Cognition and Cerebrovascular Function in Primary Hyperparathyroidism before and after Parathyroidectomy

Abstract

Purpose:

There are cognitive changes in primary hyperparathyroidism (PHPT) that improve with parathyroidectomy but the mechanism of cognitive dysfunction has not been delineated. We assessed if cerebrovascular function is impaired in PHPT, improves post-parathyroidectomy and is associated with PTH level and cognitive dysfunction.

Methods:

This is an observational study of 43 patients with mild hypercalcemic or normocalcemic PHPT or goiter. At baseline, cerebrovascular function (dynamic cerebral autoregulation and vasomotor reactivity) by transcranial Doppler and neuropsychological function were compared between all 3 groups. A subset underwent parathyroidectomy or thyroidectomy, and was compared 6 months post-operatively.

Results:

Mean cerebrovascular and neuropsychological function was normal and no worse in PHPT compared to controls preoperatively. Higher PTH was associated with worse intracerebral autoregulation (r=−0.43, p=0.02) and worse cognitive performance on some tests. Post-parathyroidectomy, mood improved significantly, but changes did not differ compared to those having thyroidectomy (p=0.84). There was no consistent improvement in cognition or change in vascular function in either surgical group.

Conclusions:

Although higher PTH was associated with worse intracerebral autoregulation, cerebrovascular function, cognition and mood were normal in mild PHPT. PTX did not improve vascular or cognitive function. The observed improvement in mood cannot be clearly attributed to PTX. Notwithstanding the small sample size, the results do not support changing current criteria for parathyroidectomy to include cognitive complaints. However, the associations between PTH, cognition and cerebral autoregulation merit future studies in those with more severe hyperparathyroidism.

Introduction

PHPT is a common endocrine condition associated with hypercalcemia and elevated or inappropriately normal serum levels of PTH. While most patients today are “asymptomatic”, lacking the classical skeletal and renal manifestations of PHPT, non-specific neuropsychological symptoms are common (1-4). Many studies suggest PHPT is associated with impaired quality of life (QOL) and depression (5-10). While less well studied, some have demonstrated reduced memory or impairment in other cognitive domains (5,11-16). More controversial is whether neuropsychological symptoms are specifically attributable to PHPT and reversible with surgical correction of the disorder. While observational studies suggest post-parathyroidectomy (PTX) improvement, three randomized controlled trials have had conflicting results (5,8-10,17). Further, a mechanism accounting for the putative neuropsychological dysfunction in PHPT has not been clearly delineated. For these reasons, neuropsychological symptoms are not currently considered an indication for PTX by most guidelines and experts (18).

We have shown that vascular stiffness is increased in PHPT and associated with the degree of PTH elevation (19,20). In separate studies, we also showed that patients with PHPT have depressive symptoms and reduced verbal memory and non-verbal abstraction that improve post-PTX (5). Cerebral hemodynamic dysfunction is associated with cognitive impairment in vascular dementia, asymptomatic carotid disease and other conditions (21-25). We therefore hypothesized that PTH-induced vascular dysfunction within the intracerebral circulation, might compromise cerebral vascular function or blood flow regulation and thereby account for the observed cognitive changes in PHPT. To address this matter, we performed a longitudinal “proof of concept” study assessing cerebrovascular function, cognition and mood in patients with PHPT or goiter, some of whom planned to undergo PTX or thyroidectomy, respectively.

Subjects and Methods

This study was approved by the Columbia University Medical Center (CUMC) Institutional Review Board. Consecutive patients with PHPT were recruited from the Metabolic Bone Diseases Unit and the endocrine surgery and general endocrinology clinics at CUMC. After written informed consent was given, three groups of postmenopausal (>1 years post-menopause), English-speaking women ages 50-85 years old without vitamin D deficiency (25-hydroxyvitamin D ≥ 20 ng/ml) and with normal renal function (GFR >60ml/min) were enrolled: 1) PHPT (N=29), defined by the presence of hypercalcemia [calcium above normal (>10.2mg/dl but <11.5) with an elevated PTH level (>65 pg/mL; normal range 10-65 pg/ml). Those with more marked hypercalcemia were excluded in order to assess the biochemical phenotype most commonly seen today. Women with inappropriately normal PTH were excluded because we hypothesized that vascular dysfunction was due to elevated PTH. 2) Normocalcemic primary hyperparathyroidism or NPHPT (n=7) was defined as having an elevated PTH and normal ionized and albumin-corrected calcium and no secondary causes of high PTH (26-28). The NPHPT group was included to distinguish the effect of calcium from that of PTH on the study outcomes. 3) The third group consisting of patients (n=7) who had goiter or thyroid nodules and planned to have complete or hemi-thyroidectomy was included to account for the placebo effect of surgery. All surgical controls had normal calcium and PTH, were euthyroid and having surgery for non-cancer diagnoses (non-toxic nodular goiter, pre-operative fine needle aspiration with Bethesda Category IV pathology or lower risk). None had Hashimoto’s thyroiditis.

We excluded those with familial or drug-induced PHPT; history of stroke or dementia; seizure disorder; malignancy other than non-melanoma skin cancer; uncontrolled hyper- or hypothyroidism; liver dysfunction, kidney disease (GFR <60ml/min); traumatic brain injury; prior neurosurgery or substance abuse. Men and premenopausal women were also excluded to avoid the confounding effects of sex or menopause status.

A total of 431 prospective participants were screened for the study: 176 patients with goiter and 253 with PHPT (including NPHPT). Figure 1 indicates the reasons for exclusion. PHPT patients who participated did not differ in age (64.9±7.3 vs. 64.9±8.7 years, p=1.0) or serum calcium (10.5±0.5 vs. 10.6±0.6, p=0.37) from those who did not.

Figure 1.

Flow diagram indicating patient screening and enrollment.

Evaluation

Participants having thyroid or parathyroid surgery were evaluated at baseline and 6 months post-surgery while those with NPHPT were evaluated only at baseline. The following tests were performed:

Questionnaire:

Information about demographics and covariates (age; race; education; PHPT characteristics; medical history; medications; and cardiovascular risk factors) was collected, as described in our prior studies (5,19,29).

Neuropsychological Testing:

Participants were tested by one, blinded, trained tester at both time points, using the following well-established and validated tests for depression, verbal and visual memory, verbal fluency, visual attention/task switching, mental manipulation, and motor coordination (22): The Centers for Epidemiological Studies Depression Scale (CESD); Letter Cancellation; Hopkins Verbal Learning Test (HVLT); Rey-Osterreith Complex Figure Test; Repetition of Phrases and Sentences test; Trail Making Tests A&B; Wechsler Adult Intelligence Scale-III Digit Symbol Subtest; Wechsler Adult Intelligent Scale III Digit Span Subtest; Grooved Pegboard Test; Line Bisection Test; Controlled Oral Word Association (COWA) Test; and Boston Naming Test (BNT). Table 1 shows the task required for each test and the cognitive domain assessed. Participant scores were compared to the standardized norms and converted to Z-scores based on age, education and/or race with the exception of Sentence Repetition, Letter Cancellation, Letter Bisection and CESD, which were reported as raw scores.

Table 1.

Description of Cognitive Testing

Neurocognitive Test	Task	Z-score Adjusted for
Verbal Memory, Fluency and Word Retrieval
Hopkins Verbal Learning Test	Listen to a list of words and recall them immediately and after 20 minutes. Then recognize these words from a longer list.	Age
Controlled Oral Word Association (COWA)	Produce words beginning with a given letter (F, A, S) in one minute	Age,Education,Race
Boston Naming Test (BNT)	Name 60 objects (line-drawn) of graded difficulty
Sentence Repetition	Repeat phrases and sentences of alternating vocabulary difficulty and predictability	N/A (Raw score)
Visuospatial Perception, Constructional Ability, Visual Memory
Rey-Osterreith Complex Figure Test	Look at a complex figure drawing and copy it.Reproduce the figure after 3 minutes and 30 minutes ("immediate" and "delayed" recall). Then recognize parts of the figure from an assortment of designs.	Age
Visual Attention, Mental Flexibility, Processing Speed
Trail-Making Test A	Draw lines to connect numbered circles, in consecutive order, that are randomly arranged on a page. Timed.	Age,Education,Race
Trail-Making Test B	Connect randomly-arranged, numbered and letter circles by alternating (1, A, 2, B, 3, C, etc). Timed.
Letter Cancellation	Cross out all the "A" on a field of pseudorandomly arrayed capital letters	N/A (Raw score)
Working Memory/Mental Manipulation
Digit Symbol	Enter the symbol corresponding to numbers 1-9 based on a key. 120 second time limit.	Age
Digit Span	Repeat increasingly longer strings of digits forwards and backwards.
Motor Speed and Coordination
Grooved Pegboard Test	Rotate 25 pegs into position in slotted holes using the dominant and nondominant hand. Timed.	Age,Education,Race
Spatial Neglect
Line Bisection	Mark exactly in the middle of a line (six 6cm or 12cm lines shown consecutively)	N/A (Raw score)

Transcranial Doppler (TCD):

Both middle cerebral arteries were insonated via the temporal windows at a depth of ~50-58mm using 2MHz probes attached to a head frame (Terumo Trifid PMD150B, Spencer Technologies) to continuously record cerebral blood flow velocity. Arterial blood pressure waveforms were continuously and non-invasively monitored using a Finapres finger cuff (Finometer Pro, Amsterdam, Netherlands). End-tidal CO₂ was measured continuously by an infrared capnometer connected to a facemask. During the testing, arterial blood pressure, cerebral blood flow velocity and end-tidal CO₂ were recorded continuously into a multichannel recorder. After 10 min of baseline recording during normal breathing (normocapnia), 5% end-tidal CO₂ was administered by facemask, to elevate CO₂ to a new plateau for 2 min (hypercapnia). Two TCA measures were assessed: 1) vasomotor activity (VMR) which assesses how much “vasodilatory reserve” is available to protect the brain from extremes of hypo or hypertension (Primary Outcome) and 2) dynamic cerebral autoregulation (DCA), which assesses current/spontaneous blood flow regulation (Secondary Outcome).

VMR was calculated as % change in mean cerebral blood flow velocity per 1mm increase in end-tidal CO₂ during CO₂ inhalation and averaged between the left and right sides. VMR<2.0%/mm HgPCO₂ is considered abnormal (30,31). Dynamic cerebral autoregulation (DCA) was assessed by TCD while subjects were breathing spontaneously. After temporal synchronization of the blood pressure and blood flow velocity wave forms, DCA was analyzed with transfer function analysis using an in-house program on a Matlab platform (MathWorks, Natick, USA), quantifying the extent to which the input signal (arterial blood pressure) is reflected in the output signal (cerebral blood flow velocity) in a given oscillatory frequency range (32). DCA was expressed as phase shift - the relative separation of signals in the frequency range in which autoregulation occurs. DCA measures the efficiency of counter-regulation of blood flow to spontaneous changes in blood pressure. The more efficient the counter-regulation of flow, the more cerebral blood flow velocity “leads” the fluctuating pressure wave. Lower phase shift indicates worse autoregulation, which has been shown to predict stroke outcomes (32). Normal PS is generally >30 degrees. TCD measures were not available in 7 participants due to inability to insonate the MCA.

Biochemistries:

Fasting samples for serum calcium (normal 8.6-10.2 mg/dl), phosphate (normal 2.5-4.5mg/dl), LDL cholesterol (normal <100 mg/dl) and creatinine (normal 0.5-1.1 mg/dl) were measured on a Cobas Integra 400 plus (Roche Diagnostics, Indianapolis, IN) with intra- and inter-assay precision <2.5% for all tests. Total intact PTH (normal 14-65pg/ml) was measured with an immunoradiometric assay (Scantibodies Laboratories, Hanover, Germany) with intra- and inter-assay precision of 4.0% and 5.8% respectively. 25-hydroxyvitamin D2 and D3 were measured using Ultra Performance Liquid Chromatography-Tandem Mass Spectrometry (LC-MSMS). Inter-day precision was 8.1% and 5.5% for 25-OH-D2 and D3. TSH (normal range 0.4-4 uIU/ml) was measured by a chemiluminescent assay using a Siemens Immulite 1000 auto analyzer (Tarrytown, NY) with an intra-assay precision of 7.1%.

Statistical Analysis

Descriptive statistics were expressed as absolute (n) and relative (%) frequency and mean and standard deviation (SD) for categorical and continuous variables. We assessed the normality of all variables with the Shapiro-Wilk test. Student’s t-test, Chi-squared test or Fisher’s exact test were used, as appropriate, to assess between-group differences in normally distributed continuous or categorical variables, respectively. We assessed non-normally distributed continuous variables with the Wilcoxon Rank Sum test (VMR, letter cancellation, HVLT delayed recall/recognition, Rey recognition, sentence repetition, CESD, TSH, creatinine, vitamin D, PTH, blood pressure, heart rate and BMI). ANCOVA was utilized to assess differences adjusted for education or hypercholesterolemia. Values are expressed as means ± SD or percentages. Within-person longitudinal changes were evaluated with linear mixed models for repeated measures. Pearson correlation was used to calculate the association between cerebrovascular function, cognition, and biochemistries. Multiple regression was carried out to determine whether PTH level is an independent predictor of autoregulation. Known confounders were included in the model regardless of significance. Regression diagnostics indicated residuals were normally distributed. Statistical analysis was performed using R 64-bit, version 3.5.1 and SAS, version 9.4. A two-tailed p-value <0.05 was considered statistically significant. Correlations were not adjusted for multiple comparisons given the “proof of concept” nature of this study. With the current sample we could detect an absolute difference of 2.2 %/mm HgPCO₂ in VMR (primary outcome).

Results

Baseline Evaluation

Those with PHPT had evidence of biochemically mild PHPT with mean±SD Ca 10.5±0.5 mg/dl and PTH 93±60 pg/ml, as shown in Table 2. All those with hypercalcemic PHPT had serum calcium within 1 mg/dl of the upper limit of normal (≤11.2 mg/dl) and 89% had PTH < 2 times the upper limit of normal (<130 pg/ml). In the PHPT group, 27.6% had a history of nephrolithiasis, 31.0% had osteoporosis, 13.8% had a history of fragility fracture and 65.5% met 1 or more surgical criteria for PTX based on 2014 guidelines. Among NPHPT participants, the rate of osteoporosis (p=0.008), but not fragility fracture (p=0.23) or nephrolithiasis (p=0.81) was higher compared to those with PHPT.

Table 2.

Baseline Clinical Characteristics and Biochemistries

Variable	PHPT, N = 29	NHPT, N = 7	Thyroid ctrls, N = 7	p value
Age (years)	64.9 ± 7.3	66.7 ± 6.2	61.60 ± 5.6	0.37
Education (years)	17.9 ± 1.7	15.7 ± 2.4	16.3 ± 2.9	0.06
Race (% Caucasian)	93.1%	71.4%	71.4%	0.17
Ethnicity (% non-Hispanic)	96.6%	100%	85.7%	0.39
BMI (kg/m2)	27.2 ± 7.0	22.9 ± 14.9	30.2 ± 18.3	0.24
PHPT characteristics
Nephrolithiasis (%)	27.6%	14.3%	n/a	0.81
Osteoporosis (%)	31.0%	100%	n/a	0.008
Fragility fracture (%)	13.8%	42.8%	n/a	0.23
Cardiovascular risk factors
Hypercholesterolemia (%)	44.4%	85.8%	100%	0.007a
Diabetes (%)	3.7%	14.3%	14.3%	0.49
Hypertension (%)	28.6%	28.6%	28.6%	1.0
Myocardial infarction (%)	3.6%	14.3%	0%	0.42
Anti-hypertensive use (%)	28.6%	14.3%	28.6%	0.75
Aspirin use (%)	17.9%	28.6%	14.3%	0.78
History of smoking (%)	32.1%	42.9%	57.1%	0.48
Systolic blood pressure (mmHg)	124 ± 29	110 ± 58	130 ± 63	0.73
Diastolic blood pressure (mmHg)	80 ± 19	67 ± 37	85 ± 40	0.17
Heart rate (bpm)	76 ± 21	60 ± 41	70 ± 45	0.46
Serum biochemistries
Calcium (mg/dl)	10.5 ± 0.5	9.4 ± 0.9	9.5 ± 0.9	< 0.001b
PTH (pg/ml)	93 ± 60	102 ± 121	36 ± 11	0.002c
Phosphate (mg/dl)	3.2 ± 0.5	3.3 ± 1.1	3.7 ± 1.1	0.11
LDL (mg/dl)	133 ± 34	121 ± 67	136 ± 67	0.56
25-Hydroxyvitamin D (ng/ml)	37.1 ± 15.4	38.1 ± 30.7	36.6 ± 30.7	0.82
Vitamin D insufficiency (%)	32.1%	14.3%	28.6%	0.88
Creatinine (mg/dl)	0.75 ± 0.2	0.77 ± 0.33	0.76 ± 0.33	0.90
TSH (mIU/L)	2.2 ± 1.3	1.8 ± 2.7	1.1 ± 2.7	0.03d

Values represent mean ± SD or percentages

PTH parathyroid hormone, LDL low-density lipoprotein, TSH thyroid stimulating hormone

^ap < 0.05 for thyroid vs. PHPT

^bp < 0.001 for PHPT vs. both other groups

^cp ≤ 0.05 for PHPT and NPHPT vs. thyroid

^dp = 0.05 PHPT vs. thyroid

There was no difference in age (p=0.37) or education (p=0.06) between the 3 groups. As expected, serum calcium was elevated in those with PHPT compared to the other groups (p<0.001 for both pairwise comparisons; Table 2). PTH levels were higher in the PHPT and NPHPT groups compared to those with thyroid disease (both p≤0.05), but the PHPT and NPHPT groups did not differ from each other (p=0.90). Vitamin D, phosphate, as well as renal function did not differ between groups, nor was there a difference in the frequency of vitamin D insufficiency (Table 2). TSH tended to be slightly higher within the normal range in the PHPT group vs. those with goiter.

In terms of cardiovascular risk factors, only the prevalence of hypercholesterolemia differed between groups (Table 2; p=0.007), but there was no difference in low density lipoprotein (LDL) levels. Rates of hypercholesterolemia were higher in those with goiter compared to those with PHPT (p=0.02), but the difference between those with NPHPT and PHPT did not meet statistical significance (p=0.07). There were no between-group differences in any other cardiovascular risk factor including, systolic blood pressure (p=0.73), heart rate (p=0.46), BMI (p=0.24), the frequency of hypertension (p=1.0), myocardial infarction (p=0.42), diabetes (p=0.49), history of smoking (p=0.48), or use of hormone replacement therapy (p=0.85), or aspirin (p=0.77). There were no between-group differences in the frequency of clinical anti-hypertensive use (p=0.75), which included angiotensin converting enzyme inhibitors, angiotensin receptor blockers, B-blockers, diuretics, and calcium channel blockers.

As shown Table 3, the mean VMR was normal (>2%) and did not differ between the 3 groups (p=0.19). Adjusting for hypercholesterolemia did not change the significance of the comparison (p=0.15). Mean DCA was also normal (>30 degrees) and did not differ between groups before (p=0.77) or after adjusting for rates of hypercholesterolemia (p=0.62). On average, none of the groups reported depression (CESD score ≥21) and there were no between-group differences in mood (Table 3). Cognitive Z-scores were normal on average (within 2SD of age-matched controls) in all groups, and there were no between-group differences in cognition except for performance on the COWA, which was worse among thyroid controls. Controlling for education level did not alter the significance of results (Table 3).

Table 3:

Baseline Comparison of Depressive Symptoms, Cognitive Z-Scores and Cerebrovascular Measures

	PHPT	NHPT	Thyroid	P-value	Adjusted P-value*
Cerebrovascular Measures
Vasomotor Reactivity (%/mm HgPCO2	3.4±1.8	4.8±3.2	3.9±3.7	0.19	0.15
DCA Phase Shift (Degrees)	31.4±27.7	40.4±56.9	33.3±63.6	0.77	0.62
Mood
CESD	12.3±10.9	12.7±22.2	12.0±22.2	0.96	0.99
Verbal Memory, Fluency and Word Retrieval
HVLT Total Recall	0.0±1.2	−0.5±2.4	−0.5±2.4	0.37	0.34
HVLT Delayed Recall	−0.1±1.3	−0.1±2.7	0.0±2.7	0.88	0.97
HVLT Recognition	0.5±0.8	0.5±1.6	0.0±1.6	0.17	0.17
COWA	−0.2±1.2	0.8±2.5	−0.7±2.5	0.02	0.03
BNT	−0.3±1.3	0.1±2.7	−0.2±2.7	0.72	0.72
Sentence Repetition	15.2±1.8	14.4±3.7	15.3±3.7	0.97	0.45
Visual Memory
Rey Copy	−1.7±2.6	−3.3±5.4	−2.2±5.4	0.27	0.29
Rey Immediate Recall	−0.2±1.6	−1.3±3.2	−0.5±3.2	0.14	0.14
Rey Delayed Recall	−0.2±1.7	−1.1±3.5	−0.6±3.5	0.40	0.40
Rey Recognition	−0.1±1.8	−0.8±3.7	−0.1±3.7	0.25	0.55
Visual Attention and Task Switching
Trails A	0.1±1.4	0.0±2.8	0.5±2.8	0.69	0.68
Trails B	0.0±1.5	−0.3±3.0	0.9±3.0	0.19	0.20
Letter Cancellation	−0.1±1.1	0.2±2.34	−0.5±2.3	0.60	0.38
Working Memory/Mental Manipulation
Digit Symbol	0.8±1.2	0.0±2.4	0.5±2.4	0.22	0.22
Digit Span	0.4±1.1	−0.1±2.2	0.1±2.2	0.37	0.37
Motor Coordination
Pegboard – Dominant hand	−0.5±1.3	−0.2±2.6	−0.1±2.6	0.59	0.60
Pegboard – Non-dominant hand	−0.5±1.4	−0.5±2.7	−0.1±3.0	0.77	0.78
Spatial Neglect
Line Bisection	−0.3±4.1	−1.7±8.4	−1.1±8.4	0.59	0.60

Values represent mean Z-scores±SD;

^*cognitive measures adjusted for education level; cerebrovascular measures adjusted for hypercholesterolemia

Within the PHPT group, we assessed if those with and without osteoporosis, hypertension or vitamin D insufficiency had worse vascular function or cognition. There were no differences in cognition, autoregulation, or VMR between those with and without osteoporosis or hypertension. In those with vitamin D insufficiency, only performance on one cognitive test (Digit span: (−0.3±0.9 vs. 0.7±0.8, p=0.007) was lower compared to those with vitamin D sufficiency.

In the entire cohort, we assessed associations between biochemical indices of PHPT severity, vascular indices and cognitive performance. VMR did not correlate with serum calcium (r = −0.23, p= 0.20) or PTH (r = 0.05, p=0.80). On the other hand, degree of PTH elevation was associated with worse DCA (Figure 2; r=−0.43, p=0.02) but not with calcium (r=−0.15, p=0.44). Within the PHPT group, PTH also correlated with worse DCA (r=−0.44, p=0.04); DCA and VMR did not correlate with cognitive test Z-scores except for a negative correlation between DCA and the Rey-Osterrith Complex Figure Recognition test (r=−0.46, p=0.01). In a regression model (Table 4) including PTH, age, calcium, creatinine and vitamin D, PTH remained a significant predictor (p=0.003) of autoregulation. The parameter estimate is interpreted to mean that for every 10 mg/dl increase in PTH, DCA phase shift decreases (worsens) by 3 degrees. The model accounted for 39.3% of the variance in autoregulation (p=0.03).

Figure 2.

The association between serum PTH level and DCA in the entire cohort. The majority of PTH levels are mildly elevated (<130 pg/ml).

Table 4.

Regression Model of Autoregulation

Variable	Parameter Estimate	Standard Error	P-Value
PTH (per 10 mg/dl)	−3.002	0.907	0.003
Age (per 10 years)	14.142	5.821	0.02
Creatinine (per 0.1)	−4.8837	3.3326	0.23
Ca (per 1 mg/dl)	7.5777	7.7720	0.34
25-hydroxyvitamin D (per 1 mg/dl)	−0.3331	0.2921	0.27

PTH level negatively correlated with Sentence Repetition performance (r=−0.30, p=0.01), Trails A performance (r=−0.26, p=0.03), and motor coordination (r=−0.27, p=0.03; grooved peg board, dominant hand) indicating higher PTH levels were associated with worse performance. On the other hand higher calcium correlated positively only with depressive symptoms (r=0.25, p=0.04). There were no other significant correlations between PTH or calcium and any other cognitive performance scores (data not shown).

Longitudinal Comparison

Biochemical indices

Twenty-three individuals with PHPT underwent PTX and 5 patients underwent thyroid surgery and had follow-up testing at 6 months post-surgery. As shown in Table 5, all patients in the PHPT group were cured by PTX. As expected, serum PTH and calcium normalized (both p<0.01) post-PTX. 25-hydroxyvitamin D (<0.01) and phosphate (p=0.01) increased within the normal range while creatinine (p=0.70), LDL (p=0.33) and TSH did not change (p=0.98). In thyroid controls there were no biochemical changes over time (Table 3).

Table 5.

Change in Biochemistries after PTX

	Parathyroidectomy N=23			Thyroidectomy N=4
	Pre- surgery	Post- surgery	Pre- post p-value	Pre- surgery	Post- surgery	Pre-post p-value
Calcium (mg/dl)	10.5±0.4	9.4±0.4	<0.001	9.5±0.5	9.3±0.2	0.53
PTH (pg/ml)	93.4±55.1	45±19.1	0.002	35.8±10.9	36.1 ±4.5	1.0
Vitamin D (ng/ml)	3 7.8± 16.1	44.6±17.3	0.003	41.6±10.6	45.3±5.7	0.82
Phosphate (mg/dl)	3.3±0.4	3.7±0.5	0.01	3.7±0.7	4.0±0.5	0.99
Creatinine (mg/dl)	0.7±0.1	0.8±0.1	0.70	0.8±0.1	0.8±0.1	0.74
LDL (mg/dl)	130±26	138±31	0.33	136±22	161±14	0.66
TSH mIU/l	1.98±1.0	1.86±1.0	0.98	0.96±0.41	1.36±2.2	0.93

Post-surgery vascular function

VMR did not change within the PTX (3.4±1.0 vs. 3.5±1.0% mmHgPCO₂, p=0.99) or thyroidectomy group over time (3.9±2.1 vs. 3.4±1.4 mmHgPCO₂, p=0.92). DCA did not change post-PTX (28.5±21.0 vs. 24.3±16.1 degrees, p=0.40) or post-thyroidectomy (33.3±8.5 vs. 21.4±15.5, p=0.54). There were no between-group differences in the change over time in VMR (p=0.72) or DCA (p=0.51).

Neuropsychological Testing

As shown in Figure 3, depressive symptoms declined by 43.1% after PTX compared to baseline (p=0.008). After thyroidectomy, the improvement (38.3%) in mood was not significant (p=0.51). The change in mood over time did not differ between groups (p=0.84). As shown in Figure 4, there was a small improvement (~ 0.5SD, p<0.05) in motor coordination (pegboard only with the dominant hand) and worsening of visual recognition (HVLT) after PTX compared to baseline. There were no changes in the thyroidectomy group compared to baseline and no between-group differences (PTX vs. thyroidectomy) in post-surgical changes on any test.

Figure 3.

Comparison of Mean Within-Person Changes in depressive symptoms in the PTX (black) and thyroidectomy (gray) groups; *indicates p<0.05 compared to baseline

Figure 4.

Comparison of Mean Within-Person Changes in (A) Verbal Memory, Fluency and Word Retrieval; (B) Visual Memory, Visual Attention and Task Switching; and (C) Working Memory, Motor Coordination and Spatial Neglect After Surgery in the PTX (black) and thyroidectomy (gray) groups. *indicates p<0.05 compared to baseline. There were no between group differences in the change over time.

Discussion

This is the first study to assess cerebrovascular function in PHPT and its relationship to cognition and PTH level. While PTH is considered a calciotropic hormone, it also has vascular effects (19,33-35). We hypothesized that PTH-induced intracerebral vascular dysfunction, might account for the observed cognitive changes in PHPT. Our results do not confirm this mechanism. While we did not find impaired vascular function or cognition, we did show an independent association between extent of PTH (but not calcium) elevation and worse cerebral autoregulation. The latter suggests this mechanism could be relevant in those with severely elevated PTH. Few patients in our study had severe hyperparathyroidism. This may have contributed to the lack of difference between those with and without PHPT. However, there were also no consistent associations between cerebrovascular function and cognition or improvement in autoregulation after PTX.

While the relationship between cerebral autoregulation and PTH is intriguing, the small sample size and lack of relationship between autoregulation and cognition, limits the overall conclusions that can be drawn in this regard. On the other hand, the association between PTH and autoregulation (regardless of a link to cognition) may have implications regarding risk for stroke in severe PHPT because TCD parameters predict stroke and stroke outcomes (36). Consistent with this, we previously showed other stroke risk factors (IMT) to be elevated in PHPT(19). The “proof of concept” nature of this study, however, requires confirmation. Future studies in those with more markedly elevated PTH (severe PHPT, renal failure or vitamin D deficiency) may be helpful.

Despite the association between PTH level and autoregulation we did not find cerebrovascular function to improve after PTX. While not unexpected given cerebrovascular function was normal at baseline, our prior data in PHPT suggested some indices that are normal (cognition, BMD etc.) at baseline improve after PTX. Studies investigating function in other parts of the vasculature after PTX have had mixed findings (37-41). Future studies comparing PHPT patients undergoing PTX to those who are observed could be helpful to determine if PTX prevents worsening of cerebrovascular function.

No other studies have assessed intra-cerebrovascular function and cognition in PHPT. Cerebrovascular blood flow (CBF) has been investigated in two small studies of moderate to severe PHPT. Neither of these assessed the relationship of blood flow to cognition and both used single photon emission computed tomography (SPECT), rather than TCD. The first (n=16) indicated most had reduced CBF at baseline that improved post-PTX (42), but changes in CBF did not correlate with depression or biochemistries. The second (n=24) showed hypo-perfusion was present in 23% of brain regions and there was a significant correlation with serum calcium and PTH levels (43). These prior results lend support the association seen in our study, though direct comparison is difficult due to differences in outcomes and PHPT severity.

As in our prior study, neuropsychological function was not abnormal before PTX. Only mood and motor coordination improved after PTX compared to baseline (5). Post-surgical changes, however, did not differ from those who underwent thyroidectomy. Because few participants had thyroidectomy, future larger studies would be needed to determine if post-surgical changes after PTX and thyroidectomy are comparable or whether the lack of between-group changes was due to a type 2 error. Many but not all prior studies indicate depression and impaired cognition in PHPT, but improvement in neuropsychological function after PTX is inconsistent (5,11,13-16,44-49). For this reason, neuropsychological symptoms are not considered an indication for PTX by most guidelines and experts. In spite of our small sample size, our results do not support changing current surgical criteria for PTX to include cognitive symptoms.

Our study has several limitations. The major weakness of our study is the small sample size. This may have restricted our ability to detect between-group differences and changes. For this reason, we also assessed results with regard to normal ranges as well as linear correlations. Moreover, we would consider the study to be “proof of concept” and most useful to guide design of future larger studies. Many potential participants declined enrollment which may limit the generalizability of results. However, a limited comparison of those who did vs. did not agree to participate found no difference.

Our study has several strengths. This is the first study to assess cerebrovascular function and its association to cognition in PHPT. We collected information on cardiovascular risk factors and controlled for them in the analysis. Further, the study design included a surgical comparator group and participants with NPHPT in order to control for placebo effects of surgery and to dissect the effect of calcium from PTH on cognition and cerebrovascular function. Those assessing cognitive and cerebrovascular function outcomes were also blinded to disease status.

In conclusion, in PHPT, cerebrovascular and neuropsychological function was normal but greater PTH elevation was associated with worse indices of cognition and cerebrovascular function. There was no improvement in vascular or cognitive function post-PTX. The observed improvement in mood cannot be clearly attributed to PTX. Notwithstanding the small sample size, the results do not support changing current surgical criteria for PTX to include neuropsychological symptoms. However, the associations between PTH, cognition and cerebral autoregulation warrant future studies in those with severe hyperparathyroidism.