Acknowledgement

This CME activity is supported by an educational grant from NicOx, Warren, New Jersey.

Disclosures

Dr. Friedewald has received honoraria for speaking from Novartis, East Hanover, New Jersey. Dr. Friedewald has received fees for consulting from NicOx, Warren, New Jersey; and AstraZeneca, Wilmington, Delaware. Dr. Ram has received fees for speaking and consulting from Genesis, Kennett Square, Pennsylvania; Cogenix, Broomfield, Colorado; MedKnowledge Group, Rocky Hill, Connecticut; and Medcon, San Francisco, California. Dr. Ram is a member of the advisory board of NiCox. Dr. Wesson has no relevant financial relationships to disclose. Dr. White has received research grant support from the National Institutes of Health , Bethesda, Maryland; and Novartis. Dr. White is a consultant for and a member of the cardiovascular safety and steering committees of Abbott Laboratories, Abbott Park, Illinois; Astellas Pharma US, Deerfield, Illinois; Biosante Pharmaceuticals, Lincolnshire, Illinois; NiCox; Nycomed, Zurich, Switzerland; Roche Diagnostics GmbH, Mannheim, Germany; and Takeda Global Research and Development, Deerfield, Illinois. Dr. White has received commissioned research support from Pfizer, New York, New York; and Teva Neurosciences, Kansas City, Missouri. Dr. Williams has clinical trial contracts with and is a principal investigator for Pfizer and Amgen, Albuquerque, New Mexico. Dr. Williams has received consulting fees and is a member of the data safety monitoring board of Pfizer. Dr. Williams has received consulting fees and is a member of the advisory boards of UCB, Brussels, Belgium; and Ortho-McNeil, Titusville, New Jersey. Dr. Williams has received honoraria for speaking from Genentech, South San Francisco, California; and Pfizer. Dr. Roberts has received honoraria for speaking from Merck, Whitehouse Station, New Jersey; Novartis; and AstraZeneca.

Disclosures

Dr. Friedewald has received honoraria for speaking from Novartis, East Hanover, New Jersey. Dr. Friedewald has received fees for consulting from NicOx, Warren, New Jersey; and AstraZeneca, Wilmington, Delaware. Dr. Ram has received fees for speaking and consulting from Genesis, Kennett Square, Pennsylvania; Cogenix, Broomfield, Colorado; MedKnowledge Group, Rocky Hill, Connecticut; and Medcon, San Francisco, California. Dr. Ram is a member of the advisory board of NiCox. Dr. Wesson has no relevant financial relationships to disclose. Dr. White has received research grant support from the National Institutes of Health , Bethesda, Maryland; and Novartis. Dr. White is a consultant for and a member of the cardiovascular safety and steering committees of Abbott Laboratories, Abbott Park, Illinois; Astellas Pharma US, Deerfield, Illinois; Biosante Pharmaceuticals, Lincolnshire, Illinois; NiCox; Nycomed, Zurich, Switzerland; Roche Diagnostics GmbH, Mannheim, Germany; and Takeda Global Research and Development, Deerfield, Illinois. Dr. White has received commissioned research support from Pfizer, New York, New York; and Teva Neurosciences, Kansas City, Missouri. Dr. Williams has clinical trial contracts with and is a principal investigator for Pfizer and Amgen, Albuquerque, New Mexico. Dr. Williams has received consulting fees and is a member of the data safety monitoring board of Pfizer. Dr. Williams has received consulting fees and is a member of the advisory boards of UCB, Brussels, Belgium; and Ortho-McNeil, Titusville, New Jersey. Dr. Williams has received honoraria for speaking from Genentech, South San Francisco, California; and Pfizer. Dr. Roberts has received honoraria for speaking from Merck, Whitehouse Station, New Jersey; Novartis; and AstraZeneca.

Objectives

After completion of this activity, physicians should:

• 1
  

  Identify patients at increased risk for nonsteroidal anti-inflammatory drug (NSAID)–induced blood pressure (BP) elevation.

  
  
• 2
  

  Recommend nonpharmacologic treatments for patients with osteoarthritis when NSAID-induced BP elevation occurs.

  
  
• 3
  

  Appropriately modify hypertensive treatment regimens when NSAID-induced BP destabilization occurs.

  
  
• 4
  

  Proactively inquire of possible NSAID use in all patients with hypertension.

Introduction

NSAIDs have long been associated with blood pressure (BP) destabilization. One of the most common uses of nonprescription and prescription NSAIDs is in the treatment of osteoarthritis, which is a common co-morbid condition with hypertension, especially in older patients. Thus, NSAIDs, which are regularly taken by millions of individuals—use that will likely increase in an aging and increasingly overweight population in the United States—are important to primary care physicians and cardiologists in the primary and secondary prevention of cardiovascular disease.

Dr. Friedewald: What is the current perspective of osteoarthritis by rheumatologists?

Dr. Williams: Osteoarthritis is common and increases in frequency and severity with age. The same process driving osteoarthritis in many of these patients, obesity, also increases with age. Up to 40% of patients with osteoarthritis also have hypertension in excess of its prevalence in the general population. Total serum cholesterol and diabetes mellitus also are more common in these patients. An underlying common factor is excess weight, which increases load across the joint, particularly the knee. Thus, obesity is an underlying cause of both osteoarthritis and hypertension, which increase with age.

Dr. Friedewald: The weight-bearing joints are mainly the hips, knees, and ankles, but in obese persons, the main joint that is affected is the knee?

Dr. Williams: That is correct. The association between obesity (body mass index >30 kg/m ² ) and osteoarthritis is greatest for the knee joint. Women are affected more than men. Osteoarthritis of the knee also progresses more rapidly in women.

Dr. Friedewald: What is the initial treatment for knee osteoarthritis?

Dr. Williams: There is general agreement that nonpharmacological interventions should be tried before medical treatment. Many factors contributing to knee osteoarthritis can be modified, particularly weight. Symptomatic osteoarthritis improves about 10% with weight loss, so patients are encouraged to lose weight and aim for ideal body weight to decrease pain and disease progression. Quadriceps strengthening also is very important for improving knee instability. Weight reduction is most important and can sometimes eliminate symptoms.

Dr. Friedewald: What is the optimal pharmacologic treatment for osteoarthritis?

Dr. Williams: There are no data supporting pharmacologic modification of the disease itself, only symptom relief. Pharmacologic interventions generally begin with drugs like acetaminophen and then progress to NSAIDs, both selective and nonselective agents, followed last by opioids. Joint replacement is sometimes necessary when these measures fail.

Dr. Friedewald: Is there any evidence of disease modification from the anti-inflammatory effects of drugs?

Dr. Williams: There is no good evidence that any of the agents currently used for long-term treatment modify the course of the disease. Tetracycline likely has some disease-modifying activity, but there is little evidence that NSAIDs or acetaminophen modify the course of the disease. Inflammation appears to play some role in osteoarthritis and its symptoms. Thus, control of inflammation has some advantages over the use of drugs for simple pain relief, so patients who have not done as well on narcotics may do much better on NSAIDs, even though the non-weight-bearing pain relief is similar. Joint use clearly improves in patients on anti-inflammatory drugs, due to both pain relief and reduction of inflammation.

Dr. Friedewald: Is acetaminophen an anti-inflammatory medication?

Dr. Williams: Acetaminophen is not an anti-inflammatory drug, although it probably inhibits cyclooxygenase to a small extent.

Dr. Friedewald: Do you prescribe drugs for osteoarthritis on an “as-needed” (PRN) basis, or to be taken on a regular schedule?

Dr. Williams: Increasingly, we are encouraging patients to use NSAIDs on a PRN basis. There is no good evidence that they require continuous use to be effective. Unlike disease-modifying drugs for rheumatoid arthritis, where long-term continuous use is important, it is probably best to match the use of drugs for osteoarthritis with patient level of activity and amount of pain. For example, many patients use these drugs only on weekends in relation to weekend physical recreation, which is encouraged in patients with osteoarthritis. Their toxicity in the gastrointestinal tract and their effect on BP is less at the lowest effective dose and usage possible.

Dr. Wesson: This is an excellent approach to osteoarthritis. Nephrologists see only those patients with renal disease who take NSAIDs. The patients we see are older with decreased glomerular filtration rates (GFRs), and they also often have cardiovascular disease, diabetes mellitus, and many other problems that are made worse by NSAIDs. Consequently, we try to keep patients with renal disease off NSAIDs when possible.

Dr. Ram: I have been associated with a number of national guidelines on hypertension, and we always comment on co-morbid conditions such as heart failure, diabetes mellitus, chronic kidney disease (CKD), and lung disease but generally fail to pay attention to its common co-morbidity with osteoarthritis. Hopefully, this will change in future guidelines.

Dr. Friedewald: There is only 1 paragraph devoted to NSAIDs in the seventh report of The Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure in its most recent report (JNC 7). This paragraph points out that older agents are the most extensively studied (i.e., indomethacin). There is a small mention of cyclooxygenase-2 (COX-2) inhibitors, and JNC 7 concludes with several sentences related to caution when prescribing NSAIDs in patients with diabetes mellitus. What is the interrelation between diabetes mellitus and osteoarthritis?

Dr. White: I participated as advisor and author on this topic for JNC 7 and for the American Heart Association (AHA) guidelines on resistant hypertension. While both committees recognized the importance of NSAIDs on BP destabilization, the amount of space devoted to the topic was abbreviated due to the numerous issues that had to be covered in both guidelines. The issue with diabetic patients is that compared to nondiabetics, they more often have renal disease and hence they are more likely to be taking an angiotensin-converting enzyme (ACE) inhibitor or an angiotensin receptor blocker (ARB), and NSAIDs interfere with their anti-hypertensive effect. Thus, the authors of both JNC 7 as well as the AHA Resistant Hypertension Working Group recommended that patients with diabetes mellitus required more caution when prescribed an NSAID.

Dr. Williams: The development of insulin resistance and the metabolic syndrome and their association with weight and diabetes are likely the common underlying threads with osteoarthritis.

Dr. Friedewald: There also is a shared inflammatory component between diabetes mellitus and osteoarthritis, but the relation is unclear. What is the effect of NSAIDs on renal function?

Dr. Wesson: There are a lot of data on the effect of NSAIDs in patients with reduced renal function, as manifest by reduced GFR and increased serum creatinine. We try to avoid using NSAIDs in patients with low GFR. For patients who have a low GFR and are already taking these drugs, we try to find a way, in consultation with their rheumatologist or primary care physician, to discontinue them.

There are at least 5 mechanisms by which NSAIDs can reduce the GFR, and they usually do so in patients in whom the GFR is already reduced or, secondarily, in patients whose GFR is normal but have some underlying cardiovascular problem that has led to a reduction in kidney blood perfusion. This reduced kidney perfusion can be due to depletion of extracellular fluid volume and/or 1 of the edematous states that are characterized conceptually by reduced effective arterial blood volume, such as heart failure and liver failure. These patients are at increased risk of a decline in GFR in response to NSAIDs, even if their initial GFR as measured by serum creatinine is normal. Most often, however, patients who have NSAID-induced GFR reduction already have reduced GFR when placed on an NSAID.

The first and most common cause of NSAID-induced GFR decline is mediated by hemodynamic mechanisms , probably related to a reduction in glomerular perfusion secondary to NSAID-induced decrease in the kidney levels of vasodilating prostaglandins. The result is an increase in the serum creatinine or concomitant reduction in the GFR, usually by only minimal amounts. There is, however, the occasional patient with a dramatic increase in creatinine or a decrease in GFR in response to even relatively small doses of commonly used NSAIDs. Fortunately, such a reduction in GFR is almost always reversible, so stopping the drug typically leads to a return to the pre-NSAID kidney function within a few days of discontinuation.

There are 4 other less common mechanisms for renal dysfunction secondary to NSAIDs. Acute interstitial nephritis , which has an allergic basis and is uncommon, is associated with nearly all of the NSAIDs. In most patients with this condition, kidney function returns to normal within a few weeks after discontinuing the NSAID or to the baseline level in patients with prior renal disease.

The third mechanism is acute tubular necrosis (ATN). This is a very unusual occurrence and has not been reported with all NSAIDs, but there is a suspicion that all NSAIDs at least have the propensity to cause ATN. Unlike acute interstitial nephritis, which appears to be secondary to an allergic response, ATN is related to the NSAID acting as a direct tubular toxin, analogous to effects of other nephrotoxic drugs such as aminoglycosides. These patients typically have the usual course of ATN, with low GFRs for a few days to 2 weeks, followed by progressive recovery.

Fourth, and even less common is a form of glomerulonephritis that resembles classic membranous glomerulonephritis, except pathological specimens typically show a white blood cell interstitial infiltrate, mainly with lymphocytes. This is important to recognize because these patients also typically improve within weeks to months after stopping the NSAID, thereby allowing avoidance of treatment with toxic agents such as glucocorticoids that are used for more common forms of acute glomerulonephritis.

The fifth mechanism by which NSAIDs can reduce the GFR is papillary necrosis . Fortunately, because it is not reversible, this is very uncommon, and the patients who are at risk are those who have low arterial perfusion of the renal papilla. The kidney is overall very highly perfused, but when comparing different regions of the kidney, the papilla is the so-called watershed area with the least perfusion, and this appears to be the reason why it is the area of the kidney most commonly associated with necrosis. Papillary necrosis has most often been described in patients with diabetes mellitus and ureteral obstruction, and occasionally in patients with hemoglobinopathies such as sickle-cell disease.

Because of these 5 mechanisms by which the GFR can be reduced by NSAIDs, nephrologists typically do not prescribe them, and we try to stop them in patients with CKD. In primary care, however, GFR reduction secondary to NSAID use is uncommon.

Dr. Friedewald: Do these renal effects occur with both selective (COX-2) and nonselective NSAIDs?

Dr. Wesson: There is debate among nephrologists whether the COX-2 inhibitors are less likely to have these renal complications, but most nephrologists regard their potential nephrotoxicity as the same as nonselective NSAIDs.

Dr. Ram: Is there a relation between kidney function and NSAID dosage and treatment duration?

Dr. Wesson: The higher the dose in general, the more likely a decrease in GFR. Thus, patients with reduced GFR should be on a low dose of NSAID if an NSAID is required. Although there are data showing that there are some dose-related hemodynamic effects, I am reluctant to give an NSAID to a patient who has demonstrated a prior NSAID-induced reduction in GFR. The other 4 forms of NSAID-related renal disease—acute interstitial nephritis, ATN, glomerular nephritis, and papillary necrosis—have no relation to dose, so any NSAID dose can lead to those 4 complications.

Dr. Williams: What is the effect of acetaminophen on renal function?

Dr. Wesson: The combination of phenacetin and acetaminophen is more likely to cause a so-called analgesic nephropathy than phenacetin taken alone. It has been hard to show, however, that chronic lone acetaminophen use causes analgesic nephropathy, although there are “hints” in some studies that it can do so. Given that data, most nephrologists would not use that combination, and many are loathe to use acetaminophen chronically (i.e., for years). There are some data also suggesting that the combination of NSAIDs and acetaminophen taken together over many years can cause a nephropathy that appears pathologically different from classic phenacetin nephropathy, and this form mostly involves the renal interstitium. Although the data are slim for an acetaminophen-NSAID combination to cause nephropathy, there is a high degree of suspicion that it does.

Dr. Friedewald: An editorial in the American Journal of Kidney Disease states, “There is evidence that elderly patients [when prescribed NSAIDs] are particularly at risk for a critical decline in the GFR under the condition of absolute or functional extracellular volume contraction and subsequent stimulation of the renin-angiotensin-aldosterone system. Therefore, it is good clinical practice to pay attention to sufficient fluid intake in patients who receive NSAIDs.” Do you agree?

Dr. Wesson: Yes, definitely . Any condition that causes kidney underperfusion—such as frank arterial volume depletion or decrease in effective arterial volume associated with edematous states like heart failure and liver failure—leads to an increased risk of NSAID-induced reduction in GFR. I often tell our house staff a true story from an article in the mid-’70s, when indomethacin was popular. Medical students who presumably were healthy and had measurable normal GFR and renal blood flow were given a 25-mg dose of indomethacin. The renal blood flow and GFR were remeasured and were again normal. Next they were placed on a very low sodium diet and given 10 mEq per day of sodium, which is a very low sodium diet, for 1 week, then had remeasurement of the renal blood flow and GFR, which were again normal. Finally, after continuing on the sodium-restricted diet for a week, they were given the same dose of 25 mg of indomethacin, and both renal blood flow and GFR dropped by 40%—and these were healthy persons . This illustrates that, under so-called normal conditions of a normal cardiovascular system and normal intravascular volume, these drugs are less likely to reduce GFR. This study showed that 1 reason the kidneys maintain good renal blood flow and GFR under the conditions of volume depletion caused by sodium restriction is the presence of very high levels of vasodilating kidney prostaglandins. When those are knocked out with an NSAID—in this instance, indomethacin—there follows a big reduction in GFR. Thus, it is important for patients to remain volume replete, not just water replete, when taking NSAIDs .

Dr. Ram: Occasional or moderate use of drugs like acetaminophen or even NSAIDs probably does not have a significant impact on kidney function. However, this does not represent the “real world” of patients, who generally take these over-the-counter (OTC) drugs frequently and regularly. These areas are ill defined, except in populations who see physicians often, such as patients with severe joint problems.

Dr. Wesson: The nephrology community fought very hard against making OTC NSAIDs available, even in small doses. Many persons mistakenly believe drugs that can be purchased without a prescription must be safe and can be taken as often and as long as they desire; consequently, many patients with NSAID-induced renal complications have taken them in the OTC form. We emphasize to our trainees that taking a good drug history must include OTC drugs, because patients often assume you are only concerned about prescription drugs when questioned about drug use.

Dr. Friedewald: What is the effect of NSAIDs on BP in patients without renal disease?

Dr. Ram: There are multiple factors in normal BP control. First, there are humoral factors such as vasodilatory prostaglandins (i.e., prostacyclin). These provide a balance between the vasoconstrictor and vasodilatory forces that regulate vascular tone. However, when the balance tilts towards vasoconstriction or vasodilatation, the BP changes. The vasodilatory mediators include prostacyclin, bradykinin, and nitric oxide. Thus, when a relative deficiency or absolute deficiency of these vasodilatory substances allows the vasoconstrictor forces to take control of vascular tone, the arterial BP rises. The vasoconstrictor substances include angiotensin, endothelin, and norepinephrine. Prostacyclin can be altered in the presence of cyclooxygenase inhibition caused by nonspecific NSAIDs and COX-2 inhibitors, providing a shift toward the vasoconstrictor forces. That appears to be the dominant physiological reason for increase in peripheral vascular resistance that occurs with both older NSAIDs and COX-2 inhibitors.

There also is a direct renal effect of NSAIDs , in addition to the 5 mechanisms Dr. Wesson outlined, and this causes more subtle renal impairments. For example, aberrations in prostaglandin E ₂ (PGE ₂ ) which is particularly associated with natriuresis and water excretion, causes up to 40% increase in sodium and water retention. This mechanism can be reset with time. Thus, there are 2 mechanisms by which the BP can be dysregulated by NSAIDs: (1) sodium and water retention in the kidney and (2) dominance of vasoconstrictor forces due to partial elimination of prostacyclin or nitric oxide due to cyclooxygenase inhibition . Both of these mechanisms can occur in the absence of kidney disease. Thus, even with normal renal function , BP dysregulation can occur with NSAID use. With regard to cardiovascular disease that has been uncovered with the use of NSAIDs, the BP effects have not been separated from increased prevalence of myocardial infarction, stroke, and heart failure.

Dr. Friedewald: Does the BP response to NSAIDs vary among persons with normal renal function?

Dr. Ram: This has not been studied very much. The mean BP can increase by 3 to 6 mm Hg upon exposure to these drugs. Such elevations could have a tremendous impact over time because they are associated with adverse end-organ effects such as with increased arterial intimal-medial thickness, left ventricular hypertrophy, and microalbuminuria. Even a modest 2 mm Hg increase in the mean arterial BP in the long term causes a tremendous increase in the prevalence of stroke and coronary artery disease. Incidentally, this fact shows that it is bad practice to “round off” the BP by a couple of millimeters because of not wanting to “label” somebody as hypertensive, which is a tendency among many health practitioners. That is not a good practice because a modest increase in BP has been shown to cause a significant increase in disease burden in the long run.

Dr. Friedewald: You emphasize the mean BP; what about systolic compared to diastolic BP?

Dr. Ram: Among patients with hypertension receiving NSAIDs, both systolic and diastolic BP are increased above pre-NSAID treatment levels. In patients aged >65 years, the effect on the systolic BP by NSAIDs is greater. This effect, however, is highly variable, with some people showing no increase in BP with NSAIDs. There may be some genetic factor involved in this variability.

Dr. Williams: Although most patients taking NSAIDs do not have BP elevation, some patients have a dramatic rise in BP, in the range of 10 to 20 mm Hg systolic. Patients placed on NSAIDs—especially patients with hypertension, established kidney disease, and older patients—need to have their BP and renal function evaluated within weeks after starting a NSAID. This evaluation should not be delayed until the customary 3 month follow-up visit.

Dr. Friedewald: The hypertensive response to NSAIDs occurs soon after they are started?

Dr. Williams: It is relatively quickly, and we do not see much compensation, at least in the studies that have been done following patients over a 3-month interval, in which patients became hypertensive and they remained hypertensive. Thus, an early assessment within a few weeks after starting an NSAID is recommended.

Dr. Ram: The effect of NSAIDs in some patients is very acute, with a dramatic rise in BP because sometimes there is 30% to 40% decrease in natriuresis. Randomized clinical trials generally exclude patients with significant hypertension, vascular disease, and abnormal renal function, whereas such patients are often included in observational studies. Thus, randomized clinical trials underestimate the NSAID-related BP risk in community-based practice.

Dr. Roberts: Is there a gender difference in the BP response to NSAIDs?

Dr. White: None of the clinical trials have shown a gender difference.

Dr. Friedewald: Are there ethnic differences in NSAID BP response?

Dr. Williams: There is no evidence that I am aware of.

Dr. White: I also agree, but there is much less known about African Americans and the effect of NSAIDs on BP.

Dr. Wesson: I agree.

Dr. Williams: Clinical trials tend to report “hypertension” in a variety of ways by different investigators, and it is often poorly defined. Thus, BP effects that are reported in clinical trials must be interpreted with caution, with special attention to the methodologies related to BP measurements.

Dr. Wesson: Among patients seen by nephrologists, it is important to follow up these patients within 1 to 2 weeks after starting an NSAID . Prostaglandins, especially PGE ₂ , are powerful natriuretic hormones, mostly in the ascending limb and the distal nephron, and when they are knocked out with an NSAID, the patient is put at risk for sodium retention. This issue is relatively unimportant in people with normal cardiovascular status and normal vascular volume because they can excrete salt. In patients with heart failure, NSAIDs can cause dramatic increases in salt retention. In addition to carefully monitoring their BP and GFR, these patients should be instructed to watch for weight gain when starting an NSAID. Weight gain after beginning an NSAID is another signal of salt retention and a volume-mediated contribution to hypertension.

Dr. Friedewald: Are NSAIDs a neglected factor in patients with “resistant” hypertension?

Dr. Wesson: Yes. I have seen many such cases.

Dr Ram: When an antihypertensive drug effect is suddenly lost, 1 of the considerations is whether the patient has started taking an NSAID, which negates the effects of β blockers and inhibitors of the renin-angiotensin system, but apparently not the antihypertensive effect of calcium channel blockers.

Dr. Friedewald: Let’s discuss the effects of NSAIDs in patients with CKD.

Dr. Wesson: Patients with reduced GFR often have an elevation of BP when starting an NSAID. Thus, patients with diabetes mellitus, hypertension, and many chronic cardiovascular diseases are at risk for increased BP response to NSAIDs, especially because the GFR is often reduced in these populations.

Dr. Friedewald: Can NSAIDs worsen heart failure?

Dr. Ram: It is possible that there could be an aggravation of heart failure in patients taking NSAIDs, either due to the direct effects on the kidney or by negating the benefits of ACEs and ARBs, which are commonly used to treat heart failure. Sodium and water retention, even in people who have normal kidney function, is quite common with NSAIDs, particularly those who take them often. Some people taking NSAIDs have ankle edema very similar to the edema due to dihydropyridine calcium channel blocking drugs in the presence of normal renal function. Thus, there are dual mechanisms in which heart failure can be worsened: (1) reversing the beneficial effects of cardiovascular drugs and (2) causing sodium and water retention.

Dr. Wesson: Nephrologists see many patients with heart failure whose sodium retention gets worse after beginning an NSAID. Patients who take NSAIDs, even those with normal kidney function, have been demonstrated to have a subtle decrease in their ability to excrete free water, and that effect is exacerbated in patients who have heart failure. Thus, although we tend to think exclusively about sodium retention manifested by edema, it also is not uncommon for patients put on an NSAID, especially those with class III and IV heart failure, to have water retention and hyponatremia.

Dr. Friedewald: What is the risk of developing hyperkalemia with NSAIDs?

Dr. Wesson: Hyperkalemia is a definite risk with NSAIDs. The mechanism, which is well defined, is due to reduced aldosterone secretion. Older patients already have decreased aldosterone secretion, which is usually subclinical. But a hypoaldosterone state can become clinically manifest by putting older patients, particularly those who have heart failure, on an NSAID, resulting in significant hyperkalemia.

Dr. Ram: It is well known that cyclooxygenase inhibition causes what is termed hyporeninemic hypoaldosteronism , which is reminiscent of type 4 renal tubular acidosis in diabetic patients, in which a serum potassium >5 mg/ml can occur because they are often already receiving an aldosterone antagonist, either eplerenone or spironolactone, for heart failure. If an NSAID is added to this hypoaldosterone state caused by an aldosterone antagonist, there is potential for significant, dangerous hyperkalemia, even in patients with apparently normal renal function . Hyperkalemia can be very dangerous in patients with heart failure exposed to NSAIDs, because they are already set up for hyperkalemia due to pharmacologic treatment.

The pathophysiological relation between heart failure and NSAIDs is conceptually sound, and there are epidemiolological data supporting a relation between NSAIDs and heart failure, such as the Rotterdam Heart Study.

Dr. Wesson: Many clinicians recognize the potential danger and try to use drugs other than NSAIDs for arthritis in patients with heart failure. Many physicians, especially primary care physicians, are less careful about using NSAIDs in these patients. It is underappreciated that NSAIDs inhibit aldosterone secretion and that a “mild” hyperkalemia due to an aldosterone antagonist already being taken can be made much worse if an NSAID is added. Although not a common situation, there is insufficient awareness among physicians of the danger of hyperkalemia in patients with heart failure.

Dr. Friedewald: Although we started this discussion with arthritis, NSAIDs are taken for many other conditions, such as headache. There are even NSAID combinations with antihistamines—“NSAID PMs”—that people take for sleep! Let’s discuss the interaction between NSAIDs and specific antihypertensive drug classes.

Dr. Wesson: The first pharmacologic approach to treating the arthritis patient who is already being treated for hypertension should be to try to treat with a drug other than an NSAID. Nephrologists love ACE inhibitors and ARBs for treating hypertension, and their antihypertensive effects are lessened by NSAIDs.

Dr. Friedewald: Why do you “love” ACE inhibitors and ARBs?

Dr. Wesson: They provide good vascular, cardiac, and renal protection. There are at least 25 years of nonhuman animal data and almost 20 years of human data suggesting that they are protective in diabetic nephropathy, and some evidence that they are renal protective in hypertensive nephropathy, meaning that patients who have reductions in GFR due to diabetes and reductions of GFR due to hypertension can have progression of their disease that can be slowed by an ACE inhibitor or an ARB. The fact that many of our CKD patients are at increased risk for myocardial infarction and stroke is another reason for using ACE inhibitors and ARBs, due to their cardioprotective effects. Thus, patients treated with ACE inhibitors and ARBs are compromised by taking NSAIDs, both from the standpoint of BP control and the other adverse effects of NSAIDs on the kidney. At any level of BP control, there is better kidney protection with ACE inhibitors and ARBs than with other classes of antihypertensive agents.

Our next choice of drug for hypertension for patients who must be on an NSAID is a calcium channel blocker. Within the nephrology community, however, there is continuing debate about the relative renal protective effect of the dihydropyridine calcium channel blockers compared to the nondihydropyridine calcium channel blockers. Patients taking dihydropyridine calcium channel blockers without an ACE inhibitor may actually have worsening of their proteinuria, with the assumption that there also is acceleration in nephropathy progression, especially in hypertensive nephropathy. Thus, in patients for whom we cannot prescribe an ACE inhibitor or ARB, we commonly use a nondihydropyridine calcium channel blocker as an antihypertensive agent because they are likely to be kidney protective, whereas a dihydropyridine, without being combined with an ACE inhibitor, would not be renal protective. The data suggest that the dihydropyridine calcium channel blocker with an ACE inhibitor is renal protective. In summary, if the patient can be maintained on an ACE inhibitor or ARB, I would prescribe a dihydropyridine calcium channel blocker; if the patient is not on or cannot be maintained on an ACE inhibitor or ARB, I use a nondihydropyridine calcium channel blocker.

Dr. Friedewald: Do you use the indications for ACE inhibitors and ARBs interchangeably?

Dr. Wesson: For the most part, yes. I like the ACE inhibitors better because they are generally less expensive, and because they have a longer scientific history. If a patient cannot tolerate an ACE inhibitor for cough or another reason, my second choice is an ARB.

Dr. Friedewald: If a patient is already on an ACE inhibitor and is starting an NSAID, you still continue the ACE inhibitor?

Dr. Wesson: Yes.

Dr. Friedewald: Is the NSAID-negating effect on ACE inhibitors and ARBs related to the BP or some other direct effect on the kidney or to both?

Dr. Wesson: Nonhuman animal studies suggest that ACE inhibitors maintain a renal protective effect even when the BP is increased by an NSAID. Translating the nonhuman animal data to humans, renal protection remains from the ACE inhibitor or ARB even if the BP rises.

Dr. Friedewald: Where do the diuretics fit into this picture? Patients on ACEs and ARBs also are often taking diuretics.

Dr. Wesson: Although an ACE inhibitor and a diuretic is a good antihypertensive combination, there are no good data showing whether this combination is more renal protective than ACE inhibitors alone. There is an unconfirmed fear that diuretics have the potential to hurt renal function because they stimulate the renin-angiotensin system. Nevertheless, I still use diuretics for their intravascular volume-controlling effects.

Dr. Roberts: If ARBs and ACE inhibitors cost the same, which one would you prefer?

Dr. Wesson: If they were of equivalent cost, I would go with the ARBs because they are backed by good data and they are better tolerated, especially since cough is not as likely to occur as with the ACE inhibitors.

Dr. White: The addition of an NSAID to an antihypertensive drug regimen destabilizes the BP in some patients taking any drug that is affected by volume retention. Thus, the antihypertensive effects of ACEs, ARBs, diuretics, and β blockers are all negatively affected by the addition of an NSAID . There are data on the calcium antagonists, and I believe it applies to both dihydropyridines and nondihydropyridine forms, showing this class is impervious to the effects of selective COX-2 and nonselective NSAIDs. Some data suggest that patients taking an α blocker are also minimally affected by NSAIDs. As for some of the older drugs like methyldopa, clonidine, and others, there are little data.

Dr. Friedewald: How much NSAID-induced BP elevation do you see among hypertensive persons with normal renal function?

Dr. White: There are different ways this has been analyzed, but it is established that the introduction of a drug like high-dose ibuprofen into a mix of a controlled hypertensive population taking drugs that may be influenced by NSAIDs—ACE inhibitors, ARBs, diuretics—causes about 20% to 25% of them to have a clinically meaningful increase, up to a 20 mm Hg rise, in the average systolic BP.

Dr. Friedewald: How quickly does this occur?

Dr. White: The time course of this is pretty quick. If an NSAID is added to the antihypertensive medicines mentioned, the BP rise may be observed within 3 or 4 weeks. It may actually go up as quickly as within 1 or 2 weeks, but this rise may take longer. If by 3 to 4 weeks there is no elevation in an individual’s BP, the patient is probably not going to develop destabilization of BP control.

Dr. Friedewald: What is your clinical approach?

Dr. White: If an individual has a good clinical response from an NSAID, but the BP rises, you might first consider manipulation of whatever antihypertensive drug or drugs they are already taking. That is a more reasonable approach, rather than removing the NSAID that is relieving pain and inflammatory symptoms. If Mrs. Jones is age 80 years, and she has been on hydrochlorothiazide and an ARB that are controlling her BP at 126/80, and I put her on an NSAID and her BP rises to 140/84 mm Hg but says, “My life is 100% better, my stiffness is better, I’m walking more, I’m happy,” I either change her antihypertensive regimen to something that is less likely to be affected or maybe up-titrate the doses of the drugs she is already on to see if that has any beneficial effect. This is all trial and error, because there are no clinical trials to test how this situation should be approached.

Dr. Friedewald: Let us assume your patient has been controlled on submaximal doses of a dihydropyridine calcium channel blocker, a hydrochlorothiazide, and an ARB, and you are going to up-titrate to counter the BP-elevating effect of an NSAID. Where would you start?

Dr. White: I probably would first increase the dose of the dihydropyridine.

Dr. Friedewald: Do you have a particular concern about β blockers when treating hypertension in patients on NSAIDs?

Dr. White: Many believe that β blockers are not as affected by NSAIDs, but I do not agree. I believe that the BP-lowering effect of β blockers is counteracted when the intravascular volume rises as may occur on an NSAID. Thus, patients who are on β blockers and not on a diuretic and are treated with an NSAID, there is a 20% to 25% chance that there will be some increase in BP. In some patients, this increase may not be clinically relevant. This is especially true for patients taking a β blocker for an indication other than hypertension, such as an arrhythmia or coronary heart disease, in whom a small rise in BP is sometimes unimportant.

Dr. Roberts: How many of the 70 million persons with hypertension in the USA take an NSAID?

Dr. White: In the middle-age population, which comprises most of the United Healthcare and Anthem Blue Cross Blue Shield databases, the number of hypertensive patients who also are on an NSAID seems relatively low—about 10% to 15% taking them chronically—the same number of NSAID users who do not have hypertension. The Medicare database, however, shows that a higher number, perhaps 30% of persons aged >65 years, take an NSAID, including both patients with and without hypertension. What we do not know, but we are in the process of assessing at this time, is among patients treated for hypertension and then prescribed an NSAID, what is the likelihood that changes are made in their antihypertensive regimen? And likewise, if they are not on any antihypertensive drug and they start an NSAID, what is the likelihood that they have to go on an antihypertensive drug?

Dr. Friedewald: Among patients who are being treated on an antihypertensive regimen—and it is generally defined that 80% or more doses taken per month is “full” compliance—less than half take their drugs that often. That corresponds to missing the antihypertensive drug at least 6 days a month, which is a lot of missed doses. On days when patients have arthritic pain, however, it is likely they will not fail to take the NSAID. Thus, we can speculate that a lot of patients with what appears to be controlled hypertension when they are in your office are taking BP-elevating NSAIDs on days when they are not taking the ACE or ARB. Unfortunately, most patients on antihypertensive regimens appear to be only 50% to 60% adherent, so the story is even worse for most patients.

Dr. White: This is the kind of thing that we practicing physicians prefer to sweep under the rug, since it is largely out of our control. Several years ago, investigators from Italy were presenting a pharmacodynamic and pharmacokinetic study of an antihypertensive agent in which they were performing 24-hour BP monitoring and evaluating effects of dosing time on BP reductions. Built into this study was a computerized “memory” cap on the bottle which allowed for a clear denotation of when patients opened their pill bottle and presumably took their BP medication. And of course the patients did not know about the memory cap. The striking results were that only 60% of the subjects took their medications, even with a nurse saying, “You are getting your monitor; please take your medication today, take it tonight,” and so on. It is a frightening finding, and I wonder how we are doing in practice if that happens in a study where people are handed their medicine by a study nurse, were paid to come to their visits, and given such explicit instructions. So your point is well taken, and what it means is that patients are probably going to be more likely to skip their drug for their asymptomatic hypertension than for their symptomatic arthritis.

Dr. Friedewald: Exactly. There are many days when they are going to be taking an NSAID when they are not taking their antihypertensive, making the NSAID-induced hypertensive response even worse.

Dr. White: We should take that into consideration. There are some drugs that have fairly long half-lives, like chlorthalidone, amlodipine, telmisartan, and aliskiren. The missed-dose studies for those drugs look a lot better than results for the shorter term drugs. Clinical trials show that if a patient takes aliskiren for 8 weeks and then does not take it for 5 days, there is still a residual antihypertensive effect.

Dr. Friedewald: Does that affect your prescribing decisions if you think you have a particularly nonadherent patient?

Dr. White: It does for me, yes.

Dr. Ram: Fortunately, if the long-acting drugs are missed at times, the BP does not go up.

Dr. White: One dilemma is what to do with patients with CKD, because we have few data on the efficacy and safety of NSAIDs in this population. Thus, for people with renal disease, we avoid this class of drugs. It does not mean that if someone has a GFR of 70 ml/min that some of these agents might not be okay. Still, they must be watched closely for changes in potassium, blood urea nitrogen, and creatinine. I do not know whether there is enough information to say that there is a threshold within the CKD patient population without advanced disease at which you should absolutely not versus possibly attempt to use an NSAID.

Dr. Ram: Because of billing and other nonclinical issues, once you circle on the form even “CKD 1,” the nephrologist avoids all prescribed and OTC NSAIDs. The moment “CKD” is placed in the chart, however minimal its severity, patients are advised not to take them because of the connection between renal dysfunction and NSAIDs.

Dr. Wesson: In general, once a patient is diagnosed with CKD, we try hard to keep them away from NSAIDs. That being said, the data regarding their risk of having the most common mechanism for a reduction in GFR (that is, the hemodynamic mechanism) suggests that the higher the GFR, the less likely they are to have a reduction in GFR. Thus, I am much more willing to risk putting patients with higher GFRs on an NSAID if I am convinced that there are few other options.

Dr. White: You, as a nephrologist, know the GFR of your patients. Do you not believe, however, that many primary care physicians do not know the GFR? They know their patient’s age and serum creatinine, but they are not calculating the GFR, so if a patient is 75 years old, has a low body mass index, and has a creatinine of 1.1 mg/dl, they are not necessarily going to withhold an NSAID from that patient. We do not know what that does to a patient in a systematic fashion; in fact, in all of the clinical trials that I have been involved with, the creatinine was the test that kept people out of the studies, not an estimated or calculated GFR.

Dr. Williams: Despite our discontinuing them, some patients with CKD resume OTC NSAIDs or obtain them from another physician. On the other hand, there are patients who, for a variety of reasons, such as volume depletion, might have a significant reduction of renal perfusion and a small rise in their creatinine and we discontinue their NSAID, and in some of those patients the NSAID can be represcribed. I think that is the group that we are talking about with a GFR of 60 to 70 ml/min, not the group with a GFR of 30 ml/min, in whom I believe none of us would be comfortable using these drugs. It is clearly a concern, and having the GFR calculated when laboratory studies are obtained can be helpful in alerting people. Some people are astonished by the fact that a creatinine of 1.2 mg/dl in an older patient can mean a significant decline in renal function.

Dr. Roberts: If all of the NSAIDs cost the same, would there be any particular advantage of 1 over another for a patient who had associated systemic arterial hypertension or in a patient who had associated renal dysfunction or in a patient with associated diabetes mellitus versus the group of patients who do not have hypertension or renal dysfunction or diabetes?

Dr. White: Of the 3 most widely used arthritis NSAIDs—naproxen, ibuprofen, and diclofenac—ibuprofen clearly induces more hypertension and probably even more renal dysfunction than the other 2. Diclofenac may cause more hypertension than naproxen or celecoxib. Of the COX-2 inhibitors, only celecoxib is available, and it does not seem to cause too much destabilization of BP in patients with most of the different kinds of hypertension and antihypertensive drugs, and it may be about the same or a little better than naproxen for induction of renal problems. The only large study I know of evaluating the diabetic population is the CRESCENT (Celecoxib and Rofecoxib Efficacy and Safety in Comorbidities Evaluation Trial), involving diabetic hypertensive patients on an ACE inhibitor, and BP was less affected by naproxen and celecoxib compared to rofecoxib. In contrast, most of the information in the normotensive population is based on clinical BPs obtained sporadically, or maybe in some cases systematically, in efficacy trials of the NSAIDs. Among those normotensive persons, there is much less interference with BP than occurs in the treated hypertensive patient populations.

Dr. Williams: There also may be some differences with respect to various NSAIDs in patients with borderline or overt heart failure. Heart failure is difficult to assess in trials as an outcome. With respect to the question of which NSAID to prescribe, if cost is not a factor, this is not a population of patients that we study because they are excluded from trials. NSAIDs should be used with caution, if at all, in patients with heart failure.

Dr. Ram: Celecoxib and naproxen appear to be less prohypertensive than other types of NSAIDs. There are some interesting data from nonhuman animal studies that COX-2 is vigorously expressed from the renal cortex in response to volume depletion. Thus, COX-2 could be a compensatory response to volume depletion, and if decapitated with a COX-2 inhibitor, acute tubular necrosis could be precipitated.

Dr. Friedewald: What is the general level of awareness among physicians about NSAIDs and their effect on BP and renal function?

Dr. Wesson: Our primary care colleagues have much less awareness of the issues that we are talking about. The real focus is to get the message out to primary care physicians and make sure they have the requisite vigilance as to the patient populations who are at risk for untoward effects from NSAIDs.

Dr. Friedewald: Dr. Wesson, with the electronic medical record system at Scott and White, you have various electronic alerts programmed into the system. For the patient with hypertension, does it display “Does the patient take an NSAID?” or, if they are put on an NSAID, does the question appear, “Does the patient have hypertension?”

Dr. Wesson: Yes, our system does that. Our main problem is that there are so many such pop-ups that we are constantly trying to determine which ones we should keep and which ones we should get rid of, because our physicians say, “Look, I’m overwhelmed and most of these pop-ups I just ignore, so just give me those that really are meaningful.”

Dr. Friedewald: You are going to keep the NSAID warning?

Dr. Wesson: Yes, certainly the nephrologists are going to argue to keep it.

Dr. Ram: CKD treatment is more and more being shifted from the nephrologist to primary care, and this type of awareness must grow.

Dr. Friedewald: The other group we need to reach are pharmacists, because so many OTC NSAIDs are purchased without physician knowledge.

Dr. White: One thing we do not capture in most of the studies, including the observational studies, is the OTC NSAID, which is never recorded anywhere, so we have no data about the magnitude of this problem. I have patients with gastrointestinal bleeding whom I find are taking an OTC NSAID. The same occurs with renal dysfunction.

Dr. Roberts: OTC NSAIDs were approved by the Food and Drug Administration (FDA) despite their potential complications, while OTC statins, which are much safer drugs, were rejected.

Dr. White: There were hearings in 2005 at the FDA that reviewed these drugs and every aspect of their cardiovascular safety. The focus was more on heart failure, myocardial infarction, and stroke, rather than BP at those meetings. The advisory committee suggested to the FDA that all prescribed NSAIDs have a black-box warning about cardiorenal problems. If they had made that recommendation for the OTC products, then they all would have been reverted to prescription-only status, because an OTC product cannot have a black-box warning label. The FDA required stronger language in the OTC NSAID labels, but they did not require any of them to be taken off the market. Of course, a person can achieve the same dose with an OTC NSAID as the prescribed drug that has the black-box warning.

Dr. Friedewald: Thank you.