Clinical evaluation of the safety and efficacy of sustained-release niacin is limited. Regular niacin appeared to be slightly more effective in decreasing total cholesterol levels [6]; however, direct comparisons of the lipid-lowering capability could not be made because of differences in doses actually consumed. The hypocholesterolemic effects of regular and extended-release niacin were similar in two uncontrolled studies [10, 16]. A recent, randomized, controlled clinical trial showed that relatively low doses of wax-matrix sustained-release niacin were effective treatment for hypercholesterolemia [17].

The Department of Veterans Affairs Medical Center, Long Beach, California, has been active in implementing accepted guidelines [5] for the detection and treatment of hypercholesterolemia, and, in keeping with these guidelines, niacin has been extensively promoted and used as a first-line agent. This pharmacoepidemiologic study is a retrospective descriptive analysis of all patients treated with controlled-release niacin during a 36-month period. We describe the safety and efficacy of controlled-release niacin in a sample of predominantly male veterans treated for lipid disorders.

Methods

Top Methods Results Discussion Author & Article Info References

The medical records, laboratory records, and pharmacy prescription records were reviewed for all patients initiated on controlled-release niacin (Slo-Niacin; Upsher-Smith, Minneapolis, Minnesota) between October 1988 and April 1991, with data collection continuing to October 1991. Demographic characteristics and social, medical, and medication history were collected from the patients' medical and prescription records. Those patients with evaluable lipid profiles had lipoprotein phenotypes defined (for this study) as follows: phenotype IIa, LDL cholesterol level of 3.36 mmol/L (130 mg/dL) or more and triglyceride level of less than 2.82 mmol/L (250 mg/dL); type IIb, LDL cholesterol level of 3.36 mmol/L (130 mg/dL) or more and triglyceride level of 2.82 to 4.50 mmol/L (250 to 399 mg/dL); type IV, LDL cholesterol level of less than 3.36 mmol/L (130 mg/dL) and triglyceride level of 2.82 to 4.50 mmol/L (250 to 399 mg/dL) or a triglyceride level of 4.52 to 11.28 mmol/L (400 to 999 mg/dL); and type V, triglyceride level of 11.29 mmol/L [1000 mg/dL] or more. Active liver disease was defined as any one of the following: 1) increases in levels of alanine aminotransferase (ALT), aspartate aminotransferase [AST], or alkaline phosphatase that were 1.5-fold or more than the upper limits of normal; 2) histologic evidence of fibrosis or cirrhosis; 3) evidence of decompensated liver disease; 4) positive test results for hepatitis B surface antigen; or 5) positive serologic test results for anti-hepatitis C. A history of liver disease was defined as previous disease (for example, jaundice, abnormal liver test results, ascites) but no current abnormal levels of liver enzymes and no histologic evidence of liver disease.

Lipid profile; levels of hepatic enzymes including ALT, AST, -glutamyltransferase, and alkaline phosphatase; and levels of glucose, uric acid, and albumin were recorded at baseline, at the end of each dose titration, and at the end of the study (the last levels were obtained while the patients were receiving controlled-release niacin). Liver enzyme levels, blood biochemistry test results, and lipid levels [18, 19] were measured by standard enzymatic methods. Low-density lipoprotein cholesterol levels were calculated using the Friedewald formula but were not estimated for patients who had triglyceride levels of 4.52 mmol/L (400 mg/dL) or more. The laboratory participates with the Centers for Disease Control and Prevention Program for standardization of laboratory tests.

Those patients identified in the medical record as having niacin-induced hepatic dysfunction as well as those patients with increased levels of AST or ALT that were threefold or more than the upper limit of normal or increased levels of alkaline phosphatase that were twofold or more than the upper limit of normal were evaluated for causality with controlled-release niacin using the Naranjo nomogram [20]. This probability scale for adverse drug reactions is based on a series of ten scored questions that evaluate the relation between an adverse event and a suspected drug. Based on this probability scale, the likelihood of a positive association between controlled-release niacin and liver dysfunction was classified into four categories: doubtful reaction—"likely related to factors other than a drug" [20]; possible reaction—"followed a temporal sequence after a drug, possibly followed a recognized pattern to the suspected drug, and could be explained by characteristics of the patient's disease" [20]; probable reaction—could not reasonably be explained by other causes; and definite reaction—confirmed by dechallenge and rechallenge of the suspected drug.

Blood lipid values, liver function test results, and standard biochemistry test results comparing baseline levels with those at the time of the final maintenance dose were analyzed by the Student paired t-test. Risk factor assessment of niacin-associated hepatic dysfunction was analyzed by chi-square analysis.

Results

Top Methods Results Discussion Author & Article Info References

Patient Characteristics

Data collection and analysis were completed in 92.5% (896 of 969) of eligible patients. Patients not included were those who transferred to other facilities, left the area, or were lost to follow-up. Baseline characteristics Table 1 indicated that patients were predominantly white men older than 50 years of age. Lipoprotein phenotype could be determined in 22.3% of the patients; the type IIa phenotype was the most common. Relative and absolute contraindications to niacin prescribing Table 1 occurred in 41.4% and 4.0% of the patients, respectively.

Instructions for dosage titration of controlled-release niacin were used in 30% (269 of 896) of the patients. The average daily dose was approximately 1.5 g with the final dose (1.67 g [0.8 g], mean [SD]) only 0.3 g greater than the initial dose (1.36 g [0.7 g]). The number of dose adjustments for controlled-release niacin ranged from 0 to 5 (0.86 [1.0], mean [SD]). Colestipol (168 patients) was the most frequently prescribed additive antilipidemic agent, followed by gemfibrozil (33 patients) and lovastatin (11 patients). Data were insufficient to assess hepatic dysfunction in patients receiving lovastatin and controlled-release niacin concurrently. Approximately one half (461 of 896) of the patients were still receiving controlled-release niacin at the end of the survey period. Of the 435 patients no longer taking controlled-release niacin, 249 had 276 documented reasons for discontinuation. The primary documented reasons for discontinuation were adverse effects, of which flushing and itching (80 patients), increased blood glucose levels (43 patients), gastrointestinal complaints (33 patients), and increased hepatic enzyme levels (33 patients) were the most common. Of the 33 patients who discontinued controlled-release niacin because of gastrointestinal complaints, aspirin use was similar (15 of 33) to that of the group as a whole (549 of 896). Controlled-release niacin had to be restarted 46 times in 70 patients in the long-term cohort (treatment duration, 30 to 36 months); 24% of the treated patients discontinued controlled-release niacin by themselves for no apparent reason every year.

Lipoprotein Response

Blood lipid values at baseline compared with those at the final maintenance dose are summarized in Table 2. The overall lipoprotein response showed the expected decreases in total cholesterol ( –19.1%),LDL cholesterol ( –24.0%),and triglycerides ( –32.5%)as well as increases in HDL cholesterol (+5.7%). Evaluable patients with mean changes in lipid levels in the long-term cohort (n = 70) were limited (total cholesterol, –27.8%,n = 35; LDL cholesterol, –29.3%,n = 11; HDL cholesterol, –0.2%,n = 14; and triglycerides, –53.2%,n = 13). For those patients in whom a complete lipid profile was available at baseline (n = 200), a 25.4% decrease was noted in levels of LDL cholesterol in patients with phenotype IIa compared with a 15.1% decrease in patients with phenotype IIb, a difference of 10.3% (95% CI, 3.3% to 17.3%; P = 0.004). Patients with HDL cholesterol levels of 1.03 mmol/L (40 mg/dL) or less at baseline had a 12.2% increase in HDL compared with a 3.2% increase in those with baseline levels more than 1.03 mmol/L (40 mg/dL), a difference of 9.0% (CI, 1.1% to 16.9%; P < 0.03). The relation between controlled-release niacin dose and changes in levels of plasma lipoprotein cholesterol is shown in Figure 1. A greater decrease in levels of mean total cholesterol and triglycerides was seen as the dose of niacin increased from 1.0 to 3.0 g/d (11.7% to 27.0% and 18.7% to 53.0%, respectively). The decrease in the mean level of LDL cholesterol was 12.1%, 26.2%, and 22.6% for niacin doses of 1.0, 2.0, and 3.0 g/d, respectively. The increase in levels of HDL cholesterol peaked at a niacin dose of 2.0 g/d (14.2%).

View larger version (20K): [in this window] [in a new window]   Figure 1. The relation between the controlled-release niacin dose and the plasma lipoprotein cholesterol response based on percentage change. HDLC = high-density lipoprotein cholesterol; LDLC = low-density lipoprotein cholesterol; TC = total cholesterol; and TG = triglycerides.

Blood Chemistry Test Results

Serum chemistry values at baseline compared with those at the final maintenance dose are summarized in Table 3. Levels of mean liver enzymes increased in a statistically significant manner over baseline but remained within the normal range. Mean changes in levels of hepatic enzymes were similar when comparing the entire cohort with the long-term cohort (AST levels, +29% compared with +33%; ALT levels, +23% compared with +20%; and alkaline phosphatase levels, +25% compared with +22%). A decrease in total cholesterol levels was associated with an increase in AST levels (r = –0.32,P < 0.001) and also with a decrease in serum albumin levels (r = +0.31, P < 0.001). No change was seen in uric acid levels, although there was a 6.7% increase in glucose levels (P = 0.0001) and a 4.7% decrease in albumin levels (P = 0.0001). Overall, the average change in laboratory tests with controlled-release niacin was not clinically significant. The relation between controlled-release niacin dose and percentage change in liver enzyme levels is shown in (Figure 2). Statistically but not clinically meaningful dose-related increases were found over a niacin dosage range of 1.0 to 3.0 g/d.

View larger version (26K): [in this window] [in a new window]   Figure 2. The relation between the controlled-release niacin dose and the percentage change in levels of liver enzymes. AP = alkaline phosphatase; ALT = alanine aminotransferase; AST = aspartate aminotransferase; and GGT = -glutamyl transferase.

Of 160 patients with diabetes mellitus, 14 (4 with new-onset diabetes and 10 with diet-controlled diabetes) required the addition of oral hypoglycemic agents while on controlled-release niacin. Oral hypoglycemic agents were the most common mode of treatment (41.9%), followed by diet alone (36.2%) and insulin (21.9%). Controlled-release niacin was discontinued in 106 of 160 (66.3%) patients with diabetes because of poor glycemic control in 40.6% (43 of 106).

Hepatotoxicity

Forty-six patients met criteria for niacin-associated hepatotoxicity. The Naranjo ratings were classified as follows: 1 definite, 19 probable, 22 possible, and 4 doubtful. The doubtful patients were excluded from further analysis. Twenty of 896 (2.2%; CI, 1.4% to 3.4%) patients met biochemical criteria for probable niacin-induced hepatotoxicity, and 42 of 896 (4.7%; CI, 3.4% to 6.3%) patients met criteria for possible or probable niacin-induced hepatotoxicity.

Most reactions (30 of 42) were mild and resulted in biochemical changes that resolved after discontinuation of controlled-release niacin or after dosage reduction. Twelve patients had vague abdominal and "flu-like" complaints that also resolved after discontinuation of controlled-release niacin. Two patients required hospitalization for niacin-induced hepatotoxicity. After dosage increases in controlled-release niacin (2.0 to 3.0 g/d and 3.0 to 6.0 g/d), both patients had marked increases in levels of hepatic enzymes, decreases in levels of serum albumin, anorexia, nausea, vomiting, and bilateral lower extremity edema that resolved after discontinuation of niacin.

The average daily dose of controlled-release niacin was greater in patients with probable niacin-induced hepatotoxicity compared with those with a rating of possible toxicity (3.1 g compared with 2.1 g, P = 0.0075). All patients with definite, probable, or possible hepatotoxicity associated with controlled-release niacin were included in the bivariate analysis of factors associated with increased risk (Table 4). Niacin-induced hepatic dysfunction was not associated with patient age or with diet or insulin-managed diabetes. Patients with hepatic dysfunction included 41 men and 1 woman equally distributed according to race. Factors associated with an increased risk for hepatotoxicity induced by controlled-release niacin included diabetes with receipt of oral hypoglycemic agents, preexisting liver disease, excessive alcohol use, and a higher mean daily dose of niacin. Three of the 4 patients with preexisting liver disease had a history of alcoholic hepatitis and the fourth patient had a history of granulomatous hepatitis. One patient with a history of alcoholic hepatitis also had a history of hepatitis B (currently negative for hepatitis surface antigen). One half (21 of 42) of the patients had at least one risk factor for the development of hepatic dysfunction. In addition, hepatic dysfunction followed a reasonable temporal sequence with concurrent controlled-release niacin in 4 patients receiving phenytoin, 1 patient receiving carbamazepine, and 1 patient receiving amiodarone.

The number of months on a given dose of controlled-release niacin before the development of hepatotoxicity ranged from 1 to 28 months (7.7 months [6.3 months], mean [SD]), whereas the number of months to recovery was 2.8 months ([SD] 2.6 months; range, 1 to 12 months). Peak chemistry values showed a fourfold increase in levels of AST, ALT, and alkaline phosphatase, a decrease (1 g/dL) in levels of serum albumin, and a 43% decrease in levels of total cholesterol.

Discussion

Top Methods Results Discussion Author & Article Info References

Despite the retrospective design of our study, the blood lipid-lowering effect of controlled-release niacin is clear. This benefit was apparent despite an environment in which optimal compliance with diet and drug therapy could not be verified. The dose-response relation for controlled-release niacin was evident up to a dose of 2.0 g/d E for all lipid values. The failure to show further decreases in levels of LDL cholesterol and increases in HDL cholesterol with dosages of 3.0 g/d may result from small sample sizes (n = 23 and 28, respectively) and poor compliance because of the decreased patient tolerance seen at higher doses. The mean decrease in levels of LDL cholesterol seen at 1.5 g/d ( –23%)compares favorably with that reported by Keenan and colleagues [17] ( –20%)and McKenney and colleagues [15] ( –22%)using different formulations of sustained-release niacin at 1.5 g/d. This decrease in levels of LDL cholesterol is greater than that reported by Knopp and colleagues [6] ( –13%)using 3.0 g/d of time-released niacin capsules. This may be explained by poor patient tolerance and compliance (68% adherence) with time-released niacin and a lower bioavailability seen with time-released niacin [21]. The decrease in levels of LDL cholesterol seen at 1.5 g/d with controlled-release niacin ( –23%)is similar to that seen with 3.0 g/d of unmodified niacin [6] ( –21%),suggesting that controlled-release niacin is twice as potent as crystalline niacin in decreasing levels of LDL cholesterol. In a subset of 160 patients, we found greater decreases in levels of LDL cholesterol occurring in patients with phenotype IIa compared with those with phenotype IIb ( –25.4% compared with –15.1%).Further investigation is needed to characterize the effects of niacin in decreasing LDL cholesterol levels in different lipoprotein phenotypes.

The increases in HDL cholesterol levels with controlled-release niacin peaked at a dosage of 2.0 g/d (+14%). Keenan and colleagues [17] also reported a plateau effect with wax-matrix niacin at a dosage of 1.5 g/d (+9%). The reported effects of various sustained-release dosage forms of niacin given at low doses (1.0 to 2.0 g/d) on HDL cholesterol levels are quite variable [15, 17, 22-24] (+9% to +41%) and are similar to those reported by Knopp and colleagues [6] (+26%) in patients taking 3.0 g/d of unmodified niacin. We found a greater increase in levels of HDL cholesterol in those patients with lower baseline HDL levels ( 1.03 mmol/L [40 mg/dL], +12.2%; >1.03 mmol/L [40 mg/dL], +3.2%). This finding supports those reported by Lavie and colleagues [24] and Squires and colleagues [23] of marked improvement of HDL cholesterol levels in patients with very low (0.67 mmol/L [26 mg/dL] and 0.88 mmol/L [34 mg/dL], respectively) baseline HDL cholesterol levels (+30%, mean dose, 2.4 g/d; and +18%, mean dose, 1.3 g/d, respectively). Lavie and colleagues [24] also noted that patients with hypertriglyceridemia (3.50 mmol/L [310 mg/dL]) had greater increases in HDL cholesterol levels compared with those patients with normal triglyceride levels (1.84 mmol/L [163 mg/dL]) at baseline (+41% and +27%, respectively). These results suggest that the effect of niacin on HDL cholesterol levels requires further evaluation and may be affected by different dosage forms of niacin, baseline HDL cholesterol levels, baseline triglyceride levels, and other factors.

Triglyceride levels decreased progressively as the controlled-release niacin dosage increased from 1.0 and 3.0 g/d (Figure 2). The greater triglyceride-lowering effects we found with controlled-release niacin (1.5 g/d, –27%)relative to those found by Keenan and colleagues [17] with wax-matrix niacin (1.5 g/d, –9%)may be caused by the fact that baseline triglyceride levels were much higher in our sample of patients (3.48 mmol/L [308 mg/dL] compared with 1.63 mmol/L [144 mg/dL]). The triglyceride-lowering effects of 1.5 g/d of controlled-release niacin were identical ( –27%)to those reported by Knopp and colleagues [6] for 3.0 g/d of unmodified niacin, again suggesting that controlled-release niacin is more potent than regular niacin.

Because our study was not controlled, many patients with relative contraindications to niacin therapy were not excluded. This allowed us to follow a large number of patients with diabetes mellitus (n = 160). Controlled-release niacin was poorly tolerated by patients with diabetes, with two thirds of them discontinuing therapy, in 40% of cases because of poor glycemic control. Henkin and colleagues [25] also found a greater discontinuation rate of niacin in patients with diabetes (88%) compared with patients without diabetes (33%). Garg and Grundy [26] found a deterioration of glycemic control with niacin therapy in patients with non-insulin-dependent diabetes mellitus and suggested that those with diabetes that was previously well controlled by diet may require oral hypoglycemic agents. This is consistent with our findings. Ten patients whose diabetes was controlled by diet alone required oral hypoglycemic agents while receiving controlled-release niacin. A recommendation that patients with diet-controlled diabetes take oral hypoglycemic agents after poor glycemic control resulting from niacin therapy appears unwise given the increased potential for hepatic dysfunction that was observed in our study. Prolonged administration of niacin results in insulin resistance [27]. The mean increase in glucose (+7%; average dose, 1.7 g/d) was similar to that found with 3.0 g/d (+7%) with unmodified niacin [28]. Although no substantial biochemical changes in uric acid were noted, small increases in uric acid have been reported for unmodified niacin [28] and sustained-release niacin [23].

Niacin has been associated with increases in levels of hepatic enzymes. We found that controlled-release niacin resulted in a dose-related increase in liver enzymes over a dosage range of 1.0 to 3.0 g/d. Keenan and colleagues [17] and McKenney and colleagues [15] also found dose-related increases in levels of hepatic enzymes with different formulations of sustained-release niacin. Increases in levels of hepatic enzymes after patients received 1.0 g/d of controlled-release niacin were similar to those found with 3.0 g/d of unmodified niacin [28] (AST, +9% compared with +12% and alkaline phosphatase, +10% compared with +10%, respectively). A new finding was a small but statistically significant decrease in serum albumin levels, which correlated with a decrease in total cholesterol levels (r = +0.31) and may indicate a slight decrease in hepatic function.

Studies of the incidence of niacin-induced hepatotoxicity have found a variance of 0 to 46% in the incidence of increased levels of hepatic enzymes [29]. The difficulty in determining the prevalence of niacin-induced hepatitis is due to differences in defining hepatotoxicity (increases in levels of enzymes, clinical disease, biopsy results, etc.), failing to establish a causal relation, differences in samples of patients studied, and differences in duration of the trial period. Our findings of a 2.2% prevalence of probable niacin-induced hepatotoxicity are similar to those reported by Blankenhorn and colleagues [4] (3.2%) in patients treated with crystalline niacin, 3 to 12 g/d.

Several recent case reports [9-14] describe hepatotoxicity associated with various sustained-release preparations of niacin. Etchason and colleagues [14] described increases in levels of hepatic enzymes in five patients treated with 3.0 g/d or less of sustained-release niacin. We also found that niacin-associated hepatic dysfunction occurs at relatively low doses (mean, 2.3 g/d). This is in contrast to hepatic dysfunction associated with higher doses of crystalline niacin. In addition, reports describe new-onset hepatotoxicity in patients switched from regular niacin to sustained-release niacin [11, 13, 14]. McKenney and colleagues [15] reported a high incidence of hepatotoxicity (12 of 23, 52%) defined as an increase in levels of liver aminotransferases to more than three times the upper limit of normal, occurring with a different sustained-release dosage form of niacin. Five patients had symptoms of hepatic dysfunction that occurred at dosage levels of 2.0 to 3.0 g/d. Decreases in levels of LDL cholesterol were similar when comparing 1.5 g/d of sustained-release niacin ( –21.9%)with 3.0 g/d of immediate-release niacin ( –21.7%).This is consistent with our findings that hepatic dysfunction and decreases in levels of LDL cholesterol are associated with lower doses of sustained-release niacin when compared with unmodified niacin, suggesting potency differences. "Generalization" [30] of the reported results [15] to other dosage forms of immediate- or sustained-release niacin cannot confidently be made without pharmacokinetic and quality-control data or comparative clinical trials.

All niacin products available without prescription, regardless of formulation, are sold as nutritional supplements. Regulatory requirements for nutritional supplements, or foods, are less rigorous than those for prescription drug products. Differences in quality control and product variability may occur with different over-the-counter niacin products because some companies follow good manufacturing practices for prescription drugs for their nutritional supplement niacin formulation, whereas others do not.

Prescribing guidelines for niacin state that hepatic dysfunction is an absolute contraindication to its use and a history of liver disease is a relative contraindication. We found an increased risk for hepatotoxicity associated with preexisting liver disease, even though levels of hepatic enzymes were within normal limits at baseline. No data are available defining what factors may predispose patients to niacin-induced hepatic injury. We found support for a dose-response relation in that patients receiving a higher mean daily dose (2.3 g/d) of controlled-release niacin were more likely to develop hepatotoxicity than those receiving a lower mean daily dose (1.6 g/d). Those patients with excessive alcohol intake (medical record notation) have a greater propensity for developing hepatotoxicity while receiving niacin. Finally, niacin-induced hepatotoxicity was more prevalent in patients with diabetes who were receiving oral hypoglycemic agents. Thus, the risk for niacin-induced hepatitis may be greater in those patients taking agents also known to cause hepatic dysfunction, such as ethanol and sulfonylurea agents.

Controlled-release niacin appears to be more potent than immediate-release niacin with respect to efficacy and biochemical changes. The toxicity of controlled-release niacin is similar to that of immediate-release niacin if used in equipotent doses. Controlled-release niacin should be avoided in patients with hepatic dysfunction or a history of liver disease, patients with diabetes mellitus, and patients who use alcohol in excess. A safe level of alcohol intake in patients taking controlled-release niacin cannot be determined from our data. Most patients respond to a dosage range of 1.0 to 2.0 g/d. The incidence of flushing and itching can be minimized by titrating controlled-release niacin to an initial target of 1.0 g/d, prescribing it with meals, and pretreating patients with aspirin as necessary. Doses higher than 2.0 g/d are more likely to cause hepatotoxicity and require more frequent monitoring. Dosage increments should not be greater than 0.5 g/d, with levels of hepatic enzymes and blood glucose evaluated at baseline, within 6 weeks of any dose increase, and every 3 months while patients receive long-term therapy. Sustained-release formulations of niacin are not the same [21, 31] and should not be interchanged. Patients should be cautioned never to switch niacin formulations unless advised by their physicians. If patients are switched from immediate-release niacin to controlled-release niacin, a dose reduction of 50% to 70% is indicated.

Presented in part at the XI International Symposium on Drugs Affecting Lipid Metabolism, Florence, Italy, May 1992.