A systematic review of body mass gain after deep brain stimulation of the subthalamic nucleus in patients with Parkinson's disease

This systematic review investigated the effects of deep brain stimulation of the subthalamic nucleus on extent and time course of body mass changes in patients with Parkinson's disease. A computerized search identified relevant articles using a priori defined inclusion and exclusion criteria. A descriptive analysis was calculated for the main outcome parameters body mass and BMI.


| INTRODUCTION
Body mass changes in both directions, weight loss and weight gain, have been reported in patients with Parkinson's disease (PD) 1 and are often observed in response to treatment. Deep brain stimulation (DBS) of either the subthalamic nucleus (STN) or globus pallidus internus (GPi) is an efficient treatment option for severe motor complications in advanced PD. Hence, DBS significantly improves motor and non-motor fluctuations, rest tremor, dyskinesias as well as quality of life (QoL), and usually results in a decrease of dopaminergic medication. [2][3][4][5][6][7] Side effects may include speech disturbance, postural instability, mood changes, and significant body mass gain independent of the DBS target region. [8][9][10] The extent and time course of body mass gain after DBS surgery have not yet been systematically addressed 11 although some potential mechanisms of body mass gain have been suggested like improvement of resting tremor and dyskinesias, 4,12 reduction in energy expenditure, 13,14 changes in eating behaviour and food intake, 12,[15][16][17][18][19][20] as well as alterations in hypothalamic adipokine release. 10,18,19,[21][22][23] The purpose of this study is to systematically analyze the extent and range of body mass gain after STN DBS. We selected available studies investigating postoperative body mass changes and conducted a systematic review to quantify stimulation-induced body mass changes.
We aimed to clarify i) the magnitude of the relationship between STN DBS and body mass gain, ii) the extent of body mass gain and BMI gain, and iii) the time course of assumed body mass and BMI gain. OR (increase))) AND ((STN DBS) OR (Subthalamic nucleus deep brain stimulation) OR (deep brain stimulation) OR (DBS) OR (GPi deep brain stimulation) OR (globus pallidus deep brain stimulation) OR (pallidal deep brain stimulation))). Search was performed for articles published between 1984 and 2017 and was restricted to English and German publications, but not to age and gender of subjects, as well as origin of publication.

| Study Selection and Data Collection
All abstracts and articles of the computerized search were independently screened by two investigators (JS, BW) for potential relevance.
Any disagreements were resolved by further examination of a third investigator (NB) and via consensus.
The following studies were excluded: reviews, letters, commentaries, abstracts, posters, case reports, correspondences to articles, and double nominations of publications in different search portals. Furthermore, animal studies, studies including DBS of the GPi, ventral intermediate thalamic nucleus (VIM) or caudal zona incerta (cZi), studies with alternative surgical methods (e.g. pallidotomy), articles reporting non-weight related outcomes, studies assessing body mass gain in another disease and studies targeting other research questions were not considered.
The included studies had to contain at least one of the following outcomes: absolute body mass before and after STN DBS or body mass changes, absolute body mass index (BMI) before and after STN DBS or BMI changes. Normal weight (BMI: 18.5-24.9 kg/m 2 ), overweight (BMI: 25.0-29.9 kg/m 2 ) and obesity were defined according to the WHO definition. THE WHO defines overweight and obesity for adults as follows: overweight is a BMI greater than or equal to 25 and obesity is a BMI greater than or equal to 30. 24 Additionally, UPDRS III and IV scores, as well as levodopa equivalent doses (LEDD) were investigated to reveal the efficacy of the DBS treatment. Moreover, sufficiently specified numerical baseline and follow-up outcome data for body mass, BMI, UPDRS III, UPDRS IV, and LEDD were required as well as data on standard deviations (SD) or standard errors of the mean.

| Statistical analyses
Results are reported as mean ± SD. Data were analyzed using Excel

| Study Selection and Characteristics
The literature search identified 206 potentially relevant articles of which 154 studies were assessed for a more detailed evaluation ( Figure 1). Following the selection process, 54 studies 1,2,4,6,8-10,12-21,23,25-58 fulfilled the inclusion criteria as described above. For the analysis, 38 studies could be included of which 18 (47%) were prospective case studies, 12 (32%) were prospective case-control studies, 4 (11%) were retrospective case studies, 2 (5%) were retrospective case-control studies, 1 (2.5%) was a cross-sectional, and 1 (2.5%) was a retrospective survey study. In sum, the 38 selected studies included 979 patients with PD and STN DBS and 287 controls comprising of non-stimulated patients with PD under medical treatment (N = 186) and healthy control subjects (N = 101).

| Effects of STN DBS on motor function
In the 25 studies (N = 696) with complete UPDRS III datasets, 4

| Predictors of weight gain following STN DBS
In search for predictive factors of body weight gain after surgery, we performed a correlation analysis of the following variables: delta weight, delta BMI, delta LEDD, delta UPDRS III, delta UPDRS IV, disease duration, age, as well as weight preoperatively. Change in weight was correlated with age (r = -0.4239, p = .031; Figure 4). Regarding the symptoms of PD, mean change in BMI was positively correlated with mean change in LEDD (r = 0.440, p = .0231) and with mean change in UPDRS III scores when 'on L-Dopa' (r = 0.502, p = .010; Figure 4).

| DISCUSSION
The focus of this systematic review was to provide a comprehensive analysis of recently published studies that investigated body mass gain after STN DBS as a starting point for the development of new approaches to prevent this clinically relevant side effect. All but one study (50) reported weight gain after DBS with no study reporting weight loss. Thus, there is strong and consistent evidence for weight gain after STN DBS affecting the vast majority of patients.
The body weight gain occurs already in the first months after DBS implantation and appears to stabilize after one year. The maximum body mass gain across studies was 5.9 kg after one month 38 and 11.1 kg one year after DBS. 13 Nine studies investigated the body mass change 3 months after the surgery showing a mean increase in body weight of 3.25 kg. There was no detailed discussion about the exact mechanisms of this rapid weight gain in these articles. Interestingly, the Swedish Obese Subjects (SOS) study, which investigated the effects of bariatric surgery on subjects with obesity, found a mirrored effect. Here, subjects with obesity lost weight very rapidly during the first months after surgery followed by a plateau phase. 60 There seem to be general mechanisms that may drive these rapid body weight changes after interventions. Further investigations are required to address these effects more precisely.
As our patient sample had normal weight prior to surgery, the body mass gain does not necessarily compensate for preoperative malnutrition or underweight which is well in line with recent evidence. 1,18,34,35 In this context, it is also important to assess the changes in body composition after DBS-surgery. However, only a few studies reported in detail on changes in body composition. There is evidence that females gained disproportionately fat mass while weight gain in men was driven by both, fat free mass and fat mass. 6,13,14,23,[40][41][42]48,54 Interestingly, patients with a greater improvement of motor dysfunction and a stronger reduction in LEDD were likely to gain less weight arguing that an optimal lead localization in the sensorimotor part of the STN is associated with a lower likelihood to develop this side effect. This finding seems to be contractionary on the first view.
On the one hand, DBS improves rigidity and rest tremor which would otherwise contribute to a higher preoperative energy expenditure that

| Impact of dopamine replacement therapy
The impact of dopamine replacement therapy on weight changes in patients with PD are only rarely studied and are still controversial. It has been shown that patients on levodopa significantly loose body weight within one year of treatment in comparison to patients on dopamine agonists. 61 Here, the effect was dose-dependent: higher LEDD at baseline in levodopa-treated patients was associated with a more rapid weight loss. 61 In contrast, other studies found no clear relationship between BM changes and LEDD although the results may depend on the type of medication. 35 For example, ropinirole had no effect on BM, whereas cabergoline and pergolide led to unintentional weight loss, and pramipexole increased BM in patients with PD. 35 Besides the role of dopamine in motivational and reward processing, catecholamines are also involved in the regulation of brown adipose tissue (BAT) thermogenesis. Recent evidence revealed that BAT-dependent non-shivering thermogenesis is involved in regulation of body weight and could increase insulin sensitivity. 62 One possible suggestion is that dopamine replacement therapy may facilitate mitochondrial UCP1-induced thermogenesis, which could potentially also influence BM.

| Physiological mechanisms
Body mass homeostasis is a complex and multifactorial process that is determined by physiological, metabolic, genetic, epigenetic, motivational, and behavioural factors. 34,35 In PD, body mass is a non-motor

| Changes in motor function
A plausible mechanism for postoperative body weight gain is the amelioration of motor sign severity. The improved motor function, i.e. due to reduction of rigidity, dyskinesias, limb akinesia, tremor and improvement of gait as well as reduction of dopaminergic medication collectively give rise to reduced energy expenditure. 4,12,[25][26][27][28][29]37,51 In contrast, other studies found no correlation between body mass gain and changes of the UPDRS III score. 14,47,51 Taking advantage of our systematic approach however, we found less or no weight gain in patients with a higher improvement of disease severity. This observation challenges the concept of reduced energy expenditure as a relevant mechanism whereas the localization of DBS electrodes may have a stronger impact on weight changes. 4,19 In keeping with this notion, the distance of the active electrode to the wall of the third ventricle in mediolateral direction is inversely correlated with body mass gain and UPDRS III score of the contralateral extremities. Patients with more laterally located electrodes had a better motor improvement and gained less body mass than patients with at least one more medially located electrode. 4 Moreover, it is well known that the CNS, and especially the hypothalamus plays an important role in the regulation of glucose homeostasis and peripheral insulin sensitivity. It has been shown that some of these neurons in these nuclei are assigned to glucoregulatory properties, which could potentially be co-stimulated due a more medial electrode position. [63][64][65][66][67] This finding suggests that the STN is involved or may even exert a regional effect on adjacent structures that are involved in the regulation of energy balance, reward, and food intake. If the active electrode is more located in the limbic subdivision of the STN, stimulation could act as an internal stressor to the limbic system. At the current state of knowledge, DBS represents a 'noise source' considered to disrupt abnormal bursting activity in the parkinsonian basal ganglia 43,68 and influences thereby motivational processing and the reward system, because the limbic cortico-basal ganglia loop is not only involved in reward processing and hedonic control of food intake but additionally interconnected with the hypothalamus, which regulates energy homeostasis. 68

| Metabolic changes
Previous studies have proposed that the weight gain could be seen as a homeostatic response to the previous disease-related weight loss 6,14,41 assuming that patients with PD may normalize their weight back to their premorbid status. As STN DBS is associated with a weight gain exceeding the previous weight loss 33 A compelling hypothesis is a reduced secretion of growth hormones with consequently decreased lipolysis. 35 Furthermore, a drop of HDL cholesterol concentrations has been observed. 32,48 Additionally, one study found an increased glucose oxidation after DBS implantation. 48 These results remain contradictory as the basal glucose production and insulin sensitivity were reported to be unchanged in a different study. 58 To sum up, there is first evidence that STN DBS affects glucose and lipid metabolism, but this is still contrarily discussed. 20,32

| Changes of brain function
Recent PET studies using 2-deoxy-2[18F]fluoro-D-glucose tracer found a correlation between STN DBS-related weight gain and metabolic changes in associative and limbic brain areas, but no correlation with sensorimotor brain regions. 15,30 These findings suggest that the STN might be involved in motivational processing related to eating behaviour. 15 Indeed, the STN is connected to the limbic system, especially to the ventral tegmental area and ventral pallidum, which are key structures of the reward system. 57  One assumption for these endocrine alterations is that the spread of stimulation current beyond the borders of the STN may influence the hypothalamic regulation of hormone secretion, energy homeostasis, 2,35,47,55 and may disrupt the melanocortin system, which has been linked to obesity. 42 The disrupted hypothalamic hormone secretion leads to alterations in the central appetite mechanism. 34,43,59 Moreover, it has been shown that cortisol levels are normalized after STN DBS and the respective anabolic effect of this normalization process has been hypothesized to drive the body mass gain. 23,43,54,76 In addition, it has been shown that cortisol levels decreased over time after DBS 23,55 and that this decrease was correlated with the position of the active electrode in the STN. The more medially the electrode was located, the greater was the decrease in cortisol levels. Furthermore, lower cortisol levels were strongly associated with weight gain and higher trait anxiety.
These results seem to be contractionary to the relationship between cortisol levels and body mass gain with patients with abdominal obesity having increased cortisol levels. An important predictor of the impact of cortisol on metabolism is cortisol responsiveness. High cortisol responsiveness is associated with greater propensity to gain body mass in comparison to low responsiveness. This difference in susceptibility isat least in partdependent on different physiological factors such as gender and pregnancy. Furthermore, inter-individual differences in stress response are also determined by genetic background. Putting this foreword, cortisol responsiveness may be different within the assessed subjects. Overall, there is strong evidence that cortisol modulates food intake and therefore impacts on body weight, but also different stressors are known to elicit different cortisol responses. 71 Thus, STN DBS may mimic the effect of chronic stress and disturbed limbic and motivational systems. 55 However, these findings are still contradictory, because one study revealed that hormone levels of the hypothalamic-adrenal-, hypothalamic-somatotropic-, hypothalamic-gonadal-axis were 3 or 6 months after DBS device implantation considered as normal. 34,43 Thus, it is still under discussion how hormonal changes may contribute to body mass gain.

| Limitations
Our meta-analysis was calculated based on incomplete data sets that in part were imprecise with regard to the reporting of several variables of interest. The sub-analyses in this article therefore included different numbers of subjects. Some relevant studies, e.g. 16,17 could not be included as the exact time point of assessment after surgery was missing.
Thus, we were not able to generate forest and funnel plots due to missing data. Given that only six studies included sufficient information about control groups, it was not possible to calculate the necessary odds ratios.
Further limitations are the limited range of clinical disease severity due to guidelines for the treatment with DBS and the impossibility to randomize groups. Some reports found no change in food intake, appetite, or hunger which could be due to the fact of inaccuracy of selfreported intake. 33 In addition, most studies covered only a limited time period (in most cases 12 months). Longer assessment periods are warranted to investigate the complete time course of body mass changes.

| CONCLUSIONS
Deep brain stimulation is an efficacious technique that greatly Therefore, the clinical implications from our results is that all patients should be informed that weight gain may occur as a consequence of DBS. Potential candidates for this treatment may be given preoperative nutritional counseling, physiotherapy, and sports therapy after the implantation to prevent rapid weight gain leading to obesity.
Moreover, larger and better controlled trails are needed to establish long-term efficacy of nutritional intervention studies.