Parkinson’s Disease Progression and Statins: Hydrophobicity Matters

INTRODUCTION

Parkinson’s disease (PD) is a progressive neurodegenerative disorder marked clinically by motor dysfunction (bradykinesia, rigidity, and tremor), and pathologically by dopaminergic neuron loss and Lewy pathology characterized by α−synuclein aggregation [1–3]. There is a growing realization that PD involves many non-motor symptoms including cognitive decline, depression, autonomic, and olfactory change [4]. Despite effective symptomatic treatment, patients continue to experience progressive disability. Developing novel and/or disease-modifying therapies for PD is an urgent unmet need, although there currently are none showing efficacy [5–7].

Among agents being studied in this active research area [8], statins have been suggested to be neuroprotective [9, 10]. Previous epidemiological studies have focused on statins and incidence of PD, yielding inconsistent results [11–15]. The effect of statins on clinical PD progression has received less attention, although a recent retrospective study showed early-stage PD patients on long-term statin therapy had less motor impairment over four years [16]. Two recent clinical trials testing lovastatin [17] and simvastatin [18], however, reported no impact on motor symptom progression, although mean percentage change in the striatal ¹⁸F-dopa uptake ratio was less with lovastatin use [17].

Whereas statins are known for inhibiting cholesterol synthesis, in vitro studies also have demonstrated that they potentially antagonize pathological processes in PD [10, 19–22]. Statins, however, vary markedly in hydrophobicity and this impacts their ability to cross the blood-brain barrier (BBB) [23]. Since cholesterol synthesis in the central nervous system occurs de novo and is largely independent of that in the periphery [24], hydrophobicity may be an important determinant of statin effects on PD progression.

Leveraging a PD biomarker study cohort with longitudinal clinical and imaging data, we investigated a potential effect of statins on PD progression.

METHODS

RESULTS

Statin effect on clinical measurements

Change from baseline

Statin non-users did not demonstrate a significant change in MDS-UPDRS scores over time (p > 0.24; Table 2 and Fig. 1A). Statin users also did not show longitudinal changes in MDS-UPDRS-I or –II scores (p > 0.15), but MDS-UPDRS-III scores decreased (improved; β= –5.7, p = 0.017; Table 2). When parsed by hydrophobicity, lipophilic statin users did not demonstrate significant changes in MDS-UPDRS scores over time (p > 0.17; Table 2 and Fig. 1A). Hydrophilic statin users, however, had significantly increased (worse) MDS-UPDRS-1 (β= 4.2, p = 0.017; Table 2 and Fig. 1A) and trend increased MDS-UPDRS-II (β= 3.5, p = 0.11) scores, but decreased (improved) MDS-UPDRS-III scores (β = -9.1, p = 0.029; Table 2).

OPEN IN VIEWER

Table 2

The effects of statin usage on progression of clinical measures and MRI susceptibility

	Change from baseline (SE)	p for change	Effect estimate (SE)	p for group difference
MDS-UPDRS-I
Statin non-users	–0.6 (0.6)	0.32	Reference
Statin users	0.8 (1.0)	0.42	1.5 (1.2)	0.22
Lipophilic statin users	–0.8 (1.2)	0.50	–0.2 (1.4)	0.90
Hydrophilic statin users	4.2 (1.7)	0.017	4.8 (1.8)	0.010
Hydrophilic vs. Lipophilic	NA		5.0 (2.1)	0.020
MDS-UPDRS-II
Statin non-users	–0.4 (0.8)	0.66	Reference
Statin users	1.8 (1.3)	0.15	2.2 (1.5)	0.14
Lipophilic statin users	1.0 (1.5)	0.51	1.4 (1.7)	0.43
Hydrophilic statin users	3.5 (2.2)	0.11	3.9 (2.3)	0.10
Hydrophilic vs. Lipophilic	NA		2.5 (2.7)	0.35
MDS-UPDRS-III
Statin non-users	–1.8 (1.5)	0.24	Reference
Statin users	– 5.7 (2.4)	0.017	–3.9 (2.8)	0.16
Lipophilic statin users	–3.9 (2.9)	0.17	–2.2 (3.2)	0.50
Hydrophilic statin users	– 9.1 (4.1)	0.029	–7.3 (4.3)	0.10
Hydrophilic vs. Lipophilic	NA		–5.1 (5.0)	0.31
LEDD
Statin non-users	64.3 (40.9)	0.12	Reference
Statin users	81.1 (63.4)	0.20	16.8 (75.3)	0.82
Lipophilic statin users	82.9 (77.8)	0.89	18.0 (87.8)	0.84
Hydrophilic statin users	72.3 (111.2)	0.99	7.4 (118.5)	0.95
Hydrophilic vs. Lipophilic	NA		–10.6 (135.7)	0.94
SNc R2*
Statin non-users	1.6 (0.6)	0.010	Reference
Statin users	3.7 (1.1)	0.001	2.1 (1.2)	0.08
Lipophilic statin users	2.8 (1.3)	0.039	1.2 (1.4)	0.41
Hydrophilic statin users	5.3 (1.7)	0.003	3.7 (1.8)	0.043
Hydrophilic vs. Lipophilic	NA		2.5 (2.2)	0.24

Data represent estimates on effect of statins on specified outcomes at 18 months while adjusting for age, gender, and disease duration. Non-statin is the reference group unless otherwise stated. Effect estimates represent beta estimates from linear mixed effect model and represent change in addition to that seen in the reference group. Values for statin users represent PD subjects taking any statin and are derived from a model considering statin use vs. non-statin use. Values for non-statin, lipophilic, and hydrophilic are derived from models considering statin use, use of lipophilic statin, or use of hydrophilic statin. LEDD, Levodopa equivalent daily dosage; MDS-UPDRS, Movement Disorders Society Unified Parkinson’s Disease Rating Scale, I, non-motor experiences of daily living, II, motor experiences of daily living, III, motor examination.

OPEN IN VIEWER

Fig. 1

Depiction of changes in (A) MDS-UPDRS-I and (B) SNc R2^* over 18 months in the study population. Data represent mean and error bars represent SEM.

Group differences

Compared to statin non-users, overall statin use did not significantly influence PD clinical progression (p > 0.14; Table 2). When parsed by hydrophobicity, lipophilic statin use also did not impact clinical progression significantly compared to statin non-users (p > 0.43; Table 2). Hydrophilic statin users, however, showed increased (worse) MDS-UPDRS-I (β= 4.8, p = 0.010; Table 2 and Fig. 1A) and trend increased MDS-UPDRS-II (β= 3.9, p = 0.10) scores, but an opposite trend decrease (improvement) in MDS-UPDRS-III scores (β=−7.3, p = 0.10) (Table 2). Compared to lipophilic statin users, hydrophilic statin users had faster progression of MDS-UPDRS-I scores (β= 5.0, p = 0.020; Table 2 and Fig. 1A). These results were not altered by further correcting the models for baseline serum LDL levels (Table 3).

OPEN IN VIEWER

Table 3

The effects of statin usage on progression of clinical and imaging metrics after adjustment for baseline serum LDL

	Change from baseline (SE)	p for change	Effect estimate (SE)	p for group difference
MDS-UPDRS-I
Statin non-users	–0.6 (0.6)	0.32	Reference
Statin users	0.8 (1.0)	0.42	1.5 (1.2)	0.23
Lipophilic statin users	–0.8 (1.2)	0.50	–0.2 (1.4)	0.90
Hydrophilic statin users	4.2 (1.7)	0.018	4.8 (1.8)	0.011
Hydrophilic vs. lipophilic	NA		5.0 (2.1)	0.020
SNc R2*
Statin non-users	1.7 (0.6)	0.008	Reference
Statin users	3.7 (1.1)	0.001	2.0 (1.2)	0.10
Lipophilic statin users	2.8 (1.3)	0.039	1.1 (1.4)	0.44
Hydrophilic statin users	5.3 (1.7)	0.003	3.6 (1.8)	0.050
Hydrophilic vs. lipophilic	NA		2.5 (2.2)	0.25

Data represent estimates on effect of statins on specified outcomes at 18 months while adjusting for age, gender, and disease duration. Non-statin is the reference group unless otherwise stated. Effect estimates represent beta estimates from linear mixed effect model and represent change in addition to that seen in the reference group. Values for statin users represent PD subjects taking any statin and are derived from a model considering statin use vs. non-statin use. Values for non-statin, lipophilic, and hydrophilic are derived from models considering statin use, use of lipophilic statin, or use of hydrophilic statin. MDS-UPDRS, Movement Disorders Society Unified Parkinson’s Disease Rating Scale, I, non-motor experiences of daily living.

DISCUSSION

Preclinical, epidemiological, and clinical trial data and their interpretations regarding the effects of statins on PD progression have been controversial, complicated by their cholesterol-lowering and pleiotropic effects. Findings from the current study are consistent with our hypothesis that hydrophobicity is also an important determinant of statin effects on PD. Consistent with previous studies [17, 18], the current results indicate that lipophilic statins are not associated significantly with clinical or imaging progression. Hydrophilic statin use, however, significantly is associated with increased non-motor (MDS-UPDRS-I) scores and imaging (SN R2^*) values. We also found a consistent trend with motor scores using MDS-UPDRS-II but not MDS-UPDRS-III scores that we previously reported to be problematic in tracking PD progression [27]. Future studies are needed to test if these findings can be replicated since the results may have both scientific and clinical implications.

Statins have been suggested to be neuroprotective [9] but previous epidemiological studies focused on statins and incidence of PD, with some reporting lower PD risk among statin users [11, 32–35] and others not [12, 37]. The interpretation of these data, however, is not straightforward. For example, two recent meta-analyses suggested a significant publication bias toward reporting protective effects of statins in PD [38, 39]. In addition, the interpretation also may be confounded by the fact that statins differ in their ability to cross the BBB. Given their BBB penetrating property, most preclinical studies reporting potential neuroprotective effects of statin are focused on lipophilic compounds [40, 41]. Lee and colleagues [42] reported that discontinuing statins, especially lipophilic ones, was associated with higher PD risk. This was interpreted as compelling evidence for neuroprotection, although the number of subjects in these studies was very small. In 2017, we utilized a large US claims database (MarketScan) and found that statin usage (especially lipophilic ones) is associated with higher risk of PD [15]. Since then, we have not noted any other studies reporting their stratified results, although some may be pending in press.

The effect of statins on clinical PD progression has received less attention overall until recently. A recent retrospective study indicated PD patients on long-term statin therapy showed less motor impairment over four years [16]. Two prospective clinical trials have tested the lipophilic statins lovastatin [17] and simvastatin [18], but neither had a beneficial effect on measures of clinical progression (MDS-UPDRS scores). The “trend” noted in one study [17] was far from significant (p = 0.14) and should have been noted as a negative result. Consistent with the overall conclusion of these trials, the current study found that lipophilic statin use did not impact clinical progression measures. In addition, we also demonstrated that lipophilic statins did not significantly change imaging progression (SN R2^* values). This is in contrast to the lovastatin study that reported less mean percentage change in the striatal ¹⁸F-dopa uptake ratio with lovastatin use [17]. The difference in regions (SN vs. striatum), imaging metrics/sensitivities (susceptibility vs. positron emission tomography), and disease duration (∼3 y [17] vs. ∼6.5) may contribute to the discordant results. It is possible that lipophilic statins may have neuroprotective effects in the early course of the disease, such as in prodromal states, but not later disease stages. Additional studies are needed to determine the effect of lipophilic statins on imaging metrics.

This study suggests that hydrophilic statin use may be associated with faster PD progression. For clinical metrics, this was evident particularly for non-motor symptoms (MDS-UPDRS-I scores), with a trend increase in reported motor symptoms (MDS-UPDRS-II scores). PD traditionally is characterized by motor symptoms but there is growing evidence that non-motor symptoms are burdensome and impact patient quality of life [43, 44]. The current results are consistent with previous studies reporting increased MDS-UPDRS-I and –II scores over time [45, 46], even though our MDS-UPDRS-II findings were at the trend level. These types of patient-reported questionnaires seem to be more reliable in capturing PD progression than rater-based exams such as the MDS-UPDRS-III [27]. It should be noted, however, that the MDS-UPDRS-I assesses only some non-motor symptoms and shows floor effects [47]. Thus, future studies should consider including scales that may evaluate a broader spectrum of non-motor symptoms.

Among MRI metrics evaluated as potential PD progression markers, SNc MRI susceptibility (specifically R2^*) has emerged as one that can distinguish PD from controls and track PD progression [28, 49]. To the best of our knowledge, this is the first study to utilize MRI susceptibility values, an objective metric, to evaluate statin effects in PD. The results indicate significantly faster R2^* imaging progression in hydrophilic statin users compared to statin non-users. Although the comparison of SN R2^* values between lipophilic and hydrophilic statin users did not reach statistical significance, the direction of the effect is similar to that of the clinical metrics (MDS-UPDRS-I). Quantitative susceptibility mapping (QSM) has been suggested to reflect iron deposition better than R2^* [28, 51]. Previous studies have indicated, however, that R2^* captures PD progression more robustly [28, 51] and thus it was the metric of choice for the current study. Replication of the current findings in other cohorts, however, is needed. Collectively, the current study suggests in a small cohort of PD patients that hydrophilic, but not lipophilic, statins may be associated with faster PD progression.

We need to be extremely cautious in interpreting this observation because the exact mechanism is unclear. It is possible that hydrophilic statins may have decreased beneficial pleiotropic effects compared to lipophilic statins on degenerating neurons, or they may impose unidentified adverse effects indirectly because they do not cross the BBB [52, 53]. It also is worth noting that the overall finding was not supported by rater-based MDS-UPDRS-III scores, although the MDS-UPDRS-I, II, and imaging data converge. In the past, we have found that using MDS-UPDRS-III scores to track PD progression is problematic [27]. This scale is entirely rater-dependent and captures only a fraction (five minutes) of motor performance. In addition, MDS-UPDRS-III is affected by drug treatment because most clinicians adjust drug dosages according to their impression of motor progression. Thus, despite our best efforts, MDS-UPDRS-III has not been a good metric for PD progression [27], and improved MDS-UPDRS-III scores over time in the current study merely may reflect higher drug usage [54]. Indeed, the two recent trials testing statin effects in PD failed to show effects on motor symptom progression reflected by MDS-UPDRS-III scores [17, 18].

The study, however, also has other limitations. First, the sample size in the statin subgroups is small and there could be other explanations regarding increased MDS-UPDRS-I scores in the hydrophilic PD group that may not relate to statin usage. The observational nature of the study also may lead to selection bias. Second, statin duration/dose was captured by participant recall that may not have been reliable. Although we adjusted our analysis with baseline cholesterol level, we were not able to obtain pre-PD cholesterol levels or its dynamic change under statin treatment before and during our study. In addition, we did not adjust for other co-morbidities that may influence the choice of statin subtype use. Third, the association of SNc R2^* values (potentially reflecting underlying pathology) to MDS-UPDRS-I scores is not straightforward since MDS-UPDRS-I scores evaluate non-motor aspects of daily living that arise from dysfunction of both dopaminergic (e.g., SNc) and non-dopaminergic systems. Some previous studies have suggested statins may reduce dopaminergic neurodegeneration [10, 19–22]. If true, it would be reasonable to hypothesize that nigral pathology (reflected by SNcR2^*) would correlate with motor function (e.g., MDS-UPDRS-II and -III scores). The trend association between SNc R2^* and MDS-UPDRS-II scores is consistent with this notion. The fact that MDS-UPDRS-III scores decreased may reflect the intrinsic limitation of the score [27] and/or the possibility that some PD patients were not medically optimized at baseline. Fourth, both clinical progression and imaging data may be influenced by the use of anti-PD drugs, which this study cannot tease out. Because the decision regarding statin usage and its subtype choice (generally by a physician other than a movement disorder specialist) is independent of anti-PD drugs usage, any bias is unlikely conscious. Fifth, reverse causation cannot be ruled out since statin use was not randomized at baseline and the statin groups were not matched equally on all clinical variables. The follow up time also was limited to 18 months. Larger randomized trials with matched groups and longer follow up times can address these important factors in future studies.

In summary, this study is consistent with previous studies showing that lipophilic statins do not influence PD clinical progression. Moreover, it suggests that hydrophilic statins significantly increase clinical and imaging progression in PD patients. Future studies are needed that integrate the insights we learned from this explorative study and may have both clinical and scientific implications.