Hypertension, Antihypertensive Medications, and Risk of Incident Primary Hyperparathyroidism

Abstract

Context:

Hypertension is associated with higher PTH levels, and specific antihypertensive medications may modulate PTH. Whether hypertension or the use of specific antihypertensive medications influences the risk of developing incident primary hyperparathyroidism (P-HPTH) is not known.

Objective:

The purpose of this study was to investigate whether a history of hypertension and the use of specific antihypertensive medications determine the risk for developing P-HPTH in a large prospective study.

Design/Participants:

A longitudinal prospective cohort of female nurses in the Nurses' Health Study I (n = 75 600), who did not have P-HPTH at baseline and completed a questionnaire assessment of lifetime history of P-HPTH were followed from 1986 to 2008. Most participants were white and postmenopausal.

Setting:

The study was a nationwide cohort study.

Main Outcome Measure:

Incident P-HPTH was assessed initially via questionnaire and then was confirmed by medical record review. Cox proportional hazards models were used to adjust for potential confounders.

Results:

We documented 347 incident cases of P-HPTH during 1 719 416 person-years of follow-up. The age-adjusted relative risk (RR) for incident P-HPTH associated with hypertension was 1.80 (95% confidence interval [CI], 1.43–2.26), and the multivariate-adjusted RR was 1.45 (95% CI, 1.10–1.91). Among participants with a history of hypertension, the use of furosemide, when compared with the use of other antihypertensive medications, was associated with increased risk for developing P-HPTH; age-adjusted RR for incident P-HPTH was 1.79 (95% CI, 1.15–2.79) and multivariate-adjusted RR was 1.71 (95% CI, 1.08–2.71).

Conclusions:

In a large longitudinal prospective cohort study of mostly older white women, a history of hypertension and use of furosemide were associated with a significantly higher risk of developing P-HPTH.

Primary hyperparathyroidism (P-HPTH) is an endocrine disorder that represents the most common cause of ambulatory hypercalcemia (1–5). The prevalence of P-HPTH in the United States is estimated to have tripled in the last 3 decades, and the burden of disease disproportionately affects women older than the age of 50 years (1). P-HPTH can cause symptomatic hypercalcemia, kidney stones, decreased bone mineral density (BMD), and increased risk for fragility fractures (6, 7). Because these adverse skeletal consequences of P-HPTH are amplified in older and postmenopausal women, it is notable that up to 2% of postmenopausal women could have P-HPTH (8, 9).

Despite the rising prevalence of P-HPTH and costly clinical sequelae, our understanding of the modifiable risk factors for P-HPTH remains poor. We previously reported that higher dietary and supplemental calcium intake was independently associated with a decreased risk of P-HPTH in a cohort of older women (mean age, >50 years) (10). The results of this large prospective analysis from the Nurses' Health Study I (NHS) (10) implicated the maintenance of higher calcium intake as a potential method to reduce incident P-HPTH based on the previously proposed hypothesis that an overall positive calcium balance may suppress excessive parathyroid stimulation and/or development of a neoplastic parathyroid process (11–13).

In addition to calcium intake, blood pressure has been implicated as a potential risk factor for hyperparathyroidism; higher blood pressure and hypertension status are associated with higher circulating parathyroid hormone levels (14–18), and P-HPTH is associated with hypertension and cardiovascular disease (19–21). Further, several antihypertensive medication classes modulate calcium homeostasis and parathyroid hormone levels. For example, thiazide and loop diuretics alter urinary calcium excretion (the former decreases calciuria whereas the latter increases it), markers of bone turnover, and parathyroid hormone levels (22–26), renin-angiotensin-aldosterone system (RAAS) inhibitors lower parathyroid hormone (14, 27–30), and calcium channel blockers may potentially raise parathyroid hormone levels (31, 32).

In this context, we hypothesized that hypertension and the use of antihypertensive medication therapies may predict the risk of developing incident P-HPTH. We tested this hypothesis by conducting a large prospective study to examine the association between the diagnosis of hypertension and use of individual antihypertensive medications and risk of incident P-HPTH in women participating in the NHS.

Materials and Methods

Study population

The NHS is an ongoing, prospective cohort study that began in 1976, enrolling 121 700 female registered nurses between 30 and 55 years of age and residing in 11 US states. The cohort is followed by questionnaires that ask about lifestyle practices and newly diagnosed diseases mailed every 2 years. The follow-up of participants exceeds 90% of the eligible person-time. The baseline sample for our analysis included 75 600 women who answered either the 2006 or 2008 questionnaire, which assessed lifetime history of P-HPTH. The study protocol was reviewed and approved by the Brigham and Women's Hospital institutional review board.

Assessment of hypertension and antihypertensive medications

Self-reported hypertension (33) and diabetes (34) were previously validated in this cohort. Self-reported use of antihypertensive medications was ascertained from biennial questionnaires. Thiazide use was queried in 1980 and subsequently in 1982, 1988, 1994, and every 2 years thereafter. Furosemide use was first queried in 1994 and every 2 years thereafter. β-Blockers and calcium channel blockers were first queried in 1988 and subsequently in 1994 and every 2 years thereafter. Angiotensin-converting enzyme (ACE) inhibitors were queried in 1988 and subsequently in 1996 and then every 2 years thereafter. The complete assessment of RAAS inhibitors in this study was limited as ACE inhibitors were the only RAAS inhibitors available for this analysis because angiotensin receptor blockers were not queried in NHS until 2008.

Assessment of other nondietary factors

Age, body mass index (BMI), smoking status (never, past, or current), physical activity (in metabolic equivalent task scores), history of diabetes, hypertension, menopausal status, and postmenopausal hormone use were ascertained from the biennial questionnaires. Self-reported weight was highly reliable (r = 0.97) among a subset of participants who underwent direct measurement of their weight (35). Physical activity reported on the questionnaires has been previously validated against physical activity diaries in a similar cohort (r = 0.79) (36). Self-reported age at menopause and type of menopause were previously validated in the NHS and were shown to be highly accurate (37). Race was self-reported and categorized in this analysis as white and nonwhite.

Assessment of dietary intake

To assess the participants' diet, we used semiquantitative food frequency questionnaires that asked about the average intake of more than 130 individual food items and 22 individual beverages during the previous year. The participants were asked to complete food frequency questionnaires in 1986 and every 4 years thereafter. Intake of specific dietary factors was computed from the reported frequency of consumption of each specified unit of food and from US Department of Agriculture data on the content of the relevant nutrient in specified portions. Nutrient values were adjusted for total energy intake to determine the nutrient composition of the diet independent of the total amount of food eaten. The food frequency questionnaire also asked about the use of calcium supplements, vitamin D supplements, and multivitamins. The intake of supplemental calcium, vitamin D, and vitamin A in multivitamins or in isolated form was determined by the brand, type, and frequency of reported use.

The food frequency questionnaire has been extensively validated (38, 39).

Assessment of P-HPTH cases

Participants were asked about a diagnosis of hyperparathyroidism on the 2006 and 2008 questionnaires. On the 2008 questionnaire, nurses were also asked about lifetime history of hyperparathyroidism. To distinguish P-HPTH from nonprimary forms of hyperparathyroidism, we subsequently obtained the medical records of all participants who gave consent. We confirmed cases of P-HPTH by elevated serum concentrations of calcium (≥10.6 mg/dL; 1 mg/dL = 0.25 mmol/L) with a concomitantly high or insufficiently suppressed parathyroid hormone level (> 50 pg/mL; 1 pg/mL = 1 ng/L) and/or pathology report of a resected adenomatous or hyperplastic parathyroid gland in the context of medical records, indicating a diagnosis of P-HPTH. Of the medical records that were reviewed, we confirmed P-HPTH in 83% of participants. Self-reports were rejected after medical record review for a variety of reasons, most commonly identification of secondary hyperparathyroidism from vitamin D deficiency or renal insufficiency (9%), misreporting of a thyroid disorder as hyperparathyroidism (4%), or insufficient data to confirm the diagnosis of P-HPTH (3%). We included in the analysis only cases of P-HPTH that were diagnosed during the 22 years between the date on which the 1986 questionnaire was returned and May 31, 2008. Participants with a history of P-HPTH at baseline were excluded.

Statistical analyses

The study design was prospective; information on hypertension diagnosis and use of antihypertensive medications was collected before the diagnosis of P-HPTH. For each participant, we counted person-time of follow-up from the date on which the 1986 questionnaire was returned to the date on which P-HPTH was diagnosed or death occurred or May 31, 2008, whichever occurred first. Information on exposures of interest that were recorded in response to the 1986 questionnaire was updated on subsequent questionnaires. We allocated person-time of follow-up according to exposure status at the start of each follow-up period.

Because antihypertensive medication use is almost exclusively a function of hypertension diagnosis, we first assessed whether self-reported hypertension diagnosis was a predictor of incident P-HPTH. The measure of association used was the relative risk (RR)—the incidence rate among women with hypertension divided by the corresponding rate in women without hypertension. We subsequently assessed whether the use of individual antihypertensive medication classes (thiazide, furosemide, β-blocker, calcium channel blocker, and ACE inhibitor) was associated with risk for incident P-HPTH in the subset of participants with a self-reported history of hypertension (n = 52,694); the sample size for this subset analysis represents all participants in whom hypertension was diagnosed over the follow-up period.

We used Cox proportional hazards regression to simultaneously adjust for several risk factors. The variables considered in these models were the following: age; BMI (<22, 22–24.9, 25–29.9, or ≥30 kg/m²); race (white or nonwhite); physical activity level (quintiles); alcohol intake (none, 0.1–4.9, 5–14.9, or ≥15 g/d); supplemental calcium intake (none, 1–500, and >500 mg/d); quintiles of intakes of total vitamin D and vitamin A, dietary calcium intake and total protein; self-reported diabetes; menopausal status; postmenopausal hormone use (yes or no), and physical examination by the participant's primary doctor in the preceding 2 years (yes or no). Because furosemide use was observed to be more prevalent among individuals with a history of congestive heart failure (CHF) and chronic kidney disease, we conducted additional multivariable analyses including adjustment for the self-reported diagnosis of these conditions. Because diagnosis of osteoporosis and use of bisphosphonate medications can potentially hasten or delay the diagnosis of P-HPTH, respectively, we conducted additional multivariable analyses including adjustment for self-reported diagnosis of osteoporosis and self-reported use of bisphosphonates. Last, we repeated our analyses after including a 2-year lag time for the exposures of interest (history of hypertension and furosemide use). We calculated 95% CIs for all RRs. All P values were two tailed.

Results

History of hypertension and incident P-HPTH

When categorized by self-reported diagnosis of hypertension (Table 1), women with hypertension were older, had a higher BMI, and had a higher prevalence of antihypertensive medication use and history of kidney stones, CHF, chronic kidney disease, and diabetes. When categorized by furosemide use, hypertensive patients who used furosemide had a higher BMI, were less physically active, had higher total vitamin A and dietary calcium intake, and had a greater prevalence of CHF, chronic kidney disease, kidney stones, diabetes, and osteoporosis, than hypertensive patients treated with a nonfurosemide antihypertensive or with no antihypertensive medication (Supplemental Table 1). There were 222 incident cases of P-HPTH over >700 000 person-years of follow-up among those with hypertension, compared with 125 cases of P-HPTH over >1 million person-years of follow-up among those without hypertension (Table 2). The age-adjusted RR for P-HPTH among women with hypertension compared with that among women without hypertension was 1.80 (95% CI, 1.43−2.26). After adjustment for BMI, race, smoking status, menopausal status, postmenopausal hormone use, physical activity, history of diabetes, dietary calcium, supplemental calcium, total vitamin D and vitamin A intake, dietary intakes of magnesium, and total protein and alcohol intake, the RR did not appreciably change (Table 2). The inclusion of individual antihypertensive medication classes (thiazides, furosemide, calcium channel blocker, β-blocker, ACE inhibitor, and other) attenuated the RR to 1.45 (95% CI, 1.10–1.91). The RR and 95% CI did not change after further adjustment for self-reported history of CHF, chronic kidney disease, osteoporosis, and use of bisphosphonate medications (data not shown). In additional lag-time analyses in which participants had to have a history of hypertension for ≥2 years before the diagnosis of P-HPTH, the RR for P-HPTH was 1.48 (95% CI, 1.13–1.92). We calculated the population-attributable fraction for history of hypertension to be 15.6%.

Table 1.

Age-Standardized Baseline Characteristics of NHS Study Participants Categorized by History of Hypertension in 1986

Characteristics	History of Hypertension
	No	Yes
No.	58 480	17 120
Age, ya	50.7 (6.9)	54.0 (6.6)
BMI, kg/m2	24.6 (4.1)	27.6 (5.6)
Physical activity, MET-h/wk	14.8 (21.8)	13.2 (17.9)
Smoking status
Never	45.2	46.5
Past	34.7	36.2
Current	19.2	16.5
Missing	0.9	0.8
Race
White	94.6	92.7
Black	0.9	2.1
Asian	0.6	0.7
Multiple	1.2	1.6
Other/unknown	2.6	2.8
Menopausal status
Premenopausal	37.1	33.1
Postmenopausal	62.7	66.7
Missing	0.2	0.2
Postmenopausal hormone use	17.4	18.3
Antihypertensive medication use
Thiazide diureticb	3.5	36.2
Furosemidec	0.9	3.7
β-Adrenergic antagonistb	3.0	22.7
Calcium channel blockerb	1.1	5.3
ACE inhibitorb	0.5	7.6
Other medication classesb	0.2	5.9
History of kidney stones, in 1992	3.8	5.5
History of osteoporosis	2.3	3.0
History of CHFd	0.8	2.3
History of chronic kidney diseasee	0.1	0.6
History of diabetes	1.6	5.9
Dietary calcium intake, mg/d	720.2 (251.1)	709.9 (248.0)
Supplemental calcium use	56.7	56.0
Supplemental calcium intake, mg/d	351.3 (424.8)	354.0 (429.1)
Alcohol intake, g/d	6.0 (1.0)	6.2 (11.6)
Total vitamin D intake, dietary and supplemental, IU/d	338.4 (250.6)	337.4 (249.3)
Total vitamin A intake, dietary and supplemental, IU/d	13 335.9 (7917.5)	13 222.7 (7778.5)
Dietary magnesium intake, mg/d	300.2 (68.7)	296.2 (67.7)
Dietary protein intake, g/d	74.7 (12.7)	76.0 (13.2)
Dietary animal protein intake, g/d	54.4 (13.6)	55.9 (13.9)

Abbreviation: MET, metabolic equivalent task. Values are means (SD) or percentages and are standardized to the age distribution of the study population. Values of polytomous variables may not sum to 100% due to rounding.

^aValue is not age adjusted.

^bAssessed in 1988.

^cAssessed in 1994.

^dAssessed in 1998.

^eAssessed in 1990 and 2002.

Table 2.

Relative Risks for Incident P-HPTH According to History of Hypertension

	History of Hypertension
	No	Yes
No. of incident P-HPTH cases	125	222
Person-years of follow-up	1 016 074	703 342
Age-adjusted RR (95% CI)	1.0	1.80 (1.43–2.26)
Multivariate-adjusted RR (95% CI)a	1.0	1.76 (1.39–2.24)
Multivariate + antihypertensive medication–adjusted RR (95% CI)b	1.0	1.45 (1.10–1.91)

^aThe multivariable model includes age, BMI categories (<22, 22–24.9, 25–29.9, or ≥30 kg/m²), race, smoking status (past, current, or never), menopausal status (premenopausal or postmenopausal), postmenopausal hormone use (yes or no), physical examination in prior 2 years, physical activity (quintiles), history of diabetes, dietary calcium intake, supplemental calcium intake, total vitamin D and vitamin A intake, alcohol intake, dietary magnesium intake, and dietary protein intake.

^bIncludes all variables in the multivariable model in addition to other individual antihypertensive medication classes: thiazide diuretics, furosemide, calcium channel blockers, β-adrenergic receptor antagonists, ACE inhibitors, and other antihypertensive classes.

Antihypertensive medication use and incident P-HPTH

Given the independent relation between hypertension status and incident P-HPTH, we evaluated the role antihypertensive therapy may play in this relationship. To do this, we restricted our analyses to women with a self-reported history of hypertension (n = 52 694). Of the 5 distinct antihypertensive medication classes examined (thiazides, furosemide, calcium channel blockers, β-blockers, and ACE inhibitors), only furosemide was found to be significantly associated with risk for P-HPTH (Table 3). The age-adjusted RR for P-HPTH was 1.79 (95% CI, 1.15–2.79), and this increased risk was only modestly attenuated after multivariable adjustment, including that for other antihypertensive medication classes (1.71; 95% CI, 1.08–1.71) (Table 3). Multivariable models that also included CHF and chronic kidney disease did not attenuate the RR; however, multivariable adjustment for self-reported osteoporosis and use of bisphosphonate medications resulted in an adjusted RR for P-HPTH of 1.64 (95% CI, 1.04–2.60). In a lag-time analyses in which participants had to have had exposure to furosemide for ≥2 years before the diagnosis of P-HPTH, the RR for P-HPTH was 1.64 (95% CI, 0.96–2.77); however, the number of P-HPTH cases in this analysis was only 16. We calculated the population-attributable fraction for furosemide use among those with a history of hypertension to be 3.8%.

Table 3.

Relative Risks For Incident Primary Hyperparathyroidism among Those with a History of Hypertension According to Classes of Antihypertensive Medication Use

	Thiazide Diuretic	Furosemide	Calcium Channel Blocker	β-Adrenergic Receptor Antagonist	ACE Inhibitor
	No. of incident P-HPTH cases	55	22	42	72	50
Person-years of follow-up	120 932	29 986	97 142	149 811	109 333
Age-adjusted RR (95% CI)	1.25 (0.91–1.72)	1.79 (1.15–2.79)	1.15 (0.82–1.62)	1.31 (0.99–1.74)	1.02 (0.74–1.41)
Multivariate-adjusted RR (95% CI)a	1.20 (0.88–1.66)	1.77 (1.12–2.78)	1.10 (0.78–1.55)	1.26 (0.95–1.67)	1.01 (0.73–1.39)
Multivariate + antihypertensive medication–adjusted RR (95% CI)b	1.21 (0.88–1.67)	1.71 (1.08–2.71)	1.10 (0.78–1.55)	1.26 (0.95–1.68)	1.05 (0.76–1.46)

^aThe multivariable model includes age, BMI categories (<22, 22–24.9, 25–29.9, or ≥30 kg/m²), race, smoking status (past, current, or never), menopausal status (premenopausal or postmenopausal), postmenopausal hormone use (yes or no), physical examination in prior 2 years, physical activity (quintiles), history of diabetes, dietary calcium intake, supplemental calcium intake, total vitamin D and vitamin A intake, alcohol intake, dietary magnesium intake, and dietary protein intake.

^bIncludes all variables in the multivariable model in addition to other individual antihypertensive medication classes: thiazide diuretics, furosemide, calcium channel blockers, β-adrenergic receptor antagonists, ACE inhibitors, and other antihypertensive classes.

History of hypertension and furosemide use

To evaluate whether the history of hypertension and furosemide use may have additive risks for development of P-HPTH, we categorized participants by hypertension status and furosemide use in the total population. We classified participants without hypertension and with no history of furosemide use as the referent category (101 cases of P-HPTH over 558 552 person-years of follow-up). When compared with this reference, participants who did have hypertension but were not treated with furosemide (194 cases of P-HPTH over 506 359 person-years of follow-up) had a significantly higher age-adjusted RR (1.88; 95% CI, 1.47–2.40) and multivariate-adjusted RR (1.57; 95% CI, 1.17–2.11) for P-HPTH. Participants who had a combination of both hypertension and furosemide use (22 cases of P-HPTH over 29 986 person-years of follow-up) had an even higher age-adjusted RR (3.39; 95% CI, 2.11–5.44) and multivariate-adjusted RR (2.79; 1.65–4.72) than with the referent group. The category of furosemide users without hypertension was too small to adequately compare (2 cases of P-HPTH over 6116 person-years of follow-up).

Discussion

The prevalence of P-HPTH has nearly tripled in the last 20 years, rising from 76 to 233 per 100 000 in women and from 30 to 85 per 100 000 in men (1). More important are the demographics of those affected by P-HPTH: women are affected 2 to 3 times more than men, and P-HPTH is most common among women older than 50 years of age (1, 5). Because P-HPTH significantly increases the risk for osteoporosis and related fragility fracture, older women who are perimenopausal or postmenopausal are at the highest risk for P-HPTH–related sequelae (1, 5). Given these evolving insights into P-HPTH demographic and prevalence patterns, a better understanding of risk factors that may contribute to P-HPTH might have substantial public health relevance. In this large longitudinal prospective study of older women, we found that a diagnosis of hypertension was associated with a 45% higher risk of developing incident P-HPTH than that of participants without a diagnosis of hypertension and that women with hypertension who used the loop diuretic furosemide had a 71% higher risk of developing P-HPTH than those who did not use furosemide. These findings may shed light on potential mechanisms that may cause P-HPTH and indicate potentially modifiable factors to prevent P-HPTH.

A link between excess parathyroid hormone and hypertension or cardiovascular disease has long been described; however, the direction of this association and mechanisms accounting for the observation remain incompletely understood (15, 18, 19, 40–46). Our findings build upon and extend these prior studies by confirming a prospective link between history of hypertension and subsequent development of P-HPTH. Women with hypertension in our cohort and no known history of P-HPTH had a 45% increased risk of developing incident P-HPTH compared with that of women without a history of hypertension. This observation was independent of many potential confounders and the use of specific antihypertensive medication classes. Our analyses suggest that the population-attributable fraction for hypertension (the fraction of P-HPTH cases that could be reduced by eliminating exposure to hypertension assuming that a history of hypertension is a causal and independent predictor of P-HPTH) is 15.6%. How hypertension may contribute to P-HPTH could not be assessed by our study design; however, prior studies have suggested that RAAS activity and abnormal calcium homeostasis may represent potential links between the pathophysiology of hypertension and P-HPTH (10, 14, 29, 47, 48). Individuals with hypertension have higher RAAS activity than normotensive individuals (49); because excess RAAS activity has been shown to stimulate PTH secretion (29, 30, 47, 48), this may represent one potential mechanism by which the history of hypertension may increase the risk of development of P-HPTH. We also previously demonstrated that higher intakes of dietary and supplemental calcium were associated with a lower risk for incident P-HPTH, suggesting that a higher calcium intake to reduce parathyroid gland stimulation may suppress adenomatous or hyperplastic transformation (10). In this regard, medications that modulate calcium homeostasis or RAAS activity could be hypothesized to alter the risk of development of P-HPTH.

For example, thiazide diuretics could theoretically reduce the risk of P-HPTH because they decrease renal calcium excretion and raise serum calcium, whereas loop diuretics may be hypothesized to increase the risk of P-HPTH because they increase renal calcium excretion (25). Recently, 2 studies independently demonstrated an association between the use of furosemide and higher PTH levels in individuals without kidney disease (26, 32). Further, furosemide use is associated with a higher risk for development of fracture (50), whereas the use of thiazides is associated with increased BMD (51). In this regard, our findings demonstrate that participants with hypertension who used furosemide had a 71% higher risk of developing P-HPTH than participants with hypertension but treated with other antihypertensive medications. Further, the combination of hypertension and furosemide use resulted in a 179% higher risk of incident P-HPTH than that of women without hypertension not using furosemide, suggesting an independent and potentially additive risk profile when these 2 risk factors were concomitantly present. It is important to note that over many years of antihypertensive medication practice guidelines, furosemide is less frequently used in the front lines of contemporary hypertensive therapy and more commonly used in resistant hypertension and conditions of volume excess such as CHF and chronic kidney disease (52). Nevertheless, the fact that furosemide use predicts increased risk for P-HPTH sheds important insights on potential disease pathophysiology. Whether the increased calciuria induced by furosemide represents a chronic stimulus to the parathyroid gland and may result in subsequent adenomatous or hyperplastic transformation and to the development of P-HPTH warrants dedicated investigation.

Furosemide also increases urinary excretion of sodium, thereby inducing a state of intravascular volume contraction. This hypovolemia can physiologically stimulate the RAAS and increased activity of both angiotensin II and aldosterone has been shown to increase PTH, potentially via their receptors expressed in parathyroid tissue (27–30, 47, 48). One recent large observational study showed that the use of RAAS inhibitors (ACE inhibitors and angiotensin receptor blockers) was strongly associated with lower PTH levels in hypertensive participants (14). In contrast, we observed no association between ACE inhibitor use and incident P-HPTH in the current study; however, our assessment of RAAS inhibitors in NHS was limited to ACE inhibitors (and not angiotensin receptor blockers or mineralocorticoid receptor antagonists), thus preventing a complete examination of RAAS inhibitor use as a risk factor for P-HPTH. It should also be noted that we observed no association between thiazide use or calcium channel blocker use and incident P-HPTH. Thiazides produce a natriuresis and volume contraction that could also stimulate the RAAS, yet their chronic use has not been associated with higher PTH levels (53), possibly because the hypocalciuric effect of thiazides counteracts the natriuretic and RAAS effects. Calcium channel blockers have been associated with higher PTH levels in cross-sectional studies (14); however, this effect was associated almost entirely with the use of the dihydropyridine category and not the nondihydropyridine category (14, 32), a distinction we could not assess in our study.

Our study has limitations. First, our study population was female and almost entirely white, and, therefore, our findings are not necessarily generalizable to men or other races; however, recent large cohort studies indicate that the demographic of our study population represents most at-risk individuals for P-HPTH (1). Second, it is possible that we missed cases of P-HPTH because our study design did not include biochemical screening of calcium and parathyroid hormone in all study participants. This may represent one reason why our overall incidence rate of P-HPTH was lower than that recently reported by Yeh et al (1), who detected cases using biochemical criteria from a large pool of participants who all had laboratory testing for calcium and PTH. In this regard, it should be noted that many cases of P-HPTH may be asymptomatic and detected by routine blood analysis, which may be more frequent in women who pursue regular preventive care or in individuals diagnosed with hypertension or starting furosemide treatment. However, we adjusted our analyses for regular physical examinations, and we conducted lag analyses implementing a minimum of 2 years between self-reported hypertension or furosemide use and the diagnosis of incident P-HPTH. Third, the 95% CIs for P-HPTH risk estimates associated with thiazides, calcium channel blockers, and ACE inhibitors do not exclude clinically and scientifically meaningful associations; future investigations that include assessment of other RAAS inhibitors and consideration of the different subclasses of calcium channel blockers may provide further insights. Fourth, we did not have direct measures of serum creatinine or glomerular filtration rate; however, we did have questionnaire-based assessment of the development of chronic kidney disease, which we included in our models, and we evaluated serum creatinine when conducting medical record reviews to exclude cases of secondary hyperparathyroidism. Finally, we did not have measurements of vitamin D levels, but we did adjust for dietary and supplemental vitamin D intake, as well as for factors associated with plasma 25-hydroxyvitamin D such as age, race, BMI, physical activity, and alcohol intake (54). Vitamin D is an important aspect to any study evaluating incident P-HPTH because prior studies have suggested that vitamin D deficiency may stimulate parathyroid adenoma growth (55, 56).

In conclusion, in a large prospective cohort study of older predominantly Caucasian women, we observed that a history of hypertension and the use of furosemide were independently associated with a higher risk of developing P-HPTH. Beyond identifying these potential risk factors of P-HPTH, our findings in conjunction with recent studies (10, 14, 26, 27) suggest that treatment of hypertension or modulation of urinary calcium or sodium handling may significantly influence the risk of developing P-HPTH. Reproduction and validation of our findings in other large prospective cohorts or in intervention studies are needed to determine whether hypertension and furosemide represent modifiable risk factors of incident P-HPTH.