A typical antipsychotic treatment induced gradually expanding white matter alterations in healthy individuals with persistent auditory verbal hallucinations—an artificially controlled pilot study

Abstract

Objective

The aim of this study was to explore the effects of atypical antipsychotics (AaPs) on brain white matter (WM) tracts in healthy individuals with auditory verbal hallucinations (Hi-AVHs).

Methods

We analyzed neuroimaging, AVH symptoms, and cognitive assessment data obtained from 39 Hi-AVHs who reported being distressed by persistent AVHs and volunteered to receive AaP treatment. We used tract-based spatial statistics (TBSS) and t tests to explore AaP pharmacotherapy effects on AVH symptoms and brain WM alterations in Hi-AVH subjects.

Results

TBSS and t tests revealed WM alterations after AaP treatment, relative to pretreatment observations. Although AaPs alleviated AVH symptoms, WM alterations in these subjects expanded over 8 months of AaP treatment, encompassing most major WM tracts by the end of the observation period, including the corpus callosum, arcuate fasciculus, cortico-spinal tracts, anterior commissure, and posterior commissure.

Conclusions

The worsening of AaP-associated WM alterations observed in this study suggest that AaPs may not be a good choice for the treatment of Hi-AVHs despite their ability to alleviate AVHs.

Keywords:

• Atypical antipsychotics
• auditory verbal hallucinations
• white matter
• TBSS

Introduction

Auditory verbal hallucinations (AVHs) are prevalent among diverse psychiatric patients, including patients with schizophrenia, bipolar disorder, borderline personality disorder, and anxiety; AVHs are also experienced by individuals in the general population [1–3]. Recent elucidation and increased exploration of the pathological features of AVHs and AVH treatment effects in the general population have provided information important for the diagnosis, pathological characterization, and treatment planning for AVHs in general populations [4–7].

Applying the strictest AVH criteria, some 0.7% of the general population has experienced AVHs; these people can be defined as healthy individuals with AVH (Hi-AVHs) [1,8]. Investigating Hi-AVHs may help to clarify the precise pathological features, and thus treatment targets, of AVHs without conflation from other symptoms, such as delusions or thought disturbances [9–11]. The occurrence of AVHs has been related to white matter (WM) alterations in the brain [12–15], and atypical antipsychotic (AaP) medications are known to alleviate AVHs in a variety of psychiatric patients, including patients diagnosed with schizophrenia, bipolar disorder, and borderline personality disorder [6,16–19]. To the best of our knowledge, no study has investigated the relationship between AaP pharmacotherapy and brain WM alterations in Hi-AVHs. In particular, we could not find a cohort study of WM alterations before versus after an extended AaP treatment period.

Dynamic monitoring of WM aberrations in Hi-AVHs can help to further our understanding of the pathological features of AVHs in Hi-AVHs and AVHs in patients diagnosed with a mental disorder. Toward this aim, in the present pilot study, we used magnetic resonance imaging (MRI) to observe the evolution of WM alterations in Hi-AVHs taking AaPs from a pretreatment baseline assessment to 6- and 8-month follow-up time points. The patients’ cognitive functioning levels were assessed to examine whether the treatment had adverse effects on cognition. We used tract-based spatial statistics (TBSS) [20] and t tests to explore the effects of AaP pharmacotherapy on AVH symptoms and brain WM alterations in a cohort of Hi-AVHs.

Methods

Study population

The protocol for this study was approved by the Tianjin Anding Hospital ethics review board. All assessments were carried out in compliance with the Declaration of Helsinki guidelines and approved by the institutional ethics committee of our hospital. All of the participants provided written informed consent for participation in this study after being informed about the potential deleterious effects of taking AaP medication and potential risks of participating in the study. Although the participants volunteered to participate the study, we included only those who self-reported suffering distress from persistent AVHs and wanted to receive AaP treatment. For safety, brain alterations were monitored with MRI examinations during the study period.

An advertising campaign was used to recruit Hi-AVHs from 1000 communities (total source population > 200,000) between January 1^st, 2016 and June 31^st, 2018; the advertisements targeted individuals identified in our institutional database, which was established in 2011 [21–23]. A total of 300 healthy volunteers with a diagnosis of AVHs according to the diagnostic criteria of Professor Johns [8] were recruited. Of these, 115 reported having suffered mental distress from persistent AVHs and volunteered to receive AaP treatment. None of the AaP-treated Hi-AVHs had a diagnosis of any mental disorder according to the Diagnostic and Statistical Manual of Mental Disorders, 4^th Edition, and none had a history of substance abuse, with the exception of moderate to severe nicotine dependence according to the Fagerstrom nicotine dependence questionnaire [24]. AVH severity was assessed with the auditory verbal hallucinations rating scale (AHRS, Haddock et al. 1999) [1].

MRI data acquisition

Each subject completed MRI examinations before starting treatment (baseline) as well as 6 months and 8 months after commencing treatment. The scans were performed with a 3.0-T MR system (Discovery MR750, General Electric, Milwaukee, WI). Tight but comfortable foam padding was used to stabilize the subjects’ head positions, and the subjects wore earplugs to attenuate scanner noise during image acquisition. A three-dimensional T1-weighted brain volume (BRAVO) sequence with 188 sagittal slices was obtained with the following parameters: repetition time = 8.2 ms; echo time = 3.2 ms; inversion time = 450 ms; flip angle = 12°; field of view = 256 mm × 256 mm; matrix = 256 × 256; slice thickness = 1 mm; and no gap.

MRI data analysis

Based on a previous study [20], we employed PANDA software [25], a MATLAB toolbox that integrates FSL (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki), Diffusion Toolkit (http://www.trackvis.org/dtk/), and MRIcron (https://www.nitrc.org/projects/mricron) image processing software. Before processing, the images were subjected to quality checks; images with excessive movements or otherwise poor image quality were discarded. Images that passed the quality control checks were subjected to brain tissue extraction, wherein non-brain tissues were removed, and brain masks were estimated. Subsequently, the images were corrected for the elimination of movements and eddy current-induced distortions. For diffusion tensor imaging (DTI), diffusion tensor fitting was completed and diffusion metrics were calculated in FSL DTI Fit. Finally, the images were subjected to spatial normalization with FSL FNIRT, wherein diffusion metrics were normalized to Montreal Neurological Institute space and segmented images were spatially smoothed with a Gaussian kernel (full width at half maximum = 6 mm³).

TBSS

As described in detail previously [20,25–27], we registered all DTI metrics maps non-linearly to align them with a target image by selecting predefined target image with a 1-mm³ resolution in the standard space. Subsequently, DTI metric images in Montreal Neurological Institute-152 space of all of the subjects were converged into a single four-dimensional image. We set a threshold for the mean fractional anisotropy skeleton image with a default value of 0.2 to exclude voxels in gray matter and cerebrospinal fluid and then created a distance map. Distances from individual voxels to the skeleton were used to project DTI metric values onto the original mean fractional anisotropy skeleton according to the TBSS protocol [20]. Finally, DTI metric images of all subjects were established in the WM skeleton for subsequent analysis.

Cognitive test

Cognitive function was assessed with the Wisconsin Card Sorting Test (WCST)[28]. WCST scores were compared over the longitudinal time points with a repeated measures analysis of variance.

Statistical analysis

The data were analyzed in SPSS Statistics 20.0 for Windows. We used paired t tests to compare WM tract alterations and AVH symptoms before versus after AaP treatment. For AHRS score comparisons, p < 0.05 was considered significant. For WM tract alterations, the data were family wise error (FWE) corrected [29] with a threshold of p < 0.05 to assess small volume corrections.

Results

Treatment status and medication compliance

Of the 115 Hi-AVHs who participated in this study and volunteered to accept pharmacological treatment with risperidone (Johns and Johns, Xi’an Yang-Sen

Pharmaceutical Co., Ltd.), 79 completed the full study-indicated course of risperidone (dropout rate, 31.30%). The main reason for dropping out was an inability or unwillingness to tolerate adverse secondary effects (ASEs) of risperidone. Specifically, 12/36 study dropouts (33.3%) cited ASEs beyond the central nervous system and 8/36 (22.2%) cited prolactin increases. The 36 participants who dropped out had experienced, on average, a 30.5% reduction in AVH symptom severity (determined by the AHRS) after 2 months of AaP treatment (mean dosage, 260.20 ± 55.0 mg, chlorpromazine equivalent).

During the first 6 months of treatment, 46/79 participants (58.22%) experienced extrapyramidal syndrome, and 5/79 participants (6.32%) were found to have elevated prolactin levels; no other obvious ASEs were reported. The mean AaP dosage taken by the 79 retained participants was 270.80 ± 55.50 mg (chlorpromazine equivalent). Most of the retained participants (52/79, 65.8%) experienced a dramatic alleviation of their AVHs (AHRS scores reduced 66.7% compared to baseline). The remaining 27 retained participants experienced a slight alleviation of their AVH symptoms (AHRS scores reduced by on average 27.5%). The retained participants who experienced robust symptom reduction and those who experienced moderate symptom reduction were similar with respect to gender, AVH history duration, AVH symptom severity at baseline, and AaP dosage after acquisition of first MRI scanning data (see Table 1).

Table 1. Demographic and clinical characteristics of the study subjects.

Characteristic	Final cohort of Hi-AVHs* N = 39	Dramatic relief of AVHs N = 52	Slight relief of AVHs N = 27
Mean age, years	25.5 ± 2.8	24.0 ± 1.5	25.0 ± 1.0
Gender, no. males: no. females	14:25	22:30	17:10
Education, years	10.3 ± 2.81	12.5 ± 2.0	11.0 ± 2.5
AVH duration, months	129.0 ± 46.3	130.5 ± 45.0	126.5 ± 11.5
AaP dosage, mg/d chlorpromazine equivalent	270.8 ± 55.5	280.5 ± 60.0	260.0 ± 40.5
AHRS
Scores at baseline	20.5 ± 2.5.	21.0 ± 3.5	20.0 ± 2.0
Scores at 6 months	6.5 ± 2.5	7.0 ± 3.0	14.5 ± 4.5
Score reduction from baseline to 6 months	68.3%	66.7%	27.5%
Scores at 8 months	6.5 ± 1.0	7.0 ± 2.5	15.0 ± 4.0
Scores at 6 months vs. at 8 months	t = 0.117	p = 0.789
WCST
Error responses at baseline	47.0 ± 15.0	47.5 ± 13.5	46.5 ± 16.0
Error responses at 6 months	46.5 ± 13.5	47.0 ± 15.5	46.0 ± 14.0
Error responses at 8 months	47.5 ± 18.5	47.0 ± 18.0	47.0 ± 17.0
Repeated measures analysis of variance^#	p = 0.885	p = 0.980	p = 0.765

* Final cohort were retained participants (N = 79, after exclusion of dropouts) with TBSS-qualified MRI scans at the baseline and 8-month AaP treatment time point.

^# WCST results are represented as numbers of error responses, which were similar across the three time points.

MRI data findings

Of the 115 participants who agreed to undergo MRI scans, only 65 completed full baseline MRI scans, and only 49 completed both baseline and follow-up MRI scans. However, only 39 of the 49 participants who completed both scans had MRI data (DTI sequence acquisition) that were fully valid for TBSS analysis. The main reason for MRI data being unqualified was too much head movement in one or both scans (see limitations text in the Discussion).

Evolution of WM alterations and cognitive performance

Compared to the baseline assessment performed when the patients first accepted AaP, we observed a worsening of WM alterations after the patients had been taking AaP medication for 6 months, mainly in the corpus callosum, anterior commissure, and posterior commissure (Figure 1). The WCST test results, however, did not show evidence of any decline in the patients’ cognitive function at the 6-month time point. Based on the WCST results, we and the subjects agreed to continue the treatment. All 39 participants with qualified MRI data for both the baseline and the 6-month follow-up scans were among those who experienced dramatic AVH symptom improvement from baseline to the 6-month treatment time point and participated in the subsequent 2-month continuation study. For the 2-month continuation study, all 39 participants were re-briefed regarding potential risks of participating in the study and, subsequently, signed a second written informed consent form. The ethics committee of our hospital also approved the continuation study.

Figure 1. WM changes from baseline to the 6-month AaP treatment time point. Green pseudocolor represents normal WM and blue pseudocolor represents sites of WM impairment.

By the 8-month AaP treatment time point, we observed extensive impairment of WM tracts encompassing all major WM tracts, including the corpus callosum, arcuate fasciculus, anterior commissure, posterior commissure, cortico-spinal tracts, and other tracts linking the frontal, temporal, parietal, and occipital lobes (Figure 2). Despite these disheartening imaging findings, the WCST scores of each subject remained within normal range. Although the Hi-AVHs remained interested in continuing the AaP treatment, we did not agree and suggested that they instead try psychotherapy [30–32] or other interventions, such as avatar therapy or transcranial magnetic stimulation [33–36]. Comparing the 8-month neuroimaging results to those obtained at 6 months, we found that the WM impairments appeared to be even more widespread at the 8-month time point than they had been at the 6-month time point, though still mainly impacting the same regions (Figure 3). Meanwhile, the subjects did not experience further improvement in their AVH symptoms from the 6-month treatment time point to the 8-month treatment time point (Table 1).

Figure 2. WM changes from baseline to the 8-month AaP treatment time point. Green pseudocolor represents normal WM and blue pseudocolor represents sites of WM impairment.

Figure 3. WM structural differences from the 6- to the 8-month AaP treatment time point. Green pseudocolor represents normal WM and blue pseudocolor represents sites of WM impairment.

Discussion

In this pilot study, we observed mediocre treatment compliance, with a dropout rate of 31.30% principally due to ASEs, and a TBSS-qualified two-scan MRI data rate of only 33.91%. The relatively high dropout rate and low qualified MRI data rate limits the representativeness of our findings. Other strategies, such as repetitive transcranial magnetic stimulation, may be considered for those participants who are not well-suited for AaP therapy due to ASEs. Meanwhile, given the challenges we faced in obtaining TBSS-qualified MRI scans, we intend to develop an improved MRI scanning protocol and methodology to improve MRI data quality with this population. Despite the disappointing dropout rate and TBSS-qualified MRI data rate, we obtained useful information regarding AaP treatment effects and accompanying WM alterations in Hi-AVHs. The main findings of this pilot study were an association of AaP treatment with marked improvements in AVH symptoms within the first 6 months of treatment together with a persistent worsening and expansion of WM alterations in Hi-AVHs over time in the absence of detectable negative effects on cognition as determined by the WCST.

WM tracts mediate information sharing among brain regions, including modulation of diverse brain processes, including sensory, motor, cognitive, executive function, and emotive processes [37–42]. Many previous studies have reported the presence of WM aberrations in association with AVHs in patients with neuropsychiatric diagnoses, supporting the hypothesis that AVHs may be the result of functional co-disturbance of multiple brain networks [43–46].

Although AaPs have been demonstrated to be effective for alleviating AVHs in patients with schizophrenia, bipolar disorder, and borderline personality disorder [16], information about AaP treatment effects on WM alterations is limited. The development of widespread WM alterations in Hi-AVHs taking AaPs while the AaP treatment is alleviating AVH symptoms is a puzzling finding. Notwithstanding, antipsychotic mediations have been reported to induce gray matter and WM structural impairments in the brains of patients with schizophrenia, while also alleviating their hallucinations [47–49]. Further research is needed to clarify the clinical significance of these effects.

Limitations

This pilot study has notable limitations. First, as noted above, we did not obtain good treatment compliance or TBSS-qualified MRI data rates, which limits our ability to make strong inferences from our findings. Second, our study design was weaker than a randomized double-blind controlled study [50,51]. We compared data across time points with paired t tests, but we did not compare data from AaP-treated Hi-AVHs to data from an untreated Hi-AVHs control group or non-AVH control group. Third, we did not take into consideration the potential confounding factors of mood status or distress severity due to a lack suitable assessment tools. We are seeking suitable tools that will enable us to include such information in future research and thus obtain a more detailed characterization of brain alterations. Fourth, our final TBSS study sample included only 39 patients. Although Hi-AVHs are difficult to enroll, we are working to secure a larger sample to enable us to acquire higher accuracy findings in a future study. Fifth, we did not have the technical resources that would enable us to complete more detailed WM analyses, such as fractional anisotropy, radial diffusivity, and axial diffusivity. Additionally, given the extensiveness of the WM tracts affected, we could not establish a direct relationship between AaP dosage and WM impairment. We intend to collaborate with a senior mathematician to help us address these questions in the future. Likewise, due to technical limitations, we did not develop analyses of dynamic WM network alterations.

Conclusion

Despite the aforementioned limitations, in this pilot study, we found that AaP treatment was associated with marked WM deterioration while alleviating AVH symptoms in Hi-AVHs. Though the scope of WM alterations increased over time, the AaP treatment did not cause detectable cognitive impairments in Hi-AVHs. Considering the rapid WM changes observed, we do not recommend that AaPs be a first-line treatment for AVHs in Hi-AVHs. Rather, we suggest that other types of intervention, such as psychotherapy or avatar therapy, be used to treat AVH symptoms in Hi-AVHs. Finally, given our technical limitations, we welcome collaboration with scholars with complementary expertise for future research.

Contributors

Chuanjun Zhuo, Chunhua Zhou, Yong Xu conceived and designed the experiments; Xiaodong Lin, Gongying Li, Feng Ji, Langlang Cheng performed the experiments; Guangdong Chen, Wei Zhang, Zhonghua Su, Zhenqing Zhang, Wenqiang Wang analyzed the data, Chuanjun Zhuo, Xiaodong Lin wrote the paper. All authors approved the final version of the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by grants from the Tianjin Health Bureau Foundation (2014KR02 to C.Z.), the National Natural Science Foundation of China (81871052 to C.Z.), the Key Projects of the Natural Science Foundation of Tianjin, China (17JCZDJC35700 to C.Z.); the National Key Research and Development Program of China (2016YFC1307004 to YX), and the Multidisciplinary Team for Cognitive Impairment of Shanxi Science and Technology Innovation Training Team (201705D131027 to YX).