Joint National Committee 7 (JNC-7) defines resistant hypertension as a persistent elevation of blood pressure (BP) above goal - ≥ 140/90 mm Hg for the general hypertensive population or ≥ 130/80 mm Hg for persons with diabetes mellitus or chronic kidney disease - for at least three months despite treatment with three or more optimally dosed antihypertensive agents, including a diuretic. The exact prevalence of resistant hypertension is uncertain but may include 5-20% of hypertensive persons in primary care settings and 15-35% of the older, higher cardiovascular risk hypertensive patients incorporated into recent clinical trials of antihypertensive therapy. Observational studies demonstrate that patients with resistant hypertension experience a higher rate of cardiovascular and renal target organ damage such as left ventricular hypertrophy, microalbuminuria, and renal insufficiency and more cardiovascular disease (CVD) events than patients whose hypertension is well-controlled. Additionally, resistant hypertension patients may be subjected to the considerable expense of multiple office visits, diagnostic testing for secondary causes of hypertension, and referral to hypertension specialists. Because multiple factors can contribute to resistant hypertension, an explicit, sequential approach to evaluation and management is essential to optimize blood pressure, reduce cardiorenal morbidity and mortality, and avoid unnecessary expense. A number of observational studies have suggested the potential efficacy of both spironolactone and amiloride when added to a 3 drug antihypertensive regimen, but to date no randomized study has directly compared the two agents. The goal of this study is to determine whether spironolactone or amiloride is the more effective fourth agent to add to a three drug regimen in patients with resistant hypertension.