Safety, Tolerance, and Efficacy of Extended-Release Niacin ...

399

ORIGINAL ARTICLE

Safety, Tolerance, and Efficacy of Extended-Release Niacin Monotherapy for Treating Dyslipidemia Risks in Persons With Chronic Tetraplegia: A Randomized Multicenter Controlled Trial Mark S. Nash, PhD, John E. Lewis, PhD, Trevor A. Dyson-Hudson, MD, Yaga Szlachcic, MD, Florence Yee, PharmD, Armando J. Mendez, PhD, Ann M. Spungen, EdD, William A. Bauman, MD ABSTRACT. Nash MS, Lewis JE, Dyson-Hudson TA, Szlachcic Y, Yee F, Mendez AJ, Spungen AM, Bauman WA. Safety, tolerance, and efficacy of extended-release niacin monotherapy for treating dyslipidemia risks in persons with chronic tetraplegia: a randomized multicenter controlled trial. Arch Phys Med Rehabil 2011;92:399-410. Objective: To test the safety, tolerance, and efficacy of extended-release niacin monotherapy on dyslipidemia in persons with chronic tetraplegia. Design: Placebo-controlled, blinded, multicenter, randomized controlled trial. Setting: Three spinal cord injury research/rehabilitation centers. Participants: Persons with chronic tetraplegia (N⫽54) and low plasma high-density lipoprotein cholesterol (HDL-C) levels. Intervention: Extended-release niacin monotherapy (48 weeks; n⫽31) on a dose-titration schedule versus matched placebo (n⫽23). Main Outcome Measures: Safety was assessed by using percentages of treatment-emergent adverse events and increased levels of hepatic transaminases, uric acid, glycosylated hemoglobin, and fasting glucose. Tolerance was assessed by using participant reports for frequency and intensity of adverse effects of extended-release niacin. Primary effectiveness outcomes were fasting HDL-C level and plasma total cholesterol (TC)/HDL-C ratio. Secondary outcomes included plasma low-

From the Departments of Neurological Surgery (Nash), Rehabilitation Medicine (Nash), Psychiatry and Behavioral Medicine (Lewis), and Medicine, Division of Endocrinology, Diabetes, and Metabolism (Mendez), Institutional Centers of Excellence at the Miami Project to Cure Paralysis (Nash), and Diabetes Research Institute (Mendez), University of Miami, Miller School of Medicine, Miami, FL; The Rancho Los Amigos National Research Center, Downey (Szlachcic, Yee); Department of Clinical Medicine, Keck School of Medicine (Szlachcic), and Department of Clinical Medicine, School of Pharmacy (Yee), University of Southern California, Los Angeles, CA; Kessler Foundation Research Center, West Orange (Dyson-Hudson); Department of Physical Medicine and Rehabilitation, University of Medicine and Dentistry of New Jersey-New Jersey Medical School, Newark, NJ (Dyson-Hudson); Veterans Affairs Rehabilitation Research and Development Center of Excellence for the Medical Consequences of Spinal Cord Injury; Medical, Spinal Cord Injury, and Research Services, Veterans Affairs Medical Center, Bronx (Spungen, Bauman); and Departments of Medicine and Rehabilitation Medicine, Mount Sinai School of Medicine, New York, NY (Bauman). Supported by a Spinal Cord Injury Collaborative Research Project (grant no. H133A111105) from the National Institute on Disability and Rehabilitation Research, U.S. Department of Education; and Kos Pharmaceuticals, Inc. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Mark S. Nash, PhD, Dept of Neurological Surgery, Miller School of Medicine, University of Miami, Lois Pope Life Center, R-48, 1095 NW 14th Terrace, Miami, FL 33136, e-mail: [email protected]. Published online January 31, 2011 at www.archives-pmr.org 0003-9993/11/9203-00255$36.00/0 doi:10.1016/j.apmr.2010.06.029

density lipoprotein cholesterol (LDL-C) and TC levels and LDL-C/HDL-C ratio. Results: Significant increases in fasting HDL-C levels (24.5%) were accompanied by decreases in TC/HDL-C and LDL-C/HDL-C ratios, LDL-C levels, and TC levels (all P⬍.05). No evidence of sustained hepatotoxicity or hyperglycemia was observed. Treatment-emergent withdrawals (12.9%) accompanied flushing (n⫽1), hypotension/presyncope (n⫽1), and diarrhea (n⫽2). One subject experienced transient hyperuricemia. Other drug-reported symptoms did not differ from those for placebo. Conclusions: Extended-release niacin monotherapy is safe, tolerated, and effective for most persons with chronic tetraplegia. Special precautions for changes in bowel habits and postadministration hypotension should be observed. Key Words: Cardiovascular diseases; Cholesterol, HDL; Cholesterol, LDL; Niacin; Randomized controlled trial, publication type; Rehabilitation. © 2011 by the American Congress of Rehabilitation Medicine ONSIDERABLE ATTENTION has been focused on the C risks for all-cause CVD and related endocrine disorders in persons with SCI. The contemporary term cardiometabolic 1-3

syndrome describes the complex array of these risks and encompasses the CVD-component hazards of central obesity, increased TG levels, low plasma HDL-C concentrations, hypertension, and fasting hyperglycemia.4 If untreated, these disorders incite atherosclerotic plaque formation and premature CVD, and constitute a threat so severe that 3 or more CMS components or a diagnosis of diabetes alone confer the same CVD risk as extant coronary disease.4 The CMS risks of central obesity,5-7 diabetes,8,9 and dyslipidemia10-12 are observed earlier in the life span and at increased

List of Abbreviations AE CMS CVD HbA1c HDL-C IRB LDL-C LFT NaF SAE SCI TC TG UA

adverse event cardiometabolic syndrome cardiovascular disease glycosylated hemoglobin high-density lipoprotein cholesterol institutional review board low-density lipoprotein cholesterol liver function test sodium fluoride serious adverse event spinal cord injury total cholesterol triglycerides uric acid

Arch Phys Med Rehabil Vol 92, March 2011

400

EXTENDED-RELEASE NIACIN IN TETRAPLEGIA, Nash

frequencies after SCI. They also show a disturbing tendency to cluster,13 which intensifies the CVD risks of physical deconditioning and proinflammatory activity commonly reported in the population. The most widely reported cause for accelerated CVD after SCI is dyslipidemia,14-18 for which the consistent feature of a low HDL-C level10,11,19-21 represents a component risk for CMS. Extensive trial-based evidence has confirmed a vital role of HDL-C in protection against all-cause CVD and consistently reports risks imposed by low HDL-C levels, irrespective of LDL-C levels,22 that are lessened by even small increases.23,24 In most cases, lifestyle interventions directed toward weight loss and hypertension management effectively serve as firstline treatments for CMS and typically focus on strategies that target decreased caloric intake and increased daily caloric expenditure.25 However, these approaches may have questionable effectiveness for persons with cervical SCI because loss of body fat may require unattainable or unreasonable levels of caloric restriction,5 and basal and exercise-induced caloric expenditures are decreased in patients with tetraplegia because of diminished active muscle mass and adrenergic dysfunction accompanying injury above the level of spinal sympathetic outflow.26 For patients in whom the first-line approaches of diet and exercise fail to modify lipid-related risks, evidence-based guidelines and current practice standards recommend pharmacotherapy. Nicotinic acid (niacin), a widely prescribed therapeutic agent, is an inexpensive broad-spectrum drug that decreases concentrations of all atherogenic plasma lipids/lipoproteins and is the most effective agent for increasing HDL-C levels.27 In crystalline (ie, intermediate-release) form, it provokes a robust cutaneous flushing, thus compromising patient tolerance when therapeutically dosed.28 However, an extended-release formulation of niacin29 administered with a prostaglandin antagonist and gradual dose escalation reduces this discomfort. Because a randomized controlled trial examining the drug management of dyslipidemia has not been reported in patients with SCI and extended-release niacin appears to be the best targeted agent for management of depressed high-density lipoproteinemia in persons with chronic tetraplegia, a randomized, blinded, multicenter trial was performed to examine the safety, tolerance, and efficacy of extended-release niacin monotherapy.

centers, and reviewed at a prestudy visit attended by all personnel from the site. All standardized forms used for the study were contained in the manual, as well as a flow diagram of procedures and data collection. Annual reviews of each site were performed, and participant/pharmacy logs were reviewed for accuracy and completeness. A study physician identified at each study site was responsible for physical examination, screening and qualification, ongoing care, monthly visits, review of laboratory values, and study discharge. The study principal investigator supervised clinical activities, communicated with study-site directors and sponsors, and oversaw logistics of supplies, blood transfers, data collection and entry, and analysis. Study drug and matching placebo were received from the manufacturer by the lead research pharmacy and distributed to the site research pharmacies. Persons meeting inclusion and exclusion criteria were referred from institutional clinics that treat patients with longterm SCI and those presenting for annual physical examination. Information was disseminated through clinic personnel and IRB-approved flyers. In 2 cases, databases were searched for characteristics being sought for study participation. These efforts led to the screening of 97 healthy men and women (age, 18 – 65y) with neurologically complete SCI at the C4 through C8 levels for longer than 1 year. Levels and completeness of injury were defined by using the International Standards for Neurological Classification of SCI. Subjects included in the trial were in good health and without evidence of acute illness. None of the participants used tobacco products. Participants had fasting HDL-C concentrations of 40 mg/dL or less, TC levels less than 250 mg/dL, and TG levels less than 250 mg/dL. Exclusion criteria (table 1) were based on failure to satisfy inclusion criteria and restrictions based on health status, lifestyle status or modifications, and medication use. In most cases, exclusions were targeted to minimize research confounds, especially HDL-C level increase by external influence. Otherwise, issues of safety were the basis for these decisions. In particular, alcohol consumption near the time of extendedrelease niacin administration may worsen the predominant AE of flushing and also increase HDL-C levels when consumed regularly. Coadministration of extended-release niacin also may be synergistic with other agents that decrease blood pressure.

METHODS

Randomization Subjects who cleared screening were randomly assigned by means of a permuted block design to 1 of the 2 treatment arms designated (1) extended-release niacin monotherapy or (2) placebo. Study drug and placebo were dispensed, at the beginning of each study month, by the research pharmacies located at each study site. Any unused drug, placebo, and coated

Participants The study was approved by institutional ethics boards at all participating centers. Informed consent was obtained before testing or data collection in accordance with guidelines of these IRBs. A trial manual was developed, distributed to the 3 study

Table 1: Study Exclusion Criteria Based on Restrictions of Health Status, Lifestyle Status or Modifications, and Medication Use Variable

Health status

Exclusion Criteria

Recurrent acute infection or illness, trauma, or surgery within the preceding 6 mo; pregnancy; previous myocardial infarction or cardiac surgery; lipid-lowering therapy within 6 mo of study enrollment; daily alcohol consumption ⬎2 drinks (2 beers, 2 glasses of wine, 2 oz of grain alcohol); and abnormal menstruation Lifestyle status or lifestyle modifications within 6 mo Voluntary or imposed dietary modifications, weight loss or gain ⬎5% of body of study enrollment mass, change in tobacco use (adoption or termination), alcohol consumption (see Health status), and exercise training for recreation or competition Medications ␤-Adrenergic antagonists without intrinsic sympathomimetic activity (except the ␣, ␤-blocker labetalol), methyldopa, thiazide and loop diuretics, parasympatholytic agents, zinc, estrogen replacement therapy, and insulin-sensitizing drugs



aspirin tablets were collected at this time and recorded in an inventory. The inventory log was transmitted to the lead center and cross-checked for accuracy as each site underwent annual site visits. Subjects randomly assigned to extended-release niacin (Niaspan) treatment underwent 48 weeks of treatment. An escalating dose of extended-release niacin started at 500mg nightly before bedtime for 4 weeks, then increased in 500-mg increments every 4 weeks to a peak dose of 2g before bedtime. This represents the maximum tolerated dosage adopted in clinical trials of primary and secondary CVD prevention.29,30 Subjects ingested a 325-mg Maalox-coated aspirin (Ascriptin) 60 to 90 minutes before taking extended-release niacin and consumed a small snack before bedtime. Those randomly assigned to the placebo study arm received a tablet with matched size, shape, color, and texture. Subjects were masked from their group assignment until after the study was completed or they withdrew from the trial. All group instructions were identical concerning nighttime administration, avoidance of alcohol and hot drinks within several hours of administration, and pretreatment with Ascriptin and a small snack. Data Collection Blood collection and sample preparation. Blood was obtained at baseline, at 4-week intervals from weeks 4 to 24, and at week 48. Blood was collected from a superficial arm vein 8 to 12 hours after last taking the drug and after an overnight fast that was initiated at 8:00 PM of the evening before phlebotomy. Blood samples were collected as follows: serum (liver function, lipid, lipoproteins), ethylenediaminetetraacetic acid-whole blood (HbA1c), and NaF plasma (glucose). NaF and ethylenediaminetetraacetic acid sample tubes were chilled after filling and maintained at 4°C. Serum tubes were allowed to clot for 20 to 30 minutes at room temperature. Samples were centrifuged at 4°C for 15 minutes at 2000g, and serum and plasma samples were frozen at ⫺70°C at each study site until shipped on dry ice to a core laboratory for assay. Samples were thawed once just before analysis. Laboratory methods. Chemistry assays (glucose, TC, TG, and LFTs) were performed on an automated analyzer (Roche Cobas-Miraa) by using commercially available reagents. Total HDL-C level was determined after sequential polyanion precipitation of apolipoprotein B– containing lipoproteins.31 LDL-C level was calculated by using the Friedewald equation.32 HbA1c was measured by using high-performance liquid chromatography (Variant Hemoglobin Testing Systemb). Drug safety. Drug safety was determined according to the prevalence of SAEs and AEs during study participation. SAEs and AEs were reported to the site IRBs in accordance with federal and institutional guidelines. IRB review was undertaken to determine whether the SAE or AE was directly related, possibly related, or unrelated to the study drug, which was communicated to all site directors and the study sponsor. UA and LFTs (aspartate and alanine aminotransferase), alkaline phosphatase, total bilirubin) were measured every 4 weeks through study week 24 and again at week 48. Three times the upper limit of the reference range was applied as the threshold criterion for an increase in LFT results, above which drug administration was discontinued. Drug tolerance. Subjects’ perceptions of drug tolerance were assessed every 4 weeks for 24 weeks and at study termination (48 weeks). Subjects were provided a sheet containing a series of blank visual analog scales on which they reported the intensity (from 0 –100) for the following AEs: flushing, rash, itching, headache, insomnia, joint pain, and muscle pain. In addition, several lines were marked “other” to permit open-

401

ended responses not contained in forced-choice options. On the same form, subjects rated, on a Likert scale (range, 1–5), how frequently in the past 4 weeks they experienced AEs, as never (⫽1), occasionally (⫽2), sometimes (⫽3), often (⫽4), or always (⫽5). Drug effectiveness. Improvement in blood lipid profile was used as the criterion for drug efficacy. HDL-C level was the primary outcome, and TC/HDL-C ratio served as a primary global risk assessment. Secondary outcomes included TC, TG, and LDL-C levels and LDL-C/HDL-C ratio. Data Analyses This study used a 2-group (extended-release niacin and placebo) single-blind design. An intention-to-treat approach was adopted before data were collected and used for the final data analysis. Study withdrawals were logged, and the cause for departure was investigated. To avoid possible inflation of the type I error rate, missing data were incorporated by using the last observation carried forward. Source data were recorded by a study coordinator at each study site and forwarded to the lead site. There, a redactor who was otherwise unaffiliated with the trial cross-checked source data from the center records and core biochemistry laboratory data bank with the entered data and then certified the database. Site directors then inspected a deidentified database copy and certified analysis. Data were analyzed by using SPSS for Windows, Version 15.0.c Frequency and descriptive statistics were calculated. Baseline values for all lipid fractions were compared between the extended-release niacin and placebo groups to determine equivalency and the need to account for such potential differences in subsequent statistical modeling procedures with independent-samples t tests. A repeated-measures model analysis of variance was used to examine the main effects of group (extended-release niacin vs placebo) and time from baseline to 48 weeks’ follow-up for HDL-C level, TC/ HDL-C ratio, TC level, TG level, LDL-C level, and LDL-C/ HDL-C ratio. Interaction effects were explored for time by group assignment by study site. The ␻´ Huynh-Feldt correction factor was used to adjust the df for the tests of significance. The ␻´ Huynh-Feldt estimate is a correction factor that is applied to the df to calculate the P value for the observed F statistic. The correction is based on the sphericity (ie, the equality of variances of the differences between measurements) of the data, which is an assumption of repeated-measures analysis of variance. Significant main effects were examined further by using simple-effects tests. UA, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and total bilirubin were analyzed by group, and percentages greater or less than normative limits were calculated. Pearson product-moment correlations were used to determine relationships between selfreported flushing and insomnia and visual analog scales and Likert scores. The criterion for significance in all tests was ␣ equals .05. RESULTS Figure 1 is a trial flow diagram showing participant screening, randomization, allocation, follow-up, and analysis. Persons (N⫽97) were screened at 3 study sites. Of these, 43 subjects were excluded for the following reasons: 26 for HDL-C level greater than 40mg/dL, 11 for TC and/or TG level greater than 250mg/dL, and 6 for inability to fulfill the study testing schedule. Of 54 persons randomly assigned to extended-release niacin and placebo treatment, 14 (8 extendedrelease niacin, 6 placebo) withdrew consent or were disArch Phys Med Rehabil Vol 92, March 2011

402


Fig 1. Consolidated Standards of Reporting Trials (CONSORT) flow diagram.

charged by the investigators. Descriptive characteristics for these participants are listed in table 2. Causes for withdrawal are listed in table 3. Four withdrawals were determined to be probably or possibly related to the study drug, whereas the rest were unrelated. Of the 4 cases, 1 (hypotension/presyncope) occurred after the first drug dose and 3 (2 diarrhea, 1 flushing discomfort) occurred after the first month of treatment. Only 1 AE was rated as an SAE, which was determined as near-syncope that transcended simple hypotension and flushing. Outcomes for UA, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and total bilirubin are shown in figure 2. Outcomes for fasting glycemia and HbA1c are shown in figure 3. Descriptive data for symptom intensities and frequencies (Likert values) for flushing and insomnia are shown in figure 4. Outcomes for lipids, lipo-

Table 2: Characteristics of Study Participants Patient Data

ER-N

No. Age (y) Duration of injury (y) Level of injury

31 32.2⫾8.2 10.4⫾6.7 C4–C7

Placebo

23 34.1⫾9.3 9.9⫾8.2 C5–C7

NOTE. Values are mean ⫾ SD unless otherwise noted. Abbreviations: ER-N, extended-release niacin; ND, not done.


P

ND .41 .52 ND

proteins, and risk ratios at baseline and 4-, 8-, 12-, 16-, 20-, 24-, and 48-week follow-up times are shown in figure 5. Although LFT results, except for alkaline phosphatase and UA levels, increased after extended-release niacin therapy, only a single value for UA exceeded 3 times the upper limit of the reference range at week 12, whereas all other interaction effects for AEs and tolerance failed to reach clinical significance. LFT results and UA levels did not significantly increase over time compared with baseline levels in the extended-release niacin group. No differences were determined between groups for any outcome variable at baseline. When values for lipids, hepatic transaminases, and UA and flushing and insomnia by group and study center over time were compared, only TC level showed an interaction effect of time ⫻ group ⫻ center (F10.1,242.1⫽2.1; P⫽.028). The ␻´ Huynh-Feldt correction factor for within-subject effects df was .72. Simple-effects tests showed that no changes occurred for subjects at the Kessler site; extended-release niacin subjects at the Miami site showed significant decreases from baseline to weeks 8, 20, 24, and 48; and subjects using extended-release niacin at the Rancho site showed significant decreases from baseline to weeks 24 and 48. The relationship between flushing and insomnia was investigated and correlated with visual analog scale scores in the placebo group at weeks 24 (r⫽.56; P⫽.005) and 48 (r⫽.52; P⫽.01) and with Likert scores in the extendedrelease niacin group at week 12 (r⫽.36; P⫽.046), reflecting

403

EXTENDED-RELEASE NIACIN IN TETRAPLEGIA, Nash Table 3: Study Withdrawals for Treatment-Emergent and Non–Treatment-Related Events No.

Study Arm

Cause for Withdrawal

Study Week

IRB Finding

2 1 1 1 1 1 1 6 Total: 14

ER-N Placebo ER-N ER-N ER-N ER-N ER-N 5 Placebo, 1 ER-N 8 ER-N, 6 placebo

Diarrhea Fracture Flushing Presyncope/flushing/hypotension Presyncope/hypotension Pressure ulcer Requested group reassignment, denied Transportation/relocation/logistics

4 36 4 1 24 8 16 4-48

Probably related (AE) Unrelated Related (AE) Related (SAE) Probably unrelated Unrelated (AE) Unrelated Unrelated 4/14 Related or probably related

NOTE. The 4 findings that associated drug use with AEs are rated as either AE or SAE. Abbreviation: ER-N, extended-release niacin.

a detachment between drug-induced symptoms of flushing and insomnia. Repeated-Measures Analyses For the primary endpoint of HDL-C level, significant effects were found for group (F1,52⫽14.5; P⬍.001), time (F4.1,214.4⫽ 26.4; P⬍.001), and group ⫻ time interaction (F4.1,214.4⫽12.3; P⬍.001; ␻´ Huynh-Feldt correction factor df⫽.59). Further analysis of these interaction effects indicated no significant difference between groups at baseline; however, HDL-C levels were significantly and progressively higher at weeks 4 through 48 in the extended-release niacin group (see fig 5). For the primary global risk outcome of TC/HDL-C ratio, significant effects were found for group (F1,52⫽13.2; P⫽.001), time (F4.6,239.1⫽34.2; P⬍.001), and group ⫻ time interaction (F4.6,239.1⫽10.9; P⬍.001; ␻´ Huynh-Feldt correction factor df⫽.66). No significant differences were found between groups at baseline; however, the extended-release niacin group showed significant decreases in TC/HDL-C ratio from week 4 to week 48 compared with the placebo group (see fig 5). For the secondary outcome of LDL-C level, significant effects were found for group (F1,52⫽7.0; P⫽.011), time (F4.3,222.2⫽14.9; P⬍.001), and group ⫻ time interaction (F4.3,222.2⫽4.6; P⫽.001; ␻´ Huynh-Feldt correction factor df⫽.61). No significant differences were found between groups at baseline and week 4, but the extended-release niacin group showed significant decreases in LDL-C levels from week 8 to week 48 compared with the placebo group (see fig 5). For the secondary outcome TC level, significant effects were found for group (F1,52⫽4.9; P⫽.032), time (F4.5,235.2⫽9.9; P⬍.001), and group ⫻ time interaction (F4.5,235.2⫽2.8; P⫽.021; ␻´ Huynh-Feldt correction factor df⫽.65). Further analysis of these interaction effects indicated no significant difference between groups from baseline through week 4, a significantly lower TC value for the extended-release niacin group at week 8, and no differences from weeks 12 to 16. However, TC level was significantly and progressively lower at weeks 20, 24, and 48 in the extended-release niacin group (see fig 5). For the secondary outcome of TG level, a significant effect was found for time (F3.6,184.4⫽3.2; P⫽.018) but not for group ⫻ time interaction or group. The overall change showed a decreasing TG level from week 16 to 48 compared with baseline. The decrease in TG level was significant for the extended-release niacin group at week 48 compared with baseline levels (t30⫽3.5; P⫽.01) while remaining unchanged in the placebo group. Predictably, decreasing TG levels tended to correlate inversely with increases in HDL-C levels (r⫽⫺.33; P⫽.07) (see fig 5).

For the secondary global risk outcome LDL-C/HDL-C ratio, significant effects were found for group (F1,52⫽12.6; P⫽.001), time (F4.4,230.7⫽29.7; P⬍.001), and group ⫻ time interaction (F4.4,230.7⫽9.2; P⬍.001; ␻´ Huynh-Feldt correction factor df⫽.63). No significant differences were found between groups at baseline, but the extended-release niacin group showed significant decreases in LDL-C/HDL-C ratio from week 4 to week 48 compared with the placebo group (see fig 5). In the extended-release niacin group, TC level, LDL-C level, TC/HDL-C ratio, and LDL-C/HDL-C ratio significantly decreased at all follow-up times compared with baseline (see fig 5). HDL-C levels were significantly higher than baseline at all successive follow-up times. For TG levels, only differences between baseline and week 20 values were significant. Dose-dependent changes spanning the 16-week drug dose titration are shown in figure 6. All values for HDL-C level increase and decrease in both LDL-C and TC levels were significantly different from baseline (P⬍.05) and each other dosing level (all P⬍.05). DISCUSSION The findings of this study provide important information for establishing practice patterns in the treatment of SCI-associated dyslipidemia. The importance of modifying HDL-C levels to decrease CVD has been recommended in consensus guidelines for more than a decade and recently has been summarized by Dyson-Hudson and Nash.33 Of particular relevance to persons with SCI were adjusted policies of the Third National Cholesterol Education Program consensus guidelines, which expanded the number of patients who are eligible for lipidbenefitting treatment by increasing the criterion score for low HDL-C level from 35 to 40mg/dL. These guidelines also introduced the concept of 10-year absolute coronary risk (“global risk”), to which low HDL-C level adds risk, and identified CMS, of which low HDL-C level is a constituent as a “coronary risk equivalent.” A study by Nash and Mendez11 observed that this guideline change substantially increased the number of persons with paraplegia who qualify for lipidaltering therapy, which was attributable largely to the current clinical goal that seeks more extensive cardioprotection through HDL-C level increase. The overarching findings of this study confirm salutary changes of extended-release niacin therapy on a wide spectrum of the primary and secondary lipid and lipoprotein outcomes, while maintaining levels of safety and tolerance both similar to and deemed acceptable in previous trials of primary and secondary CVD prevention.28-30,34,35 Arch Phys Med Rehabil Vol 92, March 2011

404


Fig 2. Effects of extended-release niacin (ER-N) and placebo on (A) UA, (B) alanine transaminase, (C) aspartate transaminase, (D) alkaline phosphatase, and (E) total bilirubin levels.



405

Fig 3. Effects of extended-release niacin (ER-N) and placebo on (A) fasting glycemia and (B) HbA1c level.

Safety Flushing accounted either directly or indirectly for 2 of 4 treatment-emergent events and the same number of withdrawals. This 6.4% event rate approximated the 7.3% study discontinuation (with aspirin) and 9% (without aspirin) reported in persons without disability undergoing extended-release niacin therapy,36 although we believe that special attention is warranted for drug treatment in persons with hypotension accompanying cervical SCI. This attention was provided by the administration of each escalation dose in the clinic setting, which allowed us to monitor responses until the maximum drug dose was achieved. In 1 of 2 cases, flushing was limited to discomfort of warmth and tingling, whereas the other was classified as an SAE involving hypotension and presyncope. Another patient experienced hypotension after reaching full dose escalation in week 24, but his history suggested other underlying causes. In all other cases, flushing was managed by using pretreatment with coated aspirin and remained stable thereafter in both severity and intensity, even during dose escalation spanning study weeks 4 to 16. Both patients involving hypotension recovered quickly when the drug therapy was terminated. Two persons experienced bowel accidents in the first month of treatment and withdrew from the trial. It should be appreciated that, because of the loss of voluntary bowel control in persons with more neurologically complete SCI, a change in bowel habits may be embarrassing and socially limiting. The 29 other subjects using the drug reported no change in bowel habits. In comparison, niacin-induced diarrhea occurred in 14% of pooled cases of persons without disability. Although the diarrhea AE could not be confirmed as a direct result of drug treatment, the symptom resolved when drug treatment was terminated, making the association credible. One person experienced intensified itching and a UA level that exceeded 3 times the upper limit of the reference range while escalating doses from 1500 to 2000mg. In this case, dose escalation was aborted and the drug dose was tapered to 1500mg, with resolution of symptoms, and then re-escalated to 2000mg without recurrence of either pruritus or increased UA levels. A slight increase in hepatic transaminase levels was observed, as noted in other trials, although these increases were not biologically significant. Fasting glycemia and HbA1c levels were stable throughout the trial, although an increased fasting blood glucose level commonly is monitored in clinical trials

involving extended-release niacin treatment. Double-blinded trials of populations with type 2 diabetes have reported an average increase in HbA1c level by 0.3%, which usually was controlled by a minor adjustment in the diabetic drug regimen.37,38 Similar minor increases in fasting blood glucose levels have been reported in persons without diabetes and in those with well-controlled diabetes.29,39 In the present study, 1 person had mild fasting hyperglycemia before treatment, but experienced no worsening of fasting glycemia or HbA1c level throughout drug treatment. Tolerance The combination of HDL-C level increase and other broad-spectrum benefits observed in this study would appear to make extended-release niacin an ideal agent for use by persons with SCI, although the need for special dosing and unpleasant symptoms often decreased patient compliance.36 Most notable of sensations was cutaneous vasodilation characterized by flushing, which was sensed as redness, warmth, itching, and tingling.40 Niacin-induced flushing occurred in 70% to 100% of patients enrolled in clinical trials and explains why many patients discontinued treatment. Management of flushing normally incorporates behavioral and pharmacologic strategies, both of which were undertaken in this trial. Behavioral approaches typically control timing of medication with a small low-fat snack, drug administration at bedtime, and abstention from hot drinks, alcohol, and spicy foods when nearing administration.40 Pharmaceutical strategies have included formulations that slow absorption rates, although at the risk for hepatotoxicity. Because flushing is mediated by release of vasodilator eicosanoids, particularly prostaglandin D2, an aspirin dose of 325mg taken 30 to 60 minutes before the drug reduces discomfort.41,42 In the present study, reports of flushing occurred in 65% of subjects randomly assigned to drug and 52% randomly assigned to placebo, although only 2 persons withdrew from the trial because of flushing, 1 after 4 weeks of drug treatment and another in week 24 after escalation from 1.5 to 2g of drug. These outcomes compare favorably with pooled statistics reporting an 82% flush rate in men undergoing multigram dosing of niacin.35 Reports of flushing in placebo subjects are common, and 20% of nondisabled patients randomly assigned to placebo in a clinical trial of extendedArch Phys Med Rehabil Vol 92, March 2011

406


Fig 4. Frequencies of (A) flushing and (B) insomnia and intensities of (C) flushing and (D) insomnia during extended-release niacin (ER-N) monotherapy.



407

Fig 5. Effects of extended-release niacin (ER-N) and placebo on levels of (A) HDL-C, (B) LDL-C, (C) TC, (D) TG, and (E) TC/HDL-C and (F) LDL-C/HDL-C ratios (global risk ratios).

release niacin reported at least 1 flushing episode.29 Flushing in subjects on extended-release niacin therapy averaged more often than occasionally, but less frequently than sometimes, whereas the intensity of flushing increased slightly from baseline to week 4 but stabilized at average visual analog scale scores of 22.4 to 30.2 (of 100) throughout the balance of treatment. Neither intensity nor frequency ratings for flushing significantly worsened with time, even when escalating drug doses from 500 to 2000mg during weeks 4 through 16. A slight nonsignificant decrease in flushing intensity in weeks 24 to 48 may reflect common reports of

tolerance to extended-release niacin, a point at which aspirin pretreatment sometimes can be discontinued. Forced-choice ratings of rash, itching, headache, insomnia, joint pain, and muscle pain were random and statistically nonsignificant. Because self-reports of insomnia were slightly prevalent in subjects randomly assigned to the drug arm, we tested for an association between flushing and insomnia but found the 2 phenomena to be unrelated. Conversely, several subjects were pleased with the warmth provided by flushing, which offset reports of cold sensations frequently noted by persons with cervical SCI. Arch Phys Med Rehabil Vol 92, March 2011

408


Fig 6. Extended-release niacin dose-dependent changes in HDL-C, LDL-C, and TC levels. *P