provided in [Table 1].
The flow rate was captured at the level of cerebral aqueduct by prospective cardiac gating using pulse oximetry. This synchronizes the MR images to the cardiac cycle of the patient. To measure the CSF flow velocity on plane images, different types of velocity encoding can be used. The data was then processed via a flow analysis program on a General Electric (GE) Advantage Windows workstation offering computation of flow velocity in cm/s [Figure 1].
The following flow parameters were studied: Peak positive velocity (cm/s); peak negative velocity (cm/s); average flow (ml/beat); positive pixel flow (ml/beat); negative pixel flow (ml/beat); average peak velocity (cm/s); and, average pixel flow (ml/beat).
If there is a mismatch between CSF flow of the patient and the velocity encoding (VENC) value, phase artifacts can arise in the region of aqueduct. In such cases, an increase in VENC value can be carried out so that there is no mismatch. VENC provided in the PC-MRI reflects the maximum flow velocity expected. If it is too low, it may produce aliasing, that is, a misidentification of a signal frequency, introducing a distortion or an error; and if it is too high, it might be difficult to detect small flow differences. In our study, the peak average velocities obtained before and after CSF tapping were in the range of 4–7 cm/s. Hence, with a VENC value of 8cm/s (up to a maximum of 10 cm/s), an accurate assessment of flow velocity was possible.
Both the velocity as well as the CSF flow anatomy can be obtained from the images. On gray scale, the color may be differentiated according to the flow direction. The cranial flow may be designated as black, whereas the caudal flow may be designated as white. Hence, the flow of CSF observed per pixel may be directly proportional to the velocity created by the pulsatile CSF flow. For the quantitative assessment of these different CSF flow parameters, additional 2–3 minutes are required [Figure 2], [Figure 3], [Figure 4].
Gait of the patients was assessed clinically before and after CSF tapping. It was assessed by the neurologists and data were collected from the patients' notes and records. The gait scale  was used for the assessment, which consists of 4 grades from 0–3: Grade 0 – normal gait; Grade 1 – discrete imbalance when turning with short steps, widened base and occasional falling; Grade 2 – frequent falls and aid needed for ambulation; and, Grade 3 – gait being impossible to perform.
Descriptive and inferential statistical analyses were performed in our study. Pre- and post-tap values of different CSF flow parameters were compared using paired t-test. The pre- and post-gait scale were compared using Mc Nemar's chi-square test. The changes in different CSF flow parameters before and after CSF tapping were compared with the gait improvement using one-way analysis of variance (ANOVA) test and post hoc Tukey test.
Data were analyzed using the Statistical Package for the Social Sciences (SPSS) version 17.0 (IBM, USA). Continuous data were represented as mean with standard deviation. For all the tests, a P value <0.05 was considered statistically significant.
Comparison of different flow parameters before and after CSF tapping resulted in statistically significant results. PC-MRI was a reliable technique for quantification of CSF flow. The average peak velocity in our study was comparable on two occasions, i.e., before and after CSF tapping. In the pre-tap group, it was 5.8196 ± 1.420 cm/s; and, in the post-tap group it was 4.1411 ± 1.0638 cm/s. The value showed a decreasing trend in the post-tap group in the patients. The peak positive velocity as well as the peak negative velocity also revealed a decreasing trend in the post-tap group in our study. The peak CSF velocity provided in different studies have demonstrated different range of values. Depending upon the MRI machine and software used, the range of peak CSF velocity may vary. [Table 2] provides the age distribution in the study. Other parameters such as average flow, positive pixel flow, negative pixel flow, and average pixel flow also decreased in the post-tap measurements and the results were statistically significant. A comparison of average flow velocity before and after CSF tapping is provided in [Table 3].
In our study, 70% of the patients showed gait improvement. Comparison of improvement in gait with a change in the average peak velocity was statistically significant with a P value of 0.001. Comparisons of changes in the peak positive velocity and the peak negative velocity with the improvement in gait were also statistically significant with a P value of 0.004 and 0.001, respectively. A P value <0.05 was considered statistically significant.
Rest of the CSF flow parameters such as change in average flow, change in positive pixel flow, change in negative pixel flow, and change in average pixel flow have been shown to be statistically insignificant.
In our study, we arbitrarily grouped the percentage change in the average peak velocity into 3 grades: Grade 1 had between 0 to 25% change; grade 2 had between 25-50% change; and, grade 3 had greater than 50% change of velocity. Thus, 32.5% patients were in grade 1, 45% in grade 2, and 22.5% in grade 3. These grades of change in the average peak velocity were assessed for determining the number of patients with gait improvement. 5 of 13 (38.4%) patients showed an improvement in the gait in the grade 1 average peak velocity group; 15 of 18 (88.3%) patients showed an improvement in the gait in the grade 2 average peak velocity group; and, 8 of 9 (88.9%) patients showed an improvement in the gait in the grade 3 average peak velocity group. The comparison of values of average peak velocity across the cerebral aqueduct with the improvement in gait in grade 2 and 3 groups showed a good correlation.
To compare the various CSF flow parameters before and after lumbar CSF tapping in NPH patients, PC-MRI was used. The main aim of our study was to compare the average peak CSF flow velocity on two occasions, before and after a lumbar puncture tapping of CSF; additionally other flow parameters such as peak positive velocity, peak negative velocity, average flow, positive pixel flow, negative pixel flow, and average pixel flow across the cerebral aqueduct were also compared.
While studying the age distribution, the mean age of 73.88 years with a standard deviation of 5.478 was noted. The maximum age in our study was 87 years and the minimum age was 64 years. In the study conducted on NPH patients by Kahlon et al., related to aqueductal stroke volume assessment, the mean age was 72 ± 8 years. In the studies conducted by Bradley and others, the mean age was 73 years (range = 54–83 years). Thus, the age distribution in the present study was comparable with that of the previously published studies. The assessment of gender in the previous studies had shown a mixed predilection. In our study, there were 13 female and 27 male patients, with a male: female ratio of 2.07.
Comparison of different flow parameters before and after CSF tapping resulted in statistically significant results. PC-MRI was a reliable technique for quantifying CSF flow. The CSF flow studied at the cerebral aqueduct in our study had a pulsatile to-and-fro movement; The average peak velocity at the cerebral aqueduct in our study was compared on two occasions, that is, before and after CSF tapping. In the pre-tap group, it was 5.8196 ± 1.420 cm/s and in the post-tap group it was 4.1411 ± 1.0638 cm/s. The value showed a decreased value in the post-tap group. The peak positive velocity as well as the peak negative velocity also showed a decreasing trend in the post-tap group in our study. The measurements were 6.0502 ± 1.4743 cm/s and − 5.564 ± 1.4208 cm/s for peak positive and peak negative velocity in the pre-tap group, whereas the measurements were 4.2937 ± 1.042 cm/s and − 3.988 ± 1.1964 cm/s in the post-tap group, respectively. This decrease of CSF flow velocities after tapping were in accordance with the findings of previous studies. The peak CSF velocity provided in different studies had different range of values. Depending upon the MRI machine and the software used, the range of peak CSF velocity may vary. In the literature, the range of peak velocity in NPH patients has been reported to be 6.6–65.7 mm/s in one study.
Other parameters such as the average flow, positive pixel flow, negative pixel flow, and average pixel flow also showed a decreased value in the post-tap measurements, and the results were statistically significant.
In our study, we compared the patients with lumbar CSF tapping, which was done on 3 consecutive days. Post-tapping MRI was done only after the third day of tapping, and gait improvement was compared. Gait was selected as the clinical symptom for assessment of neurological improvement because it was the main and first symptom of NPH to manifest. Moreover, different studies have shown the effectiveness of gait measurement in post-shunt patients. In our study, 70% of the patients showed a gait improvement. An improvement in gait was considered when there was a change in the gait scale by a grade or more. Only a single patient in our study improved by 2 grades. For computing the variables, a grade change by one grade, or a grade change by more than one grade, were clubbed together as signifying the presence of gait improvement. In the other hand, if there was no change in the grade, it was grouped in the category of absence of gait improvement. A comparison of the pre- and post-lumbar CSF tap gait scale changes has been given in [Table 4].
In our study of 40 patients, 28 patients were categorized into the ‘improvement' group and 12 into the ‘no improvement' group. The mean CSF velocities across the aqueduct before and after the lumbar CSF tapping in the ‘improvement' group were 6.20 and 4.03 cm/s, respectively, showing a difference of about 2.17 cm/s. The mean velocities before and after lumbar CSF tapping in the ‘no improvement' group were 4.9 and 4.39 cm/s, respectively, showing a difference of approximately 0.51 cm/s.
Five patients from the ‘improvement' group underwent endoscopic third ventriculostomy, out of which 4 patients showed improvement in their symptoms. One patient showed no significant improvement after shunt surgery. Two patients from the ‘no improvement' group also underwent endoscopic third ventriculostomy and showed no improvement in their symptoms. Studies on comparison of clinical symptoms with the average peak CSF velocity across the cerebral aqueduct are very limited. Studies conducted by Sharma et al., used PC-MRI for the comparison of peak CSF velocity before and after CSF tapping. According to the study, 93% of the patients had neurological improvement by shunt surgery, and showed a statistically significant decrease in the peak CSF velocity at the cerebral aqueduct in the ‘improvement' group. This study compared the peak CSF velocity across the aqueduct both prior to and after shunt surgery, whereas our study compared the gait improvement with the average peak CSF velocity across the aqueduct after three consecutive lumbar CSF drainage procedures.
A recent study by Miskin et al., reported a direct relationship of the peak velocity at the cerebral aqueduct with gait improvement and compared it before and after shunt surgery. According to that study, there were a 25% decrease in the mean peak velocity, a 24% decrease in the aqueductal area, a 14% decrease in the gait time, and a 14% increase in the functional ambulation. In our study, the comparison of gait improvement with the change in average peak velocity revealed statistically significant results with a P value of 0.001, and 70% of the patients showed improvement following the lumbar CSF drainage procedure. The comparisons of the changes in the peak positive velocity and the peak negative velocity with the improvement in gait were also statistically significant with a P value of 0.004 and <0.001, respectively.
Rest of the CSF flow parameters such as a change in the average flow, a change in the positive pixel flow, a change in the negative pixel flow, and a change in the average pixel flow were statistically insignificant.
When the grades of improvement in the average peak velocity were assessed for determining the number of patients with gait improvement, the values of average peak velocity in grade 2 and 3 groups showed a good correlation with the grades of gait improvement. To the best of our knowledge, this differentiation of the average peak velocity into the three groups, and the comparison of the mean values in each group with the number of patients with gait improvement has not been done in other studies.
The sample size in our study was of 40 patients with NPH. Most of the previous studies had a sample size similar to our study. Still, a large sample size is warranted for obtaining a very strong statistically significant result in the study.
PC-MRI is a sensitive method to support the diagnosis of NPH, and different flow parameters were comparable before and after CSF tapping. The parameters which proved to be useful for assessing clinical improvement included a change in the average peak velocity, a change in the positive peak velocity, and a change in the negative peak velocity.
Our study showed that a direct relationship exists between a change in the average peak CSF velocity and the improvement in gait. The study also showed a good correlation of a change in peak positive velocity and a change in the peak negative velocity with the improvement in gait.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4]