Hypertension

Many advances in the field of hypertension in 1997 focused on basic issues. Dietary measures were found to be effective for patients with mild hypertension. Thiazide diuretics were found to be safe and effective in diabetic and nondiabetic persons. Finally, a study of ambulatory blood pressure monitoring may change the way we study and measure blood pressure.

Diet Was Effective in the Treatment of Mild Hypertension

Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med. 1997; 336:1117-24.

Therapy for hypertension often centers on pharmacologic treatment, but attempts to decrease body weight, decrease intake of sodium and alcohol, and increase consumption of potassium and calcium are now receiving more attention [1]. The actual effectiveness of these dietary measures has not been known.

The DASH (Dietary Approaches to Stop Hypertension) trial tested whether some dietary measures altered blood pressure. All participants (459 healthy adults with mild hypertension) received a diet low in fruit, vegetables, and calcium for 3 weeks. They were then randomly assigned to receive one of three diets for 8 weeks: a controlled diet, which included one serving of fruit or vegetables and one serving of calcium; a diet high in fruits and vegetables, which contained 5 servings of fruit or vegetables; and a combination diet, which was high in both fruits and vegetables and contained some calcium in the form of low-fat milk products. Sodium intake was kept at 3 g for all volunteers, and body weights were kept constant.

At baseline, the mean blood pressure was 131/85 mm Hg. In participants receiving the combination diet, the decrease in systolic blood pressure was 5.5 mm Hg greater and the decrease in diastolic blood pressure was 3.0 mm Hg greater than the decreases in participants receiving the controlled diet. In participants receiving the high-fruit, high-vegetable diet, the decrease in systolic blood pressure was 2.8 mm Hg greater and the decrease in diastolic blood pressure was 1.1 mm Hg greater than the decreases in participants receiving the controlled diet. The diets had the greatest effect among those with the highest blood pressures at baseline.

It seems that a diet rich in fruits, vegetables, and calcium and low in saturated fat can reduce blood pressure in patients with mild hypertension. These findings support the results of previous studies that show beneficial effects of potassium and calcium intake [2, 3]. The diets used by Appel and colleagues were easier to comply with than many previously recommended diet plans. For patients with mild hypertension, dietary therapy such as that described here should be a major component of treatment.

Thiazides Were Effective and Safe in Older Diabetic Patients

Curb JD, Pressel SL, Cutler JA, et al. Effect of diuretic-based antihypertensive treatment on cardiovascular disease risk in older diabetic patients with isolated systolic hypertension. Systolic Hypertension in the Elderly Program Cooperative Research Group. JAMA. 1996; 276:1886-92.

Isolated systolic hypertension in the elderly is now routinely treated, but controversy remains about whether the use of diuretic therapy in older persons is safe. Clinicians worry about hypokalemia and adverse effects on low-density lipoprotein cholesterol levels. In addition, a previous study of patients with diabetes [4] suggested that persons who received diuretics had excess mortality.

Curb and colleagues sought to end the controversy with a clinical trial to assess the effect of low-dose, diuretic-based antihypertensive treatment on major cardiovascular events in older, non-insulin-treated diabetic patients with isolated systolic hypertension. They enrolled 4736 persons 60 years of age or older, 583 of whom had type 2 diabetes mellitus. All had isolated systolic hypertension, which was defined as a systolic blood pressure of 160 mm Hg or more and a diastolic blood pressure less than 90 mm Hg. Patients were randomly assigned to receive either placebo or 12.5 to 25 mg of chlorthalidone per day. The blood pressure of patients receiving placebo was then managed by their physicians. The treatment group received, if necessary, atenolol or reserpine in addition to the thiazide diuretic. The outcome measures were 5-year rates of all major cardiac and cerebrovascular events, fatal cardiac events, and all-cause mortality.

Blood pressure decreased for both diabetic and nondiabetic patients in the treatment group. The 5-year rate of major cardiovascular events was reduced by 34% in both diabetic and nondiabetic patients receiving chlorthalidone (number needed to treat to prevent one adverse outcome, 10 for diabetic patients and 20 for nondiabetic patients). The greatest effect seemed to be on coronary artery events. The risk reductions for various outcomes are shown in Table 1.

This study shows that therapy with a thiazide diuretic-a simple regimen-is effective and well tolerated in older diabetic and nondiabetic patients with isolated systolic hypertension.

Ambulatory Monitoring Resulted in Use of Fewer Drugs

Staessen JA, Byttebier G, Buntinx F, et al. Antihypertensive treatment based on conventional or ambulatory blood pressure measurement. A randomized controlled trial. Ambulatory Blood Pressure Monitoring and Treatment of Hypertension Investigators. JAMA. 1997; 278:1065-72.

In theory, ambulatory blood pressure measurement should be superior to the conventional measurements of blood pressure. It is reproducible because it avoids observer bias. It is also done during the day and night, mitigating some of the issues about diurnal variation and "white-coat hypertension." But does it actually improve patient care? One difficulty involved in answering that question is that almost all clinical outcomes data about control of hypertension are based on standard monitoring in the office. We must ask whether ambulatory blood pressure monitoring has a role in patient care.

These Belgian researchers compared conventional and ambulatory blood pressure measurement to determine whether the latter affects the treatment and management of hypertensive patients. They recruited 419 healthy adults whose untreated diastolic blood pressure, by conventional measurement, averaged 95 mm Hg or more. Patients were randomly assigned to receive conventional or ambulatory blood pressure monitoring. Antihypertensive drug treatment was adjusted in a stepwise fashion on the basis of either the average daytime ambulatory diastolic blood pressure or the average of three diastolic blood pressure readings taken while patients were seated. Blinded physicians based therapy on whether diastolic blood pressure was greater than (>89 mm Hg), equal to (80 to 89 mm Hg), or less than (<80 mm Hg) the target pressure.

Compared with patients assigned to conventional blood pressure measurement, more of those assigned to the ambulatory method discontinued antihypertensive drug treatment (26.3% compared with 7.3%) and fewer progressed to sustained multidrug treatment (27.2% compared with 42.7%). Blood pressure control, left ventricular mass, and reported symptoms were similarly improved. Ambulatory monitoring saved money because it resulted in less intensive drug treatment and fewer physician visits, but these savings were offset by the cost of the equipment.

Much can be inferred from this study. First, the findings on cost may not be generalizable to the United States. Ambulatory blood pressure evaluations in Europe cost about $30; in the United States, they cost up to five times that amount. In addition, the drugs used were lisinopril, hydrochlorothiazide, and amlodipine-an expensive regimen. Second, the study raises the possibility of increased use of home monitoring of blood pressure, which is relatively inexpensive. Patients may be able to monitor blood pressure at home and then use those readings to guide therapy [5]. If the results of home monitoring mirror those of ambulatory assessment, fewer patients may need pharmacologic therapy. Third, the investigators excluded patients with serious comorbid disease, so it is unclear whether these findings can be generalized to patients with renal failure or heart failure. Finally, we expect that over the next 5 years, many studies will investigate clinical outcomes on the basis of either home monitoring or more formal ambulatory monitoring, and this study may be a turning point in our approach to the diagnosis and monitoring of hypertension.

Of course, whether treatment based on ambulatory blood pressure readings prevents adverse clinical outcomes remains unknown.

Progression of Renal Disease

In primary care, many patients have low-grade renal insufficiency. What can be done to prevent advancement to end-stage renal disease? The use of angiotensin-converting enzyme (ACE) inhibitors has become standard in diabetic patients, but much less is known about prevention of progression of disease in patients with other causes of renal disease.

Angiotensin-Converting Enzyme Inhibitors Seemed To Slow Disease Progression

Giatras I, Lau J, Levey AS. Effect of angiotensin-converting enzyme inhibitors on the progression of nondiabetic renal disease: a meta-analysis of randomized trials. Angiotensin-Converting-Enzyme Inhibition and Progressive Renal Disease Study Group. Ann Intern Med. 1997; 127:337-45.

The effect of ACE inhibitors in slowing decline in renal function in patients with nondiabetic renal disease varies among studies, and the benefit of this therapy remains uncertain. The problem is exacerbated by the fact that treatment studies of specific renal diseases have tended to include small numbers of patients and to study a broad range of diseases, including IgA nephropathy and polycystic kidney disease. The largest single clinical trial to date [6] investigated an array of renal diseases and compared ACE inhibitors with placebo, and it showed an improvement in serum creatinine levels among patients taking ACE inhibitors. However, clinical outcomes, such as development of end-stage renal disease or death, were not reported.

This meta-analysis assessed the clinical effect of ACE inhibitors on the development of end-stage renal disease caused by factors other than diabetes. Giatras and colleagues included published studies, identified through a MEDLINE search of the English-language literature, and unpublished studies. All randomized clinical trials that compared ACE inhibitors with other antihypertensive agents and had at least 1 year of planned follow-up were selected. Studies of diabetic renal disease and renal transplantation were excluded. The investigators identified 1594 patients in 10 studies; the end points were progression to end-stage renal disease and death.

Among 806 patients receiving ACE inhibitors, 52 (6.4%) developed end-stage renal disease and 17 (2.1%) died. Among the 788 controls, 72 (9.1%) developed end-stage renal disease and 12 (1.5%) died. The pooled relative risks were 0.70 (95% CI, 0.51 to 0.97) for end-stage renal disease and 1.24 (CI, 0.55 to 2.83) for death. The decreases in the mean systolic and diastolic blood pressures were 4.9 and 1.2 mm Hg greater, respectively, in the patients who received ACE inhibitors than in controls.

This study provides strong evidence that ACE inhibitors are more effective than other antihypertensive agents in delaying the onset of end-stage nondiabetic renal disease without increasing mortality. There was no evidence that ACE inhibitors stopped progression to end-stage renal disease, but these agents do seem to delay the need for dialysis. This study could not determine whether this beneficial effect was due to the slightly greater decrease in blood pressure or to other effects of ACE inhibitors.

Clinicians often question whether therapy with ACE inhibitors can be continued safely in patients with more severe renal disease (serum creatinine levels > 3.0 mg/dL [265.2 µmol/L]). Early studies suggested that these drugs should not be used in such patients, but the more optimistic studies of the past 4 years have called the earlier evidence into question. Currently, we do not know whether there is a threshold of renal failure in which ACE inhibitors are contraindicated.

Stents Were Safe in Ostial Renal Artery Stenosis

Blum U, Krumme B, Flugel P, et al. Treatment of ostial renal-artery stenoses with vascular endoprostheses after unsuccessful balloon angioplasty. N Engl J Med. 1997; 336:459-65.

Renal artery stenosis is the most common cause of secondary hypertension, and its prevalence is estimated to be about 1% among persons with hypertension [7]. In the past, the usual treatment was vascular bypass surgery, which carried great risk. Then it was found that renal arteries could be dilated by angioplasty. The best results were seen with patients whose stenosis was distal to the ostium of the aortorenal artery. When the stenosis was at the ostium, the success rate of angioplasty was less than 35%. Ostial atherosclerosis accounts for about half of all renal artery stenosis, so the development of alternative treatments has remained a goal.

Blum and colleagues used angioplasty along with stainless steel stents to test whether they could maintain patency of renal arteries with stenosis at the ostium. They placed 74 stents in 68 patients with ostial lesions. The indications for stent placement were elastic recoil of the vessel after angioplasty (63 arteries), dissection of the vessel after angioplasty (1 artery), and restenosis after initially successful angioplasty (10 arteries). Twenty patients had mild or severe renal dysfunction. Patients were followed for a mean of 27 months with measurements of blood pressure and serum creatinine levels, duplex sonography, and intra-arterial angiography.

The initial technical success rate was 100%. No major complications occurred, and minor complications (local hematomas) were seen in only three patients. After 5 years, 84% of patients were free of primary occlusion. Restenosis of more than 50% of the vessel diameter occurred in 11% of patients. Reintervention resulted in a secondary patency rate of 92%. Long-term normalization of blood pressure was achieved in 16% of patients, but serum creatinine levels did not change significantly after successful stent implantation in patients with previously impaired renal function.

Stent placement now seems a safe alternative to surgical bypass or angioplasty alone in patients with ostial renal artery stenosis. Hypertension was alleviated in about one sixth of patients; this proportion is about the same as that seen in previous studies of angioplasty. However, the fact that serum creatinine levels did not change may indicate that the procedure helps delay progression of renal failure.

Stents Stabilized Renal Disease

Harden PN, MacLeod MJ, Rodger RS, et al. Effect of renal-artery stenting on progression of renovascular renal failure. Lancet. 1997; 349:1133-6.

If left untreated, atherosclerotic renal artery stenosis can eventually progress to end-stage renal disease. These researchers tested whether stenting of atherosclerotic arteries would delay or even prevent such progression.

Renal function, as assessed by the reciprocal of the serum creatinine level, was assessed before and after stent placement in 32 patients with atherosclerotic renal artery stenosis.

Immediate patency was obtained in the 33 stents placed in the 32 patients. One patient died after a procedure-related hemorrhage. At 6 months, the restenosis rate was 12%. Blood pressure decreased slightly among the 32 patients, but the need for antihypertensive drugs was unchanged. Renal function improved or stabilized in 69% of patients, and progression of renal failure was slowed significantly after the procedure. Unfortunately, mean survival was only 22 months, primarily because renal artery atherosclerosis represented only a fraction of the vascular disease in these patients. Most patients died of myocardial infarction and stroke.

This study, like that by Blum and colleagues, shows the feasibility of stent placement and shows that progression of renal disease can be slowed. However, overall prognosis for the patients studied remains poor because of the burden of other vascular disease. The next challenge is to identify patients likely to obtain long-term benefit from renal stents.

Microalbuminuria

A 24-hour urine test for microalbuminuria (defined as an albumin concentration > 30 mg/g of creatinine or as excretion > 20 µg of albumin per minute) is a well-established test for early renal disease in diabetic patients. Microalbuminuria has also been associated with vascular disease in general [8, 9]. As a result, many physicians now use tests for microalbuminuria more often.

Urine Specific Gravity May Simplify Test

Moore RR Jr, Hirata-Dulas CA, Kasiske BL. Use of urine specific gravity to improve screening for albuminuria. Kidney Int. 1997; 52:240-3.

The measurement of albumin in a spot sample of urine is prone to great error unless one accounts for the urine concentration or the amount of other solute excreted. As a result, one of the most common ways to test for microalbuminuria is to measure both albumin and creatinine in a urine sample and then report a rate of albumin excretion based on the ratio of albumin to creatinine. However, such tests are inconvenient and relatively expensive (they can cost as much as $50) because they must be sent to a laboratory for creatinine measurement. In search of a simpler method, Moore and colleagues studied 50 patients to examine whether urine specific gravity could be used in place of the urinary creatinine level to correct albumin concentration for differences in urine volume.

Urine specific gravity accurately estimated urinary creatinine concentration (r = 0.79). The authors developed a regression Equation to estimate creatinine concentration on the basis of specific gravity. The ratio of albumin to estimated creatinine in random spot urine samples correlated with urinary albumin excretion measured in a 24-hour urine collection (r = 0.98), as did the actual albumin-to-creatinine ratio (r = 0.95). For determining microalbuminuria, using 24-hour albumin excretion as the gold standard, the sensitivity and specificity of the albumin-to-estimated-creatinine ratio (88% and 93%) were similar to those of the actual albumin-to-creatinine ratio (89% and 93%). A dipstick to assess albumin concentration in spot urine samples performed poorly, with a sensitivity of 63% and a specificity of 80%.

These results suggest that the use of urine specific gravity and the investigators' methods for estimating creatinine concentration might help produce a test for microalbuminuria that is less expensive and more convenient than the test for the albumin-to-creatinine ratio. However, current office dipstick products do not perform well enough to replace the laboratory test.

Spot-Sample Test Was Adequate for Diabetic Patients

Zelmanovitz T, Gross JL, Oliveira JR, et al. The receiver operating characteristics curve in the evaluation of a random urine specimen as a screening test for diabetic nephropathy, Diabetes Care. 1997; 20:516-9.

A second study on the measurement of microalbuminuria challenges the belief that variation in urine volume must be accounted for when microalbumin excretion is measured, at least in a population with a high prevalence of microalbuminuria, such as diabetic persons (prevalence, 30% to 40%).

Zelmanovitz and associates enrolled 95 ambulatory patients with type 2 diabetes mellitus. They obtained 123 random morning urine samples and a timed 24-hour urine collection, which was used as the standard for determining urinary albumin excretion rate. Samples were classified as normoalbuminuric (urinary albumin excretion rate < 20 µg/min; n = 54), microalbuminuric (urinary albumin excretion rate, 20 to 200 µg/min; n = 44), and macroalbuminuric (urinary albumin excretion rate > 200 µg/min; n = 25). The investigators then calculated a receiver-operating characteristics curve to demonstrate the test performance of random urine samples at various cut-off points for albuminuria (Figure 1).

View larger version (11K): [in this window] [in a new window]   Figure 1. Receiver-operating characteristics curve for microalbuminuria. Random urine sample compared with 24-hour urine collection. Adapted with permission from Zelmanovitz T, Gross JL, Oliveira JR, et al. The receiver operating characteristics curve in the evaluation of a random urine specimen as a screening test for diabetic nephropathy. Diabetes Care. 1997; 20:516-9.

At the accepted cut-off for microalbuminuria, 33.6 mg/L, the random sampling technique had a sensitivity of 88% and a specificity of 90%. Among diabetic persons, in whom the prevalence of microalbuminuria may be 35%, the random test would have a positive predictive value greater than 80%. But in the general population, where the prevalence may be 5% or less, the positive predictive value would be less than 30%.

Therefore, among diabetic patients, a random urine specimen not tied to a measure of urine concentration may be adequate to screen for microalbuminuria. If random urine testing proves to be adequate, this testing may reduce the cost of management of diabetic patients.

Simple Office Tests for Microalbuminuria Were Highly Sensitive

Pegoraro A, Singh A, Bakir AA, et al. Simplified screening for microalbuminuria. Ann Intern Med. 1997; 127:817-9.

Whereas Zelmanovitz and coworkers focused on tests that detect microalbuminuria, Pegoraro and associates focused on the tests needed to establish with relative certainty that microalbuminuria is not present.

In a cross-sectional study, these investigators collected 221 urine samples from primary care clinics and a diabetes clinic. Random early-morning urine specimens were tested for albumin with the Micral-Test (Boehringer Mannheim, Mannheim, Germany), which is based on the color shift of a monoclonal antibody. They were also tested for protein with sulfosalicylic acid and chemically impregnated dipsticks (Chemstrip, Boehringer Mannheim). Radioimmunoassay was used as the gold standard.

Table 2 shows the characteristics of the various tests. The overall characteristics came from the combined tests, which cost about $0.50 and take about 3 minutes to perform. In the diabetic persons studied, a negative result on the combined test had a negative predictive value of 94%.

These studies lead to five conclusions. First, as with all diagnostic tests, the clinical performance of tests for microalbuminuria depends on the population being tested. Second, a negative result on a simple test using a dipstick plus sulfosalicylic acid strongly suggests that microalbuminuria is not present. Third, a positive result on a test for albumin in a random urine sample, done in a patient who is very likely to have microalbuminuria, can confirm that suspicion. Fourth, for persons less likely to have microalbuminuria (such as persons in the general population), a correction for creatinine level, or perhaps specific gravity, may be needed. This may become more important if the association between microalbuminuria and vascular disease is proven. Finally, and most important, the question of whether the 24-hour urinary albumin excretion test is reproducible under all circumstances remains. Patient activity, blood glucose level, the presence of other renal disease, and the use of medications are all caveats to the interpretation. Even in Denmark, where the concept of screening for microalbuminuria in patients with diabetes arose, patients have three values measured over 6 months. If two of the three values show increased albumin excretion, patients are assumed to have increased albumin excretion. A single positive test result does not lead to the label "microalbuminuric."

Electrolytes

Alterations in sodium and potassium homeostasis are common in hospital settings. Two studies have provided new clinical insight into the causes of post-operative hyponatremia and the hypokalemia that is often seen in patients in preterm labor. The role of dietary and supplemental calcium in the development of kidney stones has also become clearer.

Mechanism of Postoperative Hyponatremia Was Described

Steele A, Gowrishankar M, Abrahamson S, et al. Postoperative hyponatremia despite near-isotonic saline infusion: a phenomenon of desalination. Ann Intern Med. 1997; 126:20-5.

Postoperative hyponatremia is common, especially in women. Serum sodium concentrations of 120 mEq/L are not rare after surgery, even in otherwise healthy patients. The phenomenon is often attributed to intra-operative administration of free water in patients who have high endogenous levels of antidiuretic hormone (ADH), which occurs with stress from discomfort as anesthesia wears off. If ADH cannot be suppressed, urine cannot be diluted and less water is excreted.

Steele and colleagues noted severe postoperative hyponatremia that led to the deaths of five women who received very little free water. They then designed a prospective study to investigate the mechanism involved. They studied 22 randomly selected, healthy women who were undergoing uncomplicated gynecologic surgery. All women received isotonic solutions during and after surgery, and plasma electrolyte levels were measured at the time of anesthesia induction and 24 hours later.

Initial sodium levels averaged 140 mmol/L and decreased to a mean of 135.8 mmol/L. During the first 16 hours after induction, patients excreted hypertonic urine (peak concentration, 294 mmol/L).

The authors concluded that hyponatremia occurs because of the generation of electrolyte-free water during excretion of hypertonic urine-a kind of desalination process. The increased free water content of the body cannot be excreted because ADH is present.

This fairly simple study has important implications for the care of surgical patients. First, hypotonic fluids should not be given soon after surgery unless a patient is hypernatremic. Second, isotonic fluid should be given, and the amount should be based on the patient's hemodynamic state. Third, if patients excrete more than 2 L of hypertonic urine (specific gravity > 1.020) within 24 hours, plasma sodium levels should be checked to ensure that hyponatremia is detected before clinical complications arise.

ß-Agonists Caused Hypokalemia in Women in Preterm Labor

Braden GL, von Oeyen PT, Germain MJ, et al. Ritodrine- and terbutaline-induced hypokalemia in preterm labor: mechanisms and consequences. Kidney Int. 1997; 51:1867-75.

Ritodrine and terbutaline are ß-adrenergic agonists that are commonly used to inhibit uterine contractions in women in preterm labor. Hypokalemia has been noted to occur more frequently in these patients, and several cases of pulmonary edema have been reported. ß-adrenergic agonists are presumed to influence potassium metabolism in several ways: They directly stimulate sodium-potassium adenosine triphosphatase to promote cellular potassium uptake; they stimulate insulin release, which further promotes cellular potassium uptake; and, in the kidneys, they promote sodium and water retention and potassium excretion.

Braden and colleagues attempted to determine which mechanisms were most important clinically when ritodrine or terbutaline was infused. The prospective study consisted of three protocols. Participants were healthy volunteers in preterm labor who received, by random assignment, either ritodrine or terbutaline intravenously over 6 hours. One protocol studied cardiac arrhythmias, and 3 of 14 women who received the drugs had symptomatic arrhythmias, including multiple premature atrial and ventricular contractions. A second protocol, involving 83 women, found that hypokalemia was severe (mean plasma concentration, 2.5 mEq/L) and developed rapidly during infusion. In the third protocol, blood and urine samples were analyzed before and during infusion in 10 women. The investigators found that the major mechanisms of hypokalemia were increased cell uptake, probably primarily in the liver, and decreased renal excretion at the distal tubules. The effect of hyperinsulinemia was unimportant because potassium levels decreased before insulin levels increased. The two drugs produced similar effects, but effects were related to potency of dose.

This study was small and investigated many factors. However, the important conclusions for clinicians are that severe, symptomatic hypokalemia is common in women receiving ß-adrenergic agonists for preterm labor, that plasma potassium levels should be measured early in an infusion, and that the presence of arrhythmias should be recognized.

Dietary, but Not Supplemental, Calcium Had Beneficial Effects on Risk for Kidney Stones

Curhan GC, Willett WC, Speizer FE, et al. Comparison of dietary calcium with supplemental calcium and other nutrients as factors affecting the risk for kidney stones in women. Ann Intern Med. 1997; 126:497-504.

Dietary factors are believed to contribute to the development of kidney stones. However, in contrast to previous assumptions, these investigators previously found that a high dietary intake of calcium in men was associated with a smaller risk for kidney stones [10]. Because women are often counseled to increase dietary intake of calcium to prevent osteoporosis, these researchers studied the relation between calcium intake and kidney stones.

They used questionnaire and clinical data from 91 731 women participating in the 12-year Nurses' Health Study. After almost 1 million person-years of follow-up, 864 cases of kidney stones were found. After adjustment for other risk factors, dietary intake of calcium was found to be inversely associated with risk for kidney stones. However, women who took calcium as dietary supplements had a relative risk for kidney stones of 1.2 (CI, 1.02 to 1.41) compared with women with the lowest intake of supplemental calcium. In 67% of those who took supplements, the calcium was not taken with meals or was taken with meals that had low oxalate content.

It is now clear that physicians should not tell patients with kidney stones to restrict their dietary calcium intake to below the normal recommended levels. In addition, it is probably wise to counsel women who need calcium supplementation to take their calcium with meals.

Dr. Henrich: Department of Internal Medicine, Medical College of Ohio, 3000 Arlington Avenue, Toledo, OH 43614.

Dr. Roberts (Series Editor): Madrona Medical Group, 3199 Steller Court, Bellingham, WA 98226-7805.