You are on page 1of 7

Journal of Parkinson’s Disease xx (20xx) x–xx 1

DOI 10.3233/JPD-181305
IOS Press

Research Report

High Prevalence of Undiagnosed

Insulin Resistance in Non-Diabetic
Subjects with Parkinson’s Disease

Elliot Hogga , Kishore Athreyaa , Christina Basilea , Echo E. Tana , Jan Kaminskib
and Michele Tagliatia,∗
a Department of Neurology, Cedar-Sinai Medical Center, Los Angeles, CA, USA
b Department of Neurosurgery, Cedar-Sinai Medical Center, Los Angeles, CA, USA

Accepted 20 February 2018

Background: Reduced glucose tolerance has been long recognized as a potential risk factor for Parkinson’s disease (PD), and
increasing scrutiny is currently being placed on insulin resistance (IR) as a pathologic driver of neurodegeneration. However,
the prevalence of IR in PD is unknown.
Objective: To determine IR prevalence in non-diabetic patients with PD and to correlate IR with other metabolic indicators,
motor and non-motor symptoms (NMS) of PD, and quality of life (QoL).
Methods: Non-diabetic patients with a diagnosis of PD were identified and tested for fasting insulin, fasting glucose,

and HbA1c. Patients were also offered to take a battery of clinical tests (MoCA, NMSQ, and PDQ-39) and had their PD
medications, height, weight, and other demographic features recorded. IR was defined as HOMA-IR≥2.0 and/or HbA1c≥5.7.

IR abnormalities were correlated with BMI and demographic features, in addition to motor and NMS.
Results: 154 subjects (109 M, 45F, mean age 67.7 ± 10.5) were included in this study. Mean HOMA-IR was 2.3 ± 1.8. Ninety
out of 154 (58.4%) subjects had abnormal IR. IR was more frequent in overweight and obese subjects (61.1% and 82.8%
respectively) than normal weight subjects (41.5%). Multivariate analysis showed that BMI was the only significant predictor

of IR (p < 0.0001). There was no significant correlation between HOMA-IR and MoCA, PDQ-39, and NMSQ scores.
Conclusions: IR is prevalent in PD and it correlates with BMI. A correlation between IR with cognitive and QoL measures
cannot be determined on the basis of this sample.

Keywords: Body mass index, cognition, insulin resistance, Parkinson’s disease


INTRODUCTION indicated a high rate of abnormal glucose tolerance

in PD patients, ranging between 50 to 80% depend-
A relationship between Parkinson’s disease (PD) ing on the source [2–4]. DM2 is associated with more
and diabetes mellitus type II (DM2) has been severe motor and non-motor features of PD [5] and
described for several decades, predating the introduc- appears to increase the risk of developing PD [6, 7],
tion of levodopa therapy [1]. Multiple studies have although evidence is controversial [8].
∗ Correspondence
The key link between DM2 and PD appears to
to: Michele Tagliati, MD, Department of
Neurology, Cedar-Sinai Medical Center, 127 S. San Vicente Ave.,
be insulin resistance (IR), a condition defined by a
A6600, Los Angeles, CA 90048, USA. Tel.: +1 310 248 6704; reduced biological effect for any given concentration
Fax: +1 310 967 0601; E-mail: of insulin, which is common in aging [9]. Peripheral
ISSN 1877-7171/18/$35.00 © 2018 – IOS Press and the authors. All rights reserved
2 E. Hogg et al. / Insulin Resistance is Prevalent in PD

IR can be measured using the glucose clamp tech- was performed in a subgroup of 71 subjects for
nique or by a less cumbersome method provided by later analysis. This included Montreal Cognitive
the Homeostatic Model Assessment (HOMA) equa- Assessment (MOCA), Non-Motor Symptoms Scale
tion [10]. Peripheral IR is a reversible condition that (NMSQ), and Parkinson’s Disease Questionnaire –
not only predisposes to DM2 but is associated with 39 (PDQ-39) to assess health related quality of life.
central IR [11]. Importantly, central IR can be a either Dopaminergic medications were recorded including
a cause [12] or a consequence [13] of neurodegener- type and dose, to enable levodopa-equivalent dose
ation. (LED) calculation [18].
While peripheral IR is a reversible condition and Impaired fasting glucose (IFG), the intermediate
may be a target for pharmacologic and lifestyle man- state of abnormal blood sugar between normal and
agement therapies, the prevalence of peripheral IR in diabetes was defined by FPG concentration ≥100
PD is currently unknown, hampering interventions in and <126 mg/dl [19]. A cutoff HOMA index of 2.0,

this area. Complicating matters is that there is little equivalent to <50% sensitivity, was used to define IR.

consensus as to what constitutes normal and abnor- Subjects were considered to have IR if they either had
mal insulin sensitivity with significant variation in a HOMA≥2.0 and/or HbA1c≥5.7. We classified par-
reported measures of peripheral IR in different pop- ticipants as overweight if their BMI was 25 or greater,
ulations [14, 15]. We conducted a study to determine and obese if BMI was 30 or greater [20]. MoCA≤22
the prevalence of peripheral IR in a population of non- was considered consistent with dementia.

diabetic PD patients and its correlation to quality of Pearson R coefficient was used to determine the
life, cognitive decline and other non-motor features. strength of observed correlations between HOMA-
IR index, demographic (age, BMI) and clinical
data including LED, MoCA, NMSQ, PDQ-39, and
METHODS HbA1c. Student t-test analysis was used for group
comparisons. Multivariate linear regression analysis
One hundred and sixty idiopathic PD people was performed to identify variables that indepen-
attending the movement disorders clinic at Cedars- dently contributed to IR. A probability value of
Sinai Medical Center in Los Angeles were offered p < 0.05 was considered statistically significant.
testing for fasting plasma insulin (FPI), fasting
plasma glucose (FPG) and glycated hemoglobin

(HbA1c). HOMA index was calculated by the for- RESULTS

mula: HOMA-IR = (FPI x FPG)/405 [10, 16]. The

study obtained local Institutional Review Board One hundred and sixty people were initially
(IRB) approval. recruited. Six subjects were excluded from further
To meet inclusion criteria, individuals had to be age analysis as they were found to have fasting plasma
18 and above with a clinical diagnosis of PD accord- glucose (FPG) concentration (≥126 mg/dl) or HbA1c

ing to the UK Brain Bank criteria. Exclusion criteria levels (>6.5 mmol/mol) in the diabetic range. The
included a diagnosis of DM2, as well as an inability remaining 154 subjects (109M, 45F) were included
to undergo blood draws. Results were obtained and in the final analysis. Mean age was 67.7 ± 10.5, with
recorded with the date of the sample noted. Demo- a mean PD duration of 6.5 ± 5.5 years at time of

graphic information collected included age, gender, study, and a mean H&Y of 2.2 ± 0.7. Mean BMI was
ethnicity, PD subtype (i.e. tremor-predominant, 26.8 ± 4.8 and mean LED was 491.5 ± 537.2. 17/154
akinetic-rigid, mixed), age at PD symptom onset, subjects (11.0%) had undergone deep brain stimula-
duration of PD symptoms, height and weight. Motor tion (DBS). For additional demographic data please
severity and progression were assessed using the see Table 1.
Movement Disorder Society Unified Parkinson’s Mean fasting glucose levels for this population was
Disease Rating Scale (MDS-UPDRS) and Hoehn 95.8 ± 10.2. Fifty-two subjects (33.7%) had IFG, of
and Yahr Scale (H&Y) collected as part of routine which 32 (61.5%) also had abnormal HOMA-IR.
clinic visits. Body mass index (BMI) was calculated Mean HOMA-IR was 2.3 ± 1.8 and mean HbA1c
according to the formula weight (Kg)/height2 (m) was 5.5 ± 0.4 (n = 118). 70 subjects (45.5%) had an
and subjects were subdivided in obese (BMI>30), abnormal HOMA index, consistent with IR. In addi-
non-obese overweight (BMI>25) and lean (BMI<25) tion, 39 (33.1%) subjects had HbA1c data in the
[17]. Additional testing of non-motor symptoms pre-diabetic range (≥5.7). Of these 39, 20 had a
E. Hogg et al. / Insulin Resistance is Prevalent in PD 3

normal HOMA-IR, bringing the total subjects with values, with 65.8% overall overweight subjects and
undiagnosed IR to 90, or 58.4% of the study popula- 19% in obesity range. HOMA or HbA1c values con-
tion. BMI distribution was skewed toward overweight sistent with IR were found in 24/29 (82.8%) obese,
44/72 (61.1%) overweight, and 22/53 (41.5%) normal
Table 1 weight subjects (Fig. 1).
Demographic data
In the full study cohort, HOMA-IR positively cor-
Mean ± STD related with BMI (Pearson R = 0.42, p < 0.0001) and
Age 67.7 ± 10.5 years HbA1c (Pearson R = 0.19, p < 0.05), while it showed
PD Duration 6.5 ± 5.5 years
LED 491.5 ± 537.2 mg
a negative correlation with age (Pearson R = –0.19,
LED/Weight 7.9 ± 6.8 mg/kg p = 0.01). There was no significant difference in
BMI 26.8 ± 4.8 kg/m2 HOMA-IR between male and female subjects. Obese
FPG 95.8 ± 10.2 mmol/L subjects had significantly higher mean HOMA-

FPI 9.4 ± 6.7 ␮U/ml
2.2 ± 0.7
IR (3.3 ± 1.7 vs. 2.03 ± 1.7, p < 0.0005) and were

MDS-UPDRS ON (n = 112) 17.3 ± 9.5 significantly younger (63.6 ± 11.5 years old vs.
MDS-UPDRS OFF (n = 49) 20.4 ± 13.1 68.7 ± 10.1, p < 0.05) than non-obese subjects. Sim-
HOMA-IR 2.3 ± 1.8 ilarly, non-obese overweight subjects (BMI >25
Male (n = 109) 2.4 ± 1.9
Female (n = 45) 2.0 ± 1.2 and <30) had significantly higher mean HOMA-IR
5.5 ± 0.4% (2.5 ± 2.0 vs. 1.4 ± 0.6, p < 0.001) and were signifi-

HbA1c (n = 118)
MoCA (n = 71) 26.4 ± 3.0 cantly younger (67.3 ± 11.1 years old vs. 71.4 ± 7.9,
NMSQ (n = 71) 11.3 ± 5.1 p < 0.05) than lean subjects. Mean weight-adjusted
PDQ39 (n = 71) 157.2 ± 103.5
LED was 7.9 ± 6.8 mg/kg. HOMA-IR did not signif-
LED: Levodopa equivalent dose, LED/Weight: body weight (kg)
adjusted LED, BMI: Body Mass Index, HOMA-IR: Homeostatic
icantly correlate with weight-adjusted LED, as well
model assessment, FPG: fasting plasma glucose, FPI: fasting as with PD duration, LED, MDS-UPDRS and HY
plasma insulin, H&Y: Hoehn & Yahr Disease severity scale, MDS- scores. Multivariate analysis showed that BMI was
UPDRS: Movement Disorder Society Unified PD Disease Rating the only resulting independent variable correlated
Scale in ON and OFF medication states (11 subjects had both
with IR (p < 0.0001).
ON and OFF scores available from the same date for analysis),
HbA1c: glycated hemoglobin, MoCA: Montreal Cognitive Assess-
Mean H&Y scores were virtually identical
ment, NMSQ: Non-Motor Symptoms Scale, PDQ-39 Parkinson’s between IR and non-IR subjects (Table 2) and there
Disease Questionnaire–39. was no significant trend in HOMA-IR by disease

Fig. 1. Insulin Resistance distribution by body-type. Stacked bar chart describing the relationship between insulin resistance (IR) and body
type. Percentage of subjects who met criteria for IR (see text for definition) is represented on the vertical axis, and it is grouped by body-type,
as determined by body mass index (BMI, kg/m2 ). Normal weight was defined as BMI less than 25, non-obese overweight as BMI greater
than 25 but less than 30, and obese as BMI greater than 30.
4 E. Hogg et al. / Insulin Resistance is Prevalent in PD

Table 2 Table 3
Motor, cognitive, other non-motor symptoms, and health related Non-motor symptoms and health related quality of life sub-
quality of life scores in PD subjects with and without Insulin categorical scores in 71 PD subjects with and without insulin
Resistance resistance
IR+ (n = 95) IR- (n = 66) IR- IR+
MDS-UPDRS III (ON) 17 ± 7.8 18 ± 12.0 NS NMSQ
MDS-UPDRS III (OFF) 20 ± 12.7 21 ± 13.7 NS Gastrointestinal 3.2 ± 2.0 2.6 ± 2.0 NS
H&Y 2.2 ± 0.6 2.3 ± 0.8 NS Urinary 1.4 ± 0.8 1.4 ± 0.7 NS
IR+ (n = 27) IR- (n = 44) Pain 0.3 ± 0.5 0.2 ± 0.4 NS
Cognitive 1.8 ± 0.9 1.3 ± 1.2 NS
MoCA 26.3 ± 3.3 26.6 ± 2.7 NS
Psych 0.3 ± 0.5 0.1 ± 0.4 NS
NMSQ 10.6 ± 4.9 12.6 ± 5.3 NS
Mood 0.8 ± 0.8 1.0 ± 0.8 NS
PDQ-39 156.7 ± 102.6 158.0 ± 106.9 NS
Sexual 1.0 ± 0.8 0.9 ± 0.8 NS
MDS-UPDRS III: Movement Disorder Society Unified Parkin- Cardiovascular 0.7 ± 0.8 0.7 ± 0.7 NS

son’s Disease Rating Scale part III, H&Y: Hoehn and Yahr scale. Sleep/Fatigue 2.5 ± 1.4 2.1 ± 1.3 NS
0.5 ± 0.6 0.5 ± 0.7

MoCA: Montreal Cognitive Assessment, NMSQ: Non-Motor Miscellaneous NS
Symptoms Scale, PDQ-39 Parkinson’s Disease Questionnaire – PDQ-39
39, IR: Insulin Resistant based on HOMA≥2.0 and/or HbA1c≥5.7, Mobility 21.2 ± 24.4 22.9 ± 20.9 NS
NS: non-significant. ADLs 23.9 ± 19.6 20.7 ± 18.6 NS
Emotional 21.6 ± 17.6 22.5 ± 16.6 NS
Stigma 17.8 ± 14.0 17.7 ± 18.5 NS

severity between stages of the H&Y (data not shown). Social support 4.8 ± 10.0 6.3 ± 14.6 NS
Similarly, mean MDS-UPDRS measured either in Cognitive 23.6 ± 18.7 19.2 ± 19.2 NS
Communication 20.2 ± 18.1 22.2 ± 25.7 NS
ON or OFF state did not differ between IR and Bodily discomfort 25.0 ± 21.9 25.2 ± 21.6 NS
non-IR subjects (Table 2). The prevalence of chronic NMSQ: Non-Motor Symptoms Questionnaire; PDQ-39: Parkin-
renal disease was quite low overall and did not dif- son’s Disease Questionnaire – 39, IR: Insulin Resistant based on
fer between IR (4.4%) and non-IR subjects (3.1%). HOMA≥2.0 and/or HbA1c≥5.7, NS: non-significant.
Chronic cardiac disease, including coronary artery
disease, atrial fibrillation, chronic heart failure and
The DBS subgroup (n = 17) had a significantly
cardiomyopathy, was also relatively infrequent both
higher mean PD duration (11.9 ± 5.3 years versus
in IR (18.9%) and non-IR subjects (10.9%), while
5.8 ± 5.1 in non-DBS subjects, p < 0.0005) and a
the prevalence of hypertension was higher, affecting
significantly higher NMSQ than non-DBS subjects
35.6% of subjects with IR and 25% of subjects with-

(16 ± 4.5 vs. 10.3 ± 4.7, p < 0.005). Weight distribu-

out IR. All these differences failed to reach statistical
tion in the DBS subgroup was similar to the general

study population, with 5/17(29.4%) DBS subjects

In the subgroup undergoing more detailed clinical
in the normal weight range, 7/17(41.2%) non-obese
testing (n = 71), mean MoCA score was 26.4 ± 3.0,
overweight, and 5/17(29.4%) obese. Age, BMI,
with 23/71 (32.4%) abnormal values (25 or less)
HOMA-IR, Hb1ac, MoCA, and PDQ-39 were not
and 5/71 (7.0%) values consistent with demen-

significantly different between DBS and non-DBS

tia. HOMA-IR, BMI, PD duration, LED, Hb1ac,
PDQ-39, NMSQ, and age were not significantly
different between demented and non-demented sub-

jects. MoCA showed a significant inverse correlation

with age (Pearson R = –0.38, p < 0.001) but did not
correlate with PD duration, or HoMA-IR. Mean Despite the growing interest in the use of incretin
NMSQ score was 11.3 ± 5.1. Mean study PDQ-39 mimetics in the treatment of PD [21, 22], there is lit-
was 157.2 ± 103.5. There was no significant differ- tle information regarding the prevalence of IR in PD,
ence between IR and non-IR subjects in sub-section with available data limited to a small study show-
scores of either PDQ-39 or NMSQ (Table 3). Like- ing significantly elevated HOMA index scores in PD
wise, there was no significant correlation between patients as compared to healthy controls [23]. Our
BMI subgroups and HbA1c, PD duration, LED, study indicates that almost two thirds of non-diabetic
MoCA, PDQ-39, and NMSQ. Multivariate analy- people with PD may be insulin resistant, despite
sis including HOMA IR and age, disease duration, normal fasting glucose and, in many cases, normal
MoCA scores, HY stage and weight-adjusted LED in HbA1c. These figures suggest that IR in PD is a preva-
this subgroup failed to reveal significant predictors of lent and largely undetected problem, especially in
cognitive decline. those patients who are overweight or obese, a feature
E. Hogg et al. / Insulin Resistance is Prevalent in PD 5

reported in the early stages of PD [24]. While our when adjusted for weight, in line with prior research
data confirms previous studies showing that HOMA- showing that levodopa does not affect glucose
IR is more than double in obese than in lean subjects tolerance in PD [37].
[25], we also found a substantially higher percent- Despite some controversy [38], HOMA-IR
age (41%) of lean PD patients with IR than reported appears to be a reliable indicator of IR [39] and a more
in healthy adults [26]. This finding may be related sensitive and responsive measurement to changes in
to the abnormal fat distribution in subjects with PD, body metabolism than other traditional measures,
which show, on average, a higher visceral to subcuta- such as fasting glucose [40]. An optimal cutoff for
neous fat ratio [27], a feature that correlates with IR HOMA-IR has yet to be determined and may vary
[28]. We were unable to characterize fat distribution regionally [14, 15, 41] and by age [42]. We used
in this study, and therefore the role of abnormal fat a cut-off HOMA of 2.0 given that this represents
distribution in the high prevalence of IR in PD will 50% beta cell function, but a lower set point may

need further evaluation. Nevertheless, we propose increase sensitivity. IR is an important predictor of

that HOMA-IR may be a more relevant marker of metabolically linked conditions such as DM2 [43]
metabolic health than BMI in patients with PD, as and cardiovascular disease [44]. The risk of IR is
BMI does not capture weight distribution. Interest- known to increase with obesity, aging, physical inac-
ingly, the subgroup of PD patients who underwent tivity, and genetic predisposition [45], risk factors
DBS, a treatment frequently associated with weight also associated with the diagnosis of PD [46]. Inter-

gain [29], had metabolic measures largely similar to estingly, IR did not positively correlate with age or
the rest of our PD cohort. disease duration in our study population, suggest-
While other dementing neurodegenerative dis- ing that IR may be an independent process in PD.
eases, including Alzheimer’s disease [30], dementia While insulin resistance in aging has been associ-
with subcortical features [31], and Huntington’s dis- ated to abdominal obesity in the general population
ease [32] have been associated with IR, we found [47], other authors have suggested that changes in adi-
no correlation between HOMA-IR and cognitive posity may not fully explain age related increases in
decline. Mean MoCA scores were virtually iden- insulin resistance [48]. While we did find a strong
tical in subjects with and without IR. This is in positive correlation between obesity and IR, age
contrast to prior reports describing elevated HOMA- negatively correlated with obesity in our popula-
IR in subjects with PD-Dementia [33], but in line tion. PD patients tend to initially gain weight [24]

with other investigations showing an inverse corre- and then lose weight over time, although they may
lation between IR and incidence of dementia [34]. actually increase caloric intake as weight decreases,

As expected, advancing age predicted lower MoCA suggesting a complex metabolic picture [49]. Longi-
scores, however PD duration or severity did not. The tudinal data are therefore needed to understand how
unanticipated demographic composition of our study weight changes in PD may impact IR status over
population affected the relationship between IR and time [50].

MoCA, as BMI – the most powerful predictor of The generalizability of these results may be lim-
IR – inversely correlated with age – a predictor of ited, as this study was performed at a single center,
cognitive decline. Furthermore, the low incidence with a largely affluent population who generally has
of dementia in our study cohort and the use of a access to good medical insurance coverage. Two-

screening test, as opposed to a formal neurocogni- thirds of our study population was overweight, a bias
tive battery, may have limited the sensitivity of our that might have resulted in higher incidence of IR.
sample. In the end, the lack of longitudinal sampling In addition, both insulin sensitivity and HOMA may
prevents any firm conclusion in regard to a possible vary by ethnicity [51], with our cohort including a
link between IR and cognitive decline. caucasian majority. We cannot rule out self-selection
IR did not appear to affect the prevalence of any bias of patient at greater risk for IR due to diet, weight,
non-motor symptom, as mean NMSQ scores in our or other lifestyle factors. Additionally, given the high
study were comparable to NMSQ data in similar pop- prevalence of pre-diabetes or undiagnosed diabetes
ulations of PD patients [35]. Likewise, mean PDQ-39 in the general population of California [52], the rel-
scores in our cohort were consistent with other studies evance of our findings has to be assessed against a
[36]. IR was not associated with changes in health- backdrop of pervasive metabolic disease present in
related QoL sub-categories (Table 3). Finally, no our region. We did not find a particularly high preva-
correlation was found between IR and LED, even lence of diseases characterized by association with
6 E. Hogg et al. / Insulin Resistance is Prevalent in PD

IR in non-PD adults, such as chronic cardiovascular [9] Rowe JW, Minaker KL, Pallotta JA, Flier JS (1983) Charac-
disease and chronic renal failure [53], with the terization of the insulin resistance of aging. J Clin Investig
71, 1581.
possible exception of hypertension. However, the [10] Matthews DR, Hosker JP, Rudenski AS, Naylor BA,
distribution of hypertension was not significantly Treacher DF, Turner RC (1985) Homeostasis model assess-
different between IR and non-IR subjects. Finally, ment: Insulin resistance and ␤-cell function from fasting
although this study was not intended to address plasma glucose and insulin concentrations in man. Dia-
betologia 28, 412-419.
metabolic syndrome in PD, it is worth noting that [11] Clegg DJ, Gotoh K, Kemp C, Wortman MD, Benoit SC,
we did not collect lipid and cholesterol data for this Brown LM, D’Alessio D, Tso P, Seeley RJ, Woods SC
population. (2011) Consumption of a high-fat diet induces central
We conclude that peripheral IR is prevalent in PD insulin resistance independent of adiposity. Physiol Behav
103, 10-16.
and may be an under-diagnosed problem in a large [12] Kleinridders A, Cai W, Cappellucci L, Ghazarian A, Collins
proportion of PD patients, especially in those who

WR, Vienberg SG, Pothos EN, Kahn CR (2015) Insulin
are overweight or obese. Given the early occurrence resistance in brain alters dopamine turnover and causes

behavioral disorders. Proc Natl Acad Sci U S A 112, 3463-
of weight gain in PD and the putative role of IR in
accelerating the progression of motor and non-motor [13] Gao S, Duan C, Gao G, Wang X, Yang H (2015) Alpha-
features of PD, we believe that these data will deserve synuclein overexpression negatively regulates insulin
further attention and research. receptor substrate 1 by activating mTORC1/S6K1 signaling.
Int J Biochem Cell Biol 64, 25-33.

[14] Acosta AM, Escalona M, Maiz A, Pollak F, Leighton F
ACKNOWLEDGMENTS (2002) Determination of the insulin resistance index by
the Homeostasis Model Assessment in a population of
Metropolitan Region in Chile. Rev Med Chil 130, 1227-
Financial support for this research was solely 1231.
through internal departmental funding, including the [15] Esteghamati A, Ashraf H, Khalilzadeh O, Zandieh A,
Caron and Steven D. Broidy Chair in Movement Nakhjavani M, Rashidi A, Haghazali M, Asgari F (2010)
Disorders. Optimal cut-off of homeostasis model assessment of insulin
resistance (HOMA-IR) for the diagnosis of metabolic
syndrome: Third national surveillance of risk factors of non-
CONFLICT OF INTEREST communicable diseases in Iran (SuRFNCD-2007). Nutr
Metab 7, 26.
[16] Reaven GM (2009) HOMA-beta in the UKPDS and
The authors have no conflict of interest to report. ADOPT. Is the natural history of type 2 diabetes char-

acterised by a progressive and inexorable loss of insulin

secretory function? Maybe? Maybe not? Diabetes Vasc Dis
REFERENCES Res 6,133-138.

[17] WHO (1995) Physical status: The use and interpretation

[1] Schwab RS (1960) Progression and prognosis in Parkin- of anthropometry. Report of a WHO Expert Consultation.
son’s disease. J Nerv Ment Dis 130, 556-566. WHO Technical Report Series Number 854. World Health
[2] Barbeau A, Giguere R, Hardy J (1961) Experience clinique Organization, Geneva.
avec le tolbutamide dans la maladie de Parkinson. Union [18] Tomlinson CL, Stowe R, Patel S, Rick C, Gray R, Clarke

Med Can 90, 147-151. CE (2010) Systematic review of levodopa dose equivalency
[3] Woert MH, Mueller PS (1971) Glucose, insulin, and free reporting in Parkinson’s disease. Mov Disord 25, 2649-
fatty acid metabolism in Parkinson’s disease treated with 2653.
levodopa. Clin Pharmacol Ther 12, 360-367. [19] Nathan DM, Davidson MB, DeFronzo RA, Heine RJ, Henry
[4] Sandyk R, Awerbuch GI (1992) The association of diabetes RR, Pratley R, Zinman B (2007) Impaired fasting glucose

mellitus with dementia in Parkinson’s disease. Int J Neurosci and impaired glucose tolerance. Diabetes Care 30, 753-759.
64, 209-212. [20] Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL (1998)
[5] Cereda E, Barichella M, Cassani E, Caccialanza R, Pezzoli Overweight and obesity in the United States: Prevalence
G (2012) Clinical features of Parkinson disease when onset and trends, 1960-1994. Int J Obes Relat Metab Disord 22,
of diabetes came first A case-control study. Neurology 78, 39-47.
1507-1511. [21] Athauda D, Foltynie T (2016) The glucagon-like peptide
[6] Hu G, Jousilahti P, Bidel S, Antikainen R, Tuomilehto J 1 (GLP) receptor as a therapeutic target in Parkinson’s
(2007) Type 2 diabetes and the risk of Parkinson’s disease. disease: Mechanisms of action. Drug Discov Today 21, 802-
Diabetes Care 30, 842-847. 818.
[7] Yue X, Li H, Yan H, Zhang P, Chang L, Li T (2016) [22] Athauda D, Maclagan K, Skene SS, Bajwa-Joseph M,
Risk of Parkinson disease in diabetes mellitus: An updated Letchford D, Chowdhury K, Hibbert S, Budnik N, Zampedri
meta-analysis of population-based cohort studies. Medicine L, Dickson J, Li Y (2017) Exenatide once weekly versus
(Baltimore) 95, e3549. placebo in Parkinson’s disease: A randomised, double-
[8] Savica R, Grossardt BR, Ahlskog JE, Rocca WA (2012) blind, placebo-controlled trial. Lancet 390, 1664-1675.
Metabolic markers or conditions preceding Parkinson’s dis- [23] Morris JK, Vidoni ED, Perea RD, Rada R, Johnson DK,
ease: A case-control study. Mov Disord 27, 974-979. Lyons K, Pahwa R, Burns JM, Honea RA (2014) Insulin
E. Hogg et al. / Insulin Resistance is Prevalent in PD 7

resistance and gray matter volume in neurodegenerative [39] Katsuki A, Sumida Y, Gabazza EC, Murashima S, Furuta
disease. Neuroscience 270, 139-147. M, Araki-Sasaki R, Hori Y, Yano Y, Adachi Y (2001)
[24] Vikdahl M, Carlsson M, Linder J, Forsgren L, Håglin L Homeostasis model assessment is a reliable indicator of
(2014) Weight gain and increased central obesity in the early insulin resistance during follow-up of patients with type 2
phase of Parkinson’s disease. Clin Nutr 33, 1132-1139. diabetes. Diabetes Care 24, 362-365.
[25] Silha JV, Krsek M, Skrha JV, Sucharda P, Nyomba BL, [40] Vogeser M, König D, Frey I, Predel HG, Parhofer KG,
Murphy LJ (2003) Plasma resistin, adiponectin and leptin Berg A (2007). Fasting serum insulin and the homeostasis
levels in lean and obese subjects: Correlations with insulin model of insulin resistance (HOMA-IR) in the monitoring
resistance. Eur J Endocrinol 149, 331-335. of lifestyle interventions in obese persons. Clin Biochem 40,
[26] Ghose T (2015) Insulin Resistance: Risk Factor for 964-968.
Heart Disease and Diabetes, LiveScience Health, https:// [41] Haffner SM, Miettinen H, Stern MP (1997) The homeostasis model in the San Antonio heart study. Diabetes Care 20,
diabetes-heart-disease.html, Last updated June 26, 2015, 1087-1092.
Accessed on January 1, 2018. [42] Kurtoğlu S, Hatipoğlu N, Mazicioğlu M, Kendirici M,

[27] Bernhardt D, Müller HP, Ludolph AC, Dupuis L, Kas- Keskin M, Kondolot M (2010) Insulin resistance in obese
subek J (2000) Body fat distribution in Parkinson’s disease: children and adolescents: HOMA-IR cut-off levels in

An MRI-based body fat quantification study. Parkinsonism the prepubertal and pubertal periods. J Clin Res Pediatr
Relat Disord 33, 84-89. Endocrinol 2, 100.
[28] Frayn KN (2000) Visceral fat and insulin resistance [43] Hong Y, Després JP, Rice T, Nadeau A, Province MA,
—causative or correlative? Br J Nutr 83, S71-77. Gagnon J, Leon AS, Skinner JS, Wilmore JH, Bouchard
[29] Barichella M, Marczewska AM, Mariani C, Landi A, Vairo C, Rao DC (2000) Evidence of pleiotropic loci for fasting
A, Pezzoli G (2003) Body weight gain rate in patients with insulin, total fat mass, and abdominal visceral fat in a seden-

Parkinson’s disease and deep brain stimulation. Mov Disord tary population: The HERITAGE family study. Obesity 8,
18, 1337-1340. 151-159.
[30] Arnold SE, Arvanitakis Z, Macauley-Rambach SL, Koenig [44] Henkin L, Bergman RN, Bowden DW, Ellsworth DL,
AM, Wang HY, Ahima RS, Craft S, Gandy S, Buettner Haffner SM, Langefeld CD, Mitchell BD, Norris JM, Rew-
C, Stoeckel LE, Holtzman DM, Nathan DM (2018) Brain ers M, Saad MF, Stamm E (2003) Genetic epidemiology of
insulin resistance in type 2 diabetes and Alzheimer dis- insulin resistance and visceral adiposity: The IRAS Family
ease: Concepts and conundrums. Nat Rev Neurol. doi: Study design and methods. Ann Epidemiol 13, 211-217.
10.1038/nrneurol.2017.185 [45] Rader DJ (2007) Effect of insulin resistance, dyslipidemia,
[31] Geroldi C, Frisoni GB, Paolisso G, Bandinelli S, Lam- and intra-abdominal adiposity on the development of car-
poni M, Abbatecola AM, Zanetti O, Guralnik JM, Ferrucci diovascular disease and diabetes mellitus. Am J Med 120,
L (2005) Insulin resistance in cognitive impairment. The S12-18.
InCHIANTI Study. Arch Neurol 62, 1067-1072. [46] Abbott RD, Ross GW, White LR, Sanderson WT, Burchfiel
[32] Lalić NM, Marić J, Svetel M, Jotić A, Stefanova E, Lalić K, CM, Kashon M, Sharp DS, Masaki KH, Curb JD, Petrovitch
Dragasević N, Milicić T, Lukić L, Kostić VS (2008) Glucose H (2003) Environmental, life-style, and physical precursors

homeostasis in Huntington disease. Abnormalities in insulin of clinical Parkinson’s disease: Recent findings from the
sensitivity and early-phase insulin secretion. Arch Neurol Honolulu-Asia Aging Study. J Neurol 250, 30-39.
65, 476-480. [47] Kohrt WM, Kirwan JP, Staten MA, Bourey RE, King DS,

[33] Bosco D, Plastino M, Cristiano D, Colica C, Ermio C, De Holloszy JO (1993) Insulin resistance in aging is related to
Bartolo M, Mungari P, Fonte G, Consoli D, Consoli A, Fava abdominal obesity. Diabetes 42, 273-281.
A (2012) Dementia is associated with insulin resistance in [48] Fink RI, Kolterman OG, Griffin J, Olefsky JM (1983) Mech-
patients with Parkinson’s disease. J Neurol Sci 315, 39-43. anisms of insulin resistance in aging. J Clin Investig 71,
[34] Schelp AO, Mendes-Chiloff CL, Paduan VC, Corrente JE, 1523.

Vieira A, Marchette JC, de Souza JT, Luvizuto GJ, Nogueira [49] Chen H, Zhang SM, Hernán MA, Willett WC, Ascherio A
CR, Bazan R (2017) Amnestic dementia impairment in (2003) Weight loss in Parkinson’s disease. Ann Neurol 53,
Parkinson’s disease: The role of body composition, ageing 676-679.
and insulin resistance. Clin Nutri ESPEN 20, 47-51. [50] Goodpaster BH, Kelley DE, Wing RR, Meier A, Thaete FL
[35] Martinez-Martin P, Schapira AH, Stocchi F, Sethi K, Odin P, (1999) Effects of weight loss on regional fat distribution and

MacPhee G, Brown RG, Naidu Y, Clayton L, Abe K, Tsuboi insulin sensitivity in obesity. Diabetes 48, 839-847.
Y (2007) Prevalence of nonmotor symptoms in Parkinson’s [51] Chiu KC, Cohan P, Lee NP, Chuang LM (2000) Insulin
disease in an international setting; study using nonmotor sensitivity differs among ethnic groups with a compen-
symptoms questionnaire in 545 patients. Mov Disord 22, satory response in beta-cell function. Diabetes Care 23,
1623-1629. 1353-1358.
[36] Jenkinson C, Fitzpatrick RA, Peto VI, Greenhall R, Hyman [52] Babey SH, Wolstein J, Diamant AL, Goldstein H (2016)
N (1997) The Parkinson’s Disease Questionnaire (PDQ- Prediabetes in California: Nearly half of California adults
39): Development and validation of a Parkinson’s disease on path to diabetes. Policy Brief (UCLA Center for Health
summary index score. Age Ageing 26, 353-357. Policy Research PB2016-1, 1-8.
[37] Woert MH, Mueller PS (1971) Glucose, insulin, and free [53] DeFronzo RA, Ferrannini E (1991) Insulin resistance: A
fatty acid metabolism in Parkinson’s disease treated with multifaceted syndrome responsible for NIDDM, obesity,
levodopa. Clin Pharmacol Ther 12, 360-367. hypertension, dyslipidemia, and atherosclerotic cardiovas-
[38] Wallace TM, Levy JC, Matthews DR (2004) Use and abuse cular disease. Diabetes Care 14, 173-194.
of HOMA modeling. Diabetes Care 27, 1487-1495.