Introduction

According to the working definitions of sex and gender provided by the Institute of Medicine’s Committee on Understanding the Biology of sex and gender differences, the term sex refers to the classification of living things, generally as male or female according to their reproductive organs and functions assigned by chromosomal complement, while the term gender refers to a person’s self-representation as male or female, or how that person is responded to by social institutions based on the individual’s gender presentation [1]. However, these terms are not univocal and cannot always be used in a mutually exclusive fashion. Thus, since gender is rooted in biology and shaped by environment and experience [1], gender rather than sex is more appropriate to define the interaction of biological and social elements affecting health outcomes [1].

Sex differences in brain structure and function initiate through sex determining genes and fetal hormonal programming and have important implications for brain-based disease risk; then, sex-specific genetic and hormonal factors, as well as age-related physical changes further contribute to biological differences in expression of neurodegenerative diseases, including Parkinson’s disease (PD) [1]. In addition, a variety of broader societal factors, including role expectations and social attitudes, also have roles in the risk, course and outcome of neurodegenerative diseases. As a matter of fact, a range of behavioral and lifestyle choices associate with gender differences, including diet, exercise, smoking and caffeine are emerging as potential modifiers of the PD risk during life [2].

Since the official and systematic inclusion of sex and gender in biomedical research [3], gender differences have been acknowledged as important determinants of both the susceptibility to develop neurodegenerative diseases in general population and the clinical and therapeutic management of neurodegenerative patients [4].

The aim of this review is to gather the available evidence on gender differences in PD regarding clinical phenotype (including motor and non-motor symptoms), biomarkers, genetics and therapeutic management (including pharmacological and surgical treatment). Before the closedown, we will briefly discuss the role of estrogens in determining such differences. However, since our approach is mainly clinical, we will not include in the present review all the preclinical data available on the topic.

Indeed, improving our understanding in this field may result in implementation of strategies to identify prodromal PD cases and speed efforts to discern new directions for PD tailored treatment and management.

Methods

The authors searched personal files and PubMed for peer-reviewed articles published in English language with no time limits. The search terms “agonist”, "biomarker", “deep brain stimulation”, "epidemiology", “gender”, “genetic”, “kinetic”, “levodopa”, “men”, “motor features”, “non-motor features”, “Parkinson”, “sex”, “surgery”, “weight” and “women” were used. Additional articles were identified by searching the reference lists of identified reviews that provided insightful or comprehensive overviews on gender differences in PD. The studies and meta-analysis considered in this review are detailed in Table 1.

Table 1 Studies and meta-analyses focused on gender differences in Parkinson’s disease included in the review
Full size table

Epidemiology and phenotypic differences

Epidemiology

Confirming previous data [5], a meta-analysis including 17 relevant studies and over 2500 PD cases, determined an overall age-standardized incidence M:F ratio of 1.46 (95% CI 1.24–1.72) [6]. This meta-analysis also disclosed a high-level heterogeneity between included studies and a positive relationship between age of onset and M:F incidence ratio [6, 7]. As for prevalence, a meta-analysis including data published between 1985 and 2000 reported a significant difference in gender ratio for individuals from 50 to 59 years old, with a PD prevalence of 41/100.000 in women and 134/100.000 in men (p < 0.05) [8]. When stratified by geographic location, M:F prevalence ratios were in favour of men in both the Western countries and South America, but not in Asia, although methodological issues may account for such discrepancy [8].

By analyzing the Health Insurance drugs reimbursement databases, a recent French nationwide study largely confirmed previous data reporting an overall M:F incidence ratio of 1.49 (95% CI 1.41–1.57, p < 0.001) and an overall M:F prevalence ratio of 1.48 (95% CI 1.45–1.51, p < 0.001) [9]. Both M:F incidence and prevalence ratios were markedly influenced by age in a strikingly progressive pattern, with incidence increasing by 0.14 per 10-year age increment and prevalence increasing by 0.05 for every 10-year age increment. As expected, this pattern was more pronounced for incidence than prevalence ratios, since the latter is most likely affected by gender differences in survival [9]. Indeed, male sex is associated with an increased mortality rate in PD for two complementary reasons [10–12]. First, it is the reflexion of the overall shorter life expectancy of men in the general population. Then, the factors usually predicting higher mortality in PD, as cognitive impairment and higher postural instability and gait scores, are much more common in men [10–12].

Motor symptoms

By examining a clinic-based cohort of 253 subjects, Haaxma et al. first delineated phenotypic gender differences in a large sample of PD patients [13]. Key findings of this study were: (a) women were 2 years older than men at symptom onset and presented more likely with tremor (67%) than men (48%); (b) tremor dominance was associated with a slower decline on motor scales; (c) at symptoms onset, women had 16% higher striatal [123I]FP-CIT binding than men; (d) in women, age at onset correlated positively with parity, age at menopause and fertile life span [13]. Taken together, these findings suggest a more benign phenotype in women with PD possibly related to the estrogens status [13, 14]. Indeed, gender differences in PD presentation may be attributable to biological factors; however, health-care seeking behavior should not be overlooked [15]. A study evaluating tertiary care referral in PD showed that the expected duration from onset to the movement disorders specialist visit for women was 61% greater than for men (p = 0.003) [16]. The effect of gender remained significant when adjusting the model for disease severity and historical features (age of onset and family PD history) and supported a referral delay in women irrespective of disease features [16].

Yet, once the disease has started, evidence reports shorter time to develop wearing off and dyskinesia in women than in men, arguing against the theory of the protective effect of estrogens [17–19]. According to a cross-sectional study enrolling 617 PD patients, the prevalence of wearing off was higher in women (72.5% versus 64%, p = 0.034) with female gender conferring an increased risk for wearing off equal to 80.1% [19]. Confirming previous findings [20], a prospective, population-based study on 189 de novo, PD patients showed that gender is the most important independent predictor of levodopa-induced dyskinesia, with an almost threefold increased risk in women compared to men, irrespective of body weight [18, 21]. As a matter of fact, along with younger age of onset, female gender has been associated with a shorter time to occurrence of levodopa-induced dyskinesia (Hazard Ratio = 1.82; 95% CI 1.14–2.89, p = 0.011) with a median time to dyskinesia of 4 years in women and 6 years in men [22].

Yet, the role of gender in determining either a more benign or aggressive motor disease course is far to be clear.

Non-motor symptoms

Although methodological issues (as the use of different scales) limit the comparison of the available data, the majority of studies suggest the existence of gender-related differences in non-motor symptoms (NMS) prevalence in PD. As a matter of fact, several studies showed that feelings of nervousness and sadness, constipation, restless legs and pain are more common in women, while daytime sleepiness, dribbling saliva, reduced interest in sex and problems having sex are more prevalent in men [23–27]. Indeed, since it is known that dopaminergic treatment may affect several NMS differently [28], as a major limitation these studies only included patients on dopaminergic treatment. By administering the Non-Motor Symptoms Questionnaire to 200 early, drug-native PD patients and 93 age- and sex-matched healthy controls, we were able to show PD-specific gender differences in NMS, irrespective of disease progression and dopaminergic therapy [29]. Our study showed that men with PD complained more frequently about dribbling, sadness/blues, loss of interest, anxiety, acting during dreams, and taste/smelling difficulties compared to healthy control men, while female PD patients reported more frequently loss of interest and anxiety compared to healthy control women [29]. In contrast with previous data on treated PD patients [23, 24, 30], female PD patients did not present higher prevalence of mood symptoms compared to male PD patients. Comparison with healthy controls showed that several NMS classically present in the promotor phase and pointing to subjects with subsequent development of PD in large population studies (i.e., sadness/blues, acting out during dreams, taste/smelling difficulties) [31–33] are more frequent in male than in female patients [29]. Further supporting the importance of these findings, Liu et al. described a combination of NMS that can best differentiate PD from controls [34]. Remarkably, in both men and women, poor olfaction was the most powerful NMS predicting PD diagnosis, followed by the Montreal Cognitive Assessment battery score, but, once again, gender made a difference, since dysautonomia was a predictor of PD diagnosis only in men, while REM sleep behavior disorder only in women [34]. The large sample size and the use of multiple detailed NMS assessment tools further corroborate the importance of these findings [34] (Table 1).

Yet, the role of gender in the response of NMS to dopamine replacement therapy was not established. Subsequently, we conducted a 2-year prospective assessment of gender-related differences in the burden of NMS before and after starting dopaminergic therapy showing that sadness/blues presented a significant percentage reduction compared to baseline in both sexes, while urgency, daytime sleepiness, weight gain and increase in sex drive presented a significant percentage increase only in men possibly in relation to both disease progression and dopaminergic treatment [35]. Confirming previous findings [27, 36], male gender was a risk factor for developing both dribbling (odds ratio = 10.29) and nocturia (odds ratio = 9.90), irrespective of therapy and clinical features [35].

However, as the disease progresses, NMS appear in the form of non-motor fluctuations more frequently in women than in men. By administering the 19-item Wearing off Questionnaire to 47 PD patients (M:F = 31:16) after 4 years since the start of dopaminergic treatment, we showed that mood-related non-motor fluctuations (i.e., anxiety, mood changes and pain) were more prevalent in women [37]. These findings possibly account for the higher prevalence of mood-related NMS reported by women in studies including PD patients on dopaminergic treatment and with different stages of disease [26, 27]. Strikingly, in our study no gender differences were detected in either dopaminergic or antidepressants/benzodiazepines intake, despite the higher frequency of non-motor fluctuations evidenced in women, suggesting that non-motor fluctuations in women remain mostly underestimated and undertreated [37].

Regarding cognition, several studies suggest that, as opposite to the female prevalence of dementia (e.g., Alzheimer’s disease) in the general population [4], male gender is a robust risk factor for development of cognitive impairment and dementia in PD [38–40]. Interestingly, recent data suggest that dementia prevalence in women with PD began to increase steadily after the age of 65 years, reaching male estimates only after 80 years of age [41]. Thus, mirroring the course of motor symptoms, PD NMS and cognitive disturbances start with a more benign phenotype in women compared to men, but then present a steadily progressive worsening as disease progresses. Indeed, NMS develop differently in women and men; taste and smell difficulties are reported mainly in men and anxiety in women, respectively, suggesting that the prodromal stage of PD proceeds differently in both sexes [42]. In turn, NMS may be useful to differentiate patients at PD risk if gender is included as an important variable. However, it has to be considered that pre-existing sex differences such as in olfaction might be further exacerbated by the onset of PD.

Biomarkers

Despite few data suggested gender differences for other biomarkers in PD [43–46], the most robust evidence is available for urate.

Previous prospective and case–control studies showed that lower urate concentrations predicted PD prognosis and were inversely associated with disease severity in men but not in women [47–50]. In a postmortem study, urate levels in cortical and striatal tissue were lower in PD than in controls in men only [51]. Intriguingly, more recent data further expand the relationship between urate and gender in PD. With a nested case–control study based on 90.214 participants of three ongoing US cohorts, Gao et al. obtained data for 388 new PD cases (52% men) and 1.267 matched healthy controls (35% men) [52]. Logistic regression analysis showed that men, but not women, with higher urate concentrations had a lower future risk of developing PD, suggesting that urate can be protective against PD risk or could slow disease progression during the preclinical stage of the disease in men only [52]. In addition, by performing a meta-analysis on urate and PD risk in men and women separately, the authors pooled their data with additional 325 incident PD cases and further confirmed this gender difference [52]. The pathophysiological explanations underlying such gender specificity of urate in determining PD risk remain speculative. Other factors might offset the potential neuroprotective effects of urate in women, or estrogens may predominate in determining the lower risk of PD among women [52]. On a practical ground, these data, combined with the evidence on NMS, further support the need for gender-based strategies involving clinical and serum biomarkers to identify prodromal PD cases [34].

Genetics

In this section, the available evidence on gender differences in genes determining PD susceptibility is examined, while the large body of pharmacogenetic data was left out of the scope of this review.

While variable evidence suggests that specific polymorphisms’ expression may be influenced by gender [53-66), a number of studies support a role for LRRK2 status in either reverting or balancing the gender distribution in PD [67–71].

Mutations in the LRRK2 gene are among the most common genetic factors causing PD worldwide and particularly common in selected populations (e.g., Ashkenazi Jews and North African Berbers) [69]. LRRK2 mutations are inherited with an autosomal dominant pattern with incomplete and age-related penetrance. As a matter of fact, asymptomatic LRRK2 carriers represent the ideal setting to study prodromal PD [72]. Several studies suggest that PD LRRK2-associated PD patients are more likely women, as opposite to the gender distribution in glucocerebrosidase (GBA)-associated PD which mirrors the prevalence ratios in the general population [73]. Although a recent meta-analysis rebuts this finding and shows a 1:1 male to female ratio in LRRK2-associated PD [74], the factors associated with the possible rebalancing of the male to female ratio in LRRK2-associated PD compared to idiopathic PD are unknown. Indeed, there is a need for studies evaluating the effect of gender on both genetic and environmental factors determining the PD risk.

Gender differences in Parkinson’s disease management

Pharmacological treatment

Although therapeutic recommendations for PD take into account age, motor disability as well as the presence of disease-related complications (i.e., motor fluctuations and neuropsychiatric complications), to date no gender-oriented advice is available [75, 76]. Yet, gender is one of the pivotal determinants of development of motor and non-motor fluctuations as well as dyskinesia (see above) [17–22, 37]. In addition, no ad hoc prospective studies have been conducted so far and the available evidence on the topic can be inferred from either retrospective studies or the subanalysis of prospective data collected for different objectives.

As for the type of dopaminergic medication, evidence shows similar treatments assigned to men and women with PD, with no gender preference [77, 78]. As such, the NINDS NET-PD study, including data from 1.741 PD patients, reports similar gender ratios for treatment with levodopa alone, dopamine agonist alone or levodopa plus dopamine agonist [77]. Though, as for medication dosage, several studies demonstrate that men with PD are medicated with higher doses of either oral or infusional treatments, as evaluated with the levodopa equivalent daily dose (LEDD) [78–81]. However, when body weight is added as a covariate, the gender differences in LEDD recedes [77], suggesting the core of the matter might be the dosage adjustment according to the body weight [82, 83]. As opposite to dopamine agonists [84], several studies have demonstrated that levodopa pharmacokinetics is significantly affected by the body weight with an inverse correlation between the plasmatic levodopa concentration (i.e., the area under the curve, AUC) and body weight, which is lower in women on average. Arabia et al. observed a lower body weight (65.3 kg versus 73.9 kg, p < 0.001) with greater levodopa AUC in women with PD (6.45 µmol/l h among women versus 4.94 µmol/l h among men, p = 0.002) and reported an inverse correlation between AUC and T1/2 (i.e., half-life) and body weight (respectively, p < 0.001 and p = 0.001) [85]. However, further evidence suggest that women present greater levodopa bioavailability with higher mean AUC (42.3 ± 7 mg versus 23.3 ± 7.3; p < 0.0001) and higher mean C max (1388 ± 42 mg versus 800 ± 33 mg; p < 0.001) after administration of 100 mg of levodopa, irrespective of body weight [84, 86]. In addition, women display lower levodopa clearance levels, further justifying the greater levodopa bioavailability [87]. Recent evidence delineated the features characterizing a subgroup of patients reporting a “brittle response” to levodopa, defined as the presence of highly disabling dyskinesia after small doses (i.e., 100 mg or less per dose) [88]. Those extremely sensitive subjects are mainly women (58%) with lower body weight and body mass index (63.5 versus 79.6 kg, p < 0.001 and 22.3 versus 26.5, p < 0.001, respectively), longer disease duration and much many years on levodopa, but with lower dosage (12.6 versus 8.9 years, p = 0.003 and 9.8 versus 5.9 years, p < 0.001, respectively), compared to patients without a “brittle response” [88]. Although this study suggests new insight into the phenomenology of the response to levodopa, the genetic background of the patients with “brittle response” is overlooked [88].

Indeed, the lower female body weight alone cannot entirely account for the gender discrepancy in development of levodopa-related complication. PD is associated with a profound alteration in central control of energy metabolism determining continuous changes in body weight and composition and energy expenditure in relation to both disease progression and type of treatment [89]. Furthermore, genetic polymorphisms may also have a role in modulating the dyskinesia risk (e.g., DRD2 polymorphism has a protective effect against dyskinesia development only in men [20]). Intriguingly, not all PD patients convert to a “brittle response”, suggesting this subgroup might have peculiar features placing them at risk for maladaptive plastic responses to levodopa [89]. There is a need for prospective ad hoc studies to clarify why women with PD have higher rates of levodopa-related complications and are at risk for presenting a “brittle response” to levodopa.

Surgical treatment

Several randomized clinical trials have shown bilateral subthalamic nucleus deep brain stimulation (DBS) to be effective in PD patients with motor fluctuations [90]. Notwithstanding, this option is underused in certain groups of patients, such as ethnic minorities and low-level socioeconomic status subjects [91, 92]. Strikingly, in spite of the higher risk of developing dyskinesia and motor fluctuations in women, female gender has been repetitively associated with lower utilization of DBS in PD [91–93]. The observation that in the western world the proportion of male patients who receive DBS exceeds the usual male/female predominance of PD might have several explanations as doctors’ attitude and potential gender bias in proposing DBS, stronger fear for surgical risks among women or more initiative in men who autonomously demand for DBS [93, 94]. However, the lack of large ad hoc prospective studies prevents us from drawing conclusions on the reasons for the gender discrepancy in DBS access [92, 93]. As a matter of fact, women with PD perform DBS later than men displaying longer disease duration, more severe disease and much more dyskinesia at the time of surgery [95]. Yet, DBS provides benefit in both genders determining equal clinical improvement and reduction in medications with even greater impact on activities of daily living and quality of life in women [96–100]. Postponing DBS in PD women might have a detrimental impact on life planning. Recent reports demonstrate that, due to its efficacy on psychomotor status and treatment reduction, DBS is a safe option in the management of young PD women who wish to become pregnant [101]. However, there is the need to define strategies to prevent and control any worsening of clinical conditions during pregnancy and to consider device-related options (i.e., rechargeable battery to avoid battery replacement and subclavicular placement instead of abdominal) in women who plan to become pregnant [100]. Finally, DBS is a valid approach to relieve disability in patients with “brittle response” to levodopa (see above), who are mostly women [88].

The role of estrogens

Estrogens are likely contributors to gender differences in PD [102–104]. Although conflicting data are available, evidence would suggest a link between longer estrogen exposure during lifetime and both the decreased PD risk and milder features at onset in women [105–119]. Most women develop PD after menopause, further suggesting estrogen withdrawal has a role in disease pathogenesis [9, 13, 14]. Accordingly, preclinical evidence shows that estrogens are protective against dopaminergic damage. Animal models with estrogens deprivation show dopaminergic neuron loss, altered dopaminergic metabolism and transporter uptake, which can be partially reversed by the administration of exogenous estrogens [102–104].

A large body of evidence shows that estradiol and related compounds exert neuromodulatory and neuroprotective activities in the striatum and substantia nigra through several intracellular mechanisms that ultimately decrease apoptosis of neurons. In addition to these signal cascade effects, estrogens might impact PD pathogenesis via their influence on mitochondrial function and response to oxidative stress. Evidence demonstrates that estrogens might also prevent Lewy body deposition through specific alpha-synuclein anti-aggregation and fibril destabilization properties [102–104].

However, in contrast with the large body of preclinical evidence [104], a spoonful of studies on humans are available on the clinical effects of estrogens in PD. A small pilot study showed that estrogen replacement therapy in non-parkinsonian women increases putaminal dopamine active transporter as measured with TRODAT SPECT scan [120]. Although small trials have demonstrated mild efficacy of low-dose estrogens in improving motor disability and motor fluctuations in post-menopausal women [121–125], estrogens have not been further tested in larger cohort. Although there is an increasing interest of the research community in testing the disease-modifying effect of estrogens in different neurological conditions [126], clinical trials of estrogen face unique challenges possibly explaining the lack of data in larger PD cohort [123]. Estrogen is an endogenous compound with levels that naturally fluctuate throughout the lifecycle. Estrogen’s effects are widespread in and outside the brain and conventional study designs have difficulty assessing complex variables including variability in endogenous/exogenous estrogen exposure, and the interface between hormonal changes and the onset/progression of a chronic disease. Ultimately, as chronic estrogen exposure is associated with increased risk of breast cancer and coronary heart disease, risks may exceed benefits [127].

Conclusions

Here, we gathered evidence demonstrating the existence of gender differences in PD clinical phenotype, biomarkers and therapeutic management. Still, much work needs to be done to better understand the interaction between gender and genetics in determining the PD risk and clinical features. Several data demonstrate that PD in women starts with a more benign phenotype, likely due to the effect of estrogens. However, as the disease progresses, women are at higher risk of developing highly disabling treatment-related complications, such as motor and non-motor fluctuations as well as dyskinesia, compared with men. In addition, women have lower chances of receiving effective treatment for PD as DBS (Fig. 1). Taken together these findings challenge the definition of a more benign phenotype in women.

Fig. 1
figure 1

Synoptic diagram showing gender difference in early (left side) and advanced (right side) PD. As for early PD, women have lower prevalence and incidence, slightly higher age at onset, higher tremor dominance and striatal uptake compared to men, justifying the definition of “more benign phenotype”. As for advanced PD, women have more motor and non-motor fluctuations as well as dyskinesia and reduced access to DBS, thus challenging the definition of “more benign phenotype”. DBS deep brain stimulation, PD Parkinson’s disease

Full size image

Improving our understanding in this field may result in implementation of strategies to identify prodromal PD cases and speed efforts to discern new directions for PD tailored treatment and management. We just got the evidence that gender does matter in PD [128]. It matters in many ways we did not expect. It also matters in ways we have not envisaged yet [1].