High Prevalence of Liddle Syndrome Phenotype Among Hypertensive US Veterans in Northwest Louisiana

Abstract

J Clin Hypertens (Greenwich).

Liddle syndrome (LS) is an autosomal dominant disorder due to a gain‐of‐function mutation in the epithelial Na⁺ channel and is perceived to be a rare condition. A cross‐sectional study of 149 hypertensive patients with hypokalemia (<4 mmol/dL) or elevated serum bicarbonate (>25 mmol/dL) was conducted at a Veterans’ Administration Medical Center Hypertension Clinic in Shreveport, LA. Data on demographics, blood pressure, and select blood tests were collected and expressed as percentages for categoric variables and as mean ± standard deviation (SD) for continuous variables. Patients were diagnosed with likely LS when the plasma renin activity (PRA) was <0.35 μU/mL/h and the aldosterone was <15 ng/dL and likely primary hyperaldosteronism (PHA) with PRA <0.35 μU/mL/h and aldosterone level >15 ng/dL. The cohort included predominantly elderly (67.1±13.4 years), male (96%), and Caucasian (57%) patients. The average blood pressure was 143.8/79.8 mm Hg±27.11/15.20 with 3.03±1.63 antihypertensive drugs. Based on the above criteria, 9 patients (6%) satisfied the criteria for likely LS and 10 patients (6.7%) were diagnosed with likely PHA. In this hypothesis‐generating study, the authors detected an unusually high prevalence of biochemical abnormalities compatible with likely LS syndrome from Northwestern Louisiana, approaching that of likely PHA. J Clin Hypertens (Greenwich). 2010;12:856–860.

Liddle syndrome (LS; Online Mendelian Inheritance in Man #177200) was first described in 1963 as a hypertensive syndrome resembling primary hyperaldosteronism with hypertension, hypokalemia, and metabolic alkalosis but negligible aldosterone secretion. ¹ LS is classically viewed as a monogenic disorder that clusters in families and is a rare contributor to the population‐wide burden of hypertension in the world. ²

LS has a heterogenous genetic background, ³ characterized by prolonged half‐life of the epithelial Na⁺ channel (ENaC). Phenotypic manifestations are attributed to prolonged half‐life of the ENaC channels, resulting in an increased channel density, rather than alterations of individual channel function. ⁴ Animal data from a mouse model also suggest that accelerated Na⁺ absorption through intestinal ENaC may contribute to the generation of hypertension. ⁵ The individual tendency for hypokalemia and hypertension is known to be variable within well‐described kindreds with LS. ⁶ At present, LS is largely viewed as a rare but fascinating experience of nature, which, while it may have facilitated the discovery of ENaC and apical Na transport in the renal cortical collecting duct, may not be relevant to daily clinical practice. There is, however, a relative paucity of data describing the prevalence of LS in well‐characterized cohorts. Increasingly, random aldosterone/renin ratios are being used to characterize hypertension or screen for primary hyperaldosteronism in referral practices. An elevated random aldosterone/renin ratio of >30 (when expressed in ng/dL for aldosterone and ng/mL per hour for renin) is an acceptable screening test for increased secretion of aldosterone. ⁷ Therefore, it is expected that referral practices for hypertension are currently obtaining a large number of paired renin and aldosterone measurements, looking for such elevated ratios. Some data suggest that in African Americans, LS, or a mild variant thereof, may more commonly occur and contribute to the pathogenesis of hypertension. “Low renin” hypertension is certainly more common among African Americans, ⁸ although this may reflect subtle renal impairment in these cohorts. In 1995, Gadallah and coworkers⁹ reported 3 cases from Northwestern Louisiana in African Americans. Most of the previous reports were either in Caucasians or Asians. ⁹

The Veterans’ Administration Medical Center (VAMC) hypertension clinic in Shreveport, LA, follows a large number of patients from the North‐West Louisiana, South‐West Arkansas, and North‐East Texas (ArkLaTex) area. During our clinical practice, we became aware of 2 patients with features similar to those described by Gadallah and colleagues who received the tentative diagnosis of LS and subsequently had an excellent clinical response to amiloride. This observation prompted us to review the records of 149 consecutive recently seen hypertensive patients with either hypokalemia or metabolic alkalosis from the VAMC Hypertension Clinic, where paired measurements of aldosterone levels and plasma renin activity were available. Our aim was to investigate the prevalence of a phenotype compatible with LS in the ArkLaTex area.

Methods

The study was reviewed and approved by the institutional review boards of the Overton Brooks Veterans’ Administration Medical Center as well as the Louisiana State University Health Sciences Center in Shreveport, LA, before initiation of our investigation. We undertook a cross‐sectional review of the medical records of all patients who attended the Shreveport VAMC Hypertension Clinic and identified patients with hypertension, potassium <4 mmol/dL, and/or bicarbonate ≥25 mmol/dL. Inclusion criteria were the availability of demographic data, an electrolyte panel, plasma renin activity (PRA), aldosterone level, and blood pressure (BP) measurement. Patients with end‐stage renal disease were excluded. Our initial chart review identified 149 hypertensive patients with qualifying biochemical parameters. Serum bicarbonate and potassium were analyzed on Beckman LXi or DxC600 (Beckman Coulter, Inc., Brea, CA) using ion‐specific electrodes. BNP (normal: <100 pg/mL) was analyzed on the Beckman Access 2 by chemiluminescent immunoassay. Paired aldosterone level and plasma renin activity was measured on random specimens obtained in the clinic with the rest of the routine laboratory tests. To simplify clinical care, antihypertensive therapy before testing was not discontinued. ¹⁰ Renin activity (ng/mL per hour) was performed manually and completed with LabCorp by radioimmunoassay (RIA) methodology (reference ranges for adults with normal salt intake in upright position: 1.31–3.95; in supine position: 0.15–2.33; standardized for salt excretion >150 mEq/24 hours: 0.39–1.31). Aldosterone (ng/dL) measurements were performed on the T‐Can at LabCorp by RIA methodology (reference ranges for adults in the supine position: 1.0–16.0; upright position: 4.0–31). Select patients underwent additional work‐up for adrenal imaging and evaluation of the renal vasculature. Patients were diagnosed with likely LS in the presence of hypertension, electrolyte abnormalities (hypokalemia or metabolic alkalosis), PRA <0.35 μU/mL, and serum aldosterone <15 ng/dL. The criteria for likely primary hyperaldosteronism (PHA) were based on PRA <0.35 μU/mL and aldosterone levels >15 ng/dL with abnormal electrolytes. Data were expressed as percentages for categoric variables and mean and standard deviations (SDs) for continuous variables. Between‐group comparison was performed with analysis of variance and χ² as appropriate. Statistical significance was defined as a 2‐tailed P value <.05. Statistical analysis was performed using GraphPad Prism 5.0 (San Diego, CA).

Results

The overall cohort description is shown in Table I. This was a predominantly elderly and male population, with approximately 40% African American participants. BPs were under fair control (143.8/79.8 mm Hg±27.11/15.20 mm Hg) although at the price of heavy antihypertensive drugs regimen (3.03±1.63 drugs); 75% of patients were taking angiotensin‐converting enzyme (ACE) inhibitors. At the time of data collection, 2.7% of patients were taking no antihypertensive drugs at all and 16.7% took only 1 drug. For the rest of the cohort, approximately 20% each were taking 2, 3, and 4 antihypertensive drugs, while 10% were taking 5 drugs. Based on the aforementioned criteria, 9 patients (6%) satisfied the criteria for likely LS and 10 of the patients (6.7%) were diagnosed with likely PHA. With two exceptions, all of these patients were taking ACE inhibitors but none of them were taking spironolactone. The subcohort with possible LS was similar to the overall cohort: aged 61.1±11.6 years, overweight with body mass index 27.2±7.23, and taking a large number of antihypertensive medications (3±1.80).

Table I.

 Overall Cohort Demographics (N=149)

variables	mean± SD or percentage
Age, y	67.15±13.4
Male, %	96
African American, %	42.7
Weight, kg	92.4±23.14
BMI, kg/m2	29.6±6.73
BNP, pg/mL	348.2±483.4
SBP, mm Hg	143.8±27.11
DBP, mm Hg	79.8±15.20
Potassium, m/L	4.0±0.67
Bicarbonate, m/L	27.1±3.97
Serum aldosterone, ng/dL	8.5±10.26
Renin activity, ng/mL/h	3.36±7.31
Aldosterone/renin ratio	20.1±52.41
Creatinine, mg/dL	1.75±1.09
Antihypertensive drugs, No.	3.03±1.63
Comorbid illness, %
Diabetes mellitus	48.3
Obstructive sleep apnea	14.7
Medication class, %
ACE inhibitors	75
Diuretics	43.9
β‐Blockers	52
Spironolactone	10.9

Data are expressed as percentages for categoric variables and mean ± standard deviation (SD) for continuous variables. Abbreviations: ACE, angiotensin‐converting enzyme; BMI, body mass index; BNP, brain‐type natriuretic peptide; DBP, diastolic blood pressure; SBP, systolic blood pressure.

Illustrative comparison between the groups with likely LS, PHA, and the residual cohort is shown on Table II. The subpopulation with likely LS had significantly lower potassium, PRA, and aldosterone levels (P<.0001) than the two other groups. Both the LS and the PHA groups had abnormal serum creatinine values, with significant between‐group differences (P<.0001). The BP was nominally higher in the LS group but without a statistically significant difference across the groups. Diuretics were used in 33.3% of LS, 60% of PHA, and 40% of residual cohort patients (not significant).

Table II.

 Subcohort Comparisons

	LS (n=9)	PHA (n=10)	residual cohort (n=130)	ANOVA P value
PRA, μU/mL/h	0.16±0.11	0.53±0.73	3.82±7.75	<.0001
Aldosterone, ng/dL	3.6±1.4	34.6±18.5	6.9±6.4	<.0001
Creatinine, mg/dL	2.0±2.0	2.7±2.1	1.7±0.9	<.0001
Potassium, mEq/L	3.4±0.4	3.8±0.5	4.3±1.9	<.0001
Bicarbonate, mEq/L	27.1±3.4	28.4±2.5	27.1±4.6	.047
SBP, mm Hg	175±36	143±15	142±26	>.1
DBP, mm Hg	93±21	80±8	79±15	>.1
BNP, pg/mL	445±515	303±475	344±487	.014
Diuretic use, %	33	60	40	.99 (Chi‐square)

Data are expressed as percentages for categoric variables and mean ± standard deviation (SD) for continuous variables. Abbreviations: ANOVA, analysis of variance; BNP, brain‐type natriuretic peptide; DBP, diastolic blood pressure; LS, Liddle syndrome; PHA, primary hyperaldosteronism; PRA, plasma renin activity; SBP, systolic blood pressure.

Discussion

In a group of elderly US veterans referred for hypertension management in a hypertension clinic, we documented a high (6%) prevalence of biochemical abnormalities compatible with LS, on par with the diagnosis of likely PHA in this cohort (6.7%). The relatively advanced age of this cohort with LS phenotype was a surprising finding and difficult to explain at first look. Most reported LS cases are clustered in families and expected to be diagnosed among relatively young individuals. ¹¹ The relatively old age of participants makes a simple genetic disease somewhat less likely. It is, therefore, unclear what process(es) replicated the phenotype of LS in this cohort. Our patients had been enrollees of the US armed forces largely from the Vietnam era; significant hypertension in their late teens and early 20s would likely have disqualified them from serving in the armed forces. Partly to serve as an internal control for our study, we calculated the prevalence of likely PHA, which we found to be similar to published rates in the hypertensive population (5%–13%). ¹² BP control or the number of antihypertensive drugs was not significantly different in this LS subcohort from the rest of the patients. The slightly elevated B‐type natriuretic peptide level may reflect the volume‐related BP elevation that results from the gain of function mutation of the sodium channels in the distal tubules and collecting ducts. Alternatively, it may reflect renal impairment with Na⁺ retention in this cohort. Definitive diagnosis of LS could be made through direct genetic testing, via functional investigations of the sodium channels of the red blood cells or B‐lymphocytes, or, recently described, with in vivo measurement of nasal transmucosal potential difference. ¹³ Such tests were not immediately available in this group of patients. In practical terms, however, clinical response to a trial of amiloride or triamterene with resolution of metabolic derangements and hypertension is an acceptable clinical test for LS. We felt that 6% was an unusually high prevalence, indicating either a founder effect in this region of Louisiana, a new acquired LS‐like condition (perhaps due to environmental exposure), or potential diagnostic inaccuracy and overdiagnosed LS in this cohort. Our patient with the LS phenotype had, however, not only suppressed paired random renin activity (0.16±0.11 μU/mL per hour) and aldosterone levels (3.6±1.4 ng/dL) but also additional supporting features: hypokalemia (3.4±0.4 mEq/L) and mild elevation of serum bicarbonate (27.1±3.4 mEq/L), despite impaired serum creatinine (2.0±2.0 mg/dL). The control of hypertension and metabolic abnormalities in patients with possible LS is of practical importance, as triamterene or amiloride are predicted to be effective, whereas the aldosterone antagonist spironolactone is not. ¹⁴

This cohort was overweight, which raises the possibility of the interaction of sodium and potassium handling with visceral obesity. Patients with LS are known to show an exaggerated response to aldosterone, demonstrating retention of normal regulatory activity. ¹⁵ Higher body mass indices were shown to be predictive of urinary aldosterone excretion in normotensive individuals on a high‐salt diet. ¹⁶ Our patients with the LS phenotype had, however, suppressed PRA and aldosterone levels suggesting additional unknown effect(s).

Limitations and Strengths

Our study has several inherent limitations, including being a retrospective investigation. We did not perform formal 24‐hour urine collections to document relative potassium wasting, dietary sodium intake, or suppressed aldosterone excretion. Genetic testing to formally confirm or exclude LS was not performed, either. It also has some relative merits. The advanced age of the cohort made the diagnosis of congenital adrenal hyperplasia unlikely. The clinical likelihood of LS was strengthened by multiple supportive pieces of evidence: difficult‐to‐control hypertension with the patient taking multiple antihypertensive drugs, simultaneous suppression of renin activity and aldosterone levels, and supporting metabolic abnormalities (hypokalemia and/or elevated serum bicarbonate). Exposure to ACE inhibitors or aldosterone antagonists was unlikely to weaken our conclusions: ACE inhibitors increase PRA and spironolactone increases both PRA and serum aldosterone level and, thus, exposure to these classes of agents would bias against the diagnosis of LS in our cohort.

Conclusions

Appropriate antihypertensive treatment should be tailored to the specific underlying etiology driving the elevated BP. In this hypothesis‐generating study we detected an unusually high prevalence of biochemical abnormalities compatible with LS or LS phenotype from Northwestern Louisiana, a frequency approaching that of PHA. Additional studies are needed to explore the genetic and environmental background of this observation. Hypertension specialists should consider the diagnosis of LS in patients with resistant hypertension and hypokalemia, when paired aldosterone and renin measurements are potentially supportive with simultaneous suppression of both. A clinical trial of ENaC blockade (amiloride or triamterene) may be appropriate in select patients to improve BP and simplify complex antihypertensive drug therapy.