Effect of propofol anesthesia on resting state brain functional connectivity in Indian population with chronic back pain
Objective: Functional magnetic resonance imaging (fMRI) studies in healthy volunteers have shown alterations in brain connectivity following anesthesia as compared to the awake state. It is not known if the anesthesia-induced changes in brain connectivity are different in a pathological state. This study aims to evaluate changes in the resting state functional connectivity in the brain, after propofol anesthesia, in patients with chronic back pain (CBP).
Keywords: Chronic back pain, fMRI, functional connectivity, propofol
Propofol is a commonly used anesthetic drug that acts on specific pathways in the central nervous system (CNS) to produce sedation or anesthesia. Experiments in healthy volunteers reveal that propofol interferes with the functional interactions across the brain, predominantly, in the thalamocortical and corticocortical connections., Functional connectivity (FC) between regions of the brain during rest can be measured by recording the blood oxygenated level dependent (BOLD) signal using resting state functional magnetic resonance imaging (Rs-fMRI) tool. Distinct regions of the brain that are functionally connected at rest are termed resting state networks (RSN). Default mode network (DMN) is the most studied RSN and involves the posterior cingulate cortex (PCC), medial prefrontal cortex (MPFC), and bilateral parietal cortices. DMN has been reported to play an important role in both conscious state and under anesthesia., Positron emission tomography (PET) studies involving humans and primates have demonstrated reduced blood flow following anesthesia to the PCC, medial thalamus, and basal forebrain regions, the regions implicated in arousal and information processing.,
Previous fMRI studies have examined the effect of propofol on the functional connectivity of the brain in healthy volunteers. Patients with chronic pain have a structurally normal brain; however, studies have shown that the functional connectivity is altered.,, At present, it is not known whether the anesthetic requirement of propofol is different in patients with chronic pain. It is also not known if there is a differential effect of propofol on pain-associated altered brain networks. We hypothesized that the changes in the resting state FC in the brain following propofol administration are different in patients with chronic back pain (CBP) when compared to changes in the healthy volunteers described earlier.
The study was conducted at the National Institute of Mental Health and Neurosciences (NIMHANS), Bengaluru, India after obtaining the approval of the Institutional Ethics Committee. Written informed consent was taken from all the patients. This study has been registered at the Clinical Trial Registry of India (CTRI/2016/03/006720). Fourteen patients belonging to American Society of Anesthesiologists' (ASA) physical status I–II with CBP of at least 3 months, and aged between 16 and 65 years, were recruited in the study over a period of 6 months. These patients were scheduled for an MRI study of the spine under monitored anesthesia care/sedation as a part of their clinical evaluation. fMRI of the brain was performed in an awake state and after propofol anesthesia in the same setting.
Pre-anesthetic evaluation and monitoring during the procedure
All the subjects underwent routine clinical examination for CBP. Their pain scores ranged from 1 to 5 on the numeric rating scale (NRS) before functional imaging. Patients having a history of allergy to propofol or egg, systolic blood pressure <100 mmHg, significant cardiorespiratory morbidity, potentially difficult airway, and ferromagnetic implants were excluded from the study. Patients with structural brain lesions were also excluded. All the patients fasted for 8 h before administration of propofol. Monitoring during the procedure included heart rate (HR) assessment with an electrocardiogram as well as noninvasive blood pressure (BP), pulse oximetry for oxygen saturation (SpO2), end-tidal carbon-dioxide (ETCO2) and respiratory rate (RR) evaluation. A dedicated intravenous access was secured for administration of propofol. An intravenous infusion of dextrose normal saline (5% dextrose in 0.9% NaCl) was maintained at a rate of 100 ml/h during the study period. Oxygen was provided through a facemask at 5 L/min flow during the study.
Propofol (Profol ®, Claris Life sciences, Ahmedabad, India) was administered intravenously as a bolus dose of 1.5 mg/kg over 180 s, followed by a maintenance infusion of 1.5 mg/kg/h for the duration of imaging, using a MRI compatible syringe pump (B. Braun Space Station MRI ®, Melsungen, Germany). The infusion rate was increased, if necessary, with the aim of achieving and maintaining a modified Observer Assessment of Alertness/Sedation (OAAS) score of 3 to 2. All the physiological parameters mentioned above were monitored at 1-min intervals during the administration of the study drug, and every 5 min thereafter till the patient was awake and responding to verbal commands.
Magnetic resonance imaging protocol
Brain imaging was performed with a 32-channel head coil on a 3-Tesla MRI scanner (Philips Achieva 3.0T TX, Philips Healthcare, DA Best, Netherlands). The protocol included fluid-attenuated inversion recovery (FLAIR) axial sequence to rule out any structural brain pathology. T1-weighted (T1W) three-dimensional (3D) turbo field echo (TFE) was done in the sagittal plane (TR/TE 8.2/3.8, NE × 1, 165 slices covering the whole brain with thickness of 1 mm, FOV 256 × 256, 1 × 1 × 1). Resting state fMRI was performed covering the whole brain [TR/TE: 2000/35 ms, 2.88 × 2.88 × 6 mm, 23 contiguous slices, number of dynamics 192 and acquisition duration of 6 min 32 s, FOV 230 × 230]. The subjects were instructed to stay awake and relaxed with eyes closed and not to think of anything in particular. A similar imaging protocol was repeated under propofol anesthesia during the same scanning session.
Pre-processing of the fMRI images was performed using Statistical Parametric Mapping version 8 software (Wellcome Department of Imaging Neuroscience, London, UK) and processing was performed using the REST toolkit v1.8 implemented in MATLAB 12 (MathWorks, Natick, MA, USA). Pre-processing of the fMRI data included segmentation, realignment, coregistration, normalization, and smoothing. The first five volumes were discarded to allow equilibrium of the MR signal. Realignment (within-subject) was performed to account for head motion. This process involved realigning all the images obtained during the acquisition of the resting state fMRI to the mean image of the session. The mean image generated from the realignment stage was then spatially normalized (made to fit approximately to) to Montreal Neurological Institute (MNI) template (standard Atlas More Details) image. The parameters obtained from the normalization to MNI template were applied to every functional image acquired during the resting state scanning session (co-registration), so that every functional image of the individual was registered to the standard image. Finally, the spatially normalized images were smoothed with an isotropic 8-mm 3 full-width half-maximum Gaussian kernel. During the preprocessing steps, each patient's white matter, cerebrospinal fluid (CSF) space, and gray matter functional time series were extracted. Functional-MRI brain Software Library (FSL) sienax tool was used to calculate the total brain volume.
In this study, PCC was chosen as the region of interest (ROI) based on the earlier studies, which documented that propofol affects the FC between the PCC and rest of the brain., A 5 mm spherical seed around the coordinates −6, −58, and 28, as described by Chang and Glover, was used. FC between PCC and rest of the brain was calculated by performing Pearson's correlation of the time course between PCC and rest of the brain, after regressing out the effects of the time course of head motion (6 degrees of motion), mean time course of white matter, and cerebrospinal fluid (CSF). We regressed out white matter and CSF to remove any neuronal specific confounding effects, as well as to remove fluctuations unlikely to be involved in specific regional correlations. Thus, we ensured that the calculated signal was most likely from the neuronal tissue.
We performed one sample test for the pre-propofol state and post-propofol state correcting for changes that accounted for variations due to the total brain volume. A one sample t-test between the pre- and post-propofol sedation data of all the patients was performed with total brain volume as covariate of interest. The statistical result of this test was corrected for multiple comparisons using the “AlphaSim” implementation. This test is based on the Monte Carlo simulation in AFNI (AlphaSim command description at http://afni.nimh.nih.gov/afni/doc/manual/AlphaSim). A combination threshold of voxels'
P < 0.001 and cluster size 0.351 mm 3 was considered significant, which corresponded with a corrected P < 0.05.
Thalamus, MPFC, and PCC were further explored because these regions are reported to be regions of interests (ROIs) in the action of propofol. These regions were identified using the Automated Anatomical Labelling (AAL) atlas. We then calculated the mean functional correlation coefficients (FCC) between these regions (left thalamus => MPFC, left thalamus => PCC, right thalamus => MPFC, right thalamus => PCC) during the pre- and post-propofol sedation state. These mean FCCs indirectly measure the strength of association between these regions and were plotted using boxplot of the Statistical Package for the Social Sciences statistical software version 16 (SPSS Inc., Chicago, IL, USA). A repeated measures analysis of variance (ANOVA) was used to compare the changes in cardiorespiratory parameters before and after propofol administration.
All the 14 patients successfully completed both the pre- and post-propofol fMRI imaging. No patient required alteration in the dose of propofol infusion to maintain the clinical end point. There were 8 male and 6 female patients in the study, and the mean age of the study population was 46.9 ± 11.3 years.
For the analysis of functional data, the one sample t-test was performed to determine the regions of brain that were functionally connected with the PCC during the pre-propofol sedated state. One patient was excluded from the analysis due to motion artifact following propofol sedation. The FC pattern in both the awake and propofol-anesthetized state in the remaining patients was performed. We observed that the seed PCC was functionally connected with bilateral precuneus and thalamus. Interestingly, this spatial map resembled the map of DMN, suggesting the validity of FC of PCC to the whole brain [Figure 1]a and [Figure 2]a.
One sample t-test was done to determine the regions of brain that were functionally connected to PCC during post-propofol sedated state, and it revealed reduced regions compared to the pre-anesthetized state [Figure 1]b and [Figure 2]b. However, the 2-sample t-test between the two states did not reveal statistically significant differences.
Functional correlation coefficients
We observed an increase in the FCC values between the PCC and thalamus in the post-propofol sedation state compared to the pre-propofol sedation state. However, the increase in the FCC values between the PCC and MPFC were modest among the post-propofol sedation state compared to the pre-propofol sedation state. [Figure 3] and [Figure 4].
The changes in the hemodynamic (heart rate and systolic BP) and respiratory (SpO2 and ETCO2) parameters during the imaging period (both pre and post-propofol) are depicted in [Figure 5]a and [Figure 5]b, respectively. On repeated measures ANOVA analysis, the changes in the cardiorespiratory parameters before and after propofol administration were not statistically significant. This eliminated the possible effect of these parameters (blood pressure or carbon-dioxide levels) on the BOLD signal changes and made the fMRI changes truly representative of the drug effect.
In this study, we evaluated the functional connectivity changes following propofol anesthesia in patients with CBP. We initially explored the PCC–whole brain functional connectivity; later, we focussed on the PCC–thalamus, and PCC–MPFC functional connectivity. In the PCC–whole brain analysis, we observed that the activation of regions that were functionally connected to PCC, reduced during the post-propofol state compared to the pre-propofol state. However, in the PCC–thalamic FC and PCC–MPFC FC analysis, we observed a paradoxical increase in the connectivity during post-propofol state.
In patients with CBP, during post-propofol state, the whole brain PCC FC revealed reduced regions of activations in bilateral inferior parietal lobes (IPL). Similar spatial maps were reported in an earlier study involving normal volunteers. However, the involvement of somatosensory regions during the post-propofol state seen in that study were not observed in our study. IPL is implicated in subserving perceptual awareness and in maintaining attention. The loss of FC in IPL regions possibly explains the mechanism by which propofol induces reversible loss of awareness in humans.
The increase in the strength of connectivity between the thalamus and PCC with decreased FC in whole of the brain could also be due to several factors. First, it is likely that the random connections across the entire brain got smoothened, and the connections between certain areas such as the thalamus and PCC got strengthened. Second, the inhibitory influence on these particular connections might have been lost or decreased after propofol administration, resulting in increased communication between these two regions. The reduction in the perception of pain input following administration of anesthesia might also be responsible for a more coherent connectivity between these specific areas, with loss or suppressed connectivity in other areas. The reduction in functional network in the entire brain reflects the known metabolic suppressant effect of propofol with proportional decrease in the cerebral blood flow resulting in reduced activation of the regions that were active during the awake state. Liu et al., observed increased connectivity in the precuneus, but not in the PCC, following moderate sedation with propofol in an auditory task-based study. Similar to our study, these authors also did not observe any change in the MPFC connectivity during sedation as compared to the awake state. This study and our findings demonstrate that the precuneus/PCC complex do not function as a single unit, as has been traditionally believed.
Contrary to our hypothesis, the changes in the PCC FC after propofol administration in individuals with CBP were similar to those observed in healthy volunteers and published earlier. This suggests that CBP may not influence brain connectivity changes following propofol anesthesia. Stamatakis et al., noted that with an increase in the depth of anesthesia, as evidenced by increasing plasma concentration of propofol (0.27 to 0.67 µg/mL plasma propofol concentration), there was an increase in the connectivity between the thalamus and PCC. Boveroux et al., observed that with further increase in the anesthetic depth (to 3.20 µg/mL plasma propofol concentration), PCC connectivity with other areas appeared to be preserved. The increased connectivity observed in the present study between the PCC and thalamus following a fixed dose of propofol sedation as compared to the awake state, may be an evidence of a change to a slow, more stereotypic firing pattern that involves less processing and conveys less information compared to the awake state. The resulting loss of effective thalamocortical interactions contributes to the anesthetic-induced unresponsiveness.
Recently, Wu et al., have shown that connectivity between the periaqueductal gray matter and sensory thalamus plays an important role in the modulation of chronic pain. Therefore, we analyzed the relationship between thalamus and DMN (PCC and MPFC). Baliki et al., studied the activity in the MPFC and insula in healthy controls as well as in patients with CBP using fMRI. Pain intensity in CBP patients exhibited a significant positive correlation with MPFC activity and was not correlated with insular activity. Pain intensity after a thermal pain stimulus correlated with insular activity, but showed no correlation with MPFC activity in both CBP patients and normal subjects. This finding suggests a differential response between acute and chronic pain and that each of them involves a different functional circuitry. However, assessment of functional changes following painful stimulus was not an objective of our study. Our study involved only patients with CBP and our objective was restricted to assessing functional changes before and after propofol sedation, with the awake state serving as control. However, all patients included in this study had similar baseline pain scores.
Increased connectivity between the thalamus and PCC with relative preservation of connectivity between the thalamus and MPFC following propofol sedation suggests region-specific effects of propofol. Alternatively, it may also suggest an impact of propofol on network interaction or adaptive reorganization due to background CBP in the MPFC.
Whether chronic pain alters the requirement of anesthetic dose to produce loss of consciousness is not clearly known. Bansal et al., observed no difference in the requirement of propofol dose to produce loss of consciousness using target-controlled infusion during anesthetic induction in patients with CBP from an intervertebral disc pathology and in patients with brain tumors. This probably explains the lack of a differential response on the FC in our study in patients with CBP and previously reported studies in healthy volunteers.
Both structural and functional changes within the brain are described following transition from acute to chronic pain., Tagliazucchi et al., have shown alterations in connectivity between DMN, insula and middle frontal gyrus in patients with CBP as compared to healthy volunteers. This is explained as resulting from the processing of persistent pain and the emotional modulation of pain. However, the influence of an anesthetic drug on the FC in the background of CBP-induced changes in the brain is unknown. The results of this study show that propofol acts in a similar manner in select areas of the brain in patients with CBP, as in normal healthy volunteers.
Using eigenvector centrality to characterize brain network properties, Gili et al., have shown decreased centrality of the thalamus versus an increased centrality within the pons following low-dose propofol sedation in 15 healthy volunteers. This decrease of thalamus centrality results from its disconnection from a widespread set of cortical and subcortical regions. However, the sedation score was higher than in our study (OAA/S level of 4; the average targeted propofol plasma concentration was 1.2 μg/mL). Similarly, Liu et al., have shown differential effects of propofol sedation on specific and nonspecific thalamocortical systems. Nonspecific thalamocortical connections were significantly suppressed as compared to specific connections following propofol anesthesia. Further, Guldenmund et al., observed decreased connectivity of thalamus with the DMN, external control network (ECN), and salience network, whereas there was increasing connectivity with sensorimotor and auditory/insular cortices with propofol sedation. On the contrary, in a task-based study, Mhuircheartaigh et al., observed relative preservation of thalamocortical connectivity with the exception of connectivity to putamen. The difference in the methodology and analytical techniques, however, make comparisons between various studies difficult. Schrouff et al., observed greater effect of propofol sedation on functional integration in parietal areas as compared to frontal or temporal regions. They also observed that disruption of total integration within each network was significant as compared to the awake state, and certain connections such as ventral attentional (precuneus, anterior cingulate and IPL) connections retained strong interactions even at deep sedation, suggesting differential effects of propofol on different networks, which are findings similar to our observations.
We did not evaluate the plasma concentration of propofol during the study period to correlate it with the changes in connectivity. However, a previous study has shown that plasma concentration even under target-controlled infusion technique does not correlate with actual measured propofol concentration. We achieved the targeted clinical sedation in all patients, which was the end point in this study. Second, we evaluated the effect of only single, fixed dose of propofol in this study. Although we could not assess the effect of varying depths of sedation on FC, this technique facilitated a stable cardiorespiratory status throughout the study period. Similarly, changes in the hemodynamics and subsequent changes in the BOLD signal due to different drug concentration levels and anesthetic depths, as had been done in earlier studies, were eliminated.
This study is significantly different from previous fMRI studies examining the effect of propofol on brain FC. Unlike all previous studies evaluating the functional effects of propofol in healthy volunteers, we studied patients with CBP, which independently alters FC. We observed that the the dose of propofol required to produce clinical sedation in patients with CBP was similar to that in healthy volunteers, and the FC changes after propofol were no different from that seen in healthy volunteers documented previously. Second, hemodynamic (BP) and respiratory changes (CO2 and O2) can affect BOLD signals independently by their effect on cerebral blood flow and vascular reactivity. Including these variables during the analysis of functional data helped in evaluating the true drug effect of propofol and not the effect from cardio-respiratory changes associated with sedation. This information was missing in the earlier fMRI studies evaluating the effect of propofol on the brain.
In this study involving patients with CBP, propofol sedation resulted in increased strength of FC between PCC and thalamus with a generalized decrease in the integration within large scale brain networks. These findings are similar to earlier studies in healthy volunteers, and suggest that CBP may not alter the effect of propofol in the resting state FC pattern in the brain. Further studies are needed to evaluate the mechanism and dose requirement of propofol in patients with brain pathology.
We sincerely thank the patients who volunteered for the study, radiographers for their help in acquisition of images, Department of Neurosurgery for referring patients for this research, and the administration at NIMHANS for the funding.
Financial support and sponsorship
This work was conducted using the early career research grant from NIMHANS.
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]